Wikipedia talk:No original research

New articles based on primary sources

WP:PRIMARY currently says "Do not base an entire article on primary sources" but this does not conform to existing practice. Here's a couple of examples,

1. As discussed at WT:NSPECIES, species articles are routinely created without much in the way of secondary sources.
2. WP:PRIMARY also says that "For Wikipedia's purposes, breaking news stories are also considered to be primary sources" but numerous articles are created every day about breaking news such as natural disasters, political events, sports results and other topics which are routinely featured at ITN. For a fresh example, see 2024 Solingen stabbing which has a {{current}} banner tag to make it quite clear that it's breaking news and so quite unreliable.

So, the statement seems to be a counsel of perfection which doesn't correspond to what we actually do and so, per WP:NOTLAW, needs qualifying or softening.

Andrew🐉(talk) 17:50, 23 August 2024 (UTC)

At the time that the first sentence (Do not base an entire article on primary source) was added to the policy:

• the discussion on the talk page was about writing articles about books that were based entirely on the book itself (e.g., WP:NOTPLOT), and
• the definition of 'primary source' was much more restrictive than our current understanding.

The then-current definition of 'primary source' was:

• Primary sources are sources very close to an event. A primary source offers an insider's view of an event, a period of history, a work of art, a political decision, and so on. An account of a traffic accident written by a witness is a primary source of information about the accident. Other examples include archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; original philosophical works; religious scripture; published notes of laboratory and field experiments or observations written by the person(s) who conducted or observed the experiments; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.

Looking at the bit I highlighted, that rule, interpreted under that definition, treats breaking news as a secondary source so long as it's written by someone interviewing the witness, rather than by the witness themself.

I conclude from this that there was no intention to prevent the creation of articles about current events with the best sources we happen to have access to. Whether and how to fix it is probably worth a discussion, but I suggest that "fixing it by stopping people from creating articles about current events" is not going to be functional. WhatamIdoing (talk) 20:49, 23 August 2024 (UTC)

That explanation of the creepy way that this has mutated is enlightening, thanks. It's good that it's our policy to ignore all rules. Andrew🐉(talk) 21:46, 23 August 2024 (UTC)

I think the current belief is sort of:

• If you only have one primary source, and your single source has a particularly severe case of primary source-ness and no independence, then don't write that article. The Tale of Custard the Dragon by Ogden Nash is a lovely picture book, but you really need something more than just the book itself to write an article about that book.
• If you have a couple of sources, and they're pretty useful overall, maybe their primary source-ness is not exactly the most important quality to consider. For example, it's kind of unfortunate that when a big disaster happens, we only have breaking news to work with, but frankly, it doesn't take a WP:CRYSTALBALL to figure out that there will be proper secondary sources appearing later (and if we've guessed wrong, we can always delete or merge away the article later). Depending on exactly how you define secondary, we might even see some of that the next day. For example, one of the hallmarks of secondary sources is comparison, so if you see "This was a 100-year flood" or "This is the third biggest earthquake in this area during recorded history" (or, for the Wikipedia:Notability (species) proposal "this Sheltinack’s jupleberry shrub species is a more mauvey shade of pinky russet than the other species"), then the source is comparing it against past history, which could be argued to be secondary content, even if we might normally call the overall source a primary one.

The edit that added that "Do not" language also added this: Appropriate sourcing can be a complicated issue, and these are general rules. Deciding whether primary, secondary or tertiary sources are appropriate on any given occasion is a matter of common sense and good editorial judgment, and should be discussed on article talk pages. That's still in the policy, and I think it's important to remember that. WhatamIdoing (talk) 23:55, 23 August 2024 (UTC)

Using common sense and good editorial judgement seems to be the general idea of WP:IAR too. So we don't need all this other stuff then, right? Andrew🐉(talk) 06:17, 24 August 2024 (UTC)

The issue is that editors don't always agree on what constitutes 'common sense and good editorial judgement'. Relying on IAR alone would be a massive time sink of arguing over what exactly is an improvement to the encyclopedia. -- LCU ActivelyDisinterested «@» °∆t° 13:22, 24 August 2024 (UTC)

That passage's point is that the final shape of the article should not heavily rely on primary sources - an article in the early stage of development may likely be based on primary, but we expect that it should be able to be expanded with secondary sourcing as to otherwise meet the NOR aspect as well as notability factors. So we allow for species articles based on publication in scientific journals of their existance but anticipate more sourcing will come later. Similarly, breaking news stories will very likely use primary sourcing to describe the event, but to show enduring coverage as to meet NEVENT, more secondary sources need to be added over time. It is impossible to have a "finished" encyclopedic article based only on primary sources, but until the article has had time to mature with additional, it seems reasonable to allow primary sources to be the baseline. It should be stressed that WP:V's requirement about third-party sources must be considered here: an article based only on first-party primary sources, regardless of its state, has no business being on WP. — Masem (t) 12:38, 24 August 2024 (UTC)

I think the problem is the use of "do not" rather than being formulated around "should not". Like most of these things there are exceptions, but that doesn't mean the central point isn't valid. -- LCU ActivelyDisinterested «@» °∆t° 13:15, 24 August 2024 (UTC)

I believe that there should be scope for an article of the type “Evolution of the rules of <some sport>” which is essentially a catalogue of rule changes over the years. The main references would of course be the various rule books themselves. supplementary comments from reliable Secondary sources might be used to put the major changes into context, but minor changes to the rules which anybody could verify by comparing the two texts would merely be catalogued. 2A00:23C8:1DAE:2401:8933:B63A:8FD1:CF6 (talk) 13:05, 1 September 2024 (UTC)

This is fine… PSTS does allow us to cite primary sources for specific things. However, we do need secondary sources to establish that these rule changes are significant enough for WP to have a stand alone article about them. That is more a function of WP:NOTABILITY than of WP:NOR, but it is still important to do. Blueboar (talk) 13:32, 1 September 2024 (UTC)

It's good vague "goal" type advice. I wish we could just say that. Anytime someone tries to derive something more prescriptive out of it there are problems. Whether well-intentioned or using it as a weapon in a wiki-battle. North8000 (talk) 14:33, 24 August 2024 (UTC)

So why don't we just change "Do not base an entire article on primary sources" to "An entire article should not be based on primary sources"? Do we have a local consensus to make that change? -- Valjean (talk) (PING me) 14:39, 24 August 2024 (UTC)

That's a completely fair change without changing why we have that there. Masem (t) 16:18, 24 August 2024 (UTC)

Would you like to make that change? I have to run now. -- Valjean (talk) (PING me) 17:12, 24 August 2024 (UTC)

While I don't disagree with the proposed change, given that this sentence has been quoted repeatedly by one stalwart member of the Loyal Opposition at Wikipedia talk:Notability (species)#Proposal to adopt this guideline (an open RFC), I would prefer postponing any changes until the RFC has closed. Although there haven't been any new comments for two days, I expect to leave it open until the bot closes it (another ~14 days). I don't want the closers to deal with a Wikipedia:Close challenge on trivial grounds. (Substantive challenges would be welcome, of course, if the closing summary really is bad, but complaints like "It only stayed open for the length of time prescribed by WP:RFC rather than the length of time in the bot's code" or "They're gaming the system over at that other page" would not be welcome.) WhatamIdoing (talk) 07:27, 25 August 2024 (UTC)

That makes sense. There is no rush. -- Valjean (talk) (PING me) 21:52, 25 August 2024 (UTC)

The policy has basically worked. I haven't seen it weaponized to delete breaking news stories. As far as I know, it's encouraging people to add secondary sources to those types of articles. If someone can find an example where it's been misused, we can try to add some clarification. But there is always the risk of overreach. Trying to describe every exception will usually lead to bad guidance. Shooterwalker (talk) 16:31, 24 August 2024 (UTC)

There is a large problem of editors rushing to create articles on breaking news where there is zero clear indication that the event either is going to have enduring coverage, or that could not be covered as part of a larger news topic and serve a more comprehensive purpose. In other words, we really need to realign how editors are creating articles with respect to NOTNEWS and NEVENT, but that's not an issue with NOR, nor with this language specifically. Masem (t) 16:45, 24 August 2024 (UTC)

I think that's my point. I haven't seen people misusing WP:OR to delete things that shouldn't be deleted. There's a greater problem with WP:NEVENT, and think that referencing / directing people to WP:NOTNEWS would be more useful here. Shooterwalker (talk) 16:57, 24 August 2024 (UTC)

True, but I do think the reduction in harshness of the language (from "do not" to "should not") is generally a far better alignment with most other content policy languages. The only time we use absolutes like "do not" are for policies like BLP and COPYRIGHT which have legal ramifications. Masem (t) 17:11, 24 August 2024 (UTC)

The primary use of phrases like "Do not" and "must not" is the main MOS page, because bad grammar is bad grammar, and not really a question of judgment or POV. WhatamIdoing (talk) 07:15, 25 August 2024 (UTC)

Do not → should not is not a softening of the language, though. The former is imperative, the latter (unless as You should not) is not. Remsense ‥ 论 07:19, 25 August 2024 (UTC)

I meant that adding the implied 'you', "you do not base articles..." is stronger than "you should not base articles..." (when along the lines of the MoSCoW method), and we generally only use such absolutes like "you do not" or "must not" in those policies with some legal ramifications. Masem (t) 13:35, 25 August 2024 (UTC)

@Masem, I know that's a popular guess – it feels right for legal requirements to sound harsh – but if you actually go look at the legal policies, it isn't actually true. For example:

• Wikipedia:Child protection contains "Do not" only once, in ==Advice for young editors==, and it does not use the word "must" at all.
• Wikipedia:Copyright violations contains the imperative "Do not" once, in the nutshell. It does not contain "must not" at all. The only use of "must" is permission conveyed through e-mail must be confirmed – rather weak tea, IMO.
• Wikipedia:Copyrights contains the imperative "Do not" only once (in the WP:LINKVIO section). It does not contain "must not" at all.
• Wikipedia:Libel does not contain the words "Do not" or "must" at all.
• Wikipedia:No legal threats contains the words "Do not" only once (first sentence). It does not use the word "must" at all.

That's a mere four uses in the first five legal policies in Category:Wikipedia legal policies. There are only 10 legal policies in that category.

