Wikipedia talk:Verifiability

Active discussions
Frequently asked questions (FAQ)
Where should I ask whether this source supports this statement in an article?
At Wikipedia:Reliable sources/Noticeboard. Don't forget to tell the editors the full name of the source and the exact sentence it is supposed to support.
Do sources have to be free, online and/or conveniently available to me?
No. Sources can be expensive, print-only, or available only in certain places. A source does not stop being reliable simply because you personally aren't able to obtain a copy. See Wikipedia:Reliable sources/cost. If you need help verifying that a source supports the material in the article, ask for help at Wikipedia:WikiProject Resource Exchange or a relevant WikiProject.
Do sources have to be in English?
No. Sources can be written in any language. However, if equally good sources in English exist, they will be more useful to our readers. If you need help verifying that a non-English source supports the material in the article, ask for help at Wikipedia:Translators available.
I personally know that this information is true. Isn't that good enough to include it?
No. Wikipedia includes only what is verifiable, not what someone believes is true. It must be possible to provide a bibliographic citation to a published reliable source that says this. Your personal knowledge or belief is not enough.
I personally know that this information is false. Isn't that good enough to remove it?
Your personal belief or knowledge that the information is false is not sufficient for removal of verifiable and well-sourced material.
Is personal communication from an expert a reliable source?
No. It is not good enough for you to talk to an expert in person or by telephone, or to have a written letter, e-mail message, or text message from a source. Reliable sources must be published.
Are there sources that are "always reliable" or sources that are "always unreliable"?
No. The reliability of a source is entirely dependent on the context of the situation, and the statement it is being used to support. Some sources are generally better than others, but reliability is always contextual.
What if the source is biased?
Sources are allowed to be biased or non-neutral. Only Wikipedia articles are required to be neutral. Sometimes "non-neutral" sources are the best possible sources for supporting information (with due weight) about the different viewpoints held on a controversial subject.
Does every single sentence need to be followed by an inline citation?
No. Only four broad categories of material need to be supported by inline citations. Editors need not supply citations for perfectly obvious material. However, it must be possible to provide a bibliographic citation to a published reliable source for all material.
Are reliable sources required to name the author?
No. Many reliable sources, such as government and corporate websites, do not name their authors or say only that it was written by staff writers. Although many high-quality sources do name the author, this is not a requirement.
Are reliable sources required to provide a list of references?
No. Wikipedia editors should list any required sources in a references or notes section. However, the sources you are using to write the Wikipedia article do not need to provide a bibliography. Most reliable sources, such as newspaper and magazine articles, do not provide a bibliography.

Freedom of pressEdit

I made a possibly over-bold edit to insert a sentence (and footnote) on freedom of the press; please revert and discuss here as needed. HLHJ (talk) 21:30, 19 December 2019 (UTC)

Thanks, User:Chris troutman. Can you suggest any useful guidance for situations in which domestic media under government control are reporting on a subject of strong interest to the government? Is it needed here? We don't seem to get as many editors from countries without fairly high freedom of the press; certainly we get fewer political ones. Perhaps the info should go in the Wikipedia:Reliable sources guideline? HLHJ (talk) 21:54, 19 December 2019 (UTC)
I think local consensus makes more sense than a sweeping change to this policy. While I might claim that Xinhua is state-run media and therefore unreliable in all matters regarding geo-politics, there are others that would make the same claim about Stars & Stripes. I think article by article, local consensus can determine if a cited source is reliable for the content it supports. (I was also against the WP:DAILYMAIL decision for this reason.) Chris Troutman (talk) 22:00, 19 December 2019 (UTC)

The reliability of sources in countries with low press freedom is a very controversial topic, since it involves judging not only the source, but also the country that it is based in. Perhaps you can start by editing the independent sources supplement, since it's quite clear to me that these sources are non-independent (though not necessarily unreliable) for government-related topics. If that goes over well, consider discussing it on WT:RS, as the guideline can afford to be more specific than this policy. — Newslinger talk 22:10, 19 December 2019 (UTC)

In my opinion, that discussion is an additional argument in favour of modification on the WP:SOURCES section. Indeed, a current version says that "mainstream newspapers" are reliable source, but
  • the term "mainstream" is not defined, so some users believe that any newspapers, including the newspapers under control of non-democratic governments, are mainstream unless the opposite has not been demonstrated; in contrast, other users believe only a limited number of top rank newspapers are mainstream.
  • the blanket statement "mainstream newspapers are reliable sources" contradict to what WP:NEWSORG say: the guidelines explain that reliability of newspapers varies broadly: only some newspaper materials are reliable secondary sources, whereas others are of lower quality, and, being primary sources they should be used with attribution only, because they are reliable only for the author's opinion.
  • this section clearly ignores the fact that reliable sources may be of higher and lower quality, and, depending on a context, per WP:REDFLAG the same source may be acceptable or unacceptable.
Therefore, a solution to the problem described by HLHJ, and to many other problem would be to add the following explanations to the WP:SOURCES:
  • that the sources listed as reliable in that section may be of higher or lower quality, and they may be considered primary or secondary, as explained in the guidelines, depending on a context;
  • that the threshold of acceptability may vary, as described in WP:REDFLAG, and the sources that are acceptable for some non-controversial statement are not acceptable for some extraordinary statement.
In my opinion, that will resolve the problems with newspapers controlled by non-democratic governments, for these newspapers will continue to be considered reliable for, e.g. the results of a recent football game, but they will be considered unreliable for a description of some controversial political event in that country.--Paul Siebert (talk) 23:54, 19 December 2019 (UTC)
(edit conflict) Makes sense. I was advocating a case-by-case statement-by-statement approach. For instance, I'd say Stars & Stripes would not be an independent source for controversial statements about, say, the conduct of US soldiers, and Xinhua should probably not be used as a source for controversial statements about the conduct of the Chinese government; either would be fine for some statements. I'd consider the Daily Mail decision quite a separate issue; as I recall, the objection raised by ban supporters was not that the Daily Mail was too closely tied to some issues that it was reporting on, but that it was too loosely tied to reality to be relied upon to be accurate upon any topic at all, ever. This is a much more sweeping claim, and not one that could reasonably be made of most state-controlled media.
I can see how this could open a can or worms, and the policy does mention the idea of independence: "Articles must be based on reliable, independent, published sources with a reputation for fact-checking and accuracy" and "Questionable sources are those that have a poor reputation for checking the facts, lack meaningful editorial oversight, or have an apparent conflict of interest" (emphasis added). "Independent" is linked to Wikipedia:IS, to which I previously made an edit on this topic (in Wikipedia:Identifying and using independent sources#Conflicts of interest); revisions welcome. HLHJ (talk) 06:41, 20 December 2019 (UTC)
For the case that brought me here: Template talk:Third-party#Template-protected edit request on 19 December 2019. HLHJ (talk) 06:45, 20 December 2019 (UTC)
I didn't notice your addition to WP:IS until now. Thanks for that. — Newslinger talk 15:20, 20 December 2019 (UTC)
What, you didn't notice every tiny paragraph in that long page? No worries.  I'm not sure there are any real policy differences being discussed in this section; we don't have any examples where one person says it's reliable and another says that it isn't. It's just the structure and ordering of the content, and the phrasing of principles we all seem to agree on. "Mainstream media" is a perennial logomachy subject even off-wiki; if we were to create a Wikipedia-specific definition of the term de novo, we'd probably find it easier to agree. HLHJ (talk) 16:14, 20 December 2019 (UTC)
  • This is not an improvement. It is not clear what "government-controlled" means. One could say that the Voice of America, for example, is controlled by US government. But it is still an RS. My very best wishes (talk) 16:32, 20 December 2019 (UTC)
    On closer inspection, I think the phrasing in Special:Diff/931590746 is too harsh and too broad, but a milder and more restrictive version of the wording would reflect the controversial nature of sources like TASS (RfC). If TASS were based in a country with higher press freedom, the RfC wouldn't have been nearly as divisive. Anyway, changes to WP:IS should probably be discussed at WT:IS. — Newslinger talk 18:16, 20 December 2019 (UTC)
It tells: In countries without freedom of the press [all] domestic media are under the control of the government. There are very few countries like North Korea where all media can be summarily dismissed.My very best wishes (talk) 18:52, 20 December 2019 (UTC)
How about this?

Sources based in countries with low press freedom may be non-independent; this applies especially to state media, and to coverage of politics and other government-related topics. Consider using these sources with in-text attribution to clearly identify the provenance of the content, especially when it is opinionated or controversial.

I've notified WT:IS of this discussion. — Newslinger talk 19:39, 20 December 2019 (UTC)

First phrase is much better. Looks good to me (I would only say "News sources" since we are taling about freedom of press). Second phrase, it is also OK, however we would probably would like to avoid using such sources at all, especially if better sources on the subject exist... My very best wishes (talk) 20:28, 20 December 2019 (UTC)
Thanks for the feedback. I was actually proposing this for WP:IS, so I'm not sure how applicable this would be for WP:V. I do think that "Self-published and dependent sources" is not the ideal way to organize this kind of information on WP:V. Paul Siebert's bottom two bullet points are cautiously and concisely worded summaries of existing policies and guidelines (WP:RSCONTEXT and WP:REDFLAG), and they look good to me. — Newslinger talk 20:44, 20 December 2019 (UTC)

This is out of place here. You are talking about ACTUALLY reliable sources. Wp:RS is not about actual reliability. Otherwise you would see things like "objectivity" and "knowledge in the subject area" in this policy. Or to put it less flippantly, the discussion is dancing around the gorilla in the living room which has never gotten fixed. North8000 (talk) 20:13, 20 December 2019 (UTC)

@North8000: alas, you are not right, and your post reflects a common confusion that should be fixed. The confusion is as follows. Many people believe WP:V separates sources according to just one criterion: "reliable - non-reliable". However, in reality, it also says: The best sources have a professional structure in place for checking or analyzing facts, legal issues, evidence, and arguments. The greater the degree of scrutiny given to these issues, the more reliable the source. Be especially careful when sourcing content related to living people or medicine. That means it implicitly introduces another criterion: "high quality sources - low quality sources". These two type criteria are, by and large, independent. Thus, some articles published in top rank peer-reviewed journals are of very poor quality, whereas some blogs may be excellent in terms of fact checking and accuracy. Of course, that does not mean that we should allow usage of some SPS because it looks good. That is obvious. What is not obvious is the following:
Sometimes, some sources that meet formal criteria applied to RS are of poor quality, and they cannot be used in Wikipedia to support certain statements.
Consider this hypothetical example. NYT, which is a mainstream newspaper, and, hence, an RS, published an editorial that contained some opinion about Heisenberg's principle. Does it make it a RS for the Uncertainty principle article? Of course, a reasonable, good faith user will argue that, per WP:NEWSORG, this editorial is just a primary source about the author's opinion, or about NYT editorial board's opinion, and, taking into account that the editorial board is hardly composed of notable physicists, this article is not an RS for the Uncertainty principle WP article. However, NEWSORG are just guidelines, whereas the policy says "Mainstream newspapers are RS (period)". That creates an opportunity for bad faith or poorly educated users to argue that that NYT article is an RS, and to add some potentially poor quality material to Wikipedia. Of course, that would violate the WP:V's spirit, but it would be in accordance with the current WP:V's letter.
To fix that, I propose the following version of WP:SOURCE (its second part, which currently starts from "If available, academic and peer-reviewed publications":

Academic and peer-reviewed publications, as well as the books published by top rank universities are considered the best quality reliable sources for such fields as history, medicine, or science.

Non-academic sources may also be considered reliable, particularly if they appear in respected mainstream publications, such as books by respected publishing houses, magazines, journals, and newspapers, including electronic media.

Reliably published sources may be of higher or lower quality, and, depending one the context, they may or may not be used to support certain statements (see WP:RSCONTEXT, WP:REDFLAG, Wikipedia:Identifying reliable sources, and Wikipedia:Search engine test for further details).

In my opinion, that modification resolves the press freedom problem, as well as many others. Indeed, the proposed version is not a blanket approval of "mainstream newspapers", but it is not wholesale ban of newspapers published in non-democratic countries: according to the proposed version, a publication in, e.g. Pravda may be an acceptable RS for the report about Brezhnev's death, but not acceptable for Korean Boeing story.--Paul Siebert (talk) 22:00, 20 December 2019 (UTC)
@Paul Siebert:Paul, I know all of that and still stand by what I said. But I also agree with most of what you said. Also, your proposal is a good tweak to resolve the current issue without fixing the big problem. The big fix would be to weigh "objectivity and knowledge with respect to the items which cited it" in all evaluations of to what extent a source is reliable / how strong the sourcing is. North8000 (talk) 23:14, 20 December 2019 (UTC)
North8000 I cannot understand how your last post follows from your previous post. In your previous post, you write that actual reliability has no relation to verifiability. Yes, I agree with that. Strictly speaking, the very name of the policy implies that it deals mostly with verifiability, and "reliable source" actually means "reliably published source", so everybody can take a reference, go to a library and see that the source X really says "Y", and Wikipedia transmits this information correctly. However, if we leave actual reliability beyond the scope, then which policy is supposed to define which sources are trustworthy and which are not? I don't see how NOR or NPOV can do that. To create some additional policy dealing with actual reliability is hardly a good solution, which means WP:V should focus on both verifiability and actual reliability.
If you think the text proposed by me solves just minor problems, feel free to propose a more global solution. However, I think my approach is strategically correct: to define just the most basic principles in the policy, and to move all details (including "objectivity and knowledge with respect to the items which cited it" etc) to relevant guidelines. Currently, there is a contradiction between the policy, which contains blanket approval of some broad categories of sources, and guidelines, which are more restrictive. Obviously, that makes guidelines toothless, because policy takes precedence over guidelines. In contrast, in my version, all blanket approvals are removed, which gives more weight to guidelines, where all issues described by you can be explained in more details.--Paul Siebert (talk) 07:29, 21 December 2019 (UTC)
I agree that my edit is too harsh and broad. The text linked in the first link in this discussion might be better; not really fashed what form it takes. The COI section might also be improved by mention of sponsored supplements. I'm lost in this discussion, and may well have overlooked things. HLHJ (talk) 23:30, 20 December 2019 (UTC)
I would say, your edit is too narrow: it focuses just at some category of newspapers, whereas other newspapers may also be problematic. On the other hand, even government controlled newspapers may be reliable sources in some cases, as I explained above. Therefore, the solution is not to label some category of newspapers as unreliable, but to explain that every newspaper may be reliable or non-reliable depending on a context (which is explained elsewhere). That would be sufficient for the policy, so all what you say should be added to the guidelines.--Paul Siebert (talk) 07:29, 21 December 2019 (UTC)
  • I would oppose to changes as suggested by Paul. According to current version, "If available, academic and peer-reviewed publications are usually the most reliable sources, such as in history, medicine, and science.". Note word "usually". This is because in some cases "non-academic" publications (like books by known experts addressed to general public) can be much better than articles in peer-reviewed journals by PhD students, and especially in non-scientific areas, such as politics and history. My very best wishes (talk) 01:15, 21 December 2019 (UTC)

Answering Paul, the big issues are that in the core of policies:

  1. wp:rs defines reliable sources as merely having some trappings that tend to increase reliability. (e.g. layer of editorial review, published) These are not enough metrics.
  2. It doesn't recognize that there are degrees of reliability, nor have a way to utilize that. Everything is "binary"
  3. It refuses to recognize that biased sources makes them less reliable. E.G that they are more likely to present a distorted (= inaccurate) view
  4. It refuses to recognize that reliability is context-specific, as you explained.

The 3 part big fix would be:

  1. Introduce the concept that there is strength of sourcing (i.e. a matter of degree), and that it is context sensitive.
  2. "Strength of sourcing" would be determined by the existing wp:rs criteria plus these 2 new criteria: #1 Objectivity and knowledge regarding the item which cited it
  3. Say that more controversial (regarding it's accuracy or veracity) content requires stronger sourcing, and vice versa.

Unfortunately, I think that due to it's processes, en.Wikipedia is no longer capable of big fixes. But we can still try. :-) Sincerely North8000 (talk) 15:18, 21 December 2019 (UTC)

Re issues.
  1. I probably was not clear enough, bit that is exactly what I say.
  2. Yes, and I proposed the solution.
  3. I think that is more relevant to NPOV. I assume each source should be treated as biased, and the purpose of NPOV (not V) to find a balance between various biases. I don't think bringing this issue here is helpful, because even without that the discussion's subject is too complicated.
  4. Yes.
Re big fix:
  1. That is exactly what I have proposed (Reliably published sources may be of higher or lower quality, and, depending one the context, they may or may not be used to support certain statements). I do not think the policy should be more specific, the goal is to remove what some people see as blanket approval of some types of sources. Instead, the policy should redirect a user to WP:RSCONTEXT, WP:REDFLAG, Wikipedia:Identifying reliable sources, and Wikipedia:Search engine test, which explain reliability criteria in more details. In addition to that, I propose to make a stress on WP:BURDEN, to emphasize that the user who adds some source is expected to be ready to provide evidences that that source is reliable and can be used in this particular context. Currently, the policy implies that, but it does not say that explicitly. It should be clear that the user who wants to remove some source should not prove that the source is unreliable. In contrast, if adequate evidences of reliability and relevance are not presented in response to a justified criticism, that should be a sufficient criterion for removal.
  2. "Objectivity criteria" are too subjective. In addition, what is the problem with the current wording ("The best sources have a professional structure in place for checking or analyzing facts, legal issues, evidence, and arguments. The greater the degree of scrutiny given to these issues, the more reliable the source")?
  3. Sorry, but WP:REDFLAG already says that. In my version, I already provided an additional reference ro REDFLAGG to emphasize that.
In addition, we must keep in mind that many, if not majority of Wikipedia articles describe low importance subjects (small and remote towns, local schools, etc), for which sourcing criteria should be loose, otherwise these articles will be deleted. I personally don't see why we shouldn't allow them to exist.--Paul Siebert (talk) 16:48, 21 December 2019 (UTC)

I think we 90% agree. The 10% is:

  • We need to lean towards objective sources, not an inevitably subjective process of combining biased sources.
  • IMO my "big fix" #3 isn't already happening. Wikilawyers routinely knock out "sky is blue" statements by nitpicking source details (this would fall under my "vice versa" ....a lower sourcing bar when the statement itself is not challenged/ controversial) And Wikilawyers are keeping in really controversial stuff because the biased (=unrelaible) source has the qualifying wp:rs trappings.

Sincerely, North8000 (talk) 21:47, 21 December 2019 (UTC)

@North8000: sorry, I haven't noticed your responce.
  • Re objectivity. That is good in theory, but in theory, there is not difference between theory and practice, but in practice, there is. Who will decide which source is "objective"? Can you propose non-biased criteria for objectivity? I think, for entia non sunt multiplicando we should return to "better quality/lower quality reliable sources": as a rule, good quality sources, e.g. scholarly articles, are less biased.
  • Re #3, nitpicking source details is irrelevant, because as soon as "sky is blue" is not an extraordinary claim, any nitpicking is senseless REDFLAG refers to the content, not to sources. Only after the evidences have been provided that "sky is blue" is an extraordinary/controversial statement can we request using stricter sourcing criteria. And, to make REDFLAG workable, we should explain (as I am doing in my version) that there is no blanket approval of certain types of sources, and each of them may be, or may be not appropriate in some concrete context, as explained in guidelines. I think it would be good if you proposed some concrete wording that can make this idea working.--Paul Siebert (talk) 22:12, 27 December 2019 (UTC)

I saw the note at WT:IS. User:Newslinger, the difficulty with saying "Sources based in countries with low press freedom may be non-independent" is that you need to specify what they're non-independent of. Even the most oppressively state-controlled media outlet would still an independent source for some content. If The Official Ruritanian Press Organ, in a country that produces neither coffee nor cacao, declares that coffee is a better drink than hot chocolate, then they're still independent for that claim. Being thoroughly biased (or even, as in my example, obviously and greviously wrong ;-)) does not make a source any less independent. WhatamIdoing (talk) 04:31, 2 January 2020 (UTC)

That's a very good point, and if something gets added to any of these pages, the wording should take that into account. — Newslinger talk 15:18, 5 January 2020 (UTC)
Entirely agree with WhatamIdoing (here). State-controlled media would probably also be RS for the official opinions expressed by the controlling government. For domestic production stats, not so much. Different sources are independent of different things; to me, this implies that the reliability of sources can depend on the topic. HLHJ (talk) 05:17, 11 January 2020 (UTC)
Relevant discussion from Wikipedia:Reliable sources/Noticeboard#Saudi sourcing problems:

== Saudi sourcing problems ==

I ask editors to please be wary of some sources on subjects in which the Saudi government takes a strong interest. Sadly, there may not be reliable, independent sources of information available on many Saudi-Arabia-related subjects. This has been raised here before, at Wikipedia:Reliable sources/Noticeboard/Archive 5#ArabNews, but that was 2007.

The Saudi Arabian government exerts very close control over the domestic media; it appoints editors, issues national bans on employing specific journalists, sends out guidelines on how stories are to be covered,[1] requests that influential public figures make specific statements in support of the government on specific occasions, and so on.[2][3] People who publish the wrong thing, or fail to publish the right thing, may be disappeared, arrested, imprisoned, kept in solitary confinement, tortured, or killed.[4][2]

The result is a press that strongly resembles a government PR department, and publications that resemble press releases. With the best will in the world, I don't think that Saudi-government-controlled sources can reasonably be considered independent of the government. This includes any media outlet operating from a .sa website, and some Saudi-owned media outlets run from outside the country (Asharq Al-Awsat, for instance). In other countries in which there is little freedom of the press, and the censors are beholden to the Saudi government, the media also publish some stories which seem to come from the same copybook.

