Wikipedia talk:Verifiability

Freedom of press

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, requests that influential public figures make specific statements in support of the government on specific occasions, and so on. 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.

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.

References

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)

Proposed change to WP:ABOUTSELF

I explained my proposed change at WP:Village_pump_(policy)#Improving_precision_of_definition_of_WP:BLPSELFPUB_and_WP:ABOUTSELF. Please dicsuss there. Xenagoras (talk) 23:58, 11 January 2020 (UTC)

Infoboxes, in their various forms

Why are infoboxes, and the need that the information they convey also be verifiable, not covered in this WP guideline? 2601:246:C700:19D:1C66:B776:33B0:EE87 (talk) 15:26, 17 January 2020 (UTC)

Infoboxes are part of ledes which have a unique allowance that does not require citations as long as the material is clearly sourced in the body of the article, though still requiring them for direct quotations. Info in infoboxes should be replicated in the body, and if there's a field where it is not replicated that should be sourced. But not every line on an infobox needs to be sourced. --Masem (t) 15:32, 17 January 2020 (UTC)

What kind of information in an infobox wouldn't need to sourced and verifiable somewhere in the article? (Apart from sky colour parameter = blue). Agree it usually doesn't need to be cited. Scribolt (talk) 18:44, 17 January 2020 (UTC)

Non-English sources "preferred"

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)

? 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)

Break

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)

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)

Elaborated example

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. Vol. III. Translated by Bell, Clara; Fuller Maitland, John Alexander. Novello & Co. p. 28. {{cite book}}: External link in |volume= (help) 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; et al. 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 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). Vol. 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). Vol. 41. Bach Gesellschaft. Breitkopf & Härtel. pp. XXXIX–XL. {{cite book}}: Invalid |ref=harv (help) 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 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)

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). Vol. 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)

Some general thoughts inspired by NOENG and the Holocaust discussion

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)

No clear consensus/ discussion ongoing

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)

Recent edits

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)