For comparison, Wikipedia:Manual of Style says "Do not" 78 times and "must" 22. While I haven't checked every policy, it is likely that the 100 uses on this single page of the MoS uses this language more times than all of the legal policies combined. WhatamIdoing (talk) 21:38, 25 August 2024 (UTC)

I've seen it misused many times but not on breaking news stories. Most have been on "boring" encyclopedic information which secondary sources don't write about. North8000 (talk) 17:00, 24 August 2024 (UTC)

Should Wikipedia restrict itself to things that others have cared enough to have written about? SmokeyJoe (talk) 08:03, 25 August 2024 (UTC)

In general, yes, but not in an absolute sense. In an article about some celebrity can we use a tweet from the subject for a date of birth if no secondary source has written about it? Yeah, who cares. It's the type of information expected of an encyclopedia, the subject obviously doesn't mind it being published, and it's just not that serious a matter. Should we have an article about a contentious historical event based solely on primary sources? Obviously not for many reasons. -- LCU ActivelyDisinterested «@» °∆t° 11:38, 25 August 2024 (UTC)

Given what I've seen happen in areas of fiction with eager fans, or even back with that whole situation around MMA topics years ago, yes, allowing WP to cover topics that can only be based on primary sources and that no reliable source otherwise covers leads to WP being more like TV Tropes or fan wikis than a serious reference work. Masem (t) 11:59, 25 August 2024 (UTC)

I think it is more important to have Wikipedia:Independent sources than to have True™ Secondary sources. There will always be some questions about whether certain sources are True™ Independent sources or True™ Secondary sources, but IMO we should never create an article when the only Wikipedia:Published sources are indisputably non-independent. WhatamIdoing (talk) 21:42, 25 August 2024 (UTC)

• Time for another installment of “History with Blueboar”… originally this policy stated that WP (itself) should not be a primary source for information (whether facts, analysis or conclusions). This statement tied directly into the concept of NOR. If we add facts, analysis or conclusions that have never been published elsewhere, then WP is the primary source for those facts, analysis or conclusions.

Then someone added that WP should be a tertiary source, and as such should be based (mostly) on secondary sources. This addition wasn’t wrong… but it did not directly tie into NOR.

Then someone else decided that we needed to define these terms (primary, secondary, tertiary). And, as is typical, there was a lot of disagreement and discussion over how best to define them. The end result is what we now see in PSTS.

Unfortunately, somewhere along the way, we lost the original statement (about WP itself not being a primary source) - which was the very reason we were defining all these terms in the first place! We lost the statement that tied PSTS directly to the concept of NOR.

That omission shifted PSTS’s focus from what we say in our articles (NOR) to which sources we use in our articles.

Anyway, that’s the historical background… make of it what you will. Blueboar (talk) 13:20, 25 August 2024 (UTC)

Well, it does tie into OR. The primary source for a fact is a witness to the fact. (This can include reporters whose job is regularly to go to places to witness and report. It can include scientists who conduct an experiment to write a report. Etc.) Analyses or conclusions or interpretations based on facts not witnessed (not from personal knowledge) are not a primary source for those facts, they are secondary for those facts. (Thus, you can have mixed sources, primary and secondary.) Wikipedians are not to be such original reporters, nor the original publisher of analyses or conclusions or interpretations, or the original mixer of the two. -- Alanscottwalker (talk) 16:09, 25 August 2024 (UTC)

If the analysis or conclusion is done by a Wikipedian, then it makes Wikipedia the primary (original) source for that analysis or conclusion. The point of NOR is that we report on the analysis or conclusions of others, and don’t include our own analysis or conclusions. Blueboar (talk) 19:09, 25 August 2024 (UTC)

I don't know what point you're making that is different from mine. Wikipedians don't write on their experiences, or what they themselves think. Alanscottwalker (talk) 19:18, 25 August 2024 (UTC)

Well, we're not supposed to do that. But one does see it happen.

I do think that moving PSTS to a separate policy page would help with this. Over-reliance on a primary source, if the only thing you're writing is a simple description, is not OR as defined by the first sentence of this policy. For example, if you were the editor starting the article on Cow's Skull: Red, White, and Blue, then you could go look at the primary source (i.e., the painting) in the museum and write "it is a painting of a cow's skull on a background of red, white, and blue", and use this for your citation:

• O'Keeffe, Georgia. (1931) Cow's Skull: Red, White, and Blue. Metropolitan Museum of Art, New York, NY, United States.

You're not making anything up, so it's not original research. As soon as you want to say something about the painting being famous, or incorporating southwestern and Native American themes, you need to get a different source, but a simple, basic description of a primary source is a legitimate use, non-OR use of a primary source. Consequently, I think that having admonitions to not use the painting as your sole source for that article should (a) be somewhere in our ruleset, but (b) not be in this particular policy page. WhatamIdoing (talk) 21:55, 25 August 2024 (UTC)

No. It is OR, (the description is only verifiable with the picture as the source but verifiable is different from OR) -- a picture that no one but you cares to write about has no significance in sources, except that you are asserting it has significance because you originally think it does and thus originally publish on it. Alanscottwalker (talk) 22:55, 27 August 2024 (UTC)

No, that's not true. "A picture that no one but you cares to write about" might have no significance, but "nobody cares" does not mean that it is "material—such as facts, allegations, and ideas—for which no reliable, published source exists." The definition of OR, as given in this policy, doesn't say anything at all about whether anyone cares. WhatamIdoing (talk) 04:33, 28 August 2024 (UTC)

BTW, this painting is given as one of the examples in Wikipedia:Identifying and using primary sources#Primary sources should be used carefully. WhatamIdoing (talk) 04:34, 28 August 2024 (UTC)

Yes. It is true. The idea you are originally creating is its significance. Only you think that it has significance as far as can be told. And yes it can be used as a primary source, but that it is primary means nothing can be asserted about its significance from it alone. -- Alanscottwalker (talk) 09:38, 28 August 2024 (UTC)

No, writing an ordinary Wikipedia article about a subject does not mean "creating significance" for the subject. WhatamIdoing (talk) 04:36, 29 August 2024 (UTC)

Of course it does. Indeed probably the most important thing you are saying about the subject is this has significance, so much significance, you need to read about it as the subject in an encyclopedia. Alanscottwalker (talk) 14:05, 29 August 2024 (UTC)

We require sources to show that a topic is significant as to allow a standalone article on it. We cannot synthesis that significance if the sources aren't there for that. Masem (t) 14:20, 29 August 2024 (UTC)

I've already said that. You can't create that article. It is original, but Wikipedian's, not doing it correctly, sometimes well go on and try to create something they should not. -- Alanscottwalker (talk) 14:30, 29 August 2024 (UTC)

We require a secondary source to say "This is a significant piece of artwork".

We do not require a secondary source to say "It is a painting of a cow's skull on a background of red, white, and blue".

Note that:

• The primary source is supporting the quoted sentence.
• The quoted sentence says nothing about significance.
• The quoted sentence implies nothing about significance.
• The quoted sentence is not alleged to be the only sentence in the Wikipedia article.
• The primary source is not alleged to be the only source to "exist—somewhere in the world, in any language, whether or not it is reachable online" about this subject.
• The primary source is not alleged to be the only source used to create the Wikipedia article, or even the only source cited.
• The primary source is not alleged to be the basis for notability.

WhatamIdoing (talk) 18:15, 29 August 2024 (UTC)

It makes no sense for you to keep going off into irrelevancies, verifiability is not what this is about, but you keep confusing it with OR. The situation that has been set down is that there is no secondary source, and the only thing is the picture, the picture can't tell anyone anything about its significance including the significance of its appearance, but you writing about it in the pedia is asserting that factoid has significance (it apparently has significance originally to you, but no one else). Now sure, Wikipedians may be want to add all kinds of factoids from primary sources, they like or think important, and thus originally (only in Wikipedia) change the presentations of any subject, or create the worthiness of subjects to the world, but they definitely should not. Alanscottwalker (talk) 19:37, 29 August 2024 (UTC)

If you are arguing that we “imply” the conclusion that the painting is in some way significant merely by creating an article about it, my reaction would be: Meh… that is stretching the NOR policy a bit. I think what you are discussing is more a violation of WP:Notability than a violation of WP:NOR. Blueboar (talk) 20:21, 29 August 2024 (UTC)

No. 'We don't want what you want to share with the world, just because you (and only you) think it is important .' Nor is original research another way to say verifiability. And it is not just new "original" (literally) subjects, its well established subjects, reformed (by you) with original (because its [only] important to you) facts. Alanscottwalker (talk) 20:32, 29 August 2024 (UTC)

I don't think that's correct. We do want what people want to share with the world, because they think it is important – so long as they have reliable sources to back it up. WP:OR is about verifiability. OR is defined in this policy as the non-existence of sources (anywhere in the world, in any language) that could support the contents. It is not defined as "stuff you (and only you) think is important". WhatamIdoing (talk) 23:49, 30 August 2024 (UTC)

Of course, it is correct. "Only you" means that that there is no reliable source for significance. You practically acknowledged its correctness when you wrote, "reliable sources to back it up", which can only mean appropriate reliable sources for significance, and the writer can't be a reliable source for significance. Alanscottwalker (talk) 10:06, 31 August 2024 (UTC)

Notability (i.e., the process of qualifying for a Wikipedia:Separate, stand-alone article) does not require importance/significance. "Insignificant" subjects can and do get articles.

We require that sources cover the subject. We do not require that sources indicate that the subject has any "significance". If someone writes a book about The Least Significant Book Ever Published, then that book would be a valid subject for an article.

Perhaps we're talking about different things. I'm saying that a primary source is a valid way of verifying some statements in articles.

Perhaps you are saying that having a whole article requires some evidence of "attention from the world at large", even if that attention does not declare the subject to be significant (or, indeed, declares it to be of no importance whatsoever). WhatamIdoing (talk) 19:01, 1 September 2024 (UTC)

I don't think you are reading what I write. I'm not just talking about whole articles. And perhaps it will help you, if you realize I am using important solely in the sense of 'a matter of import'- meaning. And no, we don't put things in the pedia that have no significance, we relate the significance others through reliable sources give them. Alanscottwalker (talk) 19:13, 1 September 2024 (UTC)

I don't know. Do you think that Bennifer is 'a matter of import'? I don't, but the article has 66 sources at the moment, and I've no hope of being able to get that out of Wikipedia.

I give that article as an example of a subject that I personally believe has no significance and is not 'a matter of import' (to anyone except the individuals directly involved). Additionally, I doubt that we could find any sources directly claiming that I'm wrong. If "we don't put things in the pedia that have no significance", then this article shouldn't exist. WhatamIdoing (talk) 20:20, 1 September 2024 (UTC)

Is it are you not comprehending reliance on sources or neutrality? What you think about import is nothing we are to relate, either way. We are not to be the original publisher of whatever import you think to give something. Alanscottwalker (talk) 20:33, 1 September 2024 (UTC)

No, Alan, I didn't go off into irrelevancies. You read what I wrote, which was about a valid source for a single, specific sentence, and you jumped straight to the unsupportable conclusion that there was no secondary source in the whole world about the article's subject and no other sentence in the whole article.

The situation that has been set down is – and I quote – "if you were the editor starting the article on Cow's Skull: Red, White, and Blue, then you could go look at the primary source (i.e., the painting) in the museum and write "it is a painting of a cow's skull on a background of red, white, and blue"", and you could cite the painting for that one sentence.