The Saudi Arabian government also attempts to exert control over foreign media (see Jamal Khashoggi and Jeff Bezos#Politics). Saudi Arabia is spending large sums on overt and covert influencers (those who do not declare their conflicts of interest). It seems to be doing this to improve its public image abroad, especially in the wake of Jamal Khashoggi's death, and attract tourists.[5][6]

References

  1. ^ Campagna, Joel. "Saudi Arabia report: Princes, Clerics, and Censors". cpj.org. Committee to Protect Journalists.
  2. ^ a b "The High Cost of Change: Repression Under Saudi Crown Prince Tarnishes Reforms". Human Rights Watch. 350 Fifth Avenue New York NY 10118-3299 USA. 4 November 2019. Reuters noted that many of those detained had failed to sufficiently back Saudi policies, including the policy of isolating Qatar. A relative of Salman al-Awda told Human Rights Watch he said he believed that authorities arrested al-Awda because he hadn’t complied with an order from Saudi authorities to tweet a specific text to support the Saudi-led isolation of QatarCS1 maint: location (link)
  3. ^ Ismail, Raihan. "How is MBS's consolidation of power affecting Saudi clerics in the opposition?". Washington Post.
  4. ^ Yee, Vivian (26 November 2019). "Saudi Arabia Is Stepping Up Crackdown on Dissent, Rights Groups Say". The New York Times.
  5. ^ Massoglia, Anna (2 October 2019). "Saudi Arabia ramped up multi-million foreign influence operation after Khashoggi's death". OpenSecrets News. The Center for Responsive Politics.
  6. ^ Thebault, Reis; Mettler, Katie (December 24, 2019). "Instagram influencers partied at a Saudi music festival — but no one mentioned human rights".

How did I come across this? I decided to rescue an abandoned AFC draft on a book fair. In my ignorance, I really didn't expect the topic to be that political, at least not to the extent that I'd wind up writing about torture... (crossposted to New Pages Patrol) HLHJ (talk) 19:47, 23 January 2020 (UTC)

HLHJ, in principle I agree but this needs a carefully worded RfC identifying specific sources and the areas for which they should be considered unreliable. Guy (help!) 23:51, 24 January 2020 (UTC)
HLHJ, agree. Add to the list, Al-Arabiya and CNN Arabic, they are all propaganda machines for the Saudi regime.--SharʿabSalam▼ (talk) 00:14, 25 January 2020 (UTC)
Does this need an RfC? I genuinely don't know, I haven't spent much time here. Agreed on the need for specificity. I think that "media produced inside Saudi Arabia, under Saudi media law" and "anything on a .sa domain" are clearly-defined categories, and would avoid having to re-RfC for each new publication. Media published abroad, with anonymous correspondents in KSA, exist and can be quite independent. Complete Saudi ownership of overseas media could in theory occur without Saudi control, but I don't know of an instance. Al-Arabiya is a Saudi-controlled domestic outlet, and in my limited experience not at all reliable on these topics; CNN Arabic I have not come across. It is based in Dubai, which might come under "beholden"; my (again limited) experience is that these are sometimes a bit better. HLHJ (talk) 05:47, 25 January 2020 (UTC)
HLHJ, yes, per my comment above. We don't deprecate without an RfC. Guy (help!) 10:09, 25 January 2020 (UTC)
Sorry, it took me a moment to understand your comment. I was thinking that general policy, deprecating non-independent sources, seems to apply here; the only problem is that it may not be immediately obvious to an editor that, say, Arab News isn't independent. You are talking about an explicit "do-not-use" rule. While one could define categories of media, one could not produce a definitive list of media outlets (new ones keep coming out) or topics.
"Subjects of interest to the Saudi government" vary. People and policies win and lose government favour rapidly and unpredictably. The Saudi government changes the URLs of many webpages frequently, so it's often hard to go back and find out what they said a few months ago. Older media articles are also often only available through the Wayback Machine.
Examples of Saudi government position changes
Obviously it has an interest in portraying the Saudi government as capable, and Saudi Arabia as a thriving country in which nearly everything is going very well (and as an appealing tourist destination). I read a headline a couple says ago which said ~"Saudi Arabia excels in human rights". However, sometimes it can be more complex. For instance, until a few years ago, Saudi Arabia supported some groups of official clerics, who controlled the information ministry and the religious police. Then they ran a media campaign against them preparatory to transferring control of the ministry and stripping the religious police of most of their powers; the media were criticizing part of the government with support of a more powerful faction. Until a few years ago the Muslim Brotherhood were officially praised and members were appointed to official roles; the media followed suit. Now they are declared a terrorist organization, and condemned in the news. Relations to Qatar; once an ally to be praised, it can now be death to support them, or, sometimes, fail to oppose them actively enough. Yemen and Canada have also suffered abrupt reversals of esteem. Women driving was opposed, then supported (with the government explicitly honouring some activists in a public-opinion campaign), then it was announced that it would be permitted and and the activists who had called for it were arrested, so that activism to win concessions from the government would not be encouraged (this was in 2018; many are still in jail). Tourism was illegal in Saudi Arabia until recently, pilgrimage tours excepted; now the government is promoting it.
I'm trying to think of topics on which the Saudi government would probably be a reliable source. Generally, I'd take them as reliable sources on themselves; indications of the positions of the government, sources for self-fulfilling statements like official appointments, and sources for what Saudi media said about X. I would not use them to establish notability, any more than I would a press release. I'm not sure what other topics they'd be reliable for. Maybe I'd take them as a source on the location of Saudi cities, for instance, unless it becomes politically advantageous to claim they are further south or whatever (but not their size; they often seem to inflate statistics). Can you think of a better example?
The thing is, to know if Saudi media are an independent source on X, you have to understand all of the relevant current positions of the government, and that requires independent sources, who probably have better coverage. Any formal rule will need to take this into account. RfC phrasing suggestions? As an informal rule; if it's Saudi-controlled, don't use it for anything except WP:SELFSOURCE. HLHJ (talk) 17:44, 26 January 2020 (UTC)
I would vote support for your RfC.--SharʿabSalam▼ (talk) 13:42, 27 January 2020 (UTC)
HLHJ, do you want to start a RfC for this? All of the Saudi newspapers are controlled by the Saudi regime either directly or indirectly since there is absolutely no freedom of speech let alone freedom of journalism in Saudi Barbaria whatsoever unlike many other "Muslim" states.--SharʿabSalam▼ (talk) 16:41, 29 January 2020 (UTC)
SharʿabSalam, I think this is part of a broader discussion. The problem is not specific to Saudi Arabia (though there are obviously specific reasons for editors to be particularly aware of Saudi COIs at the moment, which is why I posted here). The media in North Korea, for instance, are not independent sources for statements that serve the North Korean government. I am continuing this discussion at Wikipedia talk:Verifiability#Freedom of press. HLHJ (talk) 04:06, 31 January 2020 (UTC)
Suggestion: "to the extent that media are not independent of the government, they are not independent sources on topics of government interest, and they should be treated as sources self-published by the government."
This seems inline with the TASS consensus, and deals correctly with the The Official Ruritanian Press Organ example. Does anyone have any examples where this rule would produce an obvious injustice or absurdity? HLHJ (talk) 04:06, 31 January 2020 (UTC)
There is no need to "treat" these as self-published sources. A self-published source can be an excellent source. Just say that, exactly like any other non-independent source, they need to be handled with careful attention to due weight. WhatamIdoing (talk) 04:26, 19 February 2020 (UTC)

Non-English sources "preferred"Edit

Prior noticeboard discussions have repeatedly rejected efforts to strengthen the guideline: "English-language sources are preferred over non-English ones when available and of equal quality and relevance". But what exactly does "preferred" mean? Especially, I want to know if and when it is acceptable to use non-English sources when there are equally reliable English ones available, such as the following cases:

  1. The non-English source has more detail and information, but the same quality, relevance, and reliability as a less in-depth English source.
  2. The English and non-English sources are of equal quality, relevance, and reliability but present different viewpoints. If the non-English source is excluded on the basis that English is preferred, I think this would be an issue with NPOV. (We would end up promoting the views of English-speakers.)
  3. The non-English source is more convenient (i.e. open access vs. paywall, online vs. print, or a source that the editor has on hand), but both have the same information.

I hope this can be clarified because I frequently cite non-English sources. My opinion is that it should be allowed in all the above cases; arguably, a free access non-English source is easier to verify than an out-of-print English book. buidhe 02:52, 29 January 2020 (UTC)

  • I've always thought of this one as a tiebreaker. If you have two sources that both contain the same information, and are both from publishers of equal reliability, the English one wins out. But If there is information in the non-English source that's not in the English one there is no harm in including both. There's seldom harm in including two references, even when one is superfluous and not in English, except in extreme cases of excessive reference lists.[1][2][3][4][5][6][7][8][9][10][11][12] If in doubt, include both. Of course, this assumes you are fluent enough in the language of the non-English source to be sure of not having misinterpreted it. Reyk YO! 03:04, 29 January 2020 (UTC)
There are many cases where the non-English source covers detail that the English sources do not, at least in the realm of Venezuelan editing. Yes, excluding the non-English sources can result in POV. For non-English and English sources to be truly of equal quality is rare, at least in Venezuelan topics. I cite the English where possible, for the reader, but add the Venezuelan when it provides detail that is not available in English sources. SandyGeorgia (Talk) 03:12, 29 January 2020 (UTC)
Buidhe and I have been discussing sourcing in Holocaust articles. Articles are being written based on German and Polish sources found via snippet view. The English edition of the same book is available and cited elsewhere in the article. I've been asking Buidhe why she has used three language editions (and other non-English sources) and she opened this discussion. We need the Holocaust articles to reflect recent, mainstream Holocaust historiography, nothing else, and we need them to be checkable by other editors and readers. If people want to add non-English sources as well (by adding an extra source to a bundle), that's not an issue. But using them instead of English is an issue, especially for anything contentious. SarahSV (talk) 06:04, 29 January 2020 (UTC)
Yes, I absolutely agree that it should be foreign language sources in addition to English ones, not instead of English ones. Reyk YO! 06:17, 29 January 2020 (UTC)
SV, it sounds like you are saying the exact same source (same book) is available in all three languages. Then English should be used, as they are equal. Perhaps in the main description of the books in the References section, a general link to or description of the book in other languages can be used for the benefit of those native speakers. SandyGeorgia (Talk) 13:02, 29 January 2020 (UTC)
I've been discussing Holocaust sources elsewhere with Buidhe, and I don't want to introduce those issues into a general policy discussion. Sandy, I may ping you elsewhere about it, because one of the affected articles is an FA.
Generally, the problem is that this policy (WP:NOENG) has been interpreted to mean that non-English sources are fine in all circumstances, so that even when a book is available in English, other-language editions can be cited in the same article, depending on what's visible on Google snippet view. It seems it has also been interpreted (by several editors) to mean that any non-English source can be used if a particular perspective is wanted, even when it's a perspective not widely held (or held at all) in current Holocaust historiography. It has led to Holocaust articles that are in every practical sense unverifiable, with citations that are hard or impossible to decipher. In short, NOENG may as well not exist.
I've consider proposing something to clarify the policy, but I'd want to take time to find words that don't introduce other problems. SarahSV (talk) 17:33, 29 January 2020 (UTC)
SarahSV, reading the rest of this discussion, I see it is well beyond the typical Venezuela situation (where there practically are no scholarly sources, and certainly not any that are decades old, and most of the best sources are in Spanish). When there is a body of scholarly research in English—potentially decades worth—there is possibly a weight issue with newer sources postulating different positions, regardless of the language. I don't think I can offer any information from a FAC perspective, as all we can do in such cases is hope that each "side" argues their position logically and without resorting to personalization (which is what is typically seen at FAC in these cases). Perhaps the Holocaust-article tension can be diffused by using an abbreviated section to discuss newer research, according due weight; I would hesitate to say we should completely exclude non-Eng positions being advanced. (I say that based on knowing the nuance that Spanish-language sources provide on Venezuelan articles, and without having looked at these articles, as I am trying to get out the door for vacation. You may decide my position does not apply here.) SandyGeorgia (Talk) 17:20, 31 January 2020 (UTC)

In my opinion, there are several reason behind NOENG.

  • First, those people who knows what Wikipedia is never treat WP articles as reliable sources (even WP policy itself stipulates WP is reliable). Indeed, anybody can edit Wikipedia, there is absolutely no reason to believe that some concrete fact or figure is correct in the current version. That is not a big problem, however, because a reader can always read the source and verify the fact by themselves. However, we assume an ordinary reader of English Wikipedia is an Anglophone reader who does not know other languages. Therefore, our ordinary reader cannot verify what not English sources say.
  • Second, we have more or less adequate tool to estimate relative reliability of English sources. Thus, we know which scientific journals are good, which are less reliable (such as predatory journals), and which are a total garbage. We cannot say the same about non-English sources without going in details of national specifics. For example, can you say which Egyptian newspapers should be considered mainstream, and which are marginal? Which materials are editorials, and which are op-ed? Can you say which Kazakh publisher house is respectable, and which is not?
  • Third, usage of non-English sources sharply restricts the number of participants of NOR/NPOW/V discussions in relevant noticeboards, because majority of users who are commenting there do not possess needed skills to join that discussion. As a result, the risk of achieving a local consensus that may be very far from what one might expect in a discussion about similar English source becomes very high, and that by no means improves a quality of WP content.

In my opinion, that does not mean non-English sources should be completely banned. They are quite acceptable for non-controversial and/or low importance subjects. But as soon as some user expressed a legitimate concern about such a source, it should be replaced with similar English source or removed altogether. Let me give an example to demonstrate my thought. Let's take an extreme case, Arab-Israel topic. Would it be acceptable to use an Palestinian newspaper as a source about a small Arab town to describe some non-controversial facts of its history? I don't see who we cannot do it. However, imagine the Gaza Strip article written based on mostly Arab or Hebrew sources? Clearly, it would have opened a can of worms, so it is always better to use sources that (i) can be independently verified by a larger number of Wikipedians, and (ii) are less connected to this sensitive topic.

That is why, and to avoid future discussions about blanket allowance/banning of some category of sources (I frequently see this type discussion on this and related talk page), I propose to return to a discussion of my version of WP:SOURCES (see above), which stipulates that all sources, except articles in peer-reviewed journals and university books, should be treated as conditionally reliable, and their reliability should be determined on a case-by-case basis as explained in relevant guidelines and in WP:REDFLAG.--Paul Siebert (talk) 18:45, 29 January 2020 (UTC)

Paul Siebert, Thanks for your comment. I would support your WP:SOURCES proposal, although I disagree that we know which scientific journals are good, which are less reliable (such as predatory journals), and which are total garbage. Regardless of what language it's in, individual assessment often has to be done in order to find that out. Foreign language journals are not necessarily less reliable than English ones. Furthermore, I worry that any effort to reduce the use of non-English sources would have the effect of worsening systemic bias on Wikipedia - already a serious problem. If there's a tradeoff between comprehensiveness or bias versus pitfalls of non-English sources, I'm inclined to favor the former.
I agree that in cases of ethnic conflict, writers from non-involved countries are often more reliable. (Nevertheless, in the case above, I would argue for including sources by both Israelis and Palestinians to achieve NPOV, because Anglo perspective is not neutral by itself). However, for the same reason you wouldn't want to write the Mexican-American war based only on US sources. buidhe 20:11, 29 January 2020 (UTC)
No, there are some formal criteria that allow rough estimation of journal's reliability. These criteria include, (i) impact factor (the higher, the better, although a comparison should be done only between journals about the same topic) (ii) journal's presence is some databases (Thompson-Reuter, Jstor, google scholar, Scopus, etc). For non-English journals, it is much harder to obtain similar information, and it is much less reliable.--Paul Siebert (talk) 20:09, 30 January 2020 (UTC)
[citation needed]? I haven't found that to be the case. Many foreign language sources are indexed on Google Scholar, jstor, CEEOL and similar databases. You can easily count holdings in university libraries, and check who is the publisher and author. buidhe 20:27, 30 January 2020 (UTC)
Inclusion of foreign sources is fragmentary and incomplete. And it is hard to estimate its reliability. Usually, when I find some English source, I can check how many times it was cited, and in which context (support, criticism, etc), and everybody who disagrees with my opinion can check that independently. That does not work with foreign sources: even if I can read, for example, Polish or Ukrainian, my opponents probably cannot, so they should trust to my analysis.--Paul Siebert (talk) 20:54, 30 January 2020 (UTC)
  • I think current policy explains it well enough. "English-language sources are preferred over non-English ones when available and of equal quality and relevance". So, (1) under no circumstances the non-English RS should be forbidden just because they are non-English (they may be less preferred, but they can be used), (2) the non-English RS should be used in a number of cases, for example, if they are freely available, provide additional details, or simply serve to corroborate claims made in other sources, English or not. This policy (essentially WP:RS) is fully applicable to Holocaust or coverage of any ethnic conflicts. My very best wishes (talk) 22:56, 29 January 2020 (UTC)
  • Concur with Mvbw. Regarding the OP's three specific questions: this seems all too evident for words,
    1. If the non-English source has more detail,
      • Either, some of that detail is germane as content for Wikipedia's article, in which case, for referencing that detail, the English and non-English sources are *not* of "equal quality and relevance", and thus, for that detail, the non-English source should be used.
      • Or, what often will be the case, because a tertiary source like Wikipedia is rather "summary" than "full detail", the detail is interesting but not of a level to make it to article content. In that case it can be a good idea to add the foreign-language document to a "Further reading" section.
    2. If the English and non-English source have a different viewpoint, then they are, of course, not interchangeable, per WP:NPOV: they are each relevant to their viewpoint, which makes that the relevance of both sources is at least different, not equal. Anyhow, WP:V can not be used to supersede WP:NPOV – if viewpoints are different we can't discriminate against one source because it wasn't written in English: a normal implementation of WP:NPOV applies, and the viewpoints need to be weighed according to that core content policy.
    3. If the foreign-language source is more accessible, it would probably be best to use both sources for broad verifiability.
This also hardly touches upon actual difficulties that might arise in practice: how do a group of editors working on an article compare a source they maybe don't really understand with one they do understand (e.g. in order to determine whether a viewpoint has enough weight to be treated on equal footing or not)? But there again, the policy is clear enough that quotations can be asked for, which, of course, might need to be translated to make a joint assessment possible.
Re. SV's "... In short, NOENG may as well not exist. ... I've consider proposing something to clarify the policy ..." I've yet to come across a case where a rather straightforward policy which is not followed very well, is better followed once its wording is expanded by clarifications etc. Instead, I'd ask straightforward implementation of the policy (quotes, translation of these quotes, etc), and insist on the point until policy is followed. --Francis Schonken (talk) 09:39, 30 January 2020 (UTC)
Francis, I agree about insisting on quotations, and that's something I will do going forward. But it doesn't solve the whole problem. Is the source part of mainstream Holocaust studies; is it up to date; is the editor who added it fluent enough in the language to understand scholarly sources so that they will spot something else in the article that may contradict the point; is it being cited correctly to allow other editors to find it; can it be accessed at all, even via inter-library loan?
What we're supposed to do on Wikipedia is present readers with mainstream, up-to-date Holocaust history. Issues that the international scholarly community wants to discuss are invariably translated into English, although it may take a few years. So when an editor goes off-piste with unusual sources, or sources that no one on enwiki can evaluate, it can be a red flag. We also have to think about the long-term health of the article. Who in future is going to maintain it if no one can find the sources? SarahSV (talk) 18:09, 30 January 2020 (UTC)
Agreed. If we apply the same criteria to non-English sources, that we must conclude that WP:SOURCES is equally applicable to them. That means "mainstream newspapers" or "books published by respected publisher houses" are RS. However, let's take such an author as Volodymyr Viatrovych as an example. He published a number of monographs in respectable Ukrainian publisher houses, including the one of the oldest Ukrainial university. He is considered a mainstream author in Ukraine by a part of Ukrainian society. And one of his major ideas, which he is vehemently advocating, is that UPA was not responsible for the Holocaust, and they had never been engaged in mass killings of Jews. Can we add this information into the Holocaust related articles as a statement of fact (or even as an attributed opinion)? Currently, the policy makes no difference between, e.g. Oxford Univ. press and Києво-Могилян. акад: both publishers are affiliated with their universities, and these universities are among the oldest (and the most respectable) in their countries. That means, formally speaking, the books published by Viatrovich are equally reliable as the books of his Anglophone critics (there is no supporters, btw). Are we ready to include the opinion that is de facto a denial of important aspects of the Holocaust into the article about the Holocaust? If we treat non-English sources as Francis propose, we should.
Second. Viatrovich is being criticized not only abroad, but in Ukraine too. How can we decide if his views are mainstream in Ukraine? Actually we cannot do that just based on google translation of quotes, we must possess some knowledge about a broader context. In English speaking worlds, we all implicitly have that knowledge (thus, everybody knows who David Irving is). In foreign cultural media, a situation is different: we don't know which national authors are broadly considered marginal, which are broadly recognized as mainstream, and if some user who speaks, e.g. Ukrainian, provided a handful of sources, we even cannot say if these sources are a representative sample, and not a cherry-picked set of mutually supporting marginal writings. That is a reason why we should not have any blanket allowance/bans of some categories of sources, and everything should be analyzed in context, as I proposed above.--Paul Siebert (talk) 20:03, 30 January 2020 (UTC)
Re. "... If we treat non-English sources as Francis propose, we should." – nonsense, don't put words in my mouth: that is neither supported by policy, nor by WP:V, nor by WP:NPOV, nor what I said, nor directly, nor indirectly. So, if this is the level you're trying to have a discussion, please go elsewhere.
Re. "He is considered a mainstream author in Ukraine by a part of Ukrainian society." – I don't believe you, you just posit that, as if it is TRUTH, no, that's not how it works in Wikipedia.
Re. "How can we decide if his views are mainstream in Ukraine?" – QED, you're just saying things. Either he is, and then you need references to reliable sources to support that (not vaguely "he published..." without a single reference), either he isn't, and then don't start on him in a policy discussion. --Francis Schonken (talk) 20:17, 30 January 2020 (UTC)
(ec) The policy says that English-language sources are "preferred". I'm not sure why people want to ignore that word. Paul is exactly right about the problems ignoring it leads to. In Holocaust studies, there has been a shift from viewing it as something imposed only by Germany on the rest of Europe, and a move to seeing it as a "series of holocausts" carried out by Germans and groups other than Germans. Allowing non-English-language sources to be of equal validity takes us down the road of inadvertently using perpetrator/collaborator sources as mainstream sources, based on texts most readers and editors can't read or even find. The international scholarly community of Holocaust historians publishes in English; even if they first publish in another language, anything that matters is likely to make its way into English. By ignoring WP:NOENG, we are saying we know better. SarahSV (talk) 20:27, 30 January 2020 (UTC)
(edit conflict) OUP is a respected publisher not just in the UK, but in the entire world, including non English speaking countries. I doubt that the Ukrainian publisher has anything like that level of international recognition, which has to be considered alongside national reputation. Furthermore, for such an extreme claim more than one source is certainly required. But there are other ways to assess quality and mainstream-ness of foreign language sources. For example, journals published by the national academy in their country can be assumed to represent mainstream opinions in their countries. I agree with case by case assessment but that should be applied fairly to sources but this cannot be used to selectively exclude non-Anglo perspectives that certain editors just don't like. buidhe
No one is excluding non-Anglo perspectives. The argument is that Holocaust sources that scholars want to read are translated and published in English. It's like MEDRS. We have to wait for the review articles. In Holocaust studies, it's better to wait for the English-language editions. German is perhaps an exception for obvious reasons. SarahSV (talk) 20:36, 30 January 2020 (UTC)
(edit conflict)Re 1. That is how I understood your words starting with "If the English and non-English source have a different viewpoint,... etc" You de facto propose to abstain from estimating relative reliability of these sources, so, if I understand you correct, according to you, problem of that kind belongs to the realm of NPOV, not V. That is exactly what I say.
Re 2. You are just confirming my point: neither you nor I live in Ukraine, and you cannot trust me, because you have no reason to believe I have sufficient knowledge for making such a statement. However, I also can ignore your criticism, because I have serious reasons to suspect you are also not knowledgeable in realities of modern Ukraine. If that dispute were about some English writer, we could easily resolve it by using objective criteria. However, we cannot do that in that case, which is a demonstration of my thesis that usage of non-English sources creates additional risks for Wikipedia reliability.
Re "not vaguely "he published..." without a single reference" Usually, my posts are redundantly long, so I am trying not to include non-essential information there. I provided a link to the article about that author, all needed references are there.--Paul Siebert (talk) 20:41, 30 January 2020 (UTC)
Yes, but that doesn not follow from WP:SOURCES letter. That is why it should be fixed.--Paul Siebert (talk) 20:44, 30 January 2020 (UTC)

