Wikipedia talk:Reliable sources/Archive 71

Archive 65

←

Archive 69

Archive 70

Archive 71

Primary source birth certificate, death records and such

WP:FAMILYSEARCH reads FamilySearch also hosts primary source documents, such as birth certificates, which may be usable in limited situations, as well as a large collection of digitized books, which should be evaluated on their own for reliability.. However, doing your own research using primary sources is original research. Could someone provide examples of "limited situations" of when it is acceptable for Wikipedia editors to use primary source death/birth certificates?

For example, this is something I noticed somewhere. Is this a type of application where it is acceptable for an editor to use primary source that maybe hosted on FamilySearch type site?

On December 31, 1924, New Year's Eve, Koepp married Zahrah Lee at the home of her parents in Hollywood, California. They had five children.  

Graywalls (talk) 07:59, 30 August 2023 (UTC)

See WP:PRIMARY. To be a source at all, as far as Wikipedia is concerned, a source must be published. To be used in Wikipedia, it must be reliable. Some primary sources are both published, and reliable, so they may be used in Wikipedia. Just reporting what a primary source says is not original research. Sometimes authors outside Wikipedia examine an array of primary sources to form their own original conclusions; that's original research.

Wikipedia:Biographies of living persons#Avoid misuse of primary sources indicates that certain kinds of primary sources should not be used in biographies of living persons, such public records that include personal details, such as date of birth.... But it may be acceptable to rely on primary source material to augment a secondary source. Jc3s5h (talk) 18:11, 30 August 2023 (UTC)

A case that I would consider legitimate (for deceased persons) is when we have conflicting accounts in older sources and a primary document that was published later. E.g. you have two books from the 1970s one says XX was born Feb 5, 1845, the other says XX was born Feb 6, 1845. Then in 2020 old church records are published and there is a record for XX that says that XX was born Feb 5, 1845 and baptized Feb 6, 1845. In that case I think it is fine to say that church records confirm the first account. What I would find problematic at best would be if sources said XX was born sometime in the 1840s and someone uses the church records to say XX was born Feb 5, 1845. -- Random person no 362478479 (talk) 23:22, 1 September 2023 (UTC)

References

[I moved the following answer from Wikipedia:No original research/Noticeboard after I realised forum shopping was going on. Jc3s5h (talk) 17:18, 1 September 2023 (UTC)]

The birth date of John Penn ("the American") is sourced to 17th-century Quaker records. Not only is it a good source, but because it is chronological, we can go through the records and observe when February 29th does or does not occur, and when the year changes, so we can deal with all the confusion between the Julian and Gregorian calendars. It isn't original research because it's just excerpting a fact from a primary source. Original research would be reading a bunch of sources and drawing new conclusions that nobody published before.

Primary sources are not supposed to be used in Wikipedia:Biographies of living persons unless the primary source is being used to augment a secondary source. But John Penn died long ago so this doesn't apply. Jc3s5h (talk) 17:11, 1 September 2023 (UTC)

I also asked there trying to get clear understanding of "original research". Things scholars often do is look at primary sources, then try to interpret them. For example, if you have to look at several records to deduce the answer, that's researching. I know its not too uncommon for fathers to name their son exactly the same as himself and you'll sometimes have John Adam Smith Jr, or John Adam Smith II and such and if it takes research to properly identify the record, I believe that's OR. My reason to ask at ORN was to establish what constitutes OR. If it was that simple, it would say "primary source records are ok if the subject is deceased" Graywalls (talk) 19:01, 1 September 2023 (UTC)

I think you'll find that it's a bit more complicated than just setting down a couple of rules.

We don't want editors to put previously unpublished facts (or claims) into articles. We do want editors to understand the source they're citing. Something like reading an entire book so you can make sure that the statement you've posted is accurate is not "original research" (=making up new claims).

NOR used to have a way of expressing this idea that might help you: 'Research that consists of collecting and organizing material from existing sources is encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia.' WhatamIdoing (talk) 20:38, 1 September 2023 (UTC)

That's from well over a decade ago. WP:PRIMARY summarizes our policy on not interpreting primary sources. So, that includes the interpretive process involved in reviewing multiple primary sources and connecting them together to identify the right one. A process that is supposed to be done by a reputable source, like the news media. There's nothing saying it's alright to do it ourselves for anything other than live people. Graywalls (talk) 22:16, 1 September 2023 (UTC)

I would say it is not acceptable to include info from primary sources when the bulk of that info isn't first introduced in secondary sources. So stating someone had 5 kids based solely on an editor finding a census record for that person would be UNDUE. JoelleJay (talk) 23:09, 1 September 2023 (UTC)

Ainsley Earhardt reliable sources birth date and birth year

https://meaww.com/fox-friends-host-ainsley-earhardt-celebrates-her-birthday-on-set-with-co-hosts-as-she-turns-47 Is this a good reliable source for Ainsley Earhardt, and can u edit her page since I’m blocked indefinitely from editing? Her DOB is 1976|9|20, and I think meaww.com is a reliable source. Do u guys think the link that I copied on here is a good enough source since it has her age, and I can’t edit her article right now bc I’m blocked from editing. Dandielayla (talk) 02:55, 25 September 2023 (UTC)