The situation that was actually set down says nothing at all about the rest of the sources or the rest of the article. It only talks about a single sentence and a single source for that single sentence. WhatamIdoing (talk) 23:43, 30 August 2024 (UTC)

Yes. You do go into irrelevancies. Because the situation I addressed added that there is no secondary source. It was extending fact pattern not jumping to anything, and also addressing the matter in the context of this discussion basing articles on secondary sources. Alanscottwalker (talk) 10:02, 31 August 2024 (UTC)

The situation I described did not say that no secondary source existed ("somewhere in the world, in any language, whether or not it is reachable online", to quote this policy). It said that no secondary source was required for the specific sentence I provided. WhatamIdoing (talk) 19:02, 1 September 2024 (UTC)

So, you are saying much that is irrelevant to my points. Which is the situation of no secondary source, in a discussion about basing articles on secondary sources. -- Alanscottwalker (talk) 19:13, 1 September 2024 (UTC)

If we're trimming paragraphs that secondary sources haven't written about, then the policy is working. It's impossible to write a reliable unbiased encyclopedia without reliable independent sources. Shooterwalker (talk) 19:08, 25 August 2024 (UTC)

@Shooterwalker, Wikipedia:Secondary does not mean independent. WhatamIdoing (talk) 21:56, 25 August 2024 (UTC)

That's true, but the venn diagram is a strong overlap in most articles. This thread is about the thin side of the venn diagram, where journalists are effectively eyewitnesses, which is a valid thing to bring up. Several editors have said that the main point of the policy shouldn't be bulldozed for the more rare / less common cases. Shooterwalker (talk) 13:56, 26 August 2024 (UTC)

I think we need to recognize the fact that our particular PST definitions are wiki-definitions, contained in a few general paragraphs. They are a key part of a policy that emphasizes that analysis of the item and any derivations from information should be done by others rather than by Wikipedia editors. Under those definitions, some sources will be clearly primary, some will be clearly secondary, but a whole lot of them will not clearly be one or the other. If one takes on the premise that some tidy perfection and completeness exists such that every source can be unambiguously classified, then it doesn't work out. North8000 (talk) 14:09, 28 August 2024 (UTC)

Well, indeed, there are sources that are both and one thing that is needed there, is to use parts of them appropriately -- among others things the general rule stresses is, be very familiar with the sources. Teaching both, 1) what part of being familiar with a source is, and 2) how to use it appropriately.

On some slightly different matters, I also note we have no definition, and probably disagreement on what 'news' exactly is 'breaking news', and how to draw that edge in the sand, and I have even seen good argument to me that the species we write on are (all) sourced to secondary. In short, that there are edge cases and uncertainties is always going to happen. Alanscottwalker (talk) 15:27, 28 August 2024 (UTC)

It often feels like the definition of primary is "source supporting an article I don't want to have", and secondary is "source supporting an article I do want to have". The idea that all sources are primary for something is not one we've done a good job of communicating to editors.

This is an imperfect example, but I've seen editors evaluate NCORP sources like this:

• It's a piece of long-form journalism in a respected newspaper.
• The third paragraph says something about last year's profitability.
• Conclusion: The whole article should be treated like a press release, because there's no way a journalist could get that information from any source except the company itself. Actions like interviewing someone at the company, poking around the corporate website, reading their press releases, etc. makes the journalist and the whole newspaper non-independent of the company.
• Alternate conclusion: That sentence about profitability means the whole source needs to be discarded as routine coverage of trivial information.

Some people are arguing this way because they're copying what they've seen other editors do the same (and get respect for it), but I think it's often just a case of WP:IDONTLIKEIT dressed up in an acceptable bit of WP:UPPERCASE. WhatamIdoing (talk) 04:51, 29 August 2024 (UTC)

It seems that we have lost the ability to look at current (including breaking) news events from what should be a 10-year viewpoint, and instead 1) want to rush to create an article on any breaking event regardless if other existing articles are better suited for that event and 2) justify that event being notable by including an excessive amount of detail included the dreaded reaction sections to make it appear that the number of sources make the event notable. Eg Arrest of Pavel Durov is an excellent example of this problem, how we have a huge article on what is a tiny step of a long process, which likely if we were writing for the first time but 10 years after it happened, would have been maybe one to two sentences in an existing article with all we know (at this point in time). The idea that we allow such stories to be created and then consider deletion or cleanup later is antithetical, as anyone that has tried deleted a news article that has shown no lasting significance after a few years knows this is very difficult to inform editors that a burst of coverage is not equivalent to being notable. Masem (t) 12:33, 29 August 2024 (UTC)

Turning each news story into an article is a problem. I think that kind of thing is best dealt with at WP:NOTNEWS or Wikipedia:Notability (events), and isn't really about whether a source is primary or secondary. Shooterwalker (talk) 13:21, 29 August 2024 (UTC)

Yes. We should not be doing those news of the day articles at all, but we are not going to stop it, apparently with anything, no matter what is written here or anywhere else. Alanscottwalker (talk) 14:17, 29 August 2024 (UTC)

I think that the defacto reality is that if it is of sky-is-blue extreme importance (like the top 1 -2 news events per week for the entire EN:Wikipedia) we break the rule and immediately make an article, even if it's from just primary sources.That exception is numerically very rare but highly visible because it takes up half of the top of the Wikipedia main page. North8000 (talk) 14:34, 29 August 2024 (UTC)

What's important to why news articles apply to this discussion is that even truly appropriate news events may likely go with non primary sources (that is, strictly covered by factual news reporting) for data, weeks, or months. Something like a major natural disaster will fall into that. We don't want changes her at NOR to interfere with such developments but at the same time make sure changes here don't open the floodgates to even more news articles that fail to have significance. (edited) Masem (t) 14:46, 29 August 2024 (UTC)

@Masem: I agree. I think that pointing out what I did, that this highly visible rare exception is merely that helps that cause. North8000 (talk) 15:34, 29 August 2024 (UTC)

@Masem, I'm not sure what you mean by non primary sources (that is, strictly covered by factual news reporting). Are you saying that something that is "strictly covered by factual news reporting" is not a primary source? WhatamIdoing (talk) 18:17, 29 August 2024 (UTC)

Ah, I meant just primary sources, not non primary ones — Masem (t) 19:11, 29 August 2024 (UTC)

Sure. As much as I think those articles are really bad for many reasons, I console myself with them being a relatively small number, although I don't care to find out what the number actually is (so don't try to disabuse me of that notion, please:)). Alanscottwalker (talk) 20:19, 29 August 2024 (UTC)

I think we need to recognize the fact that our particular PST definitions are wiki-definitions? No. Go to primary source and secondary source for the definitions. If you don’t like the definitions, fix them, with reliable sources. The paraphrasing in this policy should be read as subject to referring to the articles. The bold direct linking to the mainspace articles is highly appropriate. SmokeyJoe (talk) 10:45, 31 August 2024 (UTC)

The definitions in the Wikipedia articles depend on the subject area (e.g., legal scholars say that tertiary sources don't exist). We sort of pick and choose which definitions we want to use, so perhaps it's not unfair to say that they are our own definitions, based heavily on the real-world scholarly definitions from history and science. WhatamIdoing (talk) 19:11, 1 September 2024 (UTC)

I have no idea which legal scholars you are referring to but tertiary sources exist in legal scholarship.[1] -- Alanscottwalker (talk) 19:32, 1 September 2024 (UTC)

It is much better to acknowledge and agree that an encyclopedia is in the field of historiography, and then to use the historiography definitions.

It is a worse idea for Wikipedia to invent new definitions, based on history and science or otherwise. New definitions can’t be researched more deeply. Precedent for recurring problems can’t be resolved from examples if we use made up definitions.

Blurring or mixing history into science sounds dooms to generate more problems than it solves.

The historiography definitions are perfectly good. The science definitions of primary and secondary sources are wholly inappropriate for Wikipedia. The journalism definitions, despite someone asserting that good journalism is good scholarship, have too many points of incongruity for mixing historiography and journalism to be anything but a bad idea. SmokeyJoe (talk) 06:25, 9 September 2024 (UTC)

Wikipedia's definition of "notability" is its own. In dictionaries, the primary definition is "worthy of note", but our WP:GNG corresponds more closely with "noted". I've had to explain this a number of times, sympathetically, to new article creators who've insisted that the subjects of their articles were—using real-world terminology—notable. Often I disagreed that the subjects were even worthy of note, but I couldn't fault them for their confusion. Largoplazo (talk) 11:55, 9 September 2024 (UTC)

Yeah. It’s a long known problem, this Wikipedia neologism “notability”. I have been preferring to use “Wikipedia-notability”, to distinguish it from the real world term. It’s unfortunate. “Wikipedia-notability” means “the topic has been covered by reliable others”, which is close to “noted”. Wikipedia neologisms create newcomer barriers, and should be avoided even if only for that reason alone. SmokeyJoe (talk) 13:39, 9 September 2024 (UTC)

It should have been changed long ago, but I think the last time it came up formally, a kind of fait accompli won the day. Alanscottwalker (talk) 13:54, 9 September 2024 (UTC)

Yup… if I had a Time Machine, I would go back and strongly recommend that we call the guideline WP:Notedness (as that terminology is closer to what we mean) … but… it is too late to change it now. Blueboar (talk) 19:52, 9 September 2024 (UTC)

I would prefer something more explanatory, like WP:Requirements for a separate article or WP:Eligibility standards for articles. WhatamIdoing (talk) 20:51, 9 September 2024 (UTC)

WP:Stand alone? : 'This article Stands.' This article does not Stand." "Stand alone is a test . . ." Such might also make discussion less binary, bringing merge or redirect more to fore as compromise consensus. If only we had that time machine, but consensus can change, right? :) Just not so easy to get a new one. Alanscottwalker (talk) 14:31, 10 September 2024 (UTC)

(I have no idea how indented this comment should be.) I have to say I would strongly prefer the existing language's directness. In WP:CGR we have a problem of editors paraphrasing Livy's first pentad and calling that reliably sourced. It isn't. Having once edited on a breaking news event, I also recognise that sometimes there are no secondary sources to be citing. If anything needs changing, it should probably be contextual rather than across the board, ie weakened only when secondary source coverage on a topic is weak or non-existent. Notability shouldn't be an issue here; a plethora of independent primary sources can still establish notability. Ifly6 (talk) 04:05, 29 August 2024 (UTC)

I think that most people who hang around AFD and related guidance pages don't agree that independent primary sources can justify a Wikipedia:Separate, stand-alone article, though they seem willing to extend a reasonable (often multi-year) grace period to major news events. WhatamIdoing (talk) 04:44, 29 August 2024 (UTC)

Is NOR really a factor at AFD? Sure, individual sentences (and on occasion entire paragraphs) might get cut due to NOR violations… but deleting an entire article? I don’t think that happens all that often. It’s seen as more of an “Article clean-up” issue, and not a “Deletion” issue. Blueboar (talk) 20:32, 29 August 2024 (UTC)