BreakEdit

Re. "You de facto propose to abstain from estimating relative reliability of these sources" – no, I don't: please go elsewhere if that is the level you want to have this discussion: I'm no game for this kind of nonsense. --Francis Schonken (talk) 20:52, 30 January 2020 (UTC)

In my opinion, that is an absolutely normal discussion: I am just explaining how I understood your words. If you disagree with my interpretation (and you have a right to do so, for you know better what did you propose in reality), then the next step would be to explain my mistake to me. That would be a normal continuation of the discussion. In contrast, your proposal "to go elsewhere" is by no means normal. --Paul Siebert (talk) 21:01, 30 January 2020 (UTC)
For clarity, my reply was regarding the OP's second case (as I clearly stated), which starts "The English and non-English sources are of equal quality, relevance, and reliability but present different viewpoints ..." – if the "equal reliability" is part of the premise, then my answer was correct. That's why I said that all three of the cases presented in the OP seem "all too evident for words", and that they "also hardly touch[es] upon actual difficulties that might arise in practice" – so don't distort my words which is an unhelpful way leading to non-discussion. --Francis Schonken (talk) 21:13, 30 January 2020 (UTC)
@Francis Schonken: Well, I am not sure I distorted your words, I just put an emphasis differently. The roots of the problem is exactly in the words shown in bold: "are of equal quality". Your write ""equal reliability" is part of the premise", and that is a problem, because the word "equal" in your phrase implies some non-binary criterion of reliability. However, WP:SOURCES sets a binary criterion ("reliable - non reliable"), so, following WP:SOURCES's letter, an English book published by OUP is of an equal quality as a Ukrainian book published by Києво-Могилян. акад., which immediately moves the dispute about a conflict between these two sources to the NPOV area. Meanwhile, WP:V could do its job better by explicitly explaining that "reliability" is a non-binary term that must be understood in a proper context: some sources are more reliable that others, and some reliable sources are acceptable in some context and unacceptable in another. --Paul Siebert (talk) 21:27, 30 January 2020 (UTC)
Where in WP:SOURCES does it say that Києво-Могилян. акад. and OUP are of equal quality? I did a find in page for "quality" and it doesn't seem to say anything of the sort. buidhe 21:34, 30 January 2020 (UTC)
"Other reliable sources include:
  • University-level textbooks
  • Books published by respected publishing houses
That is a binary criterion that allows no nuances: both Oxford University and Kиєво-Могиляньска академiя are their nation's top universities, and, per out policy's letter, the books published by them are equally reliable sources.
Of course, by saying that, I am acting as devil's advocate. However, that argumentation can be used by some POV pushers, and, formally speaking, they would be right.--Paul Siebert (talk) 21:57, 30 January 2020 (UTC)
Books by Oxford University Press are better not because they are published in English. My very best wishes (talk) 22:59, 30 January 2020 (UTC)
@Paul Siebert: Please go elsewhere, you're still distorting my words. Hence, your contributions to this discussion are still very unhelpful. I don't read you apologising for your error, instead you're doubling down on self-righteousness without recognising an error. So, please go elsewhere. If that's your attitude in discussions regarding sensitive topics like holocaust etc., I think it would be better you don't edit in such sensitive areas. --Francis Schonken (talk) 07:04, 31 January 2020 (UTC)
@Francis Schonken: I could not distort your words in my last post, because I just quoted them and supplemented with my commentaries. If you disagree with my comments, feel free to explain me what you disagree with, but keep in mind that that is my opinion about your words. You may agree with that or disagree, and this my opinion may be right or wrong, but I see no reason why should I apologize for expressing my opinion. With regards to my apologies, I think by saying that that was just my understanding of your words, I made a good faith attempt to resolve a situation. In connection to that, your responses are redundantly aggressive, which is not a problem per se. A real problem is that you are not trying to explain me what is wrong with my interpretation of your words. You say "you misinterpreted me, apologize", but you are not giving me even a single clue on what that misinterpretation consisted in. Let's assume we both are honest Bayesians, who, according to Aumann agreement theorem cannot agree to disagree if they are discussing something that belongs to the domain of their common knowledge. In other words, if you explain me properly what was wrong with my interpretation of your words, the problem will disappear (maybe in few iterations).--Paul Siebert (talk) 16:32, 31 January 2020 (UTC)
Sorry, you distorted my words. And again. Please go away, you bring no useful insights to this discussion, as the first step in a useful discussion is to read attentively what someone else says, and not distort their words. --Francis Schonken (talk) 16:38, 31 January 2020 (UTC)
Well, maybe problem is not on my side? Maybe it is not me who read not attentively? However, since I respect my opponents, I am explaining my 21:27, 30 January 2020 post in different words. Yes, assuming that English and not English sources are of equal quality, your words are almost correct. I say "almost" because there is one small technical objection, which I explain elsewhere. However, although you are right, the problem is that WP:V does not explain the term "source quality": the word "quality" is found only twice in the policy text, one time in NOENG, another time in REDFLAG. That means we are de facto advised to operate with the criterion that is not defined in a policy, so everybody interprets "quality" based on their own criteria. That is a root of many conflicts, which are a central part of many arbitration cases. Moreover, it is that omission which forced ArbCom to make a decision that only peer-reviewed English sources are allowed in the Holocaust in Poland area. Had WP:V explained the term "quality" in even very general form (something similar to what I propose), many of those conflicts could be possible to avoid.--Paul Siebert (talk) 17:05, 31 January 2020 (UTC)
Please go away, you misrepresented my words, and continue to do so. Unless you have the intention to change your approach (for clarity: nothing of that is apparent in what you write)? Then please say so, and strike all your misrepresentations of my words. Thanks. Again, based on misrepresentations of what someone else says, and that is what you have been doing thus far, it seems nothing good will come from this discussion. --Francis Schonken (talk) 17:19, 31 January 2020 (UTC)
How can you speak about misinterpretation of your words in my recent posts when I provided no interpretation at all? I just said that your words could be correct had the term "quality of sources" been explained in the policy. However, I found no such explanation in WP:V. If you believe this term is explained in the policy, please demonstrate that. What you are saying is not the focus of my recent posts.--Paul Siebert (talk) 17:38, 31 January 2020 (UTC)
Re. "... you are right, the problem is that WP:V does not explain the term "source quality"" – I said no such thing, so you're still misrepresenting my words.
Re. "... I provided no interpretation at all" – even if that were correct (which it isn't), how would you parroting my words be in any way helpful in this discussion? --Francis Schonken (talk) 10:04, 7 February 2020 (UTC)
A full quote is "However, although you are right, the problem is that WP:V does not explain the term "source quality"". English is not my mother tongue, so a possibility cannot be ruled out that this sentence is not completely grammatically correct. Nevertheless, the words "although you are right, lorem ipsum dolor" usually mean: "what you say is correct, but lorem ipsum dolor". In other words I never said, and never wanted to say that the words "the problem is that WP:V does not explain the term "source quality"" belong to you: that was my own thought, and I never ascribed it to you. Therefore, it looks like you owe me apology for misinterpreting my words. However, I am not insisting on that because it could be quite possible that I was just not clear enough. In future, I will try to be more careful in citing your thought.
Can we consider the incident resolved now?--Paul Siebert (talk) 15:13, 7 February 2020 (UTC)
Re. "Can we consider the incident resolved now?" – no, you didn't retract any of your misrepresentations, and you're still trying to hijack the discussion: "please go away" is still my advice to you. It is extremely difficult to have a sensible discussion with the excess of distraction you throw into it. So, please go away. --Francis Schonken (talk) 21:03, 7 February 2020 (UTC)

The thing is that I have quite some experience with articles using multi-language sources: that experience may (or may not) be helpful to address the issues raised here. Problem is that talking about such experience is futile when all of it is translated, by a single contributor to the discussion, in a single-purpose vision that excludes all other approaches outside the prejudiced one. For now, I'd just oppose rewriting the policy to suit the needs of a limited set of articles, if that would mess up sound verifiability in other sets of articles. Maybe in Holocaust-related articles English-language sources may, in a practical way, generally be superior to those in other languages (except German). I can't judge that. In the area in which I'm most active, foreign-language sources are often more reliable than English-language derivatives. So, I'd reject any attempt to have, at policy level, a principle inscribed that language, in general, is an indicator of the quality of a source. If anything needs to be said about language and reliability of a source it should be (1) nuanced, and (2) in subsidiary guidance, not in policy-level guidance. --Francis Schonken (talk) 10:04, 7 February 2020 (UTC)

I totally agree with the first part, although the conclusion should be somewhat different. It would be incorrect and arrogant to claim non-English sources are of poorer quality. However, as you correctly pointed out, much smaller number of editors are capable of making a judgement about non-English sources. That means, we have much less opportunity for proper quality control. Therefore, it would be correct to say in the policy, that, "whereas the language, in general, is by no means an indicator of the quality of a source, usage of non-English sources is more risky, because standard tools that are used for quality control of English sources are less applicable to non-English ones, and a smaller number of users are capable to make authoritative judgement about such sources during discussion in relevant noticeboards."--Paul Siebert (talk) 15:20, 7 February 2020 (UTC)

Wikipedia:Identifying reliable sources (history)Edit

The page linked above is currently an essay, but by looking at it, I feel that it may be an option to discuss as part of this thread. Would there be an interest in exploring if it's feasible for WP:HISTRS being adopted in the Holocaust topic area as a guideline? That would require an RFC of course, but want to float the idea here, to begin with. --K.e.coffman (talk) 02:32, 31 January 2020 (UTC)

@K.e.coffman: The essay looks very reasonable, why cannot it be upgraded to the guidelines level? Does anybody know what is a procedure to convert an essay to guidelines? @SlimVirgin:, can you please advise us how can we do that?--Paul Siebert (talk) 16:36, 31 January 2020 (UTC)
K.e.coffman and Paul, I think this is a good idea in theory, but that page would need a lot of work. Parts of it don't really mean anything. I'd suggest first fixing it as an essay, then writing a section on the Holocaust, and only after that consider what it would need to get it to guideline status. SarahSV (talk) 17:17, 31 January 2020 (UTC)
Let's try.--Paul Siebert (talk) 17:38, 31 January 2020 (UTC)
User:Paul Siebert, that page is an essay because it's impossible. The objections raised when it was written have been largely ignored. I see, for example, that the critique I posted in 2013 still applies. According to that page's definition, things like pop culture awards in 2018 are "history" and only scholarly sources can be used. And it says that the minute a historian starts writing a book about any current politician, their entire BLP articles are now "history articles" and every non-scholarly, non-historian source needs to be removed. It is proof that WP:Policy writing is hard. You can find some good advice there on how to identify the best possible sources, but you need a lot of experience with Wikipedia and a good deal of common sense to avoid getting in trouble. WhatamIdoing (talk) 02:45, 18 February 2020 (UTC)

Elaborated exampleEdit

Not so long ago I wrote the BWV Anh. 167 article. Here are some sources that write about that composition:

  1. Mus.ms. Bach P 659 at Berlin State Library website. (that is the primary source discussed in all quotes below)
  2. Spitta, Philipp (1899). Johann Sebastian Bach: His Work and Influence on the Music of Germany, 1685–1750. III. Translated by Bell, Clara; Fuller Maitland, John Alexander. Novello & Co. p. 28. Containing:

    ..., of which Bach ... wrote out the whole of the first twelve pages ...

  3. "D-B Mus.ms. Bach P 659". Bach Digital. Leipzig: Bach Archive. 2020-01-14. Containing:

    Scribe (in detail): H. W. Ludewig (...); from p. 13 (middle): J. S. Bach

  4. The so-called "English" version (click "EN" top right of page) of RISM No. 467065900 (Mus.ms. Bach P 659). Containing:

    Notes: ... Schreiber: bis p. 13 (Mitte): H. W. Ludewig (...); der Rest ist von J. S. Bachs Hand (mittel/spät).

  5. (German original of #2 above) Spitta, Philipp (1880). Johann Sebastian Bach (in German). II. Leipzig: Breitkopf & Härtel. p. 509. Containing:

    ..., welche Bach ausschließlich der ersten zwölf seiten, ..., ... abgeschrieben hat.

  6. Dörffel, Alfred (1894). "Vorwort" [Preface]. Kirchenmusikwerke: Ergänzungsband [Church music: Complementary volume]. Bach-Gesellschaft Ausgabe (in German). 41. Bach Gesellschaft. Breitkopf & Härtel. pp. XXXIX–XL. Containing:

    ... 1–13 paginirt, so weit der Copist geschrieben hat; von Seite 13 der zehnten Linie an, also noch 14 Linien von der Seite bis zum Ende hat Bach Alles selbst geschrieben: Seite 13 die grössere untere Hälfte, Seite 14–39.

  7. D-B Mus.ms. Bach P 659 at Bach Digital website (German version of #3 above):

    Schreiber, detailliert: H. W. Ludewig (...); ab S. 13 (Mitte): J. S. Bach

  8. Plain German version of #4 above, RISM No. 467065900 (Mus.ms. Bach P 659), containing:

    Bemerkungen: ... Schreiber: bis p. 13 (Mitte): H. W. Ludewig (...); der Rest ist von J. S. Bachs Hand (mittel/spät).

  9. (My) literal translation of #4 = #8 above:

    Notes: ... Scribe: up to p. 13 (middle): H. W. Ludewig (...); the rest is in J. S. Bach's (middle to late) handwriting.

  10. (My) literal translation of #5 (highlighting difference with published translation #2):

    ..., which Bach copied, apart from the first twelve pages, ...

  11. (My) literal translation of #6:

    ... pages numbered 1 to 13, as far as written by the copyist; Bach has written everything from the tenth line of page 13, which is another 14 lines on that page, until the end: [that is] the larger lower half of page 13, [and] pages 14 to 39.

If only plain English published sources would be used, that is #2 and #3, then 50% of the sources write that Bach wrote the first 12 pages of the manuscript and 50% of the sources say he wrote nothing before the middle of the 13th page, in terms of the WP:BALANCE policy. Then one would have to write, in the BWV Anh. 167 article, something like:

According to the Bach Digital website, Bach wrote everything after the middle of the 13th page of the P 659 manuscript, while according to Philipp Spitta Bach only wrote the first 12 pages of that manuscript.

Instead, the BWV Anh. 167 article has (Kyrie–Gloria Mass for double choir, BWV Anh. 167#Bach's manuscript):

The first part of that manuscript was written by one of Bach's scribes, ..., while Bach himself completed the handwritten score ...

The reason is clear: according to WP:RSUE, other-language sources can only be suppressed if the English-language source (or: sources) is/are "... of equal quality and relevance". Since that is not the case, while one of the English translations does not match its German original, no source can be missed. Then the picture becomes completely different: the Bell & Fuller Maitland translation becomes, on this point, a negligible minority vision (in WP:WEIGHT terms), in fact a simple error – the overwhelming majority of relevant reliable sources say it differently. The primary source (#1) is available, and if one has seen a few Bach manuscripts even that confirms, without that being useable directly in Wikipedia, that all secondary sources except the Bell & Fuller Maitland translation have it right. That is saying nothing about the general quality of that translation: anyone can make an error.

So, no for the wording proposed above, containing "... usage of non-English sources is more risky ..." (etc): in the given example using the published English translation is "more risky" – no such dismissive thing can be said about foreign-language sources at policy level. There is no "quick and dirty" way to get rid (or even subtly undermine the value) of foreign-language sources in this way, and certainly not at policy level. --Francis Schonken (talk) 17:42, 7 February 2020 (UTC)

  • I think that Francis has given a great example of where applying this guideline makes no sense. Although I'm not sure if, in this context, very many editors would insist on applying the letter of the rule. buidhe 18:07, 7 February 2020 (UTC)
You clearly misunderstood: I followed policy (& guidelines). Why would you say I didn't? I followed them to the letter, I followed them generally, and according to current detail. So please, make clear in what sense I wouldn't have followed policy and guidelines in general and to the letter, instead of offering vague accusations, not supported by facts. --Francis Schonken (talk) 19:09, 7 February 2020 (UTC)
Francis, I doubt you followed policy's letter, for the policy does not define the terms "quality" and "relevance". Although your evidences are quite adequate (from the common sense point of view), non-English sources cited by you are not better that the sources #2&3, according to WP:V's letter. Therefore, formally, there was no need to use them.--Paul Siebert (talk) 19:32, 7 February 2020 (UTC)
What a nonsense. Please go away. --Francis Schonken (talk) 20:38, 7 February 2020 (UTC)
@ Buidhe: I still wanted to add that I find the WP:V policy (& related policies, & subsidiary guidelines) quite enough to handle these issues (i.e. those of the elaborate example): for the above example the guidance fits as a glove, i.e. no unnecessary detail or attempt at micromanagement, clear principles that can be applied in a wide variety of circumstances, enough support to address the case. --Francis Schonken (talk) 20:38, 7 February 2020 (UTC)
Francis, thank you for writing out that example. Just one point: I don't think anyone has suggested that non-English-language sources are inherently problematic. The difficulty in Holocaust articles is that non-English sources are cited where English sources exist and are at least as good or better. The Polish edition of a book is cited when there is a more recent English edition. Sources are cited in a way that suggests the editor citing them may never have seen them. One source is cited when in fact the text came from another source. And sources are cited that do not exist as described. So it is a very big problem, and trying to find and understand those sources has absorbed significant volunteer time.
The current situation means that those articles, for all practical purposes, are not verifiable, often on sensitive issues. Therefore, it would be good to find a solution. Your suggestion above, that we get stricter about enforcing the part of WP:NONENG that requests quotations, will help. There is a template for that—{{Request quotation}}—that should probably be used more often. SarahSV (talk) 18:14, 7 February 2020 (UTC)
Re. "I don't think anyone has suggested that non-English-language sources are inherently problematic" – incorrect, someone suggested the wording "... usage of non-English sources is more risky ..." above ("more risky" is of course one form of "inherently problematic").
Re. "The difficulty in Holocaust articles is that non-English sources are cited where English sources exist and are at least as good or better. The [other-language] edition of a book is cited when there is a more recent English edition." – well, that's exactly the way things stand with the (original) Spitta 1880 source (#5 in the elaborate example) and its "more recent English" translation (1899, #2 in the elaborate example). So, in general, I don't think it is correct to say that such issues are only to be resolved in Holocaust-related articles: they are encountered in other topic areas too. For clarity, in general, the Bell & Fuller Maitland translation is considered superior to Spitta's original: the translators consulted Spitta on some inconsistencies in his original, leading, generally (apart from the occasional mistake, of which I highlighted one), to a translation that was superior to the original first edition (leading to an errata list in subsequent editions of Spitta's German version).
That was just a general remark, I mean, I do think there are specificities, but, in general, comparable issues can be encountered accross quite different topic areas. Similarly, I recognise what you write about "Sources are cited in a way that suggests the editor citing them may never have seen them. One source is cited when in fact the text came from another source. And sources are cited that do not exist as described. So it is a very big problem, and trying to find and understand those sources has absorbed significant volunteer time." etc. As it happens, I could give examples of any of these things you describe, and, if it is any consolation to you, I recently turned down a GAC of a classical composition article for *exactly* these problems (after having invested considerable time trying to disentangle the cause of the issues that article exhibited). --Francis Schonken (talk) 21:39, 7 February 2020 (UTC)
(edit conflict)Actually, Francis Schonken's example fully demonstrated my point: if a user sustains a burden of proof, and it is clear from the evidences presented that the non-English source has a high quality, then that source can and should be used. Obviously, it should take more efforts to prove that non-English source is good (just because other users are not familiar with some national specifics, and more explanations are needed), but that should not prevent us from usage of that source.
However, what I cannot understand is the following: what is a reason for discussing non-English sources separately? If we explicitly introduce and explain the term "quality of source" (which is not defined in the policy yet), and stipulate that the burden of proof is always on those who adds a source, the problem with non-English as well as with many other types of sources will be resolved automatically.--Paul Siebert (talk) 18:23, 7 February 2020 (UTC)
SarahSV, what you are saying is correct, but it is hardly relevant to the NOENG issue. If a snippet view is presented as a source (English or non-English, doesn't matter) then there is a risk that the statement was taken out of context. If some source is cited because it is easily available, than may be an indication of a not sufficiently serious approach of a user to editing. The Holocaust article has 20,000 readers daily, and we should do our best to use the best quality sources for it. However how can we expect editors to use best quality sources if the very category "quality of sources" is not defined in the policy?--Paul Siebert (talk) 18:30, 7 February 2020 (UTC)
It's worth nothing that I'm not discussing the Holocaust article. I'm discussing lots of other articles about the Holocaust and related issues. I haven't given examples because this is a policy discussion, and it wouldn't be appropriate to highlight particular edits. SarahSV (talk) 18:35, 7 February 2020 (UTC)
That hardly change anything. I gave the statistics for the Holocaust article just to demonstrate that the topic is important, and "cheep editing" using snippet views, or advocating some sources just because they are more easily available is hardly a good style. My major point was different: do you agree that by defining the concept "reliable source's quality" we may resolve a large number of problems, including the problems related to the Holocaust topic?--Paul Siebert (talk) 18:57, 7 February 2020 (UTC)