No. Also: there is an open request about your attempts to edit on this topic at WP:ANI#Dandielayla, Ainsley Earhardt, and competence. Coming to a page like this to ask people to proxy edit for you while you're page blocked is a terrible idea. It is time for you to walk away from the Ainsley Earhardt article. Find something else to edit, Wikipedia has many, many other articles. MrOllie (talk) 03:02, 25 September 2023 (UTC)

Oh ok, and I know. How long does the block last, and I could walk away from the article until they edit it so I didn’t have to worry anymore. All they have to do is put a reliable source, and the birth date. I can try to edit other articles, and see if it works. Dandielayla (talk) 03:14, 25 September 2023 (UTC)

Also, you've already asked if this source was reliable on your talk page. As Kcj5062 stated, Fox News talk shows are not reliable. See WP:FOXNEWS. If you cannot find any reliable sources, I strongly suggest you finally stop and leave the Ainsley Earhardt article alone. Please consider finding other articles you may have interests in. Regards. 🛧Midori No Sora♪🛪 ( ☁＝☁＝✈) 03:14, 25 September 2023 (UTC)

Oh ok 👌🏻, and I will find other articles to edit on. I have other article interests, and I realized I’m blocked from editing her page. I will let them fix it so I don’t have to worry anymore, and they could put the reliable source along with her birth date since I can’t really do that. Dandielayla (talk) 03:17, 25 September 2023 (UTC)

Adding YouTube.

Is it possible where they talk about some examples of ''Unacceptable user-generated sources", we add YouTube as an unacceptable user-generated source as well? SRKFan1965 (talk) 17:26, 9 October 2023 (UTC)

@SRKFan1965 It's a little more complex, YT is a platform. A CNN-clip from CNN's YT-channel is as reliable as cnn.com. See WP:RSPYT. Even sources like Twitter and Facebook can have WP:ABOUTSELF use. Gråbergs Gråa Sång (talk) 17:34, 9 October 2023 (UTC)

Discussion on changes to the header of WP:Reliable sources/noticeboard

I've started a discussion at Wikipedia talk:Reliable sources/Noticeboard#Cutting most of the header, to discuss a proposed change to the head of the Reliable sources noticeboard. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:46, 9 October 2023 (UTC)

Page move

I have edited a title name page and after publish is still shows the old name. How can i fix this? Mironlaurentiu (talk) 12:45, 12 October 2023 (UTC)

@Mironlaurentiu See WP:MOVE and WP:COMMONNAME. Gråbergs Gråa Sång (talk) 17:57, 12 October 2023 (UTC)

Discussion on changes to the edit notice of WP:Reliable sources/noticeboard

I've started a thread at WT:RSN#Suggested changes to the edit notice, to discuss a proposed change to the edit notice of the Reliable sources noticeboard.-- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:10, 24 October 2023 (UTC)

Citing the person being biographed as a realiable source?

I see that the person being written about is considered the authoritative source, but I don't see how to properly cite such information. Heideana (talk) 09:01, 28 October 2023 (UTC)

It is certainly not true that the subject of a biography is necessarily considered an 'authoritative source' - this very much depends what they are being cited for, and biographies, like any other Wikipedia article, are expected to be based largely around content cited to sources independent of the subject. Since I assume your question concerns to the Martin Burckhardt biography, and you appear to have a conflict of interest, you probably shouldn't be editing the article directly anyway, and I'd recommend you instead do as I've suggested on Talk:Martin Burckhardt, and concentrate on finding such independent sources. We can figure out issues with citation etc when we have something to work with. AndyTheGrump (talk) 10:48, 28 October 2023 (UTC)

Léo collection

Business Today. 30 October 2023. Archived from the original on 30 October 2023. Retrieved 30 October 2023.530 crores not 561 crores Give any reliable sources for 561 crores ? 2A02:8440:3112:6B9E:0:46:E36A:2A01 (talk) 15:13, 30 October 2023 (UTC)

Hamas charter

The page doesn’t include their current charter from 2017.which actually replaces the 1988 one. E.hamam (talk) 12:29, 2 November 2023 (UTC)

This page is to discuss change to Wikipedia's Reliable Sources policy, you should raise this question on the articles talk page. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:43, 2 November 2023 (UTC)

Can you use Federal Bureau of Prisons as a source, and how?

An editor added information on Enrique Tarrio's place of incarceration, which is fine. However, they added their source as an external link, and as I tried to convert it to a proper citation, I discovered that the URL for searching prisoner info doesn't update with a specific prisoner when you search. How to resolve this? Thanks. Wes sideman (talk) 13:27, 2 November 2023 (UTC)

You can add instructions after the cite but still within the ref tags. However if the only source is primary it may not be WP:DUE, especially as the other reference for the location of this incarceration is a blogsite. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:48, 2 November 2023 (UTC)

Based on that, should I remove it entirely? Wes sideman (talk) 13:36, 3 November 2023 (UTC)

WP:BLPPRIMARY excludes using public records such as the prisoner info, so it should be removed. As to the other source the opening of WP:BLP says "Be very firm about the use of high-quality sources.", and I wouldn't count it as a high quality source (or even a reliable one). -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 13:40, 3 November 2023 (UTC)

Bachelor thesis

The text does not say anything about a bachelor thesis, is this considered reliable? I guess not, but maybe this should be added to the text? PJ Geest (talk) 16:55, 8 November 2023 (UTC)

As the current text says PhD dissertations may be reliable, and masters are only reliable if they can be shown to have had significant scholarly influence, I doubt bachelor thesis would be considered reliable unless you can show it's an exceptional circumstance. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:26, 8 November 2023 (UTC)

Court records

Are these considered a "reliable source" or "original research"?