I've never seen t be directly a factor at AFD. But this area of NOR overlaps with GNG, and wp:notability is the main factor at AFD. North8000 (talk) 20:39, 29 August 2024 (UTC)

It looks like it's mentioned in something on the order of 5% of AFDs. Here are a few current/recent examples:

• Wikipedia:Articles for deletion/List of second-level administrative divisions by population
• Wikipedia:Articles for deletion/Electronic daily devotional
• Wikipedia:Articles for deletion/History of Saturday Night Live (1975–1980) (2nd nomination)
• Wikipedia:Articles for deletion/Arrest of Pavel Durov
• Wikipedia:Articles for deletion/Petteway v. Galveston County
• Wikipedia:Articles for deletion/Granite Mountains (northern San Bernardino County, California)
• Wikipedia:Articles for deletion/Socialist Workers Thailand
• Wikipedia:Articles for deletion/List of Israeli Mizrahi and Sephardi Jews
• Wikipedia:Articles for deletion/Petteway v. Galveston County

WhatamIdoing (talk) 00:00, 31 August 2024 (UTC)

In spite of WP:PRESERVE, I don't think "articles that would merit inclusion if written competently should never be deleted for their present state" is actually defensible as a position. It's understandable that WP:BLOWITUP cases are rare because an editor adopting it as their pet project to delete as many OR (etc.) article as possible will do result in harm—but it's mystifying to me that it's rarely acknowledged that some articles are a net negative and should not be allowed to remain on the site for potentially years in the hopes that they will be rewritten. (The retort of "so fix it" falls a bit flat if one actually accepts the calculation that deletion would be a net improvement, hence is a fix for it.) Remsense ‥ 论 04:27, 30 August 2024 (UTC)

Indeed, it's as if some people want to believe a Wikipedia article is not actually published and that webhosting is some kind of improvement. Alanscottwalker (talk) 09:55, 30 August 2024 (UTC)

@Remsense, I wonder if you could describe the kinds of articles that are a net negative, and don't qualify for deletion. Obviously something that qualifies for (e.g.) {{db-hoax}} or {{db-no content}} would be negative for readers, but I'm sure that isn't what you're talking about. WhatamIdoing (talk) 00:32, 31 August 2024 (UTC)

What springs to mind immediately are articles that are about a viable subtopic/intersection—let's make one up, Karl Marx and nationalism. Reams have been written about this intersection to the extent that it can be distinguished from related subtopic articles, even. But if someone started this with no sourcing, it would be a net negative unless someone completely rewrote it and included sources that could actually be made use of by the readership. Remsense ‥ 论 00:46, 31 August 2024 (UTC)

Given an uncited article, why delete it, instead of spamming in a couple of refs? You spend a few minutes in your WP:BEFORE search finding books like these:

• Nimni, Ephraim (1991). Marxism and Nationalism: Theoretical Origins of a Political Crisis. Pluto Press. ISBN 978-0-7453-0730-5.
• Anderson, Kevin B. (2016-02-12). Marx at the Margins: On Nationalism, Ethnicity, and Non-Western Societies. University of Chicago Press. ISBN 978-0-226-34570-3.
• Szporluk, Roman (1991). Communism and Nationalism: Karl Marx Versus Friedrich List. Oxford University Press. ISBN 978-0-19-505103-2.
• Snyder, Timothy (2018). Nationalism, Marxism, and Modern Central Europe: A Biography of Kazimierz Kelles-Krauz, 1872-1905. Oxford University Press. ISBN 978-0-19-084607-7.

and you drop them in the article. If it says something reasonable, or if you can quickly WP:STUBIFY it back to something reasonable, then why would we want to choose WP:DELETE instead of WP:SOFIXIT? WhatamIdoing (talk) 19:25, 1 September 2024 (UTC)

Where NOR is a reason for deletion, it is the extreme case of it that is covered by WP:N. SmokeyJoe (talk) 10:41, 31 August 2024 (UTC)

A source can be used if it is published, reliable and citable without OR. The P/S/T classification system is a malicious invention designed to make that harder to understand. Zero^talk 07:09, 31 August 2024 (UTC)

Primary and secondary source distinction is a mature, and very useful analytical tool of historiography. Refer to the articles. WP space should not be redefining real world terms.

Historiography is the right field to choose to put Wikipedia into. It is history, it is not science or journalism. SmokeyJoe (talk) 10:39, 31 August 2024 (UTC)

@Zero0000, we really didn't invent PSTS to complicate matters. It's just that having it on this particular page is a sort of a historical accident. WP:NOR basically started with that Usenet crank trying use Wikipedia to host his debunking of Einstein (remember him?). So we said, in a rather fancy way, that Wikipedia is not supposed to be a primary source, and if you want to publish your new ideas about physics, you need to do that some place else.

Then not everyone knew what "primary source" meant, so we had to explain what a primary source is, and then people ask that since Wikipedia isn't a primary source, what is it?, and bit by very reasonable bit, half a sentence turned into a whole section that is really more about notability and NPOV than about whether editors are SYNTHing a bunch of cherry-picked sources to prove that modern physics is wrong. But it's here now, and it might take divine intervention to get it moved elsewhere (or, as I suggested a while back, put into its own policy). WhatamIdoing (talk) 19:36, 1 September 2024 (UTC)

@WhatamIdoing: I'm under no illusions about the difficulty of backing off from the mess in this article, which has been a bugbear of mine forever. My comment, though written in a flippant manner, arises from real concern over how the P/S/T divide obscures rather than clarifies the simple principles of source usage. Of course I know that that wasn't the intention. I'll post a longer critique soon. Zero^talk 02:07, 2 September 2024 (UTC)

WP:PSTS is the intellectual basis for NOR in the affirmative, even if the early authors didn’t realise. SmokeyJoe (talk) 03:06, 2 September 2024 (UTC)

Indeed, it is. Just to use the given subject and cherry-picking. This is not the place for you to publish what you think about relativity, its for you to faithfully relate what others have said through qualified reliable sources.

You can't originally publish, or originally publish on, the Einstein letter (primary source) you found in your research in Wikipedia's relativity article (that is original research).

You can't create a new article alone about that letter (that is an original purported secondary source the Wikidian made).

You can't publish what you think about that letter (that is you creating a purported secondary, or secondary and tertiary source originally made by the Wikidian).

Cherry-picking is what the Wikipedian does (but should not) -- unless qualified reliable secondary and/or to a lesser extent qualified reliable tertiary sources have already picked it up and examined it, it is Wikipedian cherry picking. And you can't (originally) misuse primary, secondary, or tertiary sources, in your writing, here, so it is good for you to have some idea of what such qualified sources are, and what they can and cannot support.-- Alanscottwalker (talk) 09:48, 2 September 2024 (UTC)

But you also mustn't misuse primary, secondary, or tertiary sources for WP:V purposes, so it would be good for readers of WP:V to have some idea of what such qualified sources are, and you mustn't misuse primary, secondary, or tertiary sources for WP:N purposes, so it would be good for readers of WP:N to have some idea of what such qualified sources are, and you mustn't misuse primary, secondary, or tertiary sources for WP:BLP purposes, so it would be good for readers of WP:BLP to have some idea of what such qualified sources are, and you mustn't misuse primary, secondary, or tertiary sources for WP:RS purposes, so it would be good for readers of WP:RS to have some idea of what such qualified sources are, and you mustn't misuse primary, secondary, or tertiary sources for WP:NPOV purposes, so it would be good for readers of WP:NPOV to have some idea of what such qualified sources are. This concept, like the concept of Wikipedia:Independent sources, is bigger than a single policy. That's why I think it should have its own page. WhatamIdoing (talk) 01:29, 4 September 2024 (UTC)

Readers of V are to know all three central content policies. Readers of NPOV are to know all three central content policies. Readers of BLP policy must know all three central content policies. So, it is not better to multiply, if anything, it is better to integrate. In these policies we are generally trying to answer three fundamental questions in a situation where "we" is anonymous, individually unaccountable for what we do, but must work together to present our work: how do we know what we know, how do we prove what we know, and how do we present what we know. But the answers to those questions by their nature are not going to be separate, they are going to be interrelated aspects. Alanscottwalker (talk) 13:52, 4 September 2024 (UTC)

“it is better to integrate”. Yes. Two important policies, V and NOR, are better combined. Merge them into WP:Attribution, which contains PSTS. PSTS, although maybe not Tertiary, is fundamental to NOR, and fits seemlessly into V. I regret opposing WP:A. It was the right idea, badly implemented.

WP:NPOV is a bit different. It should remain a separate policy, in some ways the most important policy. It is the fundamental of #1 in WP:Trifecta. SmokeyJoe (talk) 12:35, 5 September 2024 (UTC)

I agree with the diagnosis, but I'm not sure about the cure. There are a lot of concepts that have taken on their own technical meaning on Wikipedia. Notability is one of them, and so is PSTS. I'm not sure where the right place is for them, but there's nothing stopping someone from WP:BOLDly writing an essay if they think it's the right course. I wish I could think of a better solution but it escapes me right now. Shooterwalker (talk) 16:28, 4 September 2024 (UTC)

As with wp:notability, step 1 would require acknowledging the unacknowledged way that Wikipedia actually works (when it works). Which is editors making editorial decisions influenced by policies, guidelines and other considerations. (vs.binary flow-chart blocks) For the types of situations described above this would be:

1. Explaining what wp:nor seeks to avoid. And understanding that there are matters of degree of this. (we call the safer non-controversial types "writing in summary style", and the ones at the other end of the spectrum are bright line policy violations)
2. Explaining PST and how secondary means that somebody else has done the synthesis rather than the wiki editor.
3. Then per Wikipedia:How editing decisions are made editors make the decision on what to put in, influenced by the above

Sincerely, North8000 (talk) 19:32, 4 September 2024 (UTC)

I don't think that PSTS is about "what wp:nor seeks to avoid". NOR seeks to avoid having editors make stuff up, whether by making it up completely ("Fairies came to my house last night") or by SYNTHing it up ("These 37 cherry-picked sources, when assembled by me to produce claims that none of them WP:Directly support, prove I'm right and Einstein is wrong"). You don't need to know anything about PSTS to discover that these are wrong.