Regarding "risky", Francis Schonken, I was probably not clear enough. Under "risky" I meant the following: imagine a similar situation with, e.g. Sibelius, when one Finnish user is advocating some edit based on a Finnish source about that composer. That user provides evidences of the same kind as you did above. All other user feel uncomfortable, because the proposed edits contradict to what English sources say. It may be quite possible that Finnish sources are more reliable about a Finnish composer, but a possibility cannot be ruled out that that Finnish user just incorrectly uses them, or even took some sources that are considered fringe by Finns themselves. Can we check that independently? theoretically, yes, but in reality it is very hard to do. I believe now you understand what I wanted to say: as soon as it has been demonstrated that some foreign source is a high quality source, it is quite ok to use it, but the procedure that determine its quality should be somewhat more strict (more evidences should be presented). By the way, you provided a good example of how should it be done.--Paul Siebert (talk) 19:11, 7 February 2020 (UTC)

Elaborated example II (relating to accessibility and non-accesibility of sources)Edit

When I started to edit the Pietro Torri article somewhat over a month ago it contained the grand total of one single source ([1]):

Groote, Inga Mai (2003). Pietro Torri, un musicista veronese alla corte di Baviera. Verona: Della Scala. p. 120. ISBN 8885099734.

One of several problems with this source is what has been written above: "... sources are cited that do not exist as described": according to Google Books the publication only has 118 pages ([2]), while the reference, as found in the Wikipedia article, cites "p. 120" of the edition..., so also "sources are cited in a way that suggests the editor citing them may never have seen them" seems to apply. As a whole, the backbone of the English Wikipedia article on Torri seems to be a (partial) translation of it:Pietro Torri by someone who never saw another source for the article than a Wikipedia article in another language (which may, or may not, have had far laxer rules on sourcing than current English Wikipedia standards).

So, at that point, I could have pasted some tags in the article, like {{Refimprove}} etc, and move on, leaving it to someone else to clear out the apparent mess (like I had done here). I tried to do better for the article on Torri (... which was going to "absorb[ed] significant volunteer time", using the words written above). One of the first things I did was to update and expand the content of the {{cite book}} template used for the Groote reference:

Groote, Inga Mai (2003). Pietro Torri, un musicista veronese alla corte di Baviera. Sette note (in Italian). I. Broz, Barbara (Appendix: "I musicisti veneti in Europa ai tempi del Torri"). Verona: Della Scala. ISBN 8885099734. OCLC 681975493.

But that was still dancing around the main issue regarding this source: it is virtually inaccessible, i.e. probably only found in specialist libraries (not even a summary or the shortest of quotes to be found on the internet), and in Italian, which I don't understand very much. I do think it is not a crucial distinction whether the source is inaccessible for language or for difficulty to get hold of a copy: for practical purposes I could not use the Groote source in a WP:V logic when I wanted to improve & expand the Torri article. So I "demoted" the Groote publication from reference to "Further reading" (Pietro Torri#Further reading). There was not a single footnote referring to this publication in the article, so, as such, my decision to not use this publication as reference (while still keeping it available if someone who has access to it in the future can still pick it up and rejoin it with the sources used as reference) did not deteriorate the (at that point virtually non-existing) sourcing of the article. This is not an appreciation of the reliability of the Groote source: without having access to it a reliability discernement is not really possible (but more about that later). I found 14 useable sources (Pietro Torri#Sources): one in Latin [sic], one in French, one in Italian, five in German, and the remaining six at least partially in English. And another inaccessible source, this one in German, which I added to the "Further reading" section. Plus several dozens of sources only mentioned in the (currently) 270 footnotes (Pietro Torri#References). And, checking these sources, I saw several of them using the Groote publication as source (so, yes, that kind of confirms that the Groote publication would normally be a "reliable source" in Wikipedia's WP:V logic).

Again, I had no problem applying core content policies such as WP:V, and their subsidiary guidance such as WP:RS, in the form they currently are – but none of these can, of course, be applied without using sound jugdgement (Wikipedia is not written by trained monkeys using guidance that can be interpreted without exerting sound judgement): whether my judgement regarding the content of the Torri article was sound enough is of course open to further appreciation & improvement of the article (that's a basic premise of any wiki system). The point I'm trying to illustrate is that despite problems described above as "Sources are cited in a way that suggests the editor citing them may never have seen them. ... And sources are cited that do not exist as described. ... trying to find and understand those sources has absorbed significant volunteer time.", that despite problems of this nature, WP:V (and related guidance) seems to work fine when trying to address the difficulties. --Francis Schonken (talk) 08:08, 8 February 2020 (UTC)

Actually all this discussion is essentially senseless taking into account this. I haven't noticed that you de facto supported my core idea. In that context, your extremely rude tone is quite forgivable. Good job, and let's hope this change will stay.
Several comments to your above post.
I still think that Verifiability's core idea is that everybody should have an opportunity, at least, theoretically, to find the cited source and to see that Wikipedia correctly transmits what it says. In connection to that, if a source is really hard to find, that is a problem. Although that does not mean it is bad, but, since just a few people may have access to that source and make an independent judgement, the risk is greater that that judgement may be incorrect, and it is virtually impossible to check. In that sense, yes, non-English sources are more problematic, although that does not automatically imply their poorer quality.
When we discuss quality, there are two aspects: a real quality, and a possibility to assess the source's quality. If the quality of some type of source is easy to check (for example, everybody knows how to check quality of scientific publication: if a journal is good and/or the article is widely cited, the article is a good quality source), the usage of this type source poses a low risk. In contrast, in the example provided by you, your rationale looks quite convincing, however, imagine a situation when you have a dispute with some POV pusher who rejects your arguments, ans says that all your sources are garbage. How many other user will be capable of making an independent conclusion on that matter during, for example, an RSN discussion? I doubt you will get more than 1-2 comments from uninvolved users. That makes a situation very shaky: your rationale (which looks convincing to me) may be rejected, and a totally wrong view of that POV pusher may prevail. That is why I think foreign sources, for which no obvious procedure of quality assessment exists, may be very risky to use.
In connection to that, it may be reasonable to think about guidelines that explain how non-English sources should be used.--Paul Siebert (talk) 14:33, 8 February 2020 (UTC)
Re. "... imagine a situation when you have a dispute with some POV pusher who rejects your arguments, ans says that all your sources are garbage. How many other user will be capable of making an independent conclusion on that matter during, for example, an RSN discussion? I doubt you will get more than 1-2 comments from uninvolved users. ..." — well, I don't have to "imagine" such situation, I've encountered such situation more than once. Apart that it's not necessarily a "POV pusher" who uses the tactics: e.g. an example that comes clear to my mind was rather an editor who tried to take WP:OWNership of a set of articles. In a RSN discussion they attacked a series of standard reliable sources on the topic, only accepting more exotic, nevertheless no less reliable, sources, but "by coincidence" they had access to a specialist university library that held these more exotic sources. As it happens, as I afterwards found out, the more exotic sources said the same as the standard reliable sources, but by that time they had closed themselves (as an involved party!) the RSN section, rejecting the standard sources. Some other editors followed the RSN discussion, commented, but stood by "disinterested" when the involved editor closed the discussion in their own self-serving favour.
Well, the first thing that has to be said about such situations is that they in part touch upon "core content policies" (and their subsidiary guidance), but also, in part, upon behavioural policies and guidance ("POV pushing" and "WP:OWNership" issues are largely behavioural problems) and that, in Wikipedia context, mixed "content" and "behavioural" issues are sometimes difficult to address. The tactics used to evade scrutiny (& sanction) are very easy: when brought to a behavioural forum (like ANI), editors who exhibit this problem start to talk, in great detail, about their content achievements ("look what a valuable content contributor I am" – thus diverting attention from the behavioural problem); when, conversely, the problem is brought to a content forum, like RSN, their dialogue switches immediately to "look how good I behave", etc, again explained in endless walls of text, thus diverting attention from the content issue that needs resolving.
Nonetheless I think Wikipedia's principles which try to separate content issues from behavioural issues, each with their own set of noticeboards to handle them, is fundamentally sound – if every discussion would devolve into a mixed content/behaviour discussion things would definitely be worse. So, on this point, I don't think an update or rewrite of the WP:V policy is needed, and for updates to behavioural guidance this is not the right place (nor do I think that, as such, behavioural policies are currently in need of updating/rewriting). --Francis Schonken (talk) 16:02, 8 February 2020 (UTC)
Francis, unfortunately, RSN may be totally unhelpful, because it frequently happens that the same persons comment on RSN and on the article's talk page. When a dispute requires special knowledge, uninvolved editors frequently prefer not to participate. I can give you some recent examples if you want.
To resolve such situation by presenting it as a behavioral issue is not a universal solution. That is an "admin-specific" approach. Sometimes, active admins abstain from diving into that type issues, and suggest DR, because, in their opinion, that is just a content dispute.
In general, I noticed that many users who have obvious content disputes are trying to convert them into behavioral ones, because admins prefer to deal with the latter type problems. That is an additional indicator of problems with content policies, which do not provide us with adequate tools for resolving content disputes.--Paul Siebert (talk) 20:41, 9 February 2020 (UTC)
  • Comment Okay, first off I think it's rather sneaky of Francis to be discussing, however indirectly, the Talk:Mottainai issue here without having left any notification on the talk page. a source they maybe don't really understand with one they do understand is a pretty clear reference to this. a single contributor to the discussion, in a single-purpose vision that excludes all other approaches outside the prejudiced one is also quite obviously a reference to "version C" on the relevant page, the "single contributor" being me (even though Nishidani and Ryk72 were both involved as well). In the case in question, Francis was (still is?) trying to cite an unreliable English source that very obviously gets the relevant point wrong, apparently because it got its information from Wikipedia, based on the claim that English sources are inherently better than non-English ones, even when there is talk page consensus that the non-English ones are more reliable under the circumstances. The fact that the comments made on this page are fairly direct mirrors of the personal attacks he was recently called out over on ANI just makes it all the worse... Hijiri 88 (やや) 04:53, 17 February 2020 (UTC)
  • Side note: The "p. 120" thing is probably the number of pages, which the Italian Amazon reports as being 120 in number. It is incorrect, but it's not an unheard-of mistake. WhatamIdoing (talk) 02:51, 18 February 2020 (UTC)

Some general thoughts inspired by NOENG and the Holocaust discussionEdit

It seems I am starting to understand the core of the problem, and it seems that it is a part of a bigger problem: the proposed format of the solution is incorrect.

The core idea of WP:V, as I see it should be: "everything what Wikipedia says can be verified by a reader, who can go by a reference, read the source, and to verify if the source X really says "Y"". (Disclamer. The above description is not a summary of the current version of WP:V, it is a summary of my vision of the policy). In connection to that, the major criterion of a reliable source is its stability: it should not be, for example, a blog, which can be deleted or modified by an owner at any moment. In that sense, RS is any material that, once published, can be, potentially, available to everyone, and cannot be altered of removed from a public access. That should be a core of WP:V.

However, WP:V moves further, and it introduces some categories of sources that are considered reliable and sources that are not reliable. It implicitly introduces the term "quality of sources" (although that term is never explained), and the term "mainstream sources". I see two problems with that.

  • First, the idea that some formal category of sources is acceptable, whereas some other category is not is questionable. For example, a category "academic and peer-reviewed publications" includes such sources as Nature or American Historical Review and a vast number of predatory journals with impact factor of 0.5 or less (or even with no impact factor). Nominally, all of them are peer-reviewed, but a quality of peer-review is dramatically different. I myself authored scientific papers in journals with different impact factors, and I reviewed manuscripts for different journals, and I know that for a journal with impact factor of 30, 15, and 3, the acceptance criteria differ dramatically. I suspect journals with lower impact-factor may publish a total garbage. However, the current version of the policy makes no difference between them, and they all are RS.
Similarly, the category "Mainstream newspapers": what does this category include? A previous discussion demonstrated that user's opinia form a wide spectrum, from "any newspaper that has not been demonstrated to be fringe" to "newspapers of record". Such an ambiguity is a constant source of conflicts. Yes, we have guidelines that explain, at least partially, some details. However, guidelines are just recommendations, whereas the policy says "mainstream newspapers are allowed. Period." That de facto nullifies all what WP:NEWSORG says, because NEWSORG says that some materials in some newspapers are RS, but WP:V says "mainstrean newspapers are RS, AND "In the case of inconsistency between this policy and the WP:RS guideline, or any other guideline related to sourcing, this policy has priority. That makes everything NEWSORG says about newspapers null and void.
In contrast, so called self-published sources are considered non-reliable. However, a difference between self-published and non-self-published sources is vague: it is hard to determine if some publisher is "established", which is a core criterion that discriminates self- from non-self-published source. Meanwhile, some "self-publised" opinia of established experts (which are non-RS, per V) are more trustworthy than publications of obscure "scientists" in nominally peer-reviewed predatory journals (which are RS per V).
  • The term "quality of sources" is used twice in the policy, however, nowhere in the policy it is defined. That fact is a source of incessant contend disputes ("your source is garbage" - "no, YOUR source is garbage"), and that had already lead to an odd ArbCom decision to introduce sourcing restrictions to the Antisemitism in Poland topic. In my opinion, this decision is an attempt to compensate problems with that policy.

I propose to abandon the flawed (in my opinion) strategy, which consists in an attempt to allow/prohibit some category of sources based on purely formal criteria. I also propose to fix a problem with an obvious conflict between the policy and guidelines, which de facto cancels many good ideas described in guidelines.

Since this post was partially inspired by the Holocaust topic, let me give an example of how can the dispute between Buidhe and SarahSV be resolved within a paradigm proposed by me. To the best of my understanding, Buidhe says "Non-English sources are good, let's use them", whereas SarahSV says "Non-English sources are hardly better, and everything important about the Holocaust that deserves attention has already been translated to English, so there is no need to use non-English sources". (I apologize in advance if I summarized your view incorrectly, but the point is not to make a precise summary, but to show a difference in opinia). Meanwhile, if we abandon a strategy that is based of formal categories of sources, we can say: "Best possible sources should be used. Academic, peer-reviewed sources are preferable. For each category of sources, their quality should be determined before they are used in Wikipedia. The burden of proof that some source meets our quality standards rests with those who adds it. Our guidelines provide detailed description of the procedure of quality determination. Period".

That would resolve the NOENG issue quite easily: NOENG are quite acceptable, but, keeping in mind that it is much more difficult to determine its quality, some specific criteria must be applied to them, and these criteria should be described in relevant guidelines.--Paul Siebert (talk) 18:07, 7 February 2020 (UTC)

  • I think that you make some good points here. I would clarify my views slightly: There are plenty of garbage sources out there in any and all languages. If I cite a source, it's because I think it's reliable for the particular claim that it supports. I'm generally happy to explain why I think so to any editor who has a good-faith query. Although I agree with you that it's hard to define exactly what is a reliable source without context, I can't support this proposal. I think that rough outlines of what is likely to be a reliable source is really helpful to new editors who don't know how to determine quality of a source. Furthermore, not having much guidance at all leaves the high probability of false WP:LOCALCONSENSUS developing on some page to include/exclude sources when not merited. Having more centralized guidelines reduces that risk and also the time wasted on warring and rehashing disputes on lots of individual pages. buidhe 18:24, 7 February 2020 (UTC)
  • "everything that deserves attention has already been translated to English, so there is no need to use non-English sources". This is absolutely wrong. I am not even sure how anyone can claim it. A lot of important information has been reliably published in languages other than English and not translated to English. That's why we have WP:NOENG which tells we can use such sources. Moreover, we must use such sources to follow WP:NPOV in a number of cases. Even in terms of providing good content the cultural diversity is great. My very best wishes (talk) 19:08, 7 February 2020 (UTC)
I see Paul added "everything important about the Holocaust" [3]. Still wrong. There is no reason to blacklist any non-English sources that were published in Israel, for example. My very best wishes (talk) 19:21, 7 February 2020 (UTC)

WP:SOURCES revisitedEdit

There is no agreement to include the changes you did. Most especially on a policy page, but on any page, please do not make changes while discussion on this issue is ongoing. If I have missed a clear, agreed-upon, community-wide consensus please provide a link. A discussion is not a consensus. Littleolive oil (talk) 16:30, 8 February 2020 (UTC)

Where is the discussion ongoing?
What is your stance in that discussion? --Francis Schonken (talk) 16:38, 8 February 2020 (UTC)
  • I've reverted myself since I don't like to edit war. But I am concerned by claims of consensus when I don't see any, and by moving around content while discussion is on going.
You linked to a discussion that has been on and off for a while in which no formal consensus was agreed upon yet you claimed consensus. I don't like to see experienced editors claiming consensus when there isn't one and adjusting policy page content when a discussion, even a sporadic one, is ongoing is not great. This is a foundational policy one I and others have spent years watching and sometimes working on so that it is as clear, as it can be, and single editors should, in my opinion, such as it is, get agreement for changes. I won't discuss this further, but the page is watch listed, as it has been for years.Littleolive oil (talk) 16:55, 8 February 2020 (UTC)
I see no clear answer to my quite simple questions:
  • Where is the discussion ongoing?
  • What is your stance in that discussion?
--Francis Schonken (talk) 17:07, 8 February 2020 (UTC)
  • By way of context, here is the diff where the "mainstream newspapers" content was first added, and here is the related discussion from 2007. That history shows some excellent work in building consensus, but it doesn't seem to me that the "mainstream newspapers" language was ever subject to much specific discussion -- or indeed, that the section as a whole was ever reflective of a consensus any broader than the community of talkpage participants. I personally have no strong opinions on the matter -- but FWIW, given that the policy landscape has stabilized since 2007, I see little added benefit from addressing this particular issue here rather than at WP:RS; incorporating implementation details into a core policy seems suboptimal generally, and as noted in the recent discussion, "mainstream" is problematic in many contexts. -- Visviva (talk) 18:34, 8 February 2020 (UTC)
  • I think "high-quality" (suggested version) is significantly worse than "mainstream" (old version) because the meaning of "high-quality" is poorly defined. But should it be "mainstream", i.e. should the sources reflect the prevalent current thought in society? Not necessarily, because something could be a small but notable minority view. Therefore, I would suggest to completely exclude this "particularly if it appears in respected mainstream publications". But things like that should indeed be discussed on an RfC since there is obviously a disagreement. My very best wishes (talk) 19:07, 8 February 2020 (UTC)

Well, I see that several users (including myself) expressed a justified criticism of the current version of the policy, and several others reverted proposed changes, but the only clearly articulated reason is that this change has not been properly discussed. To facilitate a discussion, let me summarize again the problems with current version. The current version, where I took liberty to restore the "under discussion" template says:

Editors may also use material from reliable non-academic sources, particularly if it appears in respected mainstream publications. Other reliable sources include:

  • University-level textbooks
  • Books published by respected publishing houses
  • Magazines
  • Journals
  • Mainstream newspapers

Editors may also use electronic media, subject to the same criteria. See details in Wikipedia:Identifying reliable sources and Wikipedia:Search engine test.