Context: a state makes available extensive court records, including police reports, search warrants, lawsuits etc. However they're not deep-linkable, i.e. it's not possible to link directly to a person's arrest report, but anyone can easily go to the court website and search for the person's name and get their arrest report. RobinLikeTheBird (talk) 09:41, 23 November 2023 (UTC)

The problem with court records is often WP:BLPPRIMARY, or, if it's not about living people, WP:PRIMARY. In the non-living case, they may have some use. Gråbergs Gråa Sång (talk) 10:05, 23 November 2023 (UTC)

The process you'd have to go through (e.g., go to the court website and search) has nothing to do with whether or not it's a reliable source. If it's a reliable source, then it's not original research, per the actual definition (first sentence of Wikipedia:No original research). In RS == not OR. WhatamIdoing (talk) 00:44, 29 November 2023 (UTC)

Jailer collection falsely reported - 650 crores

Jailer collection has been falsely reported. Real collection is 650+ crores (Indian money). Citation: Hindustan times, times of India, India.com , business today.

article links: https://www.hindustantimes.com/entertainment/tamil-cinema/jailer-box-office-collection-650-crore-worldwide-rajinikanth-film-101694587768433-amp.html

https://m.timesofindia.com/entertainment/tamil/movies/news/after-rajinikanth-and-nelson-dilipkumar-anirudh-ravichander-gets-a-cheque-from-jailer-makers/amp_articleshow/103362405.cms

https://www.businesstoday.in/amp/trending/box-office/story/jailer-worldwide-box-office-collection-day-20-rajinikanths-film-may-break-baahubalis-record-marches-towards-rs-650-crore-396030-2023-08-29

https://www.india.com/entertainment/jailer-box-office-collection-day-34-rajnikanths-film-collects-rs-650-crore-amid-jawan-mania-check-detailed-report-tamil-news-6319161/amp/ Vj789 (talk) 09:08, 2 December 2023 (UTC)

@Vj789 If this is about the en-WP article Jailer (2023 Tamil film), the place to discuss it is Talk:Jailer (2023 Tamil film). Gråbergs Gråa Sång (talk) 09:15, 2 December 2023 (UTC)

I’m referring to this page. List of Tamil films of 2023 Which is totally misleading. I want the content to reliable and truthful in wiki hence I submitted my citation links from reliable Indian print and web media. Vj789 (talk) 09:19, 2 December 2023 (UTC)

I see you found that talkpage. Gråbergs Gråa Sång (talk) 09:36, 2 December 2023 (UTC)

Getty Images

Is Getty Images considered a notable source? WebSummitEditor (talk) 21:19, 10 December 2023 (UTC)

Getty Images sells photos from many different sources, your question needs a lot more context to answer. But in general just because an image appears on Getty doesn't make it notable/reliable, it could be but as I said that will require a lot more information. -- LCU ActivelyDisinterested «@» °∆t° 21:33, 10 December 2023 (UTC)

@ActivelyDisinterested: hey there. thanks for the reply. I recently attended Web Summit in Lisbon and want to work on updating both information and photos from the conference to Wikipedia. I noticed some of the sessions don't have articles, but they do have Getty photos mentioning some information from the panels such as titles, company names, etc. Could this be used to update information on Wikipedia? WebSummitEditor (talk) 00:30, 12 December 2023 (UTC)

Would it not be easier to reference the events schedule on the website? Each session has its own page, such as this one, also all the speakers are listed as well. -- LCU ActivelyDisinterested «@» °∆t° 00:40, 12 December 2023 (UTC)

Ok thanks. That's better. I appreciate it. WebSummitEditor (talk) 01:55, 12 December 2023 (UTC)

Proposed merger of WP:SELFSOURCE to WP:ABOUTSELF

FYI

 – Pointer to relevant discussion elsewhere.

Please see: Wikipedia talk:Verifiability#Merge WP:SELFSOURCE to WP:ABOUTSELF. Summary: a fairly complicated section in the WP:RS guideline near-exactly duplicates one in the WP:V policy, and they have separate shortcuts.  — SMcCandlish ☏ ¢ 😼  16:13, 13 December 2023 (UTC)

Use of AI content generation by previously reliable sources

In the past year, we've seen USA Today,[1] Buzzfeed,[2] CNET,[3] Sports Illustrated,[4] Men's Journal,[5] and many other sources publishing AI-generated content, often deceptively under the byline of a fake author. And in at least 1 case, this content included medical claims and medical advice.[6] This is obviously just the tip of the iceberg as such AI-outsourcing is only going to become more and more common as AI-technology advances. This represents a serious threat to Wikipedia's reliability as many sources we previously assumed are reliable can no longer be given such assumptions. What can we do to address this problem? Nosferattus (talk) 16:03, 28 November 2023 (UTC)

I fear all we can do is start taking note of when they started using AI content farms, and stating all articles after that date are WP:GUNREL. Outside Wikipedia individuals may try and pressure sites to not go down that route. -- LCU ActivelyDisinterested «@» °∆t° 22:10, 28 November 2023 (UTC)

At some level, this is nothing new: sources sometimes change reputation (in either direction), often as a result of changed ownership. I remember when Red Herring (magazine) was a reliable source; now it's not. The source about Sports Illustrated says they've started labeling third-party content, which is going to help us with that particular source.