I did a quick search for comments in the Wikipedia: namespace that mention the word secondary this year. Two-thirds of them were in AFDs. WhatamIdoing (talk) 00:48, 7 September 2024 (UTC)

I think that my 6 word summary of the goals of a core policy is inevitably going to have issues. The slightly longer version would say that the sourcing type distinctions are a component of wp:nor. And of course, wp:nor seeks to avoid certain things. North8000 (talk) 01:52, 7 September 2024 (UTC)

• It can both be the case that there is widespread misunderstanding of primary versus secondary sources, which is a media literacy concept and not a Wikipedia concept, and that the practice is that many articles rely more on primary sources than the policies and guidelines advise. However, that's not necessarily a problem that needs to be proactively solved since every article and everything in the project is a constantly evolving and changing work in progress. I would say more primary sources is just the natural state for a newsy item, and over time, secondary sources should replace them. So it's reasonable for the policy advice to say "don't base an article entirely on primary sources," but if those are the best sources available, like with other guidelines, exceptions and discretion exist. Andre🚐 21:00, 9 September 2024 (UTC)
• The policy is fine. There are a multitude of issues with relying on primary sources on a site that anyone can edit. The fact that it gets ignored is not a reason to disregard the policy. Articles that rely on primary sources are some of the worst in terms of quality excluding stubs and the like. Traumnovelle (talk) 19:47, 15 November 2024 (UTC)

  Yes, exactly. If anything this policy should be strengthened and better-enforced. JoelleJay (talk) 01:41, 3 December 2024 (UTC)

Routine measurements

EDIT: Proposal withdrawn, see my reply to @Zero below

I propose that Routine calculations be amended to include routine measurements. I'm not talking about de facto unverifiable cases like "I have exclusive access to object A and have measured its size to be x  ×  y  ×  z, so I should be allowed to include this fact here because anyone could verify this (IF they ever get access to the same)". What I mean are things like measuring the distance between two geographical points on a map that can be verified with minimal effort by literally anyone. For example, it's common for {{Routemap}}s to include distances between stops/stations, even though this information is rarely provided explicitly by transportation authorities (or any other reliable sources) — but they do provide detailed maps of their routes and editors are using those to measure the distances. 58.136.41.76 (talk) 06:41, 14 September 2024 (UTC)

Making measurements from a map has been discussed before, and it is interpretation of a document. In my mind, a measurement, routine or otherwise, would involve measurement of the original object, such as measuring the distances between a series of bus stops with a Total station. Jc3s5h (talk) 11:01, 14 September 2024 (UTC)

Which would you say is closer to the concept of OR, that "interpreting a document" anyone can repeat if they have a minute, or walking around with a specialized tool very few other people could duplicate the results with? I would say in this case a map would be a secondary source generated by experts in their field (which would be "read", not "interpreted") and Mother Nature would be the primary source. Or would you seriously expect there to be stuff like peer-reviewed papers or NYT articles on distances between random map points that would be closer to the concept of a secondary source than the map itself is? I don't disagree that what you describe is a purer form of taking measurements, but we're talking about this in the context of OR. 58.9.132.210 (talk) 12:18, 14 September 2024 (UTC)

Readers are meant to be able to verify anything they read in an article. Measurements that can only be verified using anything more rarified than a ruler is likely totally going against our verification policy. This is distinct from citing a source stating the measurement, so there would logically seem to be greater restrictions as to what can be considered verifiable.Remsense ‥ 论 12:23, 14 September 2024 (UTC)

When I said "read", not "interpreted" I meant that if maps were made by describing locations, distances between them, etc. with words instead of pictures, one presumably wouldn't say quoting such statements was "interpreting" them or there was a great burden to verify them (because they speak for themselves as a secondary source). Just because the data is represented in a different way doesn't mean that accessing and re-representing it has to be conceptually very different. A "direct quote" in this case would be a map image, perhaps with a line drawn between points A and B, and that can be put in your own words (or in this case numbers) in exactly the same way a prose statement can be re-represented by an editor. And a ruler is exactly all you need in this particular case, the fact that nowadays it's more likely to be a software-based ruler is conceptually irrelevant. 58.9.132.210 (talk) 12:51, 14 September 2024 (UTC)

I agree with you and think about this a lot. I think the basic process I've found viable to systematize this is outlining every claim apparently made by piece of media, "converted" into bullet points (likely in shorthand if we're actually doing it, of course). I think it's obvious we should be able to say a map with a key and sufficient precision can make specific claims analogous to those a paragraph could make.Remsense ‥ 论 12:59, 14 September 2024 (UTC)

I don’t think using a map and key to calculate distances is “Original research”… however, such calculations should be phrased as being approximate. Blueboar (talk) 13:17, 14 September 2024 (UTC)

Every measurement is approximate, so I think it depends. Generally, it would seem best to hew to less clunky wording unless there's a specific reason for precision to matter. It is generally considered unreasonable for a reader to assume an author means to say that New York is exactly 700.0 miles (1,126.541 km) from Chicago. Remsense ‥ 论 13:23, 14 September 2024 (UTC)

@Remsense Exactly. Besides, if a reputable newspaper or a scientific paper would state "the distance between station A and station B is 2.34 km", in most cases it's highly unlikely that they'd be basing this off independent terrain measurements, they'd take this information from a map just like any Wikipedia editor or reader can. So such statements would provide zero added value in terms of verification. 58.9.132.210 (talk) 13:21, 14 September 2024 (UTC)

WP:V calls for everything written on Wikipedia to be, not a copy of but, therefore, in a sense, editors' interpretations of information found in documents. Measuring distance on a map (albeit it has to be a map of a small enough area for the scale to be sufficiently precise for any measurement in any direction within its edges) might be considered to be at that level of interpretation, rather than at the level that WP:OR is concerned with. Largoplazo (talk) 13:53, 14 September 2024 (UTC)

Right now, the policy says

"Source information does not need to be in prose form: any form of information, such as maps, charts, graphs, and tables may be used to provide source information. Any straightforward reading of such media is not original research provided that there is consensus among editors that the techniques used are correctly applied and a meaningful reflection of the sources."

A map is a source, just like a book is. There are reliable and unreliable maps, just as there are reliable and unreliable books. The question here is what "straightforward reading" of a map means, but the principle is really no different from understanding a book. It depends on what type of map it is. Straightforward reading of a geological map might be that some region is primarily basalt, while straightforward reading of a railway map might be that there is a track between A and B. If a map is professionally designed to be spatially precise, such as a large scale map by a national survey agency, taking straight-line distances and directions (to reasonable precision) from the map is straightforward reading. However, taking the lengths of roads and rivers is not straightforward reading (unless they are printed on the map) because not even the best maps show all the little wriggles and measuring a wriggly line is error-prone. The essential point is that if a map is reliable for a datum, then you can cite the datum to the map. Of course, none of this applies to an unreliable map, which is an unreliable source end of story. Zero^talk 14:26, 14 September 2024 (UTC)

Well jiggedy gee, Wikipedia sure is a useful place, because TIL that I'm a genius who can spend an hour making impassionate pleas about something already covered perfectly well in the section right above the one I was on about... I think probably what happened was that since I don't think of maps when I see the word "media", I misinterpreted the section title to mean that it'd be about A/V, etc. and didn't give it proper attention. (Maybe the title could be reworded as something like "Acceptable media and data formats" to accommodate geniuses like myself?)

Apologies to everyone whose time I wasted with this stunt! Consider my proposal withdrawn, unless someone can think of other cases with measurements that might be good to mention in the policy. I didn't really have anything else in mind aside from distances from maps. 58.9.132.210 (talk) 16:43, 14 September 2024 (UTC)

Sensible, well-intentioned suggestions are always welcome.

Also, please consider visiting Special:CreateAccount so that your personal information (e.g., which ISP you're using) isn't visible to everyone on the internet. WhatamIdoing (talk) 02:25, 15 September 2024 (UTC)

BTW, @Zero, since you're an admin, can you do something about the edit summaries on [2] and [3] here? This has nothing to do with me, I just noticed these when looking at history and thought this probably shouldn't be there, (I pinged a RevDel admin, but they haven't responded.) 49.228.98.239 (talk) 08:19, 15 September 2024 (UTC)

"No original research" listed at Redirects for discussion

The redirect No original research has been listed at redirects for discussion to determine whether its use and function meets the redirect guidelines. Readers of this page are welcome to comment on this redirect at Wikipedia:Redirects for discussion/Log/2024 October 12 § No original research until a consensus is reached. C F A 💬 20:39, 12 October 2024 (UTC)

"cite reliable, published sources that are directly related to the topic of the article"

A disagreement has arisen at Wikipedia:Featured article candidates/Can't Catch Me Now/archive1, with a user insisting that mentioning that the artist's previous album received critical acclaim and was primarily produced by the same producer as this song is a NOR violation. They insist that a source that mentions those details must also mention the newer song or else those details cannot be included in the Background section. I have never seen this part of the guideline be interpreted that way. Can people familiar with the guideline help us with a neutral opinion? Thanks.--NØ 05:49, 13 October 2024 (UTC)

I think this should be at WP:NORN. Zero^talk 12:36, 13 October 2024 (UTC)

Did not know that existed. I will take it there. Thanks.--NØ 13:10, 13 October 2024 (UTC)

A little "thought bubble" in my userspace - perhaps this might be useful?

Hi all,

I've just created this little essay.

Useful? Redundant? Something else? Your opinions requested.

Shirt58 (talk) 🦘 10:52, 21 October 2024 (UTC)

Naming what's depicted in an image (including paintings) is complex. Generally, if it's obvious (e.g., anyone familiar with the area, or looking at a map of the area, would come to the same conclusion), then editors are satisfied. Otherwise, I'd suggest looking around the next time you're in the museum for a sign that describes it (or maybe a page on their website), and using {{cite sign}}. WhatamIdoing (talk) 03:22, 23 October 2024 (UTC)

"WPSECONDARY" listed at Redirects for discussion

The redirect WPSECONDARY has been listed at redirects for discussion to determine whether its use and function meets the redirect guidelines. Readers of this page are welcome to comment on this redirect at Wikipedia:Redirects for discussion/Log/2024 October 31 § WPSECONDARY until a consensus is reached. TeapotsOfDoom (talk) 22:27, 31 October 2024 (UTC)

"WP;OR" listed at Redirects for discussion

The redirect WP;OR has been listed at redirects for discussion to determine whether its use and function meets the redirect guidelines. Readers of this page are welcome to comment on this redirect at Wikipedia:Redirects for discussion/Log/2024 November 1 § WP;OR until a consensus is reached. TeapotsOfDoom (talk) 19:06, 1 November 2024 (UTC)

Parallel citations to primary sources

Should citations of secondary sources – especially ancient ones – include parallel citations of primary sources? What I call a parallel citation of a primary source is something such as:

• Mouritsen Politics in the Roman Republic (2017) p 121 n 40, citing Cicero, Pro Sestio, 97 or
• Cornell Beginnings of Rome (1995) p 331, citing Livy, 6.11.7

The parallel portions are the portions underlined above. In both these cases, the secondary source is actually citing those primary sources in analogous way (as with most works in classical studies, the citations are abbreviated).

I guess there are also three positions here: (1) people prefer including them, (2) people don't prefer inclusion or exclusion, and (3) people prefer removing them. I've normally written with (1), except when constructing the parallel citations is tedious, but other people's contributions show (2) is probably modal. That said, one or two people have yelled at me for pressing for parallel citations' inclusion, and I guess they might hold (3).