I see several problems with that version.

  • The sentence ""Editors may also use material from reliable non-academic sources" is a pure tautology. Indeed, taking into account that the section's title is "What counts as a reliable source", this sentence de facto says "if sources are reliable, they are reliable". This nonsense discredits the policy, for it is supposed to be at least logically non-controversible.
  • "...particularly if it appears in respected mainstream publications" As the previous talk page discussion demonstrated, the work "mainstream" is a source of big problems, because some users believe that refers only to top sources in each category, whereas others believe that everything that is not considered completely discredited is mainstream. Therefore, this word either should be properly defined or removed.
  • The list starting from "University-level textbooks..." This list is a source of constant conflicts, for it makes guidelines essentially senseless. Indeed, the policy says that the sources belonging to the "mainstream newspapers" category are reliable. In contrast, WP:NEWSORG (a part of guidelines) says that only a fraction of newspapers content is reliable for statements of fact (although frequently contains errors), whereas other materials are not reliable for statements of fact (reliable only for opinion of the author). In a situation when policy says that newspapers are reliable, and NEWSORG says that a significant part of newspapers materials are not, the this policy has priority. That means that everything what NEWSORG say can be ignored.
  • The last important problem is that each category in this list (University level textbooks, etc.) are very unhomogeneous. For example, "Magazines" include such sources as "Scientific American", "Playboy", and "Hustler". Do we really imply all of them may be reliable? Furthermore, we already had a discussion about the term "mainstream newspapers", and that discussion hadn't come to any logical end. The same can be said about "Books published by respected publishing houses": what does "respected" mean? Who can say that? "University-level textbook": which university? A level of Harvard university and, e.g. Samarkand university differ dramatically, but each of them publish books (it is sad, but low rank universities publish a lot of garbage).

Clearly, a literal interpretation of the above list is a source of big problems, which means it is supposed serve just as an example of what may be (under some circumstances) reliable. In other words, it is more like a soft recommendation. However, that is hardly an appropriate style for a policy. Examples of that type are more appropriate for guidelines.

Literally every item in that list is poorly defined, and that is a source of constant conflict disputes. I already cited the Antisemitism in Poland arbitration case. ArbCom introduced specific sourcing restrictions to stop incessant edit war in that topic, and I am sure the problem with the above quoted segment of the policy is one of the reason for that step. I think this part of the policy should be fixed, and if there will be no rationally explained objections to that, I am ready to propose the way to fix it (actually, what Francis Schonken has done was a step in a right direction, but that version can be further improved).

One more point. Currently, the policy refers to the "Search engine test" essay. Whereas I agree with many ideas of that essay, I think it is not completely correct that the reference to an essay is included in a policy. Maybe, we should take some steps to convert that essay into guidelines first? --Paul Siebert (talk) 01:39, 9 February 2020 (UTC)

A possible quick fix is to change "Other reliable sources include:" to "Other reliable sources may includes works from the following types of sources, subject to consensus:" which means editors should not just assume that for example any magazine will work, and there is no need to be more explicit on any of these definitions. If there are specific cases of existing guidance on classes of work (eg MEDRS/SCIRS for journals), they can be linked. And of course, linking the WP:RS/N as a "if you have any questions about a source..." guidance. --Masem (t) 01:56, 9 February 2020 (UTC)
@Masem:, yes, although I would say " may include some works". However, there is still a space for misinterpretations, which may be used by civil POV-pushers. I already proposed this version:

Academic and peer-reviewed publications, as well as the books published by top rank universities are considered the best quality reliable sources for such fields as history, medicine, or science.

Non-academic sources may also be considered reliable, particularly if they appear in respected mainstream publications, such as books by reputable publishing houses, magazines, journals, and newspapers, including electronic media.

Reliably published sources may be of higher or lower quality, and, depending one the context, they may or may not be used to support certain statements (see WP:RSCONTEXT, WP:REDFLAG, Wikipedia:Identifying reliable sources, and Wikipedia:Search engine test for further details).

I would like to know what is wrong with that version.--Paul Siebert (talk) 02:20, 9 February 2020 (UTC)
Paul, please don't continue with this. This is happening because of the dispute at Soviet gas van. That can't be allowed to cause a change to a core content policy. As for what's wrong with your version, it literally doesn't say anything: something may or may not be something else, and may or may not be used for this or that. The section is fine as it is. SarahSV (talk) 03:34, 9 February 2020 (UTC)
SarahSV, that is not correct. In reality, whereas I had a content dispute with another user about some poor quality source, that source has already been removed from the article, and that removal was supported by other users. The content dispute has already been resolved, which means the changes I am proposing are not dictated by my desire to win a content dispute.
By the way, if I am not wrong, you yourself are here because of your content dispute about some Holocaust related topics, and that is perfectly ok.--Paul Siebert (talk) 04:35, 9 February 2020 (UTC)
Regarding "something may or may not be ", in my opinion, that is better than a falsely concrete version that we currently have: it looks concrete, but its concretness turns all good ideas described on NEWSORG into a garbage, because all of that can be ignored per policy.
--Paul Siebert (talk) 04:51, 9 February 2020 (UTC)
Paul, I think you've misunderstood the policy. It includes "mainstream newspapers" as one of the types of source that counts as an RS. Whether a particular newspaper article is an RS depends on the context. You learn how to use sources by gaining editing experience and knowledge of the sourcing guidelines and best practice. SarahSV (talk) 05:01, 9 February 2020 (UTC)
Where does it say in the policy that its advice is easily superseded by what experienced users will tell you, and by sourcing guidelines? On the contrary (and saying this I assume that WP:RS was not written by "inexperienced" users), the policy says, in the WP:V#Reliable sources noticeboard and guideline section: "In the case of inconsistency between this policy and the WP:RS guideline, or any other guideline related to sourcing, this policy has priority."
So what you propose is unworkable:
  • Either, in the the WP:V policy the above "WP:V-supersedes-WP:RS-and-similar-guidance-in-case-of-inconsistency" quote is replaced by something like "The advice in this policy is only a broad direction, with exceptions documented in guidelines and overviews such as WP:RS and WP:RSP, and further restricted by the outcome of RfCs such as WP:DAILYMAIL." Then whatever nonsense (like that whatever "mainstream newspaper" may or can be considered a "reliable source") can be kept in the policy, but this is really a very unhelpful solution for newcomers, who generally should be directed ASAP to core content policies such as WP:V, and when they arrive here should not be getting a serving of "well, this is policy, but only in name, while any practical advice it contains has to be taken with a grain of salt."
  • Or, WP:V indeed supersedes other related guidance, but then its wording should (as much as possible) be fool-proof, consistent with consensus (as e.g. the consensus at the end of the WP:DAILYMAIL RfC), and no longer send newcomers in the wrong direction by questionable advice like that whatever mainstream newspaper "may" be used as a reliable source.
--Francis Schonken (talk) 10:25, 9 February 2020 (UTC)
Francis, I can give you a real example of a content dispute between me and another experiences editor. This dispute had already been resolved, so I can discus the policy change without a risk to be accused of an attempt to modify policy to win a content dispute.
The user attempted to add some very obscure newspaper as a source supporting some fact, and that was not in accordance with guidelines. When I objected to that, the user argued that mainstream newspapers are RS per V, and, as you correctly noted, the policy has priority. It required more than a month of disputes to fix the problem, and the problem was resolved just because majority of participants of the dispute supported me. Ii could not be able to implement what the guidelines say without a support by other users. That means WP:V makes guidelines non-workable, and it allows incorporation of garbge sources.
I understand that my proposal is far from perfect, but it still better than the current version.----Paul Siebert (talk) 13:47, 9 February 2020 (UTC)
No policy, not even WP:BLP, exists in an absolute superseding state that has no wiggle room; all polices have grey areas that the policy language should be helping to set up for consensus-based discussions - or more specifically, P&G are meant to be descriptive and not prescriptive, nor are meant to be read as if they were bureaucratic rules and nitpicking on the language. Policies should have far less grey area than guidelines, but they are not absent of such undefined areas. To me, that's where this present language is at (at least, with my additional wording): it is not saying that we absolutely allow, for example, all mainstream newspapers are RSes, but that they are a common source of RS, which 100% accurately describes practice. --Masem (t) 14:54, 9 February 2020 (UTC)
Masem, I am not saying that mainstream newspapers should not be allowed. Moreover, I myself am using them occasionally. However, in my opinion, the policy should stress that usage of mainstream newspapers should be strongly content specific, and different articles in the same newspaper may be reliable or not. Actually, guidelines perfectly explain that. However, the current text of the policy contradicts to the guidelines, thereby allowing some tendentious editors to use it as a pretext for using poor quality newspaper publications in contentious areas.
The current version of the policy works fine when all users who participate in editing are good faith users prone to consensus. The problem is that I am working in areas where content disputes are frequent, and I am seeing that the policy does not allow us to resolve content disputes as easily as it is supposed to. Im majority cases, the dispute is resolved de facto' by majority users vote, which is not good. A good policy should not work in that way.--Paul Siebert (talk) 15:40, 9 February 2020 (UTC)
And, remember, I came here first with the question "what should be considered as mainstream newspapers?", and we still have no answer, for different users see it totally differently (from "only newspapers of records" to "any non-discredited newspaper").--Paul Siebert (talk) 15:45, 9 February 2020 (UTC)
I think I'm still in agreement from a far earlier discussion (not just here) that "mainstream" is a term of art, and poorly defined to use here. At least for purposes of WP:V, I would replace that with "national and regional newspapers" with the implicit "that meet the basics of RS, with editorial oversight". I mean that's not perfect but its better defined that "mainstream". If anything, we should have sections at WP:RS that goes into detail of how to identify a reliable type of each of these sources, where we have given more specific information (eg like for journals: peer-review, non-predatory ones, and with deference to MEDRS and SCIRS for certain topics).
But I feel a lot of what's being talked about here is wiki-laywering over the exact language by other users, which should not be done. --Masem (t) 16:00, 9 February 2020 (UTC)

Actually, I have an idea. What if we expand the policy a little bit to make more direct and clear reference to guidelines? For example, if instead to the current version we write something like that:

Editors may also use material from non-academic sources, particularly if it appears in respected mainstream publications, including:

.....

.....

  • mainstream newspapers, as described in WP:NEWSORG;

.....

and do the same for each item in the list, that would resolve a situation. By doing that, we demonstrate that the policy provides just general principles, and the decision about reliability should be made on case by case basis as described in guidelines. BTW, similar idea was expressed by Blueboar below.

Regarding wikilawyering, I would love to avoid that, but in that case I am working as "devil's advocate" in attempt to eliminate a possibility of usage of possible gaps in the policy by POV pushers. I also find a situation with Antisemitism in Poland case a worrying sign. I think, by fixing the policy, we could create a situation when these sanctions could be made redundant.--Paul Siebert (talk) 16:09, 9 February 2020 (UTC)

Francis, you referred above to "what you propose", and the indenting made it look like a reply to me. Just noting here that I don't have a proposal. Paul, there was a long-term dispute between you and My very best wishes about Soviet gas vans and now you're both on this page. I think you've misunderstood this policy and what a policy is. We do, as a matter of fact, accept "mainstream newspapers" as RS. The policy has stated that for many years, and it reflects what actually happens on WP. Policies must be descriptive as well as prescriptive; a policy that is no longer descriptive is not a policy, no matter what it calls itself. Your dispute was caused by someone wanting to use, as you put it above, a "very obscure newspaper" for what you regarded as an important point. So there's your answer. SarahSV (talk) 18:59, 9 February 2020 (UTC)

@SlimVirgin: I think you misunderstood – this was indeed a direct reply to your comment above that comment by me. Re. "... descriptive ..." – the correct description of the current situation (after WP:DAILYMAIL, similar consensuses, and site–wide removal of a selection of such "mainstream newspapers" as sources from Wikipedia, etc.) is that there is no blanket approval of "mainstream newspapers" any more (some are accepted, others are not, and for some of them the rejection to ever still use them as a reliable source is site-wide); see alsoWP:RSP. So, "reliable sources include: ... Mainstream newspapers", as it is currently in the policy, is a remnant of a distant past that is no longer descriptive of the current situation, and thus misleading (especially as it claims to be so that "this policy has priority" as it is in WP:V#Reliable sources noticeboard and guideline) – it sends newcomers in the wrong direction, and as I have come to realise through your point-missing reply: it is especially dangerous for experienced editors who continue to have discussions in terms that no longer correspond to current reality of adequate guidance, which out of justified necessity ignores WP:V on this point.
So, indeed, whatever way you turn it, you "propose" to continue an "unworkable", unstable and untenable situation, that causes more trouble than it solves because it is no longer "descriptive" of the current situation.
So either the policy-level blanket approval of "mainstream newspapers" is removed, or a new wording qualifying the statement thus that it can no longer be misunderstood as a blanket approval of "all" mainstream newspapers, but only of those mainstream newspapers which are, as a correct *description* of the current situation, still widely accepted. As the second is unlikely to find agreement on short term, the first solution seems best until an appropriate updated wording is approved. --Francis Schonken (talk) 19:53, 9 February 2020 (UTC)
This is || this close to wikilawyering on language. I read the section, and I do not see any absolute allowance for any "mainstream newspaper", just that mainstream newspapers may be considered reliable sources, subject to the implicit IAR/concensus-based discussions all policies have. Daily Mail may be "mainstream" but we've clearly deprecated it; does this policy need to spell that out? No. There's a tiny tweak of wording I've suggested to otherwise keep this descriptive without losing any other intent. --Masem (t) 19:57, 9 February 2020 (UTC)

SarahSV, you are not right. The content dispute that I used to have had already been resolved, and I am pretty satisfied with its outcome. In contrast to another participant of that dispute, I openly declared the fact that I was a party of the dispute (see above on this talk page). I am not aware of any policy, giudelines, or other rules that prevent users who were a party of a content dispute to discuss the policy. In addition, I am here primarily because I see some general problems with the policy, and I think that the ArbCom decision about Antisemitism in Poland case (a conflict I was not a party of) is an indication of problems with WP:V.--Paul Siebert (talk) 20:14, 9 February 2020 (UTC)

I also think that mainstream newspapers should be allowed, but it is subject to some additional limitations and restriction described in NEWSORG. It should not be a blanket approval, and I agree with Francis in that. In addition, if we allow "mainstream" newspapers, we should define (in policy or in guidelines) what does that term mean. --Paul Siebert (talk) 20:18, 9 February 2020 (UTC)

(ec) Francis, thank you for the reply, which perhaps points to the source of the misunderstanding. The policy states: "Editors may also use material from reliable non-academic sources, particularly if it appears in respected mainstream publications. Other reliable sources include: university-level textbooks; books published by respected publishing houses; magazines; journals; mainstream newspapers".
That does not imply "a blanket approval of 'all' mainstream newspapers". Note: "may" also use, "particularly if it appears in". What counts as an acceptable use of a "mainstream newspaper" depends on context and whether primary or secondary. Some mainstream newspapers are disliked and distrusted on WP; none are banned that I know of. Everything depends on the particular case. SarahSV (talk) 20:19, 9 February 2020 (UTC)
Sarah, that does not sound as blanket approval to you and majority of reasonable good faith users. However, I faced a situation when some obscure newspaper was declared "mainstream" (actually, the policy gives not a single clue on how to determine mainstreamness), and an oo-ed material was used to support a controversial statement. Formally, that was in accordance with our policy, although contradicted to guidelines.
In addition, if Francis needs your explanations on what the policy says, that means the policy allows double interpretation. I believe you are not going to claim that Francis Schonken or I are not experienced users, and if we interpret policy in that way, there is no reason to conclude our interpretation is wrong, and your interpretation is correct. A good policy does not allow double interpretation, otherwise it is not a policy.--Paul Siebert (talk) 20:25, 9 February 2020 (UTC)
Paul, what you're doing here is a good example of why isolated experiences should not be allowed to inform policy. That dispute went on for months because of your failure to find secondary academic sources, follow them, then ask the community to support that version. The newspaper in question was not in any sense "mainstream". SarahSV (talk) 20:35, 9 February 2020 (UTC)
SarahSV, your description of the dispute is obviously wrong. During that dispute, I and a couple of other editors found ALL good reliable sources on that subject, and performed a careful analysis of those soures. We did our job quite well, but another user attempted to add two unrelated primary sources and one junk newspaper, and that was definitely not our fault. And, please, do not return to that issue again: that content dispute had been successfully resolved.
And, my experience is by no means isolated. Besides that dispute, I was involved in other disputes (or was a witness thereof), and, based on that combined experience I conclude there are some problems with WP:V. Thus, I was closely watching the development of a conflict between several users over the Holocaust in Poland topic, and I even attempted mediation. That was very informative, because I came to a conclusion that the whole conflict (which led to ban of one user and topic ban of another) could have been avoided had the policy been written better.
In addition, in my editing, I am using, almost exclusively, only top quality sources, and I accumilated a very significant experience in identification of reliable sources. I am, arguably, among just few Wikipedians who can claim that my skills have been independently confirmed, see this peer-reviewed publication that says about me. I have a long history of editing, and long hisotry of conflict disputes, and that my proposal is a result of that my experience/--Paul Siebert (talk) 21:22, 9 February 2020 (UTC)
Re. "... none [i.e., mainstream newspapers] are banned that I know of ..." – I don't know how many more times I have to link to WP:DAILYMAIL and WP:RSP. The Daily Mail is a mainstream newspaper and it is, according to the WP:RSP terminology, deprecated as a source for Wikipedia content (except, like any source, where it passes WP:ABOUTSELF). The site-wide consensus about that can be found at WP:DAILYMAIL, confirmed by Wikipedia:Reliable sources/Noticeboard/Archive 255#2nd RfC: The Daily Mail. Other similarly deprecated mainstream newspapers are listed at WP:RSP. Not only are they listed thus, and does the consensus exist, also edit filters usually prevent further use of such mainstream sources, and further, some editors have taken it upon themselves to remove these mainstream newspapers as sources from Wikipedia where they had been used before they were deprecated. Every once and awhile when such removal encounters resistance, it pops up at WP:RSN where rarely ever (in fact not a single case as far as I can remember) an exception is granted. Further *you knew this*, you tried to "fight" the WP:DAILYMAIL decision after the discussion had closed on consensus, triggering a second close. Instead of generously implementing that consensus in guidance, you now try to make us believe you didn't know.
Sorry to point out this disconnect between policy and the reality it should describe. Whatever way it is turned, this disconnect is unhelpful. It is unhelpful for newcomers. It is unhelpful for experienced users. --Francis Schonken (talk) 21:36, 9 February 2020 (UTC)
"Deprecated" doesn't mean it can't be used, and not only ABOUTSELF. If someone we're writing about has a byline or an interview, we can use that, although make sure it's not the kind of interview that might have been altered significantly. Again, that we can use mainstream newspapers doesn't mean they can all be used for anything. SarahSV (talk) 21:43, 9 February 2020 (UTC)
Now you're just wikilawyering – several mainstream newspapers are deprecated, and can no longer be considered as reliable sources in Wikipedia (with edit filters preventing their use installed), that is a reality about which the current state of denial in the WP:V policy is an unhelpful & counterproductive disgrace. So stop wikilawyering about some minimal exceptions to that general thrust of consensus decisions: seems rather you're still resisting a generous implementation of consensus on this point. I think you've chosen the wrong battle there, per what you write above about "descriptive": the consensus on these deprecated sources is broadly implemented, and a few editors, however experienced, should not try to keep the correct description of the current situation out of the policy, clinging to guidance that is no longer descriptive of consensus & practical site-wide implementation. --Francis Schonken (talk) 22:07, 9 February 2020 (UTC)
A point: evidence was raised during the DM deprecation discussions that shows that the DM falsified interviews and byline-written articles. (not simply taking things out of context, such as Lawrence Lessig had recently with the NYTimes). This is why the deprecation case on the DM is so strong (Brietbart's closely behind it due to similar falsification), and why it really only should be used when the DM is part of the news story itself. However, this is a side point to this discussion. --Masem (t) 22:19, 9 February 2020 (UTC)
I'm tired of the personal attacks, so I think I'll leave this to the rest of you, but meanwhile I offer you a counter-example in Kingsley Amis (permalink), note 38, interview with Hilly Amis in the Daily Mail. It could certainly be replaced by a better source for the points it supports. But is that article ("Here, in a rare interview") always and necessarily a bad source because of the publisher? What about Martin Amis's tribute to Elizabeth Jane Howard, published in the Daily Mail with his byline? In my view, just about anything could be an RS in WP if appropriate; the word "appropriate" is important. SarahSV (talk) 23:16, 9 February 2020 (UTC)
Absolutely agreed. "Appropriate" is a key word. That means that word should be explained in the policy in very details, not just mentioned in passing. We should devote a paragraph or so to describe how appropriateness should be determined.--Paul Siebert (talk) 05:06, 10 February 2020 (UTC)

WP:SOURCES, one more attemptEdit

SarahSV, in my opinion, even if some editor is rude, that is quite forgivable if their contribution to Wikipedia is generally positive. Therefore, I beg you to ignore personal attacks (or something that you see as a personal attack), and continue our discussion. I came to a conclusion that I should have to explain my position better, and here I am trying to explain my major ideas, which had been inspired by the Holocaust in Poland topic and a discussion that involved several editors, including Piotrus. First of all, I myself prefer to use only top quality English sources for editing, mostly peer-reviewed publications. As I already explained, my approach to source selection was described as good in this peer-reviewed publication, so I have a right to claim that I know how to find good sources, and that fact is based on what a reliable source says. However, I fully realize that my approach to source selection is not universally applicable. Indeed, Wikipedia is a very non-homogeneous collection of articles, it includes diverse topics starting from quantum mechanics, gene editing, or the Holocaust to baseball teams, recent Trump's activity, or local high schools in the middle of nowhere. Obviously, lion's share of Wikipedia articles couldn't be possible to write using the sources selected according to my approach. And I fully realize that that is absolutely ok.