It might help if we encouraged editors to Wikipedia:Use common sense more. Sure, telling people that they can reject a source because it feels unreliable could be a gift to POV pushers, but we really do need editors to use some Wikipedia:Editorial discretion. If an article reads like it was written by a machine (or a humorist: volleyball "can be a little tricky to get into, especially without an actual ball to practice with"), then editors should look for better sources. WhatamIdoing (talk) 00:58, 29 November 2023 (UTC)

The problem that misses is that AI can generate some very believable nonsense, and that issue is only going to get worse as AI models improve. The models are improving at generating content people will click on, not well researched and fact checked journalism. Those models are training to fool people's commonsense. Pointing editors towards WP:RS and reliable, independent, published sources with a reputation for fact-checking and accuracy when they ask for advice is showing how they should exercise editorial discretion. -- LCU ActivelyDisinterested «@» °∆t° 02:49, 29 November 2023 (UTC)

AI-generated content is presumably inconsistent with a reputation for fact-checking. I suppose that, theoretically, someone could generate an AI article and then (pay to) fact-check it, but that seems odd. It'd probably be faster and cheaper to write it properly in the first place. Also, given the (lack of) quality in the ones that we're catching, it's unlikely that they're even reading them before posting them.

The "believable nonsense" will presumably also not align with other sources, so it could be rejected as UNDUE. That should help with the most popular subjects, but it will be more difficult if it's the only source that has the desired detail (e.g., a birth date). WhatamIdoing (talk) 05:36, 29 November 2023 (UTC)

The problem with believable nonsense it that people will won't to believe it, getting readers involved increases the average length of their visit and so add revenue. And as AI content becomes more prevalent AI content will be based and align more with other AI content. -- LCU ActivelyDisinterested «@» °∆t° 11:49, 29 November 2023 (UTC)

• I think that "this article smells weird" is a reason to examine its publisher more clearly, investigate whether there's coverage indicating a sudden shift to their reputation, and to raise the issue on WP:RSN if there's a dispute so other people can help investigate; certainly if I were adding a source that I felt smelled weird, so to speak, I would probably find a better source. But when it comes to actual disputes over sourcing it's not enough - as someone who edits a lot of controversial articles, I can definitely say that encouraging editors in a dispute to go "I reject that source; it reads as AI-generated to me" would lead to problems. When a source's RS status comes into dispute, we rely on a reptuation for fact-checking and accuracy as opposed to this source smells bad to me for a reason. The good news is that (as those sources show) it usually doesn't take long for people to realize a source is using AI-generation and for its reputation to collapse as a result. --Aquillion (talk) 08:23, 29 November 2023 (UTC)

  Completely agree the smell test will also become nothing more than "this looks badly written" as the AI models improve in their ability to fool real people. This was why I suggested keeping track of when sites start using unlabeled AI content, as it would give editors knowledge of when addition caution might be needed. -- LCU ActivelyDisinterested «@» °∆t° 11:44, 29 November 2023 (UTC)

  I wonder if we could source a List of websites accused of using AI-generated content. WhatamIdoing (talk) 22:58, 29 November 2023 (UTC)

  That was my general idea (although I would suggest something outside mainspace, otherwise it has to follow content rules), add dates it started, what to look out for if they do label them, etc. It justs becomes another exception in the same way opinions pieces are handled -- LCU ActivelyDisinterested «@» °∆t° 23:42, 29 November 2023 (UTC)

  We should source it when we can, of course, but an out-of-mainspace list could be useful to editors. WhatamIdoing (talk) 00:39, 1 December 2023 (UTC)

  Definitely should add sources, but if it's in main space we can't add additional commentary specific to Wikipedia concerns (as they wouldn't be encyclopedic content). -- LCU ActivelyDisinterested «@» °∆t° 15:19, 1 December 2023 (UTC)

I think we need to distinguish between articles where an AI is fed an initial set of facts and then asked to draft an article (example prompt: The Farmingdale Bulldogs beat the Syosset Alligators by 54 to 3 on Saturday September 4 at Farmingdale. Write a sports story about the game.), and articles that are generated by AI based solely on its training data set (example prompt: Write a sports story about the Farmingdale Bulldogs beating the Syosset Alligators). In my experience, if you feed an initial set of facts into a large language model, such as GPT 4, and prompt it to write an article, it will generally preserve the facts. This is what Gannet was experimenting with for high school sports coverage, as described in the futurism.com article. The writing may be lousy or boring, but I don’t see that as necessarily compromising the reliability of the information in the article. On the other hand, as is now well known, anything generated solely by an LLM’s training data set is “per se” unreliable. Nowa (talk) 13:48, 3 December 2023 (UTC)