And then at the higher level, if consensus exists for 1, 2, or 3 should we write anything on it? Ifly6 (talk) 00:04, 12 November 2024 (UTC)

@Ifly6, GA isn't supposed to be concerned with citation formatting – see Wikipedia:Good article criteria#cite note-3 and the brightly highlighted text in Wikipedia:What the Good article criteria are not – so why is this question even coming up? WhatamIdoing (talk) 06:07, 14 November 2024 (UTC)

Yeah, I was wondering this too. What's this about? EEng 06:29, 14 November 2024 (UTC)

Okay, omit the GA review portion then. First, it doesn't matter. I mentioned it only to avoid accusations that I am inventing hypotheticals for drama. Second, quoting the GA criteria doesn't address an underlying point – which is substantive as to what is being cited and not a question of formatting – whether inclusion of parallel citations is unacceptable reliance on primary sources, ie original research. Ifly6 (talk) 07:09, 14 November 2024 (UTC)

But it does matter. Are you saying you just thought up this question out of nowhere? If so, no one's gonna want to spend their time on it. Or did it come up in some article you're working on? If so, show us and maybe we can get somewhere. Your question -- whether inclusion of parallel citations is unacceptable reliance on primary sources, ie original research -- is vague and confusing unless it's grounded in some actual example. EEng 07:26, 14 November 2024 (UTC)

If so, no one's gonna want to spend their time on it – Dubious–discuss? I love a good hypothetical. 8-)

Ifly6, I think the answer to your question is in WP:SAYWHEREYOUGOTIT. You can't be "relying on" Livy if you actually read it in Cornell. WhatamIdoing (talk) 18:52, 14 November 2024 (UTC)

I think the parallel citation could be useful. A reader who doesn't have the source cited could look up the same classical passage in some other modern work and see if it supports what the Wikipedia article says. Jc3s5h (talk) 20:17, 14 November 2024 (UTC)

I agree that the parallel citation (eg , citing Livy, 6.11.7) is defended by WP:SAYWHERE inasmuch as the claim is there. But should such parallel citations be included in the first place? Ifly6 (talk) 23:53, 14 November 2024 (UTC)

Such citations do no harm, and might be occasionally helpful… but I don’t think we need a rule about them. Whether to parallel cite is a decision that can be left up to individual editors… it should be neither encouraged nor discouraged by policy. Blueboar (talk) 00:04, 15 November 2024 (UTC)

What circumstances do you think are suitable for their removal? Ifly6 (talk) 00:19, 15 November 2024 (UTC)

Offhand, I'd consider removing it if the cited work were unpublished (e.g., citing personal communication), unimportant to the content (citing some routine reference work), or unknown (e.g., citing a book no one knows about – in fact, I'd expect it to be pretty close to famous, or extremely relevant, like "Alice's book, citing Bob's autobiography" for a statement about Bob). I'd probably also remove it if the ref is used for multiple different things, and only one of them was citing the named source.

I would not normally use this style in scientific subjects, either. I prefer just "Systematic review of efficacy for Wonderpam", not "Systematic review of efficacy for Wonderpam, citing Original Pilot Study". WhatamIdoing (talk) 05:51, 15 November 2024 (UTC)

Those are all reasonable reasons on first glance. Musing, I would think that most sources routinely cited in classics (Livy, Dio, Plutarch, Polybius, etc) would fall between those: they are published, they are important to the content because they are usually the only primary source (or one of few), and they are definitely not unknown. Ifly6 (talk) 06:06, 15 November 2024 (UTC)

No, I said that I mentioned it only to avoid [emph] accusations that I am inventing hypotheticals for drama. The origin of this question was in this GA review (myself reviewing), where I encouraged parallel citations – to get ahead of a possible reply that this is not in the GA criteria, (1) I passed the article, (2) encouragements are not requirements, (3) imo if there is anywhere a parallel citation is reasonable, it is in a statement that some author says Livy says XYZ, and (4) GA reviewer instructions state You may also make suggestions for further improvements, if appropriate. – and was then told Again, that [including parallel citations] is not what Wikipedia is here to do. If someone's interest is piqued, the secondary sources are there for them to consult. They in turn contain citations referencing the primary sources. That is how Wikipedia works. Ifly6 (talk) 00:00, 15 November 2024 (UTC)

Well, obviously you can't force the nom to take your suggestion, but WP:SAYWHERE plainly authorizes the voluntary use of the style you recommended. WhatamIdoing (talk) 05:55, 15 November 2024 (UTC)

Award controversy vis a vis with the recipient and SYNTH

Does it constitute WP:SYNTH to indirectly discredit a subject's award by mentioning the controversy on the subject's page given the citation does not mention the subject at all.

To summise. Channel 1915 mentioned on Sam Verzosa's page that the Gusi Peace Prize was under controversy because the organizer of the award is allegedly posing as an ambassador to artificially inflate his own and by extension the award's prestige. This implies that Verzosa potentially received a sham award. But the problem is the given citation Spot.ph does not mention Verzosa by name. The article only directly paints Barry Gusi in a negative light and none of the recipients.

@Channel 1915 has accused me of censorship over this and I need some third party feedback on this. Thanks! Hariboneagle927 (talk) 05:55, 19 December 2024 (UTC)

This page is really for discussions of the high level-policy wording, not for problems with specific articles. I'd suggest continuing to seek consensus on the article talk page itself, and if you get no interest there asking at WP:VPM. MichaelMaggs (talk) 13:49, 19 December 2024 (UTC)

Okay, I've settled the disagreement with the editor. I'll consider my options to how to move things forward. Hariboneagle927 (talk) 13:57, 19 December 2024 (UTC)

I forgot Wikipedia:No original research/Noticeboard. That might be the best place. MichaelMaggs (talk) 13:58, 19 December 2024 (UTC)

Editor-created images based on text descriptions

An editor is creating multiple images themselves, uploading them to Commons, and then posting the images to Wikipedia articles. Let me explain the situation and why I am posting about it here, instead of the OR Noticeboard. This editor is fairly new and seem to be doing this interpretive editing in good faith with the images being clearly-marked either as a "digital reconstruction" or as a "digital remake". They are basing their images on written published sources...but. But these images are created interpretations, they are not published elsewhere, they are being published directly onto Wikimedia platforms, in Wikipedia's voice, so yes, this would all seem to be original research. But there is nothing that specifically addresses this issue in the No original research guideline, so it seems to me that some specific wording about this image issue should be added to the Original images section, something along the lines of:

Created images of historical items that are previously unpublished interpretations unavailable elsewhere are original research and even though these images are or might be based on text descriptions in published sources any such images should be removed.

Also, in light of developing technology and AI, this is a type of issue that can come up anywhere within the Wikipedia/Wikimedia platforms and we need to get out ahead of it. Similar images could possibly pollute our historical information streams with created images. I mean, have you seen the "AI suggestions" that are now prominently displayed on the first line of Google Searches? What I've been seeing is that they are often thin paraphrases of Wikipedia articles, so just wait until the "AI Suggestion" bot gets a hold of created interpretations of historical items based only on written descriptions - the images will be repeated and repeated... Anyway, would appreciate any thoughts on adding a line about "created images, even though only based on written descriptions..." etc., etc. Thanks, Shearonink (talk) 06:19, 24 December 2024 (UTC)

We've already seen examples of AI images uploaded based on text descriptions. I'd support an addition to the guidance along these lines. Nikkimaria (talk) 14:46, 24 December 2024 (UTC)

Nikkimaria I wasn't aware of the AI images, but I suppose is possible in this particular instance that these images are actually AI-generated... Upon re-reading my suggestion above I think it should be/could be re-crafted into:

Created images of historical items that are previously unpublished and are unavailable elsewhere are original research. Even if these images are or might be based on text descriptions in published sources these images are interpretations not actual historical images of these items and any such images should be removed from Wikipedia pages.

Is this new sentence better? - Shearonink (talk) 17:24, 24 December 2024 (UTC)

I'd suggest splitting up the issue into two new sentences, one in each paragraph of OI. First: "Original images that are previously unpublished interpretations of text descriptions are considered original research and should not be included." Second would be something specifically relating to AI-generated images, which probably warrants a separate discussion. Nikkimaria (talk) 17:35, 24 December 2024 (UTC)

Ah ok, that makes sense. - Shearonink (talk) 18:45, 24 December 2024 (UTC)

We might need a community-wide RfC regarding AI-generated images (and even editor-created drawings/sketches/cartoons/paintings, particularly of BLP subjects, in general). I don't believe we have any sort of policy or guidance that addresses these issues. Some1 (talk) 18:01, 24 December 2024 (UTC)

Possibly? But (shrug) I do not agree. These images, whether they are editor-created or they are AI-generated...these images are clearly original research, which is against a stated Wikipedia policy and which, seems to me is inextricably connected to the policy of Verifiability. Others may disagree - and that is fine - but I do not think a RfC is needful or warranted. - Shearonink (talk) 18:45, 24 December 2024 (UTC)

There was a big discussion about user generated images of people awhile ago, I'm not going to dig around for it now. I don't believe it come to any conclusive consensus, most of the images were removed but I don't believe all of them. If I remember correctly the contention was between having artists impression to improve articles with, and concerns over quality and OR. Take that with some salt, as I wasn't overly impressed with some of the images and so may have a biased memory of it. If no-one else brings it up I'll try and find a link to it once I have a bit more time. -- LCU ActivelyDisinterested «@» °∆t° 18:59, 24 December 2024 (UTC)

Thanks - Shearonink (talk) 19:21, 24 December 2024 (UTC)

Here are some discussions I'm aware of, not sure if any of them are the one you're thinking of: Wikipedia_talk:Manual_of_Style/Images#WP:USERG_portraits, Wikipedia:Village_pump_(policy)/Archive_190#AI-generated_images, Wikipedia:No_original_research/Noticeboard/Archive_49#Cartoon_portraits. Plus there's WP:AIIMAGE. Nikkimaria (talk) 19:35, 24 December 2024 (UTC)

Thanks Nikkimaria, those discussions are very useful - I'm reading through all of them now. Will take me quite a while... - Shearonink (talk) 22:40, 24 December 2024 (UTC)

These images, whether they are editor-created or they are AI-generated...these images are clearly original research Oh, I definitely agree with you. I had to revert an amateurish drawing of a BLP subject recently and the editor was quite upset about it. It'd be nice to have something written in policy that directly addresses these issues (AI-generated images/editor-created drawings/sketches/cartoons/paintings, etc). I believe the only place to add something into policy is the Village Pump via RfC. Some1 (talk) 20:09, 24 December 2024 (UTC)

Perhaps you are right, that a broad RfC at the Village Pump is needed. I do think it is important to have some sort of policy/guideline in place sooner rather than later...