What is not ok is the following. The current policy does not explicitly prohibit to use different quality sources in the same article. Formally, it is quite possible to add an article from a popular kid's magazine to the Einstein-Podolsky-Rosen paradox article. Removal of such a source is a matter of editorial consensus, which makes a situation shaky. I think you yourself would object adding a claim to some Holocaust related article that Ukrainian nationalists had never been engaged in killing Jews during WWII. Meanwhile, such claims can be found in nominally scholarly peer-reviewed articles that are being published in modern Ukraine, for example, by Volodymyr Viatrovych. Obviously, for important topics and important articles, for which top quality sources are available, the level of sources should be as high as possible, and the sources should be of comparable quality. However, for less important topics, for which no good sources are available, lower quality sources can be used, unless that causes no justified objections from some editor(s).

Unfortunately, our policy does not allow us to implement that approach easily. Sometimes, enormous efforts are needed to get rid of garbage sources that are being advocated by some POV pusher. In my opinion, good faith users should be provided with better tools for comfortable editing, and our policy does not serve this goal as efficiently as it could. Thus, during a recent content dispute, one admin (a good admin, by the way) seriously proposed me to start an RfC to remove some recently added questionable source despite that fact that it was added in violation of WP:REDFLAG, and no clear support of that addition was obtained during the RSN discussion. That means even experienced admins sometimes do not interpret the policy in the way it is supposed to be interpreted according to you. That means, it must be made more clear.

If you believe I am not right, please explain me why. Otherwise, let's discuss possible ways to improve it.--Paul Siebert (talk) 04:58, 10 February 2020 (UTC)

Paul, for what it's worth, I do approve stressing that tabloid newspapers are lower on the totem pole. But since you mention the recent problematic Arbcom ruling, the issue which I raised (and which did not get any clear ruling) was that sometimes newspapers are fine for non-REDFLAG issues; in either case what really irked me back then was than an editor who restored a low quality source was summarily blocked for few days with next to no warning - but that's really was (is) an issue of battleground mentality and admin abuse of powers, not something that is that relevant here. --Piotr Konieczny aka Prokonsul Piotrus| reply here 10:21, 10 February 2020 (UTC)
@Piotrus: in my opinion, it was not only a battleground mentality issue. Currently, a user who adds some source of questionable quality, and a user who removes it are in more or less equal position. They are considered parties of a content dispute. The policy should allow a user who wants to improve quality of Wikipedia to remove questionable sources if that user feels the source is not completely relevant. The burden of proof should be on those who adds such a source, and if the party that adds that source fails to demonstrate the proposed source had an adequate quality and is relevant, its re-adding should be not allowed. If that norm existed in the policy, there would be no need in AE in the case you are referring to.--Paul Siebert (talk) 16:59, 10 February 2020 (UTC)
There are no excuses for rudeness on Wikipedia; we have a civility policy, a behavioral policy, as important as any of our editing policies. This is a collaborative project and collaborative behaviors matter. While people do make mistakes, I certainly have, either deliberately or otherwise, in terms of behavior, we cannot and do not excuse that behavior because someone is a good contributor, otherwise Wikipedia would be a mad house of people venting freely whenever it suited them. There are nuances to policy which not everyone understands. It behooves all to listen carefully especially when editors "speaking" are those who wrote or helped write the original policies. Littleolive oil (talk) 22:49, 10 February 2020 (UTC)

Recent editsEdit

My very best wishes, I reverted your recent edit because you're part of a long-term dispute that hinges precisely on the wording you removed. Please wait for consensus to form; if it has already formed and you disagree, you can try to change it here on talk. Or if it's a key issue, open an RfC, but please first seek advice and consensus labout how to word it. SarahSV (talk) 19:25, 8 February 2020 (UTC); edited 20:05, 8 February 2020 (UTC)

Sure, you are very welcome to revert my edit. That's why I indicated in my edit summary that I agree with previous (consensus) version (now restored by EEng). I think this old wording is fine, no RfC required. I did not suggest any changes on any policy pages and not sure what dispute you are talking about. My very best wishes (talk) 00:41, 9 February 2020 (UTC)
  • Could this be resolved by including a more prominent pointer to WP:NEWSORGS in the text? Blueboar (talk) 20:29, 9 February 2020 (UTC)
Yes, Blueboar, I already proposed how can it be done in the above section.--Paul Siebert (talk) 20:50, 9 February 2020 (UTC)

Changing a policy page with out consensus and while under discussionEdit

In all of the years I have been on Wikipedia I have never seen an agreement that allows for a policy to be changed without consensus-BOLD does not apply here. And it is controversial to make any kind of substantive edits to aspects of an article, especially a policy page, while that aspect is under discussion. If that has changed please link to it. Littleolive oil (talk) 21:35, 9 February 2020 (UTC)

This section in bold (mine) is not acceptable. If there is community wide consensus for this please link to it. Littleolive oil (talk) 21:38, 9 February 2020 (UTC)

This section is the subject of a current discussion. Please feel free to join in. This doesn't mean that you may not be bold in editing this section, but that it would be a good idea to check the discussion first.
Policy pages are not immune to the wiki process; see WP:PGBOLD for more on this "older but still valid method". -- Visviva (talk) 21:42, 9 February 2020 (UTC)
WP:PGBOLD, which is policy, has: "... you should not remove any change solely on the grounds that there was no formal discussion indicating consensus for the change before it was made. Instead, you should give a substantive reason for challenging it and, if one hasn't already been started, open a discussion to identify the community's current views."
EEng, Littleolive oil and Alanscottwalker: you've all reverted "solely on the grounds that there was no formal discussion indicating consensus for the change before it was made" – without giving "substantive reason"s why one version or the other would make the best policy-level guidance on the WP:V page. Just giving you the opportunity to participate in this debate on content of the matter, because as long as you don't you're in breach of policy. --Francis Schonken (talk) 07:21, 10 February 2020 (UTC)
Should not is not must not and anyway I reverted on the grounds that the discussion showed not only lack of consensus but substantial opposition. As for your lectures about breaches of policy, you really need to cool your jets. EEng 08:41, 10 February 2020 (UTC)
As I indicated when I reverted, the policy change is disputed and I suggested why it is disputed. Moreover, as far as can be seen two users are proposing different changes. If those two users ever come to agreement, it can be put to the community and the usual way to do that is RfC. This is a policy that has had multiple widely advertised and participated in requests for comment to change it. -- Alanscottwalker (talk) 10:37, 10 February 2020 (UTC)
Francis, you sort of picked one phrase out of context to give an impression contrary to what WP:PGBOLD actually says, which is that any significant change in policy is going to require a lot of support and discussion. The phrase that you picked was basically urging people to give a bit more of a rationale when reverting a bold change. North8000 (talk) 20:02, 10 February 2020 (UTC)
    • Let me clarify what is inappropriate here. Content on a Wikipedia policy is under discussion. Oddly, an editor who is part of the discussion has added a tag that allows editors to make changes to the policy while these discussions are ongoing. This is ludicrous and disingenuous. Bold means make an edit but if that edit is contested back away. Bold does not give us permission to add content while that content is under discussion; bold does not mean I am going to give myself and others permission to add or change content while that content is under discussion. Doing so is a recipe for non-neutrality. Any substantial changes in content on a policy need community wide agreement. We as a community have to live with changes in everyday editing and we as editors must decide as a group what works and what doesn't. Policy is complex and has a basis in the history of the encyclopedia and the editors who over time thought deeply about what a specific policy meant. No single editor can give permission to add content to a policy page while that content is under discussion, and changes in the policy itself needs community wide consensus which honors our editors and history. Littleolive oil (talk) 18:25, 10 February 2020 (UTC)

WP:WPEDIT Policies and guidelines are supposed to state what most Wikipedians agree upon, and should be phrased to reflect the present consensus on a subject. In general, more caution should be exercised in editing policies and guidelines than in editing articles. Minor edits to existing pages, such as formatting changes, grammatical improvement and uncontentious clarification, may be made by any editor at any time. However, changes that would alter the substance of policy or guidelines should normally be announced on the appropriate talk page first. The change may be implemented if no objection is made to it or if discussion shows that there is consensus for the change. Major changes should also be publicized to the community in general, as should proposals for new policy pages...

  • I support the interpretation that policy pages, even core policy pages like this one, may be boldly edited without prior discussion. But I also support the idea that those changes may also be reverted for lack of discussion and, specifically, I agree with EEng that "should not" does not mean "must not". Bold changes need to be allowed so that minor and/or uncontroversial changes can be made without a full-blown talk page discussion every time. But if those or other changes are major, controversial, or controverted, then people need to be able to rely on the policy's existing status quo while they're being discussed, especially since discussions of changes to policies, particularly core policies, can often drag on for weeks (and are more often unsuccessful than successful in making a change). Requiring that a bold change be kept in the policy text (even with an "under discussion" tag) is just asking for misapplication of the temporary language while the discussion is taking place and, frankly, could easily be used to game a discussion at an article. Best regards, TransporterMan (TALK) 18:47, 10 February 2020 (UTC)
  • BOLD policy changes can be made, but they can also be reverted, which at that point, if it is a serious idea (eg not nonsense by drive-by IPs) it should be discussed on the talk page and not re-reverted or edit warred. If you know the addition you are going to make will likely be controversial, then you should start at the talk page first. --Masem (t) 18:49, 10 February 2020 (UTC)
Masem just said it succinctly & perfectly. North8000 (talk)

Moving onEdit

Taking up part of an idea of Paul Siebert, and some of my own suggestions above, I've:

  • added the qualifier (exceptions include tabloid journalism; see also WP:NEWSORG and WP:RSP) to the "Mainstream newspapers" entry in the little list of "other" reliable sources
  • removed In the case of inconsistency between this policy and the WP:RS guideline, or any other guideline related to sourcing, this policy has priority. — WP:V should not claim supremacy if everyone in the discussion above agrees that subsequent reliable sources-related guidance can modify what are only broad principles in the WP:V policy.

[4]. --Francis Schonken (talk) 07:21, 10 February 2020 (UTC)

You're now starting to get WP:POINTy. By this stage of the proceedings it should be clear to you that a large number of editors feel that much hangs on even superficially minor changes to this particular page, and that your bull-in-a-china-shop approach is unhelpful. EEng 08:45, 10 February 2020 (UTC)
EEng, I am not sure that you are right. My impression is that majority of participants of that discussion agree that there are some problems with that policy.
Francis, I think the words "In the case of inconsistency... etc" should stay, because they set a hierarchy between the policy and guidelines. Instead, the wording of the policy should not contain too many details that are more pertinent to guidelines: the policy sets general principles (and may contain examples), whereas all details are specified in the guidelines. Thus, the good policy language should be (that is not an exact wording, I am just giving a general idea) " academic and peer-reviewed publications are usually (or as a rule) the most reliable sources, see WP:SCHOLARSHIP, WP:MEDRES for further details". Accordingly, the guidelines are explaining details, including the examples of nominally peer-reviewed publications that are NOT good sources. In that case, there should be no ground for a conflict between a policy and guidelines: the policy does not claim ALL peer-reviewed publications are good, and quidelines do not question the general principle that peer-reviewed publications are, in general, the best sources, although they explain in details possible exceptions to that general rule.
In connection to that, let me try to summarize what is really important in the policy.
  • First, a good policy should explain what does "Reliably published" mean, and the current version is doing that reasonably well: "published" means "made available to the public in some form". However, that is still not an optimal wording. A really reliable source is supposed to be available on a permanent basis, in other words, any posts in some blog are unreliable not only because they may contain false information, but because they may be removed or altered at any moment. In connection to that, one goal of WP:V is to minimize a risk of dead links: the source that is likely to disappear from a public access should not be considered a RS. I propose to add the words on a permanent basis to the policy, and to further explain that principle.
  • Second, and a totally different point is as follows. Among the sources that are deemed permanently available to public (a.k.a. "reliably published") we should select the sources that (i) are relevant to some concrete article (per User:SlimVirgin), and (ii) contain reasonably correct and reasonably non-fringe information. The latter may be considered as a preliminary screening of sources before they pass through a more stringent NPOV screen. Accordingly, the policy should explain general principles of relevance determination, and general ideas of what should be considered as a reasonably good source. The current problem with the policy is that the principles of relevance determination are not described at all (and I even don't know if we have any guidelines on that account). Another problem is that, whereas the policy is explaining the criteria that allow us to discriminate between good and bad sources, the list of categories of reliable sources ("Journals, newspapers, etc") de facto cancels what the previous sentence (" The best sources have a professional structure ...etc") says: who cares about "professional structure" when the policy says that "Journals are reliable"? (Again, here I am acting as devil's advocate). That may be fixed if each item from that list will be presented just as examples of what a RS could be, and all further explanations are moved to guidelines.
Thus, if we say something like that (if you proposed better wording it would be great): "peer-reviewed publications are, as a rule, best quality sources. See guidelines for further details", that would be fine.
If we say "mainstream newspapers are usually reliable sources; see guidelines for further information on how to determine if a newspaper is mainstream, and how newspaper materials should be used" (if I understand it correct, something of that kind was proposed by User:Blueboar), that also will be fine.
If these (or similar) reservations will be added to the policy, the contradiction between the policy and guidelines will be resolved. The policy will not contain too much details, but it will affirm important principles that are further specified in guidelines.--Paul Siebert (talk) 17:33, 10 February 2020 (UTC)
majority of participants of that discussion agree that there are some problems with that policy is not the same as "there was consensus that Change X should be made to fix it". EEng 15:54, 11 February 2020 (UTC)
Yes, of course. Moreover, I believe BOLD edits of policy texts should be avoided. However, the fact that many users provide different interpretations of what WP:SOURCES say, and many of them agree that the policy should be improved is an indication that the discussion about policy improvement should be continued. In connection to that, I would like to obtain your constructive comments on the above text.--Paul Siebert (talk) 16:59, 11 February 2020 (UTC)
I’m on my way to a Brazilian wedding, Ecuador, and maybe the Galapagos, so I’m afraid y’all will have to work this out on your lonesomes. EEng 19:26, 11 February 2020 (UTC)

Procedural (or not)Edit

Thanks for all the procedural advice, which is, however, partially lost on me: not as if this is the first time I initiate some guidance, or contribute significantly to policy pages. I even wrote the basis of the how-to essay on "How to contribute to Wikipedia guidance". So all the lecturing on procedural issues largely diverts from the topic at hand. For instance, @EEng: I'd really like to know how you think on whether or not it is is time to implement recent consensuses (i.e. last two years or so) on broadly barring tabloid journalism as reliable sources from Wikipedia on this policy page. And if so, how?, and if not why not? – yeah, it's really interesting to read that this is not a "must" but only a "should", but that leaves the more interesting part of the question unanswered. Same question, of course, for Littleolive oil, Visviva, Alanscottwalker, TransporterMan, North8000: you've all reverted for procedural reasons and/or commented on procedures on this talk page (so that the whole topic is now drenched in procedural comments). All that is now exceedingly clear, ie how you think on procedure, for me, and for anyone reading this section. I'd suggest Wikipedia talk:Policies and guidelines (and to a lesser extent Wikipedia talk:How to contribute to Wikipedia guidance) if you want to change policies and other guidance on how Wikipedia's policies are written. This is not the place to hash that out. But let's move on regarding the topic at hand: to me it is exceedingly clear that some form of updating is recommended regarding all too easily recognising all sorts of mainstream newspapers as reliable sources in this policy. All thoughts and suggestions on that topic are welcome. --Francis Schonken (talk) 09:40, 13 February 2020 (UTC)

Francis, you seem to be lecturing us as if we were doing what you've been doing here! :-). The best to you either way. Sincerely, North8000 (talk) 13:50, 13 February 2020 (UTC)
Your refusing to listen to others when they demonstrate your accusations are wrong and what you are doing is being disruptive is unfortunate -- if you have an actual question you are free to ask. Mainstream news is not tabloid, see [5] -- mainstream with respect to news is, inter alia, demonstrating the practice of the recognized and established mainstream standards for news reporting. Tabloids practice sensationalism, and worse, not mainstream news reporting. -- Alanscottwalker (talk) 15:57, 13 February 2020 (UTC)
  • The problem here is that the term “Mainstream” was introduced into WP policy as an opposite to the term “Fringe”. In that context, many tabloids are “Mainstream”. Blueboar (talk) 16:23, 13 February 2020 (UTC)
    • Huh? It's rather common sense that tabloids would be on the fringe of news-journalism. Alanscottwalker (talk) 16:32, 13 February 2020 (UTC)
  • Francis Schonken As long as any editor appears to misapply procedure, policy, or guidelines or attempts to adjust policy pages with out community wide input, push-back should be expected. Littleolive oil (talk) 19:07, 13 February 2020 (UTC)
First, the "fringe - mainstream" issue seems belongs to NPOV, not V.
@Littleolive oil: maybe, we will better return to the discussion about improvements of the policy?
@Alanscottwalker: I am probably repeating myself, by let me remind you that, on that talk page, I already asked what does "mainstream" mean, and an opinia spectrum was surptisingly wide: some users argued that only respectable newspapers similar to newspapers of records should be considered mainstream, other believe that any newspaper that is not widely recognized as fringe should be considered mainstream; also, some people seem to believe that that term is intuitively clear. In my opinion, that ambiguity is a permanent cause of edit wars, and the sooner we fix that, the better.
I also believe that the attempts to formalize it, for example as "mainstream newspapers except tabloids" is not a solution. Thus, what should we do with non-English mainstream newspapers? Add "mainstream newspapers except tabloid and Chinese/Russian newspapers"? If someone will find some other exception, will we add it too? And so on, and so forth.
In reality, WP:NEWSORG gives us a good example of a thoughtful and well nuanced approach to newspapers. That is the rare case when guidelines are much better written than the policy. In that situation, the policy should be changed to summarize general principles of WP:NEWSORG, without going into details, and provide a link to the guidelines for other details. Instead, it says "Mainstream newspapers may be used as reliable sources". This primitive approach de facto cancels all reasonable nuances explained in NEWSORG, and excluding/adding some category of newspapers cannot fix that situation. I think, the situation may be fixed if we do the following:
(i) make a stress on NEWSORG as follows:
"Mainstream newspapers may be used as reliable sources as described in WP:NEWSORG.
(ii) Furthermore, it is not clear why journals, magazines and newspapers should be separated. It is also unclear why newspapers are acceptable only when they are "mainstream", but no such reservation is added for journals or magazines. Obviously, the same type criteria are applicable to journals, newspapers and magazines. It makes sense to combine "Journals, magazines, and newspapers" into a single category, and instead of three separate items, write:
"Mainstream newspapers, journals and magazines may be used as reliable sources as described in WP:NEWSORG.
(iii) Since the usage of sources is content dependent, it makes sense to stipulate that newspapers cannot be used as a single source of sensational claims per WP:REDFLAG.
"Mainstream newspapers, journals and magazines may be used as reliable sourcesas described in WP:NEWSORG. Isolated publications of that type cannot be used as a source for controversial or sensational claims, per WP:REDFLAG.
These are incremental improvements of the current text, and even the first proposal seems to much more clear than the current wording.
If Francis believes tabloids should be mentioned explicitly (I would probably agree with that), my proposal allows doing that in the guidelines text.--Paul Siebert (talk) 23:53, 13 February 2020 (UTC)
Of the three, I could support something like two, but just put as described in WP:Reliable sources. Many RfC's provide alternatives. As for your initial premise that policy erases guideline, that is a crude approach, guidelines elucidate and explicate policy, and reading them together is good faith construction. Alanscottwalker (talk) 12:03, 14 February 2020 (UTC) Alanscottwalker (talk) 11:57, 14 February 2020 (UTC)
@Alanscottwalker: The link to WP:Reliable sources is already present in the policy text, the link to NEWSORG specifies which concrete part of guidelines is relevant to newspapers.
Regarding "premise that policy erases guideline", that is a crude interpretation of my words. I meant that this concrete sentence in the policy is written poorly, and, as a result, it de facto nullifies all correct and useful ideas described in some concrete section of guidelines (WP:NEWSORG). Therefore, in contrast to a normal practice, when guidelines mush be brought in accordance with policy, this particular phrase in the policy should be brought into accordance with guidelines.--Paul Siebert (talk) 01:27, 18 February 2020 (UTC)
When policy and guidance conflict, they need to be brought into sync with each other. That may mean changing the policy, it may mean changing the guideline, it may mean changing BOTH. A centralized discussion needs to be held, and consensus will determine what needs to change, and how. Blueboar (talk) 02:56, 18 February 2020 (UTC)
What you are saying is too obvious to be incorrect, and too general to be useful. Sorry.--Paul Siebert (talk) 04:33, 18 February 2020 (UTC)

ABOUTSELFEdit

The section on self-published sources could do with tightening given recent fights at, e.g., Knights of Columbus (edit | talk | history | protect | delete | links | watch | logs | views). I suggest adding the following below the list of exceptions:

Self-sourcing should be used only to add a level of detail to content otherwise drawn from reliable independent secondary sources. Content along the lines of "$SUBJECT states $OPINION (source, $SUBJECT stating $OPINION)" should be avoided, especially if they are predictable reactions to criticism.