Reading the Futurism article it seems clear the Gannet articles are little more than statistical data (scores, etc) and AI generated verbiage that has little to do with the actually games.
"A suffocating defense, helped Franklin South County handle Bloomington North 4-0 on Aug. 30 in Indiana girls high schools soccer action."
"A suffocating defense, helped Nahunta Brantley County handle Garden City Groves 30-0 in a Georgia high school football matchup."
"A suffocating defense, helped Upper Dublin handle Kennett 21-0 in Pennsylvania high school football action on Aug. 25.
"A suffocating defense, helped Ontario handle Centerburg 35-0 in an Ohio high school football matchup on Aug. 25."
These are simply table entries with additional AI wordsalad created to pump page views. Good for score and date maybe, but no more than a statistical database would be. I feel it would be a very bad idea to using any wording on the nature of events of the games from sight articles. Also given the very low quality nature of the articles it seems very doubtful that any human fact checking happened before publishing. -- LCU ActivelyDisinterested «@» °∆t° 23:47, 3 December 2023 (UTC)

Maybe including some directions on what is and may not be reliable by entry would be a solution for such a situation. -- LCU ActivelyDisinterested «@» °∆t° 23:48, 3 December 2023 (UTC)

@Nowa: The most problematic AI-generated article to date was one that was created by feeding it "an initial set of facts". The Men's Journal article about low testosterone was created by feeding the AI all of Men's Journal's previous articles about testosterone and asking it to generate a new article. Because of this, it was assumed the article was "safe" and no one bothered to fact-check it. The problem was, it built the article out of outdated information and gave medical advice to the reader based on medical claims that had since been disproven or supplanted with more modern medical research. It then presented this information as if it were current and factual. Nosferattus (talk) 22:18, 12 December 2023 (UTC)

I expect the issue comes with attribution. An AI author is meaningless -- it's the editor then that needs to be on the byline, as they are responsible for the veracity of the article (and accountable to any errors). They are also accountable for the amount of work in the piece that was strictly AI-generated versus rewritten or supplemented (although this is no different from the largely invisible editor's role now at newspapers, in which they may rewrite substantial parts or all of another person's stories themselves, or else drop contributors from bylines, etc.)

In other words, I don't see an AI writer as an issue as long as we know who is accountable to the material when it's published, and whether that person is reliable. SamuelRiv (talk) 01:40, 4 December 2023 (UTC)

Any source that uses AI to create content should be downgraded to generally unreliable, with the exception of articles proven to have been written by humans and/or having had proper editorial oversight. Cortador (talk) 10:13, 6 December 2023 (UTC)

If it's a matter of oversight, then what does it matter if the prose was produced by an AI or some rookie contributor or intern? And if a human intern burns out and ditches after two weeks, how does the newspaper prove that they wrote it? SamuelRiv (talk) 15:32, 6 December 2023 (UTC)

No company is going to thoroughly fact-check AI-generated articles, as it defeats the entire purpose of using AI, which is to avoid the expense of paying humans. If a source is using AI, it effectively means they no longer care about being reliable. Nosferattus (talk) 22:25, 12 December 2023 (UTC)

I agree with you (if you're generating content via AI to save money, you're not going to spend money on fact checkers), but it suggests a simplification to the guideline: We have said (for approximately forever) that one of the characteristics of reliable sources is that they have a reputation for fact checking (or similar processes, such as peer review). All we would need to do is to add a small note saying "AI-generated content has a reputation for not being fact checked". It would not require additional sections, new rules, or anything else. WhatamIdoing (talk) 00:55, 13 December 2023 (UTC)

So by this assumption if my company spends money on AI to generate computer code, they're not gonna pay for QA/debugging now? Considering we've been in this new generation of generative AI for less than a year, I don't how the editorial use of such content as a whole has garnered any stable reputation. SamuelRiv (talk) 06:26, 13 December 2023 (UTC)

The analogy doesn't work, not having debuggers would cost your company money (because you software wouldn't work). There is no such incentive for fact checking AI generated content. The impetus behind it is to drive site traffic at very little cost, shown by responses similar to "We hired a third party to drive SEO, and they used AI to generate content" by both CNN and Sports Illustrated. -- LCU ActivelyDisinterested «@» °∆t° 14:31, 13 December 2023 (UTC)

Also, the story for a software company is:

• I'm paying for coders and testers. Maybe I could pay fewer coders if we use AI.

The story for most media outlets is:

• I'm paying for writers, but we don't actually fact-check anything (or only if the editor thinks there's a legal risk, and even that's done as minimally as necessary and as inexpensively as possible). Maybe I could pay fewer writers if we use AI.

Hint: If you think that everything's being fact-checked, then look at websites for hiring employees or the websites of fact-checking organizations such as https://www.snopes.com/careers/ and see how few such jobs are listed. There are 85,000 full-time professional journalists in the US. There are very few fact checkers. WhatamIdoing (talk) 05:17, 14 December 2023 (UTC)

Proposed simplification of WP:RS/AC

I propose removing the struck text below, because it is repetitive and even misleading. The guidance on this can be very simply stated.