Also. Personally, I am of the opinion that the guideline against OR should stand, whether or not the OR is text or the OR is an image. If the text is not found in a published reliable source then it simply cannot be used. The same would seem to be true and should be true for an image - after all, they are both content. - Shearonink (talk) 22:40, 24 December 2024 (UTC)

I don't see "previously unpublished" as being the right criterion -- surely the publication needs to be in a reliable source. But I'm not sure even that's enough, because in my experience publications that are careful about the text they publish are sometimes a bit loosey-goosey about images. I feel we should allow images such as the lead image in Matthew the APostle, and not editor-created stuff, but I'm not sure exactly where the line should be drawn or how to describe it. EEng 04:53, 25 December 2024 (UTC)

Ok, I see your point. Perhaps this is better:

Original images posted onto Wikipedia pages that are previously unpublished in reliable sources and that are created interpretations of text descriptions are considered original research and should not be included.

Shearonink (talk) 14:57, 25 December 2024 (UTC)

Should we start a workshopping section below? I'm thinking something along the lines of:

AI-generated images and user-created artwork (such as original drawings, sketches, paintings, cartoons) posted onto Wikipedia articles that are previously unpublished in reliable sources are considered original research and should not be included.

Some1 (talk) 16:03, 25 December 2024 (UTC)

Sounds good. I agree a workshopping section should be started, either here or over on the Village pump:technical. - Shearonink (talk) 17:14, 25 December 2024 (UTC)

Try this:

Drawings, sketches, paintings, cartoons, AI-generated images, and so on not previously published in reliable sources are original research and therefore should not be used in articles.

This removes some stuff in the previous proposal that's actually irrelevant (e.g. doesn't matter whether or not it's "posted to Wikipedia" -- it's OR either way). I also removed the bit about "user-created" since if it's not from an RS, it's OR no matter who created it. I considered changing should not to must not, but I'm not absolutely certain there aren't acceptable edge cases. EEng 16:52, 25 December 2024 (UTC)

Yes, you are absolutely right about the OR *but* I think the proscription against "user-generated" images should be explicitly and plainly stated even if it does seem somewhat repetitious to us experienced editors. - Shearonink (talk) 17:14, 25 December 2024 (UTC)

OK:

Drawings, sketches, paintings, cartoons, AI-generated images, and so on (whether created by editors or by others) not previously published in reliable sources are original research and therefore should not be used in articles.

EEng 17:37, 25 December 2024 (UTC)

The only further adjustment I might make would be to state "must not" - or to add "must not - instead of "should not". "Should not" implies possible permission, "must not" says no way, don't do it. So maybe something like...

Drawings, sketches, paintings, cartoons, AI-generated images, and other illustrations (whether created by editors or by others) not previously published in reliable sources are original research and therefore must not - and should not - be used in articles.

Shearonink (talk) 18:13, 25 December 2024 (UTC)

I mean, the lead image for Gisèle Pelicot is fine, and ran on the front page. File:Light dispersion conceptual waves.gif is fine, and featured. File:Chloralkali membrane.svg is fine, and featured. File:Visit of the Mandelbulb (4K UHD; 50FPS).webm is fine, and featured on enWiki, faWiki, and commons. File:Pi-unrolled-720.gif is a good illustration, featured here and on commons, and a former featured picture of the day. The current draft would exclude high quality and encyclopedic images like these. GreenLipstickLesbian (talk) 07:34, 26 December 2024 (UTC)

Was this edit the one where you removed the "amateurish" drawing of a BLP? That image was created by a notable professional artist. If so, that's at least twice this year that an editor has complained about a drawings being "amateurish" when they should be saying something a lot closer to "by a professional artist whose artistic style is not my personal favorite". WhatamIdoing (talk) 19:18, 25 December 2024 (UTC)

I didn't realize the person was a "professional artist" when I first removed the image [4] from the infobox, so "amateurish" might not be the right word to use now. Hindsight is 20/20. But my point still stands, that these types of user-created drawings/sketches/cartoons/etc. are subjective and based on what one's own interpretation of what the BLP subject looks like. They shouldn't be used in BLP articles, especially as the lead image. Some1 (talk) 19:28, 25 December 2024 (UTC)

I've always been a little wary of this type of WP:OR, or at least WP:SYNTH. But I haven't been sure how to address it, since it is normalized on certain pages. Shooterwalker (talk) 00:54, 25 December 2024 (UTC)

Shooterwalker - "It is normalized on certain pages"... Really? Which ones, what subject areas? - Shearonink (talk) 04:18, 25 December 2024 (UTC)

I've seen it in articles on fantasy fiction (Hobbit and so on), obscure Norse gods, and stuff like that, where editors think it's OK to make up their own artwork based on textual descriptions. Example: [5]. Such stuff needs to be hunted down and shot on sight. EEng 04:30, 25 December 2024 (UTC)

Yikes. And agree... - Shearonink (talk) 14:52, 25 December 2024 (UTC)

Re: not previously published in reliable sources

Image used on Female, Human, Male, Man, Secondary sex characteristic, Woman

The image to the right (at least on my screen, it's to the right) says that it's an "Original image" and the editor's "Own work. Taken at City Studios in Stockholm, September 29, 2011". The image was not "previously published in reliable sources", but I think everyone here agrees that the image is valuable and usable on Wikipedia articles. How do we write one or two sentences (or even a paragraph) that discourage the use of:

• 1) AI-generated images
• 2) User-created artwork (drawings, paintings, sketches, cartoons, etc.)

that have not been previously published in RS but without being too restrictive? Because unfortunately, I can see proposals like these failing if the community thinks the proposals are too restrictive. Some1 (talk) 18:29, 25 December 2024 (UTC)

Why do we need an editor to make up their own image? There aren't appropriate images in (what appears to be) a reliable source like https://openstax.org/details/books/anatomy-and-physiology-2e ? And looking closely, the male is shows as having "breasts", which (correct my if I'm wrong) is not appropriate, and the female appears to be wearing nail polish on her toenails. Both of these are arguments for why we shouldn't have our amateur selves making our own illustrations.

Having said that, however, I added the red annotations to this:

and used it in my favorite article, Phineas Gage, with an explanatory caption. I would argue, naturally, that this is OK and not OR, but I'm not sure exactly how to distinguish it from what we've been talking about excluding. EEng 18:43, 25 December 2024 (UTC)

This example is simply the same as adding a caption to a reliably-sourced image. You've added an explanation not altered the image and changed its meaning. - Shearonink (talk) 21:39, 25 December 2024 (UTC)

I guess, but I did have to put 2 and 2 together a bit to do it. H marks Dr. Harlow's house, but since the map is 20 years after the events the article discusses (during which time Harlow had moved away), you'll see the house itself is actually labeled "Dr. Hazelton" on the map. Hazelton bought Harlow's practice, and the the location of the "Dr. Hazelton" house is in the same spot as a house labeled "Dr. Harlow" in an earlier, low-quality map I didn't want to use in the article; it also matches texual descriptions in RSs of the location of Harlow's house. So I felt comfortable annotating the map as I did, but strictly speaking one might consider it a step or two toward the OR boundary. EEng 00:39, 26 December 2024 (UTC)

Writing this correctly and concisely is going to be very difficult, even if we agree that it's the right thing to do. WhatamIdoing (talk) 19:15, 25 December 2024 (UTC)

Oh yeah, ain't that the truth. - Shearonink (talk) 21:39, 25 December 2024 (UTC)

A broad rule covering all AI-generated images and user-created artwork might not pass, but if the proposals were more narrow and focused, e.g. AI-generated images not previously published in RS, or user-created artwork (drawings, cartoons, sketches, paintings, etc.) depicting BLP subjects, etc. I could see them getting support. Some1 (talk) 19:43, 25 December 2024 (UTC)

I think the key to user generated images is to make it’s provenance clear in the image caption. Disclaimers such as “AI generated image of X” or “Artistic depiction of Y” resolve most questions. Blueboar (talk) 20:39, 25 December 2024 (UTC)

Oh I dunno about that Blueboar...AI-generated or user-generated images that do not exist in reliable sources? Why are these images not OR? In my original post an editor is creating images of historic flags. One example is described in historical texts as a black F on a white field...but no mention of measurements or fonts or the size of the F etc., etc., so what one editor might create will be different from another editor...who's to say which version is more correct than another? We can't. So we shouldn't. - Shearonink (talk) 21:39, 25 December 2024 (UTC)

Because the purpose of an image is to illustrate the article, not to convey accurate information. We have long held that our NOR policy does not really apply to images.

That said, NOR does apply to the caption (ie text accompanying the image)… like any other text in our article. Blueboar (talk) 22:26, 25 December 2024 (UTC)

...so what you seem to be saying is that anyone can draw up an image at any time and place that into any article because images are mere illustrations? "We have long held that our NOR policy does not really apply to images." Regardless, with editorial consensus, guidelines and policies change around here all the time. - Shearonink (talk) 23:09, 25 December 2024 (UTC)

Yes, anyone can create an image and stick it in an article as an illustration.

That doesn’t mean this image will remain in the article. If others think the article would be better with a different image (or even no image at all), simple consensus can determine this. Blueboar (talk) 00:12, 26 December 2024 (UTC)

Blueboar, we've long turned a blind eye to a certain amount of OR in images (see my own confession above), but a blind eye it is. With the advent of AI, making it way more tempting to generate ill-founded image trash, I believe the time is coming soon that we need to come to Jesus on this question, and clarify more explicitly what is and isn't acceptable. I think your asserted distinction between "illustration" and "information" is a dodge. Even if an image's caption says "Artist's conception" or "AI-generated image", we are putting WP's imprimatur on that image. Having said that, I'll repeat my statement above the the lead image in Matthew the Apostle is fine with me, even though it implies WP's imprimatur for that image, so clearly I don't have a completely crisp idea of the exact criteria apply. EEng 00:39, 26 December 2024 (UTC)

Artistic depiction of Brigette Lundy-Paine

Depends on what X and Y are. Captions wouldn't help in cases such as: [6]. According to Lundy-Paine's Instagram page, they look nothing like that "portrait". (Anyway, I'm just using that as an example; the cartoon portraits on BLPs problem has already been solved, I believe. But who knows when the issue will arise again.) Some1 (talk) 21:11, 25 December 2024 (UTC)

Nothing says we can’t swap one image for an image we think is better. That just takes a simple consensus to achieve. Blueboar (talk) 00:02, 26 December 2024 (UTC)

I confess to being flabbergasted by your position here. You're not seriously suggesting that that cartoon image would be acceptable in that article, under any circumstances, are you (other than if, somehow, the article needed to discuss the cartoon itself)? By your reasoning NPOV-violating text, or NOR-violating text, or V-violating text, would be OK in an article since, ya know, someone can someday swap it out in favor of better text. EEng 01:03, 26 December 2024 (UTC)

Sometimes a cartoon is fine (as an example, the Bayeux tapestry is essentially a cartoon and we use it to illustrate our article on Edward the Confessor)… at other times a cartoon wouldn’t be the best choice. I certainly would prefer a more realistic drawing/painting over a cartoon - if one is available… and a photo over that. However, I would happily bow to consensus if other editors thought the cartoon was the best.