This will help forestall arguments about exactly how much of an article has to be self-sourced before it's "based primarily on such sources". Self-sourcing is supposed to be a qualified exception, not a default allowing a subject to be framed primarily according to their own self-image or that of their fans. Guy (help!) 13:59, 12 February 2020 (UTC)

To combat US-centrism I propose your example be changed to “€SUBJECT states €OPINION (source, €SUBJECT stating €OPINION)”. EEng 12:33, 13 February 2020 (UTC)
This is an area I've been interested in for a while. I think I agree with what you are saying but I wanted to ask how you think it would apply to some examples I brought up here in the past [[6]]. One example is an article subject has been described as left/right wing by sources. In an interview the source disagrees with those terms and describes themselves as centrist. Would you consider that adding a level of detail even though that level was to contradict third party claims rather than support them?
My other example related specifically to something I think is wrong with the NRA page but probably applies to may articles about organizations that engage in advocacy in controversial areas. The NRA's opposition to, as an example, universal background checks, gets lots of RS coverage. But that RS coverage rarely says why the NRA opposes UBCs, or only says it with minimal detail. It's my feeling ABOUTSELF should allow inclusion of the NRA's stated stance on the issue even if that means the content is solely sourced to the NRA. Like the previous example I think what makes this an allowable ABOUTSELF case is that other sources have raised the issue/controversy to prominence but haven't offered a full telling of the article subject's POV. Do you think your proposed change would make it harder, easier, no different to include such material in an article per ABOUTSELF? My feeling is your interpretation of ABOUTSELF would make these inclusions harder rather than easier. Springee (talk) 14:20, 12 February 2020 (UTC)
It would seem to make sense to bring the idea of "unduly self-serving" that is used for BLP-released SPS here too. In other words, we are only pulling from a group's own sources as either to have content expected compared to other articles (uncontestable details like date of founding for example), or in response to directed commentary about them, and otherwise not to pull things that are simply to promote the group or its cause. I don't like the term "predictable" because that could be used to block reasonable counter-statements to a criticism against it. For example, a group gets put onto the SPLC's hate group list; their response to state they are not a hate group would be "predictable" (no one wants to be on SPLC's hate group list) but would be wholly appropriate to include if the SPLC's hate group listing is also included; that's needed for neutral coverage. --Masem (t) 15:32, 12 February 2020 (UTC)
Masem, The issue there is WP:MANDY. Example: very few racists self-identify as such. We only care about what reliable independent secondary sources say. If they note that the person says they are a centrist, while describing them otherwise, then we can say that, but if the RS say they are left or right, the person disagrees, but no independent source finds that worthy of note, then neither should we. Guy (help!) 09:16, 14 February 2020 (UTC)
This is the problem when editors take labels said by RSes as fact and thus try to remove the subjective nature of the for current groups and living people. On the MANDY point, if all we have in counterpoint is a "no I am not", yeah we don't need that, but if instead they have a long winded explanation to differentiate what they consider themselves, then a super brief summary seems reasonable. But this is very much context specufic depending on if there's a chronological events in play, etc. I am absolutely against disallowing these as there are appropriate times they are needed for impartial coverage. --Masem (t) 12:12, 14 February 2020 (UTC)
IMO there are a lot of appropriate uses for "self-sourced". The type discussed by Springee is one of them. Also there is a lot a basic objective information about an organization of the type that secondary sources usually do not cover. For example, the classes offered by a university or programs offered by an organization, or locations of their facilities. Another is the official published policies or mission statements of an organization which authoritatively direct their people. Finally, as we move into larger and more decentralized entities, more self-published material becomes appropriate. To illustrate with an extreme example, you can't say that any material written by a human should be excluded from the human race article because it is self-sourced. :-) More typically is info from a Wikipedia Signpost article about the WMF to be excluded as "self-sourcing" by/ within Wikipedia? North8000 (talk) 20:11, 13 February 2020 (UTC)
You probably mean articles about, for example, some college, which are based predominantly on the information from the college's web site? I would agree with that, but it seems Guy meant something else. As far as I understand that, his proposal was more about a controversial information. In connection to that, I would propose a small modification:
"Self-sourcing should be used only for adding a non-controversial content. Otherwise, SPS should be used only to add some detail to the content obtained from independent reliable secondary sources. Content along the lines of "$SUBJECT states $OPINION (source, $SUBJECT stating $OPINION)" should be avoided, especially if they are predictable reactions to criticism."
--Paul Siebert (talk) 00:11, 14 February 2020 (UTC)
I disagree with this... neutrality says we should give ALL significant viewpoints... and when there is controversy about a person or group, The opinion of that person or group as relates to that controversy is de facto significant. The key is to provide in text attribution. Let the reader know not just what the various opinions are, but who holds them. Blueboar (talk) 01:00, 14 February 2020 (UTC)
No, that's not how Wikipedia works. We include that which is verifiable from reliable independent secondary sources. Going to the subject's self-published materials and finding all the places where they dispute the independent view of what they do is WP:OR and, if reliable sources say one thing about them and they say another, WP:UNDUE. Newspapers give the last word to the subject ("X denies being a racist"), so we can cite that from the third party source. But if the third-party sources don't bother with the denial for whatever reason? We shouldn't either. We're not arbiters of fact, we don't know whether it's excluded because it's obvious bullshit, for example. Guy (help!) 09:16, 14 February 2020 (UTC)
@JzG: that is not original research, because the material was published (SPS). However, that is definitely WP:UNDUE, because an SPS writing about self is, by definition, a single source, and, if the same information cannot be found elsewhere, that view is by definition fringe. However, if the same information is reproduced elsewhere, that information is not a fringe view anymore, but that immediately changes a situation to the one described by you: the same content may be drawn from reliable independent secondary sources. That makes usage of SPS redundant.--Paul Siebert (talk) 17:25, 14 February 2020 (UTC)
It seems you mix "significance" and "importance". Obviously, under "significant", NPOV means not any important claim, but the majority or significant minority views. If just a single source claims that some asteroid will hit the Earth tomorrow, that information is important, but is it significant per NPOV? No, if other sources do not support that claim. A viewpoint expresses by an SPS about themselves is by definition an opinion expressed by a SINGLE source, so it can hardly be considered even a significant minority. Therefore, there is no contradiction with NPOV here. However, if the same information is confirmed by some reliable secondary source, it may be considered at least a significant minority view, but in that case, IgZ's inclusion criterion is met. Actually, what I am proposing is a softer version of what JzG proposed: he proposes to allow SPS only as a source of supplementary information about themselves, whereas I propose to allow usage of SPS as sources about themselves if the topic is not controversial.--Paul Siebert (talk) 01:23, 14 February 2020 (UTC)
The section already states, "Self-published and questionable sources may be used as sources of information about themselves, usually in articles about themselves or their activities, without the self-published source requirement that they be published experts in the field, so long as: the material is neither unduly self-serving nor an exceptional claim; it does not involve claims about third parties; it does not involve claims about events not directly related to the source;there is no reasonable doubt as to its authenticity; and the article is not based primarily on such sources."
And WP:Primary is clear about the types of sources articles should usually be based on (as in generally consist of).
I don't see what more needs to be stated in the WP:About self section. Flyer22 Frozen (talk) 02:04, 14 February 2020 (UTC)
Frankly, I myself don't see a really urgent need to add anything, but this:
"Self-published and questionable sources may be used as sources of non-controversial information about themselves"
could be an improvement. I agree that "the material is neither unduly self-serving nor an exceptional claim" means essentially the same, but from my experience I know that many people do not read policy carefully, and they prefer just to take some sentence out of context.--Paul Siebert (talk) 02:16, 14 February 2020 (UTC)
Flyer22 Frozen, it needs to be included because of the endless shit fights we get in articles like Knights of Columbus (edit | talk | history | protect | delete | links | watch | logs | views) where it took months of argument and an eventual topic ban to get an editor to stop sourcing endless self-serving trivia from affiliated websites. Consider a controversial university, for example. Should a third of the article be drawn from its own website's lists of sporting achievements, course lists and the like? You can (and people have) interpreted ABOUTSELF as permitting exactly that.
Paul Siebert, I think your proposed wording leaves a gap you can drive a coach and horses through. You could turn it round, though: Self-published and questionable sources may only be used as sources of neutral and non-controversial information about themselves. Otherwise we'll end up arguing on Talk whether marketing claims are "controversial" or not, for example. Alternatively:
Self-sourcing should be de minimis and should not be used for any claim likely to be controversial or challenged
That gets to the heart of it, which is that an article where most of the sources are the subject's own or affiliated websites, is doing something badly wrong, and this would be equally true if it were 25% or 15% of the sources and content. The default must be independent sourcing, and the qualified exception for self-sourcing is not designed to allow a subject's self-image to dominate over that of independent views. Guy (help!) 09:31, 14 February 2020 (UTC)
To illustrate one point, if the President of XYZ organization says "our board of directors has determined that our official position is that the moon is made of green cheese" is that uncontroversial information on what their position is? I think yes, and that that is important information about the XYZ organization, and that that is an important distinction. North8000 (talk) 13:38, 14 February 2020 (UTC)
I agree. I read "controversial" to apply not to the content of the claim but how reliably we can say the claim really reflects the views/opinions/etc of the subject. Springee (talk) 14:07, 14 February 2020 (UTC)
Let me give a wholly hypothetical example: say a business comes under fire as being a front for prostitution, the bulk of the sources talk about it like that, and don't cover the fact that the business, in its claims and validated through, say, business license documents, that it is a modeling agency and refutes the prostitution front, but no one is covering that. The current trend on WP, which is to make any entity or person that is seen in a poor light to have their article become a checklist of every negative thing that can be said about that entity/company, would completely ignore the "modeling agency" side and outright state, in WP's voice, the business is a prostitution ring. There is zero harm, in a case like this, to say "Business XYZ purports to be a modeling agency,(source) but it is considered to be a front for prostitution.(source list)" Now, in reality, I would be really surprised for legit journalism to not note the modeling agency facet, but modern journalism , particularly the writers often involved with the entities and people within the scope here, often omit basic tenets of traditional journalism , and those base claims/facts often get lost so that the writers can get to finger pointing. WP is absolutely not in that same business, which is why we have some common sense allowances for ABOUTSELF, or YESPOV, or similar. But, as I mentioned above, this is a very context-depend situation where ABOUTSELF should be considered. --Masem (t) 14:46, 14 February 2020 (UTC)
Yes, and that example is where the accusation is of an objective fact. In the modern post-journalism era, ( :-) ) the topic can be a creation of the writer. E.G. "some say that the XYZ organization is dog haters because they failed to sponsor a float in the dog parade". In short, the newspaper is acting as a participant, not a coverer. I would think that an official response by the XYZ organization should be include-able. North8000 (talk) 16:05, 14 February 2020 (UTC)

@JzG:@Springee:@North8000: Let me explain my point again. My major idea is that sourcing criteria cannot be absolutely uniform for such articles as, e.g, The Holocaust or Uncertainty principle and, such articles as, e.g., Cheerleading or Troy High School. Obviously, for lower importance and/or less controversial topics, sourcing criteria may be loose (of course, they are still should be minimally allowable by the policy). Otherwise, some articles could be impossible to write.

It is impossible to define in advance if some topic is controversial or important. However, that still can be done. If some users object to some text, and their objections are logical and based on good quality reliable sources, then we can speak about a controversy. In other words, a properly sourced edit is deemed non-controversial until no properly sourced objections have been presented. If some statement has been contested, and the criticism is well sourced, best available sources should be used to resolve a dispute, and, accordingly, usage of SPS should be reduced to the lowest possible level. The advantage of that approach is obvious: we do not need to write a separate rule for every case or every type of sources. Instead, it literally says: "you guys can do whatever you want until there is no content disputes, however, in a case of a serious content dispute, sourcing criteria automatically become stricter." That my proposal fully addresses Guy's concern, for his major idea seems not to limit SPS usage in all cases, but to limit their usage in controversial cases. However, my proposal also addresses a broad spectrum of problems, for, from my experience, lion's sgare of content disputes or edit wars develops according to a standard scenario: each party is pushing their own lousy source and rejects lousy sources that is being advocated by their opponent. In a situation when any edit war immediately makes sourcing criteria more stringent, majority of edit wars could be avoided.

@Masem: "Business XYZ purports to be a modeling agency,(source)" is by no means a controversial claim. Only a claim "Business XYZ is a modeling agency" is controversial. --Paul Siebert (talk) 17:08, 14 February 2020 (UTC)

  • In an article about a person or group, that person’s or group’s view of themselves is hardly UNDUE. Blueboar (talk) 18:22, 14 February 2020 (UTC)
    • Well, again, this depends on context. There's a recent BLP case of a running candidate for office who some editors felt was being shunned by the press such that her views in certain political areas were not being reported at all, and editors wanted to turn to her social media and campaign pages to include. This is where "self-unduly" comes in, as its not WP's job to speak where sources haven't spoken at all. But the cases I see where SELFPUB is fine is where the media has commented in the same area of material (the purported intent of an organization, the political beliefs of a person) and we're using SELFPUB to put what the entity/person counters or asserts instead. There is a huge grey line here, some cases where MANDY applies, others where it might not, but my caution is trying to wholly disallow SELFPUB. --Masem (t) 18:27, 14 February 2020 (UTC)
Let's think logically. What is considered a fringe view? It is an idea that departs significantly from the prevailing views or mainstream views in its particular field. When some SPS says something about itself, three scenarios are possible:
1. What this SPS says is in accordance with mainstream or significant minority views. That happens when some good secondary sources are available that say essentially the same;
2. The information presented in the SPS is not supported by other sources, but it is some non-controversial claim, and it is unlikely that it will be challenged;
3. The information presented in the SPS is some exceptional claim that contradicts to what other sources say.
The scenario #1 is exactly what JgZ's amendment describes, this SPS may be used per JgZ as an optional source.
The scenario #2 is an example of a non-controversial usage of an SPS, which is allowed per my proposal (which is more liberal than JgZ's one). However, if some user expresses a concern about that content and presents some sources, this case immediately becomes a scenario #3.
The scenario #3 describes the case when the SPS may contain a really important information, but, since it is, by definition, a single source, that automatically makes it fringe. (It is fringe, because it contradicts to what all other sources say, otherwise that would be a scenario #1).
@Blueboar: you continue not to understand. The person's opinion about themselves is obviously important, however, in a situation when that opinion directly contradicts to what all other sources say, that opinion is, by definition, a fringe opinion, so that SPS cannot be used. In a situation when what the SPS says is supported by other sources (better secondary sources), that SPS cannot be considered a fringe source, but it is redundant, and it can be used just as an auxiliary source.--Paul Siebert (talk) 18:52, 14 February 2020 (UTC)
When it comes to FRINGE, it should be when the media and a person are talking a wholly separate topic not related to that person, outside of criticism of that person's stance on that topic. That is: if some person has a wacky diet ("Eat only Twinkies! The sugar will help metabolize your fat cells!") and the media is critical of this diet and no one else supports that, that's clearly FRINGE. But as soon as the discussion is about the person themselves, it is wholly inappropriate to toss out statements made only by the person that are not reflected by the media covering it as FRINGE in that sense. If we are to be impartial, it may be necessary to include those statements briefly. Per UNDUE we don't need to give that more than a phrase or sentence if only SELFPUB is being used, but inclusion should be there for the person to "defend" themselves, depend in context like the MANDY issue. --Masem (t) 19:17, 14 February 2020 (UTC)
That looks like a straw man argument. Obviously, when some person says in their own SPS, that they are eating only Twinkies, and no other sources say otherwise (about t hat person), that SPS can hardly be considered fringe: if this topic is not important, it would be ridiculous to expect that other sources will deserve even a minimal attention to that. That case falls under a scenario #2. However, if, so some reason, that topic is important, and secondary source agree that that person has a normal diet, but the person themselves says in their SPS that they are eating only Twinkies, this view should be considered fringe. And that is a scenario #3.--Paul Siebert (talk) 19:57, 14 February 2020 (UTC)
Paul, where is your proposal that you were referring to under #2? Whatever it is, I think "more liberal than JgZ's one" is an understatement because IMO JgZ's is pretty extreme in that it would flatly preclude inclusion in that case. North8000 (talk) 20:49, 14 February 2020 (UTC)
Sorry, I was, probably, not clear enough. I propose a general principle, according to which the WP:V rules are interpreted liberally until some concrete user presents a justified criticism of some statement/sources. It other words, SPS may be used about themselves, until the content has not been contested by some user. If some SPS says "Mr X (an author) is eating only Twinkles" (let's continue using Masem's example), and other sources say nothing about that (for example, because that fact is some minor aspect of Mr X's life), that is perfectly ok. However if Mr X's diet is a subject of controversy, and secondary sources do not support Mr X's claim, that material becomes controversial. Per my proposal, anybody who finds a secondary source that contradicts what SPS says can remove that SPS, for it cannot be used to support controversial claims.
In other words, "Self-sourcing should be used only for adding a non-controversial content" discriminates between controversial (scenario #3) and non-controversial (scenario #2) cases.--Paul Siebert (talk) 21:07, 14 February 2020 (UTC)
I think that it is good proposal. I think it needs bit more emphasis/clarification that "controversial" means with respect to veracity of the material itself. So "I doubt that he eats only twinkies" is grounds for a sourcing-based removal, "I don't want his stated twinkee preference to be covered" is not grounds for a sourcing-based removal.North8000 (talk) 21:28, 14 February 2020 (UTC)
Not exactly. In reality, when some not-really-good-faith user does not want twinkee preference to be covered, that user may say "I doubt that he eats only twinkies". Therefore, that argument is hardly acceptable. However, an argument "I doubt that he eats only twinkies, because that contradicts to what the source X says about him", or "I doubt that he eats only twinkies, because the sources X, Y, and Z says it is impossible" are quite acceptable arguments.--Paul Siebert (talk) 21:48, 14 February 2020 (UTC)
Paul, I've not seen not-really-good-faith users do that. But you broke with the analogy. They'd need to say "I doubt that he said that he eats only twinkies" in which case they would sound silly. Much easier to deliberately mis-read your policy (if not clarified) & say "it's controversial because it's self-sourced. So out it goes." North8000 (talk) 22:24, 14 February 2020 (UTC)
@North8000: if a policy says that SPS are acceptable for that purpose, the argument "it is controversial because it is PSP" is not working.--Paul Siebert (talk) 04:28, 15 February 2020 (UTC)
Paul Siebert, that largely already happens. The issue at the moment is more about the de minimis part. We say that the majority of an article shouldn't be self-sourced, but people argue over whether that means 40% or 30% can be self-sourced. In practice, self-sourcing should only be used for simple things like adding an exact date where RS only have the year. We should not be listing all the things a person or organisation does, sourced to their pages promoting those things. Any statement of the form "X says " sourced to X saying Y on their own website is placing Wikipedia editors in the position of arbiters of both fact and significance. Guy (help!) 22:14, 14 February 2020 (UTC)
Not arbiters of fact. Assuming that the statement “X says Y” is verifiable by a Self published source where X actually does say Y, that source (where X actually does say Y) is the MOST RELIABLE source possible for that statement. Blueboar (talk) 22:56, 14 February 2020 (UTC)
@JzG: this and that are both SPS. Do you really think they are equally reliable. The question is not in 40% or 60%, that approach is intrinsically flawed. All of that is content dependent, and the criterion should be other user's objections. If, for example, you see that some article is using too many materials from SPS, you may request to replace them with secondary sources, or to remove that content at all. In the latter case, that can be done if you explain why there is a reason to expect that that material is not trustworthy. If our policy will allow us to do that, that is all what we need.--Paul Siebert (talk) 04:28, 15 February 2020 (UTC)
@Blueboar: MOST RELIABLE is the source with the greatest degree of scrutiny given to the issues, and with the most "professional structure in place for checking or analyzing facts, legal issues, evidence, and arguments". Unless that has been proven, we cannot say anything. And, you totally ignore a possibility that some obscure SPS may just lie about itself.--Paul Siebert (talk) 04:28, 15 February 2020 (UTC)
Not for attributed statements about what X said. For that, the most reliable source is the Primary source where X actually said it. The fact being verified is “X said Y”... not “Y” itself. To more clearly illustrate: suppose X writes “the moon is made of green cheese” on his website. Now, we can not cite X’s website to say “The moon is made of green cheese”... X is not reliable for this (unless X is a noted lunar expert)... BUT, if we say “X claims that the moon is made of green cheese”, then X’s website is 100% reliable. We are not stating that the moon is made of green cheers (as a fact), we are stating that X makes this claim (as an opinion). What we have to verify is that X actually made the claim, and to verify that X made the claim, the website where he states it (the primary source) is the single most reliable source we can find.
Now, the fact that X said this may or may not be significant. That determination depends on context and which article we are working on. In an article on the moon, what X thinks is probably not worth mentioning (again, unless X is some sort of lunar expert). However, in the Bio article on X, his views on the moon might well be highly significant. Especially so if others criticize his views. Blueboar (talk) 15:52, 15 February 2020 (UTC)
Yes, arbiters of fact and significance. We decide what is true and should be self-sourced and what is bogus or trivial. Whereas, per policy, we should be leaving that to third parties. Look at Liberty University. Roughly a fifth of the sources are the university's own website or the student newspaper, which is entirely controlled by Falwell's office. Pretty much everything about Liberty is inherently controversial. We're using it as a source for things like Melania Trump getting a standing ovation - why? That makes no sense. They'd cheer a hatstand if it was associated with the Trump family. We're using it for marketing claims like the number of countries represented among the student body. Why? We should not be including marketing claims. Guy (help!) 07:19, 15 February 2020 (UTC)
Witness that Melanoma Trump is, after all, a walking, talking coatrack. EEng 11:27, 15 February 2020 (UTC)
Which reminds me of WP:MISSION that is not policy but is about addressing a very common problem (another recently fixed instance). We see that type of promotion everywhere... —PaleoNeonate – 11:13, 15 February 2020 (UTC)
Now that we're down to asserting coverage by secondary sources to be the final arbiter of whether it deserves inclusion in Wikipedia, I think we bumped into a bigger more fundamental problem. If we ignore that concept, we'd get articles full of crap, PR department work and bias-by-overpresence. If we went with that concept 100% we'd delete an immense amount of encyclopedic information and articles, and the balance of articles would be determined by wiki-lawyering. The operative core wp:weight part/answer to that of the core policy wp:npov is unusable with respect to "two sides" issues ("in proportion to the prominence of each viewpoint in the published, reliable sources") and non-existent for mere-inclusion issues. North8000 (talk) 13:51, 15 February 2020 (UTC)

@JzG: Liberty University is a good example, let me continue with it. If my proposal is implemented, any materials taken from university's web site or university controlled newspapers are allowed, but thay can be removed if someone provides serious arguments demonstrating that that material is controversial. Thus, in this particular case, if you can support you words "pretty much everything about Liberty is inherently controversial" (I believe, you have all needed evidences), the modified policy would allow you to remove that content, and noone would be able to restore it. However, that would be done not based on some purely formal criteria, but because that SPS is controversial. Why that is important? Because materials from university's web sites are usually not inherently controversial. Thus, I doubt materials from Harvard University web site are controversial. Actually, I can imagine a situation when that could be controversial: for example, in the article about some scandalous event in Harvard. However, that is more an exception. The idea you are advocating is not good because it is based on formal criteria that are equally applicable to a wide range of sources, each of which may be of totally different quality and reliability. And that is a fundamental flaw of the policy, which makes no difference between, for example, Nature and Indian Journal of Pharmacology. According to our policy, both of them are peer-reviewed publications, but in reality, a difference in the level of publications in these two journals is greater than the difference between Harvard and Liberty universities.