A statement that all or most scientists or scholars hold a certain view requires reliable sourcing that directly says that all or most scientists or scholars hold that view. Otherwise, individual opinions should be identified as those of particular, named sources. Editors should avoid original research especially with regard to making blanket statements based on novel syntheses of disparate material. Stated simply, any statement in Wikipedia that academic consensus exists on a topic must be sourced rather than being based on the opinion or assessment of editors. Review articles, especially those printed in academic review journals that survey the literature, can help clarify academic consensus.

Bon courage (talk) 10:02, 17 December 2023 (UTC)

and even misleading. how? Softlem (talk) 14:20, 17 December 2023 (UTC)

By saying "individual opinions should be identified as those of particular, named sources", which subtly undercuts the (more complex) situation as set out in policy (WP:YESPOV). Besides, this guideline should be about identifying reliable sources, not straying into discussion of how to present material. Bon courage (talk) 14:24, 17 December 2023 (UTC)

Does even your cut-down version really belong in this guideline? WhatamIdoing (talk) 22:51, 23 December 2023 (UTC)

Arguably, since it's saying what is reliable for something. I think maybe the cut-down is too terse. So how about:

A statement that all or most scientists or scholars hold a certain view, or that there is "consensus", requires reliable sourcing that directly says that.

Bon courage (talk) 02:53, 24 December 2023 (UTC)

I can use a medical school textbook to write "HIV causes AIDS", but to write "Most scientists believe that HIV causes AIDS" (which is a weaker statement), I'd have to have a source that comments on the prevalence of the belief among the stated population. Somehow, that seems backwards.

Is this meant to be restricted to opinions (i.e., matters about which different experts could legitimately hold different viewpoints)? It makes sense to me to follow this rule for a statement like "The Tiger in the Smoke is widely regarded as the best book Margery Allingham ever wrote", but I'm not sure that it makes sense for a statement about facts (i.e., matters about which there is a universal right or wrong answer). WhatamIdoing (talk) 04:32, 24 December 2023 (UTC)

"I can use a medical school textbook to write "HIV causes AIDS", but to write "Most scientists believe that HIV causes AIDS" (which is a weaker statement)" ← you have put your finger on the bizarre nature of this guidance. But I was seeking merely to simplify the wording, not change it into being sensible. I'd be up for that too. Bon courage (talk) 04:48, 24 December 2023 (UTC)

I'm also thinking this doesn't belong in RS. The section struck out is discussing NPOV not RS, and what's left is wrong. If scientific/medical literature says that, as an example, "HIV causes AIDS", then there is no need to find sources that say that is the scientific consensus. There shouldn't be a requirement to 'prove' that vaccines are good, the world is round, or that the holocaust happened.
Maybe this could be altered to advice for when the scientific consensus isn't absolute ("The majority of physicists believe that <M-theory|Loop quantum cosmology> best describes gravity at a quantum scale"). Although I fear something like that would likely be misused. -- LCU ActivelyDisinterested «@» °∆t° 14:00, 24 December 2023 (UTC)

Maybe I am misreading your comment, but it seems like there is quite obviously a necessity to prove the world is round. What other reason, besides proof, could we possibly have for thinking (or saying) so? jp×g🗯️ 23:25, 25 December 2023 (UTC)

Given that every relevant expert says the world is round, do we really need a source that says "Every relevant expert believes the world is round"? Why not accept sources that (a) assert that the world is round, and (b) are reliable for that assertion?

This is most likely to be a problem for fringe science, when you want to write something like "The multi-level marketing company claims that string theory explains why televisions are bad for your mental health, but scientists say this is nonsense". WhatamIdoing (talk) 05:09, 26 December 2023 (UTC)

Methinks you're talking the same language but not the same dialect. There's the question of scientific knowledge - I ran 100 experiments, and 89 times, I received result X. A source interprets that result and produces some statements, arguments, and contentions which we want to import and distill into a cohesive structure that is legible. There's epistemology, metaphysics, philosophy, ways of knowing. I don't know what I know but I choose to believe because I need to believe things to make other things make sense, and go about my life interacting with objects. But verifiability is a different kind of knowledge. And original research has long been a pillar of that, because it rests on the idea that you need to show your work, and make sure it's cited to an academic standard of library science bibliography knowledge. It's quite possible to have knowledge that is verifiable but not to that exacting standard. The gray area is editorial interpretation and consensus. What this passage or statutory stanza of the guideline was attempting to express, is the idea that one shouldn't tread a less established topic, and make conclusions that haven't been made yet by the academic consensus. Citogenesis would be a risk of doing that. It's certainly happened. Of course, there is a need for some guesswork and rules of thumb to determine an academic consensus and it's often not because you're writing the sentence "the academic consensus is X." It's much more subtle. Andre🚐 05:22, 26 December 2023 (UTC)

That was my second point, but I couldn't think of a way of writing that that couldn't be taken the wrong way. Just saying "where the academic consensus isn't clear" is just something which every pro-fringe or POV pusher will make a timesink. -- LCU ActivelyDisinterested «@» °∆t° 13:30, 26 December 2023 (UTC)

Yes that's a misreading of my comment, we don't need to have a source to say that "the world is round" is the scientific consensus. We shouldn't be say in wikivoice "Most scientists believe the world is round[1]", and have to find a source that specifically says "most scientists believe the world is round". The section could be taken for saying that this is something which must be done. -- LCU ActivelyDisinterested «@» °∆t° 13:22, 26 December 2023 (UTC)

This is an aspect of SYNTH or OR, ie stating an academic consensus exists by doing one's own review of the literature and concluding, absent a review stating this, that it is the consensus academic view. So maybe it should be moved to one of those other policies. Andre🚐 23:36, 25 December 2023 (UTC)

All forms of OR involve verifiability in reliable sources, but I think we need a little room for editorial decisions here. For example, given these facts:

• HIV causes AIDS.
• AIDS denialists (for which POV Wikipedia:Fringe theories applies) don't believe this.