My point is that this decision (cartoon vs painting vs photo) is purely an ascetic one, governed by what is available and consensus discussion at the specific article. Our “No Original Research” policy is not an issue. Blueboar (talk) 02:51, 26 December 2024 (UTC)

You're simply asserting that NOR isn't an issue. V says that All material in Wikipedia mainspace, including everything in articles, lists, and captions, must be verifiable, and though images themselves aren't listed explicitly I don't see how they can escape being considered part of All material in Wikipedia mainspace. As I've mentioned, a blind eye has been turned to this heretofore, but I think that time is coming to an end. EEng 04:19, 26 December 2024 (UTC)

The Edward the Confessor infobox image[7] doesn't fall into the examples that we've been discussing, considering it is a photograph of the Bayeux Tapestry that was presumably taken in a museum; a regular editor like you or me didn't draw/paint/use AI to generate an image of Edward the Confessor then upload it onto Wikipedia. Some1 (talk) 12:54, 26 December 2024 (UTC)

This discussion has me wondering though... what's stopping non-notable artists from uploading their non-notable artwork onto Wikipedia (with their watermarks and all) in an attempt to promote themselves? "[Musician] doesn't have an image? Welp, better draw one really quick and upload it onto Wikipedia to put on their article!" I don't think this phenomenon (with the watermarks) has happened yet on Wikipedia, surprisingly enough. Some1 (talk) 13:19, 26 December 2024 (UTC)

Nothing stops non-notable artists from uploading their art. And if the rest of us feel that this art improves an article, it will remain in the article. Conversely, if the rest of us feel that it does not improve the article, nothing stops us from removing that art, or replacing it with different art.

This is how WP works. Things get added, and things get removed… and when there is disagreement over an addition or removal we discuss it and try to reach a consensus. Blueboar (talk) 14:12, 26 December 2024 (UTC)

I'll repeat yet again that you could use that kind of reasoning to allow all kinds of policy-violating material to be added to an article. That alone can't be the justification. EEng 15:49, 26 December 2024 (UTC)

You are missing my point… we don’t need to amend a policy in order to justify removing things we don’t think improve an article, and we don’t need amend policy to give us permission to add things we think do improve an article. If an image does not improve an article, we remove/replace it. If an image does improve an article, we keep it. If there is disagreement about a specific image, we discuss it. It really is that simple. Blueboar (talk) 19:37, 26 December 2024 (UTC)

While this is technically true, it's also very helpful to have something to point to other than I just don't like it when explaining why you don't think a particular addition is an improvement, and in some discussions about this I've seen the current OI wording being pointed to as meaning that user-created images of any kind are a-okay (example). Nikkimaria (talk) 19:46, 26 December 2024 (UTC)

Men have breasts. They aren't gender or sex specific. GreenLipstickLesbian (talk) 07:08, 26 December 2024 (UTC)

OK, but how about the toenail polish -- is that part of female anatomy? EEng 15:49, 26 December 2024 (UTC)

Yes, of course. GreenLipstickLesbian (talk) 21:10, 26 December 2024 (UTC)

Of course, if anyone wants to find and upload better, or even just different, images in the Standard anatomical position, please feel free. I remember when those photos happened. It was a years-long process that ultimately involved hiring professional models. The modelling agency had a lot of trouble finding anyone who met our criteria (e.g., normal-ish body weight, not heavily tattooed, without heavy tan lines) and was willing to do it. We didn't get everything we wanted (e.g., natural body hair, absence of nail polish), and we were only able to get one woman and one man, but there was nothing else available back then, and this was a substantial improvement. WhatamIdoing (talk) 05:30, 27 December 2024 (UTC)

Useful image of historical subject from Iguanodon, based on stated written sources.

There are all sorts of highly educational editor-created diagrams that we really don't want to lose including diagrams of medical, biological and chemical structures, physics explanations, historical maps, and many more. The vast majority couldn't be replaced with any (recent) published image due to copyright restrictions. Provided that such diagrams are based on written reliable sources, are factually accurate, educationally useful, and agreed by consensus they should be encouraged, not prohibited. MichaelMaggs (talk) 14:44, 26 December 2024 (UTC)

Yes, just so long as we can agree that diagrams are completely different from cartoons of living people. EEng 15:49, 26 December 2024 (UTC)

Agreed. User-created cartoons of living people are normally neither factually accurate nor educationally useful. MichaelMaggs (talk) 16:01, 26 December 2024 (UTC)

I'm just a tiny bit doubtful about that as an absolute rule. I suggest, for example, that a cartoon for Jaiden Animations would be more accurate, educational, and relevant than a photo of her, since she is largely notable for her autobiographical self-portraits. Of course, in that case, I'd want an authentic self-portrait from the artist herself.

What I'm certain of is that some editors deeply loathe any representation of a person that is not "realistic" in style. Caricatures have to be "accurate" and "realistic" in some sense, else they aren't recognizable. But these editors want something "life like". WhatamIdoing (talk) 05:45, 27 December 2024 (UTC)

editors want something "life like" Like an actual photograph of them, preferably, yes. Some1 (talk) 05:49, 27 December 2024 (UTC)

• one aspect I'm seeing above to make sure is treated as acceptable are user made images that are recreations of existing published images in copyrighted sources that can be remade in a copyright free version. Commonly this is for graphs from journal articles where the data is available. Masem (t) 16:09, 26 December 2024 (UTC)

Primary

Moved from Wikipedia talk:Notability § Primary

 – This is the correct venue. dxneo (talk) 23:10, 26 December 2024 (UTC)

Not sure where I should place this discussion, but I hope I'm at the right place. It is often said that interviews are "primary" sources, meaning they are not reliable per WP:PRIMARY. However, most of the times we get personal information (birth dates, birth place and backstory) and upcoming release dates for movies and music from interviews (late-night shows and so on) and they always turn out to be accurate. I think if the interview was published by a reliable source then it's most definitely reliable, because if another publication quotes that interview, no one would say it's not reliable. Not sure if I make much sense, but any objections? dxneo (talk) 23:10, 26 December 2024 (UTC)

Re "they always turn out to be accurate": [citation needed]. Sources directly from the subject of a biography, such as in interviews, can be reliable for uncontroversial factual claims such as birthdates per WP:BLPSELFPUB, but should not be used for evaluative claims nor when there is good reason for skepticism regarding the claims. Even for birthdates, people can often falsify these out of vanity or out of pressure from whatever industry they're in to be a different age. —David Eppstein (talk) 23:40, 26 December 2024 (UTC)

@Dxneo, Primary is not another way to spell 'bad'. You should avoid trying to build an entire article exclusively on primary sources (though this is pretty common for discographies), but you may use reliable primary sources to fill in ordinary or expected details. If you are at all uncertain about the material, consider using WP:INTEXT attribution: "In an interview with Music Magazine, the musician said she was born in California" or "According to Joe Film, the movie will be released in September 2025". WhatamIdoing (talk) 05:51, 27 December 2024 (UTC)

Images whose authenticity is disputed

I understand why we have an exception for images in this policy - we have a limited selection of free images, so we need to rely on user-uploaded content. So if someone uploads a photo they took of a celebrity, that is fine to include in the article since it's not considered original research.

But what happens if someone claims their image is of a certain celebrity but other editors dispute it? It seems like we have limited recourse within policy to handle that. The image doesn't really illustrate or introduce unpublished ideas or arguments, it is just an image of a person that has possibly been mislabelled. Would it make sense to revise the first paragraph of WP:OI to the following? The last sentence is new:

Because of copyright laws in several countries, there may be relatively few images available for use on Wikipedia. Editors are therefore encouraged to upload their own images, releasing them under appropriate Creative Commons licenses or other free licenses. Original images created by a Wikimedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments, the core reason behind the "No original research" policy. Image captions are subject to this policy no less than statements in the body of the article. Additionally, images whose authenticity is disputed may be removed in accordance with consensus.

– Anne drew 16:42, 27 December 2024 (UTC)

Good point and good start. But "removed in accordance with wp:consensus" is unclear. Do you need a consensus to remove? Do you need a consensus to keep? North8000 (talk) 20:10, 27 December 2024 (UTC)

In the case of a living person, we need to take extra care to “get it right”… therefore we would default to needing a “consensus to keep” if there were a disagreement over the image. Blueboar (talk) 20:20, 27 December 2024 (UTC)

I don't we need to write it out on the page, WP:ONUS is pretty clear. Traumnovelle (talk) 21:05, 27 December 2024 (UTC)

I don't think the addition is necessary: In all cases, for all images, for all material, for any reason, consensus can force removal.

One of my touchstones for this policy is a dispute years ago with a since-blocked AIDS denialist. Look through Wikipedia talk:Verifiability/Archive 40#RfC: Do images need to be verifiable? for one of the discussions. AFAICT he wanted certain images removed from Wikipedia because the existence of a photomicrograph of a virus undercut his story that these viruses don't exist, but since that's not a policy-based reason, he generally asked for images to be removed if there was no source to authenticate the contents. We had "images whose authenticity were disputed" – but only by a POV pusher. I would not wish to give him, or POV pushers like him, a rule that says that disputing authenticity is his best path to removing the image. I think we could safely predict that this addition would turn into a WP:BEANSY recommendation to partisan editors to dispute the authenticity of all unflattering photos of their favorite politicians.

What I'd suggest instead, in these cases, is relying on WP:PERTINENCE, which says "Images should look like what they are meant to illustrate, whether or not they are provably authentic." In the situation described above, we have an image of a BLP that editors dispute. Why do they dispute it? I'd guess it's because it doesn't look like the person. I'd bet that most of us have had the experience of a photo not turning out the way we expect, and even though we know with absolute certainty who is pictured in the photo, we couldn't say that the photo is representative of the person. That might be the only thing that's going on in this photo: Right person, but odd angle, odd expression, odd lighting – and the result is that the image doesn't look like what it's mean to illustrate, and therefore should be rejected per MOS:IMAGES. One doesn't even have to dispute the authenticity to do this: just say that it doesn't look like what you/readers expect, and the guideline therefore rejects it. WhatamIdoing (talk) 22:12, 27 December 2024 (UTC)

Thanks for the detailed reply! I hear you on using WP:PERTINENCE in these cases. That's what I leaned on in a recent discussion around a disputed photo of a tornado. It just seemed like a bit of a workaround, having to first dispute the verifiability of the caption, and then separately the pertinence of the image itself. I think expecting editors to formulate an argument like that using multiple policies/guidelines is asking a lot.

But maybe I'm overthinking this. To your point, it's already the case that consensus can remove disputed photos. But I do think there would be some value in explicitly stating that WP:OI isn't intended to help retain inauthentic photographs. For what it's worth, this isn't a one-off issue. In a similar discussion this month, I incorrectly relied on WP:OR to support the removal an image with disputed authenticity. – Anne drew 22:30, 27 December 2024 (UTC)