However, due to unhomogeneity of Wikipedia, we have to use all types of formally reliable sources (otherwise, some articles will be impossible to write). To do that efficiently, we need a flexible mechanism that would allow us to remove some formally acceptable sources in cases similar to the one provided by you. And we will never achieve this goal by inventing additional ad hoc formal criteria. The approach should be totally different, similar to what I propose: lower quality sources (which minimally fit our RS criteria) should be allowed until some user provided serious counter-arguments. The counter arguments may be either that the source is too controversial (with proofs), or that the quality of that source is significantly lower than the quality of other sources in that concrete article. Again, as I already explain, that approach may allow us to quickly finish lion's share of all edit wars.--Paul Siebert (talk) 02:02, 18 February 2020 (UTC)

I've just noticed North8000 brilliantly explained the same problem in somewhat different words. --Paul Siebert (talk) 02:07, 18 February 2020 (UTC)
  • I’m still fundamentally opposed to this. It is when a subject is controversial that we MOST need to present what the subject says about themselves. That is basic NPOV. We don’t present what they say as being accurate... but we do need to present what they say. The way we do this is through in-text attribution. We don’t give it UNDUE weight, but we DO give it SOME (minimal) weight.
of course there are limits to this... First and foremost, Outlining the subject’s opinions about themselves is often limited to the article about the subject (Bio article when the subject is a person, or the equivalent when the subject is a company/entity). That’s fine (we don’t put Alex Jones’s 9/11 conspiracy theories in the article on 9/11... but we do summarize them in the article on Alex Jones.
And... when outlining what someone says about themselves, the best way to verify that they said it is to cite the primary source... their own self published source. This is WHY we created WP:SPS in the first place. Blueboar (talk) 02:46, 18 February 2020 (UTC)
It seems that you are opposing not to my idea, but to your own interpretation of it. I already wrote that, despite the fact that this policy is called "verifiability", it combines two separate concepts, "verifiability" sensu stricto and "reliability" (i.e. quality of the source). These two things are totally independent. In my opinion, anything that have been made publicly available on a permanent basis meets the "verifiability" criterion. In that sense, any SPS is verifiable if it cannot be easily removed from a public access, or easily altered. Therefore, any university web site can be used as a verifiable source about itself, but, in some cases, this information is not reliable for the statement of fact. In controversial cases, it can be used not as a statement of fact, but as an opinion (if I understand it correct, that is exactly what you say). In other words, it should be perfectly ok to use such sources for statements like "a university X claims that Y". However, and I believe you agree with that, a really neutral article about a university Y cannot be based mostly on claims taken from its own web site. To prevent that, we should not write a separate policy and stipulate that not more that X% of information can be taken from SPS. Even an attempt to stipulate reduction of such materials to a reasonable minimum will not work, for every user has their own vision of reasonability. Instead, we should stipulate that such materials are considered as RS, but any user has a right to remove them if some adequate rationale (explanations + sources) for removal is provided.
In other words, using programmer's slang, instead of setting some "static" criteria, they should be made "dynamic".--Paul Siebert (talk) 04:51, 18 February 2020 (UTC)
Paul Siebert North8000's argument makes perfect sense if we think we're a directory that is supposed to cover the list of courses that Liberty runs. We're not. We're supposed to cover a subject entirely by reference to reliable independent secondary sources. ABOUTSELF is a qualified exception, but has morphed over time into a default for4 self-serving content that RS don't bother to cover, leading to arguments over a wide range of subjects centring on how much of an article can be self-sourced. Right now we imply up to half. That's crazy.
Blueboar I think you are wrong. When a subject is controversial we absolutely must stick with what reliable independent sources say. Should we defer to the Catholic Church's view of its handling of sex abuse cases? Of course not. Should we defer to homeopaths when describing CEASE therapy? That would be insane. Wikipedia is not a newspaper. The subject has no right to the last word. We adhere to what reliable independent sources say, and if they all say the subject is wrong, we don't "balance" that by quoting the subject saying they are right, because of course they say that. Do you think we should be using David Duke's websites to represent his point of view? Why would we send people to a neo-Nazi website when we already have secondary sources for what he says? He may be the most reliable source for how he wishes to be represented but he is the least reliable source for any facts, because everything he says is self-serving. And that applies to every controversial subject. Scam journals, promoters of quackery, degree mills, pyramid schemes, all present themselves as legitimate businesses unfairly discriminated against by The Man. And by deferring to that self-image, we violate WP:NPOV.
Go back to WP:V: "Articles must be based on reliable, independent, published sources with a reputation for fact-checking and accuracy.". What's being argued here is that the word "independent" should be ignored, and that self-published should be allowed as an alternative to published sources with a reputation for fact-checking and accuracy. That's a pretty fundamental repudiation of canonical policy. It means essentially discarding WP:RS, which says that "Articles should be based on reliable, third-party, published sources with a reputation for fact-checking and accuracy. This means that we publish the opinions only of reliable authors, and not the opinions of Wikipedians who have read and interpreted primary source material for themselves." In other words, we defer to independent sources rather than quote-mining the primary source in order to reflect a subject's views on anything.
That's what Wikipedia is supposed to do. What's being argued here is that the qualified exception for ABOUTSELF should be interpreted so broadly as to make RS effectively meaningless. Guy (help!) 09:52, 18 February 2020 (UTC)
Guy, you seem to be blending covering their views vs considering them as a source for the "facts" which they claim. If you include self-sourced "John Smith said that the moon is made of green cheese", the statement is about what John Smith said, not a statement about the contents of the moon. John Smith is reliable about the former and not the latter. North8000 (talk) 13:05, 18 February 2020 (UTC)
Exactly - for “Smith said” or “according to Smith” type statements, what needs to be verified is the fact that Smith said it ... and THAT is best done by citing the self-published primary source.
There are LOTS of restrictions on when it is appropriate to mention what Smith said... but when we do mention it, it is appropriate to cite the primary source. Blueboar (talk) 13:44, 18 February 2020 (UTC)
And to this end, when it is appropriate to use, is case-by-case. "Unduly self-serving" applies - we're not going to include material sourced this way just because we can, but only when it relevant to what context has been established through other reliable sources, including criticism directed towards that person. MANDY would still apply to certain basic accusations but not universally. When to include is a consensus-discussion that needs to happen, but it is important that the self-pub sources are not immediately thrown out, else that give no discussion. And I'll go back to that when SELFPUB is used, UNDUE must be also weighted closely; the Twinkie diet example would be reasonably to include a one sentence summary of the gist of it, but not a multi-paragraph "scientific" explanation of it , presuming no RS otherwise has covered it.
A lot of this comes down to that the bolded statement above in Guy's reply is that based word. It does not mean we exclusively use independent sources, but that I should the majority (like, 90% or more) to use independent sources, so that the article is based on them. That allows excursions to dependent sources when necessary to provide the comprehensiveness that is required from NOR and NPOV for those few details that tend to get ignored in the more common media. For example, we frequently dip into SEC filings to estimate the number of employees a publicly traded business has when that business is already readily covered in independent sources. SELFPUB has limitations and definitely when self-promotion is a concern, but SELFPUB is like spot-patching a paint job, filling in the spaces not already covered by the independent reliable sources when appropriate. --Masem (t) 14:05, 18 February 2020 (UTC)
Sure... but Guy’s proposal is to not allow self published sources when the subject has said something controversial. That is wrong. I agree that we need independent sources to establish that the controversy exist, and rises to the level of significance that we mention it in the first place... but once that is done, we can (and should) rely on primary self-published sources to present the subject’s own views on that controversial topic. We don’t say (or even imply) that the subjects views are accurate, but we should present their views. Blueboar (talk) 15:37, 18 February 2020 (UTC)
I agree with both of you, and think that your posts describe good practice that is within the letter and intent of current policy. I think that Guy's proposal would reduce or eliminate that, and I'm not in favor of such a proposal. North8000 (talk) 16:24, 18 February 2020 (UTC)
I am definitely in agreement with Blueboar and N3k and think better effort can be had by creating a guideline or essay of when it is and isn't appropriate to include SELFPUB, and when included, to what degree. --Masem (t) 18:19, 18 February 2020 (UTC)
I also agree with BB, N8K and Masem and think the essay idea makes a lot of sense. I do get Guy's concern regarding articles that seem to draw too much from SPS even if the specific claims aren't controversial but I would rather err on that side of inclusion vs excluding SPS responses to controversies etc. Springee (talk) 19:50, 18 February 2020 (UTC)

I have been threatening to talking about creating Wikipedia:Based upon for years, and now is apparently the time. I have followed the idea behind WP:LIKELY, under the belief that we benefit from having a shared understanding of what these common words, used across multiple policies, mean on wiki. In this case, I think we need to have a shared understanding of what it means for an article to be based upon something. For example, if most of the article's content comes from self-published sources, then it's "based upon" self-published sources. Or whatever. Breast cancer awareness cites more scholarly books than anything else, so it is "based upon" scholarly books. Schizophrenia cites mostly medical journal articles, so it is "based upon" medical journal articles (and its structure is "based upon" WP:MEDMOS's suggested order). It's really not as complicated as some people might make it out to be. If people are basically satisfied with this explanation, then please feel free to link it wherever it might be helpful.

User:JzG, I have not adopted your suggestion that an article can't be "based upon" independent sources until 90% or more of the content comes from them. This is probably a desirable goal for articles about large businesses or major politicians, but I don't think that it is practical for short articles (e.g., most notable academics). So I have proposed a simple "majority" as the bare minimum for calling something "based upon" a particular source type. If you want to propose a higher standard for articles (e.g., "nearly all of their content is taken from"), then that means RFCs at the core policies, to change the core policies.

I predict that most editors will think that this is all perfectly fine, until their favorite oxen get gored. If we're serious about basing articles upon independent sources, then that limits our ability to create articles about "neglected" subjects (such as academics). Again, the place to relax (or to raise) that requirement wouldn't be at WP:Based upon; it would be at WP:V, WP:NOT, and WP:NOR. This page reflects and explains the requirements that have been in the core policies for years. It does not create any new ones. WhatamIdoing (talk) 04:23, 19 February 2020 (UTC)

Wikimedia Project Grant Proposal on *Disinformation*Edit

I'm proposing a Wikimedia Foundation Project Grant to study *disinformation* and provide actionable insights and recommendations.

Please check it out and endorse it if you support it.

Meta:Grants:Project/Misinformation_And_Its_Discontents:_Narrative_Recommendations_on_Wikipedia's_Vulnerabilities_and_Resilience

Cheers! -Jake Ocaasi t | c 20:09, 19 February 2020 (UTC)

Recent changesEdit

This revert was presented as a restoration of an old consensus version. However, as far as I understand, the reverted change was made year ago, and I am not aware of any objections. That means the reverted version can be considered a long standing consensus version too.

I am not going to revert Visviva for two reasons. First, I think, BOLD should not be applied to policy page: if a possibility of revert cannot be ruled out, it is always better to discuss the change on the talk page. The policy is something that is supposed to be stable, so the less changes, the better.

Second, what Visviva is saying is correct: the version reverted by Visviva "would justify deleting >90% of the encyclopedia". However, that is only a part of truth. The second truth is that these 90% of Wikipedia content is not the most precious part of it. In reality, it is remaining 10% (i.e. articles like World War II 30,000 views per day, Global warming 14,000 views daily, or United States 45,000 views daily), which make Wikipedia a respectable and trustworthy resource.

In other words, Loose sourcing criteria allow Wikipedia to grow rapidly, to cover a broad range of subjects, and to recruit myriads of amateur editors. However, the very same liberal sourcing rules make creation of a top quality content much more difficult, and I suspect that is a reason why many professional experts are not too enthusiastic about editing Wikipedia.

We have a dilemma:

  • We cannot apply too stringent sourcing criteria to all Wikipedia articles, because that will lead to deletion of a significant part of the content. Moreover, almost every article starts as a poor quality stub article, and only after some time it may (or may not) become a good article. By deleting poorly sourced content, we may eliminate a possibility of a future growth of Wikipedia.
  • However, the very same loose sourcing criteria that make Wikipedia growth possible may have a very negative impact on good articles. Indeed, the sourcing rules applied uniformly allow anybody to add some material from a local newspaper to some good article that is bases upon good scholarly articles of university textbooks. It may take enormous time and efforts to purge good articles from poorly sourced content (thus, one admin seriously proposed me to start RfC to remove a garbage source from an article that was based on monographs and scholarly articles).

All of that means that the policy should set floating criteria for sources. These criteria should allow us to use marginally acceptable sources (local newspapers, magazines, movies, etc) for low importance or stab articles, and to prohibit usage of questionable sources in high importance/good/featured articles, or in the articles that are based upon, for example, scientific articles published in peer-reviewed journals. Any thoughts?--Paul Siebert (talk) 20:22, 19 February 2020 (UTC)

Keep in mind, outside of select case of policies with legal implications (BLP, COPYVIO, NFC), P&G are to be descriptive and not prescriptive. To that end, neither version before after said diff are good as both are "hard" instructions. I am terrible at wordsmiting these but something along the lines of "Articles should be based primarily on reliable, independent sources." would be the proper descriptive wording of practice. --Masem (t) 20:49, 19 February 2020 (UTC)
See also Wikipedia talk:Verifiability/Archive 67#Must and should and Wikipedia talk:Verifiability/Archive 68#Unresolved discussion; the change was contested at the time. Also, if you check WP:Based upon, you'll find that multiple policies say something like this, and whatever we end up with, we should have a solid shared understanding of what we actually mean, and that understanding should be consistent across all of the pages. Whatamidoing (WMF) (talk) 15:21, 20 February 2020 (UTC)
While it's certainly true that various pages like WP:IS that are prominently marked as not-policy have used "must" language in this context, I don't see what bearing that has on WP:V. Any reader will understand that those pages do not serve as authority, and cannot substitute for an actual argument for deleting content (or entire articles). In contrast, Verifiability is generally considered to apply to all content without meaningful exceptions (and is often cited even as an exception to the mother of all rules). Essays and supplements can do their own thing, but major substantive changes to WP:V should not be made without substantial and widespread community input. -- Visviva (talk) 16:22, 20 February 2020 (UTC)
User:Visviva, I believe you will find that the "must" language is also present in WP:NOT, which has been a policy for a very long time. It's all very well to say that you don't care about non-policies, but this language exists in other core policies, too. WhatamIdoing (talk) 18:58, 20 February 2020 (UTC)
I can tell I've been spending too much time with the shouty-acronyms crowd on AFD lately, because I was deeply tempted to respond with OTHERSTUFFEXISTS. That is an interesting point; indeed WP:NOT is fairly awash in "must" statements, which nobody seems to take very seriously. But I would say that there's still a big difference between a policy like V that has always been understood to lay down a single inviolable rule that all content must follow, and a policy like NOT that is basically a laundry list of more-or-less-negotiable boundaries. -- Visviva (talk) 07:16, 21 February 2020 (UTC)
FWIW, it seems to me that hortatory language is exactly right here; the "Base articles on" language makes clear that it is describing preferred user conduct. In contrast, language of obligation, especially "must," is dangerously prone to misinterpretation. (If anyone doesn't believe that people turning policy upside-down to create rationales for deleting encyclopedic content is already a problem, check out how many AfD participants think the GNG is a biconditional.) TBH, I don't understand what actual problem the change was intended to solve; assuming good faith, it appears to have originated from a couple of editors attempting to make a stylistic tweak and accidentally making a massive change to the substance of WP:V, which is absolutely not how such an important page should be modified. -- Visviva (talk) 16:22, 20 February 2020 (UTC)
I find your use of the word “hortatory” quaint. Most people would say ”bordello” or ”house of ill repute”. EEng 16:59, 20 February 2020 (UTC)
wikt:hortatory? WhatamIdoing (talk) 19:01, 20 February 2020 (UTC)
Think about it. EEng 20:14, 20 February 2020 (UTC)
I would strenuously dispute the second point. Wikipedia's high-traffic articles are IMO the weakest argument for the project's importance, because they mostly just cover material that could be perfectly well handled by a latter-day World Book Encyclopedia. Lots of people look at those articles because lots of people search for those terms, but if we disappeared tomorrow there would be plenty of other resources to take our place. Wikipedia's real value proposition is in the long tail. The unfortunate tendency of policy discussions to focus on a handful of unrepresentative high-profile or controversial articles is IMO a classic example of how hard cases make bad law. -- Visviva (talk) 16:22, 20 February 2020 (UTC)

The common sense version that is used in reality (except when there is wiki-lawyering) is that when the material is more questionable or questioned, stronger and more "by the book/ideal" sourcing is required, and when the material is less questionable or questioned, weaker or less "by the book/ideal" sourcing is OK. Since 100% "by-the-book / ideal" sources are maybe about 10% of the sourcing in Wikipedia, this practice is very important, and IMO any wording that would tend to preclude it is not a good idea. North8000 (talk) 19:16, 20 February 2020 (UTC)

Visviva, I think what would be most helpful is if you explained what you think would be appropriate. This sentence is about using sources with two qualities: (1) independence from the subject, and (2) a reputation for fact-checking and accuracy.
It says that articles should be WP:Based upon (at which link you will find my interpretation of what it means for any article to be 'based upon' anything) sources with those two qualities. For example, an article based upon sources with these two qualities would include any article about a business whose content is mostly cited to newspapers. An example of an article that is not based upon sources with these two qualifies would be any article about a business whose content is mostly cited to that business's own website.
There are basically three options available:
  1. We have an article on a subject, and the article is based on what the subject says about itself;
  2. We have an article on a subject, and the article is not based on what the subject says about itself; or
  3. We don't have a separate article on that subject (see WP:FAILN for options on how to include it "in Wikipedia" without being "a separate article").
What do you think Wikipedia should do? Can you think of any subject for which you think the best possible approach is to base our article on what the subject says about itself? Is there any situation in which you think it would be best to base a whole article on sources without a reputation for fact-checking or accuracy? WhatamIdoing (talk) 19:24, 20 February 2020 (UTC)
I would agree (and I think nearly all Wikipedians would) that articles should move iteratively from worse to better sourcing. I'd also agree that ideal sourcing comes from independent secondary or tertiary sources with a reputation for accuracy, and so forth. But I think the pre-existing language makes this expectation clear. The "must" language concerns/alarms me because instead of merely setting an expectation for user conduct, it significantly raises WP:V's minimum threshold. Specifically, by its plain meaning, the "must" language requires all articles to have gold-standard sourcing for at least the bulk of their content at all times. If enforced, this new requirement would dramatically raise the minimal criteria for article sourcing, and put all manner of perfectly good and useful articles in jeopardy. (I'm setting aside the problem that there tends to be considerable disagreement on which sources actually have the required reputation, so nobody could ever really be sure if a particular article would be safe from a WP:V-based deletion under this new standard.)
Wikipedia has succeeded -- to the extent it has -- precisely because the threshold criteria for contributing content have been low. People submitted whatever they could manage to put together, other people made whatever improvements they could, and the wiki process worked its magic. And for nearly the whole lifetime of the project, WP:V has stood for a consistent minimum standard that allows this process of iterative improvement to function: other users must be able to verify the contributed information somehow. The simple and low nature of that standard has been responsible for a good part of Wikipedia's success, so it shouldn't be altered without very good reasons, in addition to wide and deep discussion. Here, in contrast, there don't seem to have been any stated reasons for the change at all. And as you've noted, when discussion finally did occur, it didn't show much support for the change (I'd actually say it showed pretty strong and well-reasoned opposition).
Of course this new language has been relatively harmless for the relatively brief time it's been there, but I think that's mostly because hardly anybody reads policy pages beyond the first screen anyway. ;-) If somebody ever did read this passage and decide to take it seriously, they could do a lot of damage, since WP:V is after all non-negotiable. -- Visviva (talk) 06:46, 21 February 2020 (UTC)
Return to the project page "Verifiability".