Sometimes you want to write a sentence like "HIV causes AIDS", which is easy enough: It's the prevailing scientific consensus that HIV causes AIDS, so you write "HIV causes AIDS" and cited a WP:MEDRS source for it.

However, sometimes you want to write something that shows a relevant contrast, like "AIDS denialists claim that HIV is harmless, but the prevailing scientific consensus is that HIV causes AIDS". According to this, it's not enough for the statement to be obviously true, based (e.g.,) on the contents of every medical school textbook that mentions HIV; you have to be able to cite a source that specifically mentions the near-universal prevalence of this belief among relevant experts. WhatamIdoing (talk) 05:06, 26 December 2023 (UTC)

I see what you mean, and I'd favor a carveout to exempt from OR, any rationally based perception that any reasonable person with BLUESKY background knowledge and a high school education, could determine or deduce, using the ordinary tools of logic, arithmetic, readings of maps and charts, etc. This is what WP:SYNTHNOT refers to as simple synthesis or simple juxtaposition. Obviously, people routinely go much farther than this on math and science articles, including doing things that I, with a bachelor's degree and some graduate coursework in computer science, am mystified by. But I think you're thinking of things more like the AIDS example, and I agree. There should be a generous carveout for such situations. Andre🚐 05:11, 26 December 2023 (UTC)

5pillarsUK

The editor and deputy editor of this site are heavily anti-Israel and i wonder if their work can be accurately used to describe the complex Arab-Israeli conflict. Steveonsi (talk) 02:00, 9 January 2024 (UTC)

Hi, @Steveonsi. Generally, if you want to talk about a specific source, it's better to take that question to Wikipedia:Reliable sources/Noticeboard. But since you're here, I think it's useful to think about the goal of having a balanced Wikipedia article, which can be done (and sometimes might need to be done) by using sources that, if looked at individually, are biased. E. B. White recommended that schools "should strive for a well-balanced library, not a well-balanced book", and the same principle applies here. We can use works that strongly take a particular viewpoint, so long as the end result is a well-balanced Wikipedia article. It is often better to use several sources, each of which argue strongly for their own viewpoint, than to pick one source that is superficially neutral. WhatamIdoing (talk) 04:05, 9 January 2024 (UTC)

I'm not seeing anything that would make 5pillarsUK an unreliable source, it has its own bias but not anything that can't be handled by the ideas set out in WP:RSBIAS. The advice of WP:HEADLINES seems relevant as well, as much of their actual reporting seems less 'eye catching' then their headlines (certainly not something limited just to 5pillarsUK). -- LCU ActivelyDisinterested «@» °∆t° 11:38, 9 January 2024 (UTC)

POV journals

In the WP:SCHOLARSHIP section, our guidance says: "Care should be taken with journals that exist mainly to promote a particular point of view.". Do we have any further guidance on how we might evaluate whether a journal exists mainly to promote a particular point of view? Barnards.tar.gz (talk) 11:33, 14 January 2024 (UTC)

Some journos do declare such POV in their mission statement or the like, but if not I think one would have to defer to verifiable majority consensus within the field in question. Walkersam (talk) 21:57, 24 January 2024 (UTC)

Local weekly column section

If the discussion was to be around the specific source, I would be posting at RSN although I want to discuss the general idea of it. In Draft:Mary L. Hamlin, this source http://pineconearchive.fileburstcdn.com/220916PC.pdf (page 27-29) is used extensively. It is not clearly delineated as a "column" section but opening like "in my previous column..." indicates it is a column. Obviously, a local township paper can devote an in-depth coverage on extra hyper-local matters such as buildings or people in the said township although, should such columns be part of notability determination and should it be appropriate to develop a significant amount of article from such local columns? Graywalls (talk) 18:57, 20 January 2024 (UTC)

Whether it's a column is irrelevant. The question that could be relevant is whether hyper-local sources matter.

The answer depends on whether you believe the purpose of notability standards is to exclude "unworthy" subjects from Wikipedia, or to accept subjects about which we can realistically write an article (per WP:WHYN). If the former, then you should find as many excuses as possible to reject and downplay as many sources as possible. The word "indiscriminate" is pretty good for that purpose. If it were a business, you could claim that this one is insufficient by itself per WP:AUD, and all others are insufficient per any reason you can think of, though that doesn't apply to BLPs. If the latter, then if the source is actually reliable (e.g., the local newspaper does not have a reputation for making up "alternative facts"; see WP:NOTGOODSOURCE for a handy list) and you can actually write an article beyond the level of a doomed permastub, then you should not really care whether it's a local newspaper. WhatamIdoing (talk) 00:44, 2 February 2024 (UTC)

WP:LEAKS

I see a hatnote acknowledging this redirect to #News organizations, presumably meant to address citations of Wikileaks and similar orgs, yet having read this section nothing within seems to me helpful in clarifying the topic. The content regards assessing the relative reputability of News Orgs and how to handle the various types of content they may publish ("leaks" not being among them), whilst it seems the issue at hand for "Leak Publishers" would rather be what qualifies as a "News Organization." And it strikes me that online in 2024 that is a more difficult question to answer than ever - perhaps deserving of some attention? Walkersam (talk) 21:52, 24 January 2024 (UTC)

agree it should be better and rsp on wikileaks should be done again because there is a lot that was not talked like the hoax new york timess article [7] the WikiLeaks#Editorial_policy the Promotion of false conspiracy theories the missing and fake files and not verifying their own documents [8] Softlem (talk) 14:40, 27 January 2024 (UTC)

It looks like @Sceptre made a proposal in 2008 about information leaked to news organizations. You can probably find the related discussion in the archives of this page. WhatamIdoing (talk) 00:57, 2 February 2024 (UTC)

and unless confirmed by a reliable source, information derived from leaks should not be used in articles. needs to be policy and editors should only cite the RS Softlem (talk) 12:57, 2 February 2024 (UTC)

If you word it that way, someone will think you can't cite a newspaper article that talks about leaked materials. WhatamIdoing (talk) 17:59, 2 February 2024 (UTC)

Maybe and unless reported by a reliable source. Editors should cite the RS for leaked content and analysis. If RS newspaper quotes a leak or says it means something, it can be cited

Leak publishers should only be cited for ABOUTSELF or if theyre accepted RS. Publisher that mainly publishes leaks is leak publisher. Publishers that publish things like newspapers that can have leaks are not leak publishers Softlem (talk) 03:55, 3 February 2024 (UTC)

The "unless reported" language should work.

For the most part, I assume you want something more like "leaked documents should not normally be used or cited directly in articles". It's not the "information derived from leaks" that matters (the news reports will include "information derived from"). We're normally concerned with editors deciding which bits to use without the implicit guidance of reliable sources. The ideal arrangement is to have the leaked documents reviewed by reliable sources (e.g., a journalist), and to have Wikipedia editors relying on the reliable sources instead of the original leaked documents.

Some leaked documents might be accepted as reliable themselves (e.g., if their authors publicly acknowledge that it's a true copy and the document would have been considered reliable if it had been posted by those authors – say, a government report that was leaked today but would normally have been published in a few weeks anyway). I'm not sure that's very common. Most leaked documents would generally be considered primary sources. WhatamIdoing (talk) 06:21, 4 February 2024 (UTC)

For the most part, I assume you want something more like "leaked documents should not normally be used or cited directly in articles". It's not the "information derived from leaks" that matters (the news reports will include "information derived from"). Yes

The ideal arrangement is to have the leaked documents reviewed by reliable sources (e.g., a journalist), and to have Wikipedia editors relying on the reliable sources instead of the original leaked documents. Thank you thats what I mean

Most leaked documents would generally be considered primary sources. Right Softlem (talk) 15:07, 5 February 2024 (UTC)

there should be WP:LEAKS essay Softlem (talk) 15:10, 13 February 2024 (UTC)

Draft:LEAKS Softlem (talk) 18:44, 13 February 2024 (UTC)

Provided my feedback at your Teahouse thread - I think this is an excellent effort! Thanks for turning your attention to this. Walkersam (talk) 00:16, 15 February 2024 (UTC)

added feedback from Wikipedia:Teahouse/Questions/Archive_1215#WP:LEAKS and moved essay and updated WP:LEAKS and added nutshell to WP:Reliable_sources#News_organizations Softlem (talk) 13:24, 19 February 2024 (UTC)

Marketing Agency As News Source?

Throughout a lot of social media, the "Remezcla" brand is represented as a neutral news outlet when, in fact, this is actually a marketing firm that is operating as a digital ad agency for a large venture capital firm. A snippet from this article should prove illustrative: "magazine and creative agency, Remezcla, announced today a sizable investment from My Code". Its actually great business, but don't think that qualifies them as a legitimate source on par with a newspaper or more traditional news where the content is not outwardly designed as a marketing vertical. What do you guys think? Chicano Culture (talk) 19:44, 12 January 2024 (UTC)

The page you link is a press release, from Remezcla. It is clearly not 'neutral' when discussing itself. As for Remezcla being cited for anything else, we'd probably have to look at the particulars: i.e. what exactly is being cited, and what proposed text is it being cited for. Given the self-description, it may well be questionable as a source for anything possibly promotional. AndyTheGrump (talk) 19:59, 12 January 2024 (UTC)

Yes, sorry about calling it an article. Thank you for replying so fast. Basically, as you noted, they themselves are saying everything they do is of a promotional nature. Thus, the press release (and there are tons like it) seems very contrary to the quotidian use of their articles as a 'news source'. See, for example, Bad Bunny. I would not be surprised if they wrote their own wikipedia. Chicano Culture (talk) 21:07, 12 January 2024 (UTC)

The Remezcla article seems to have been created by an established Wikipedia contributor, though I'd agree that it is problematic in that it cites too many primary sources, and doesn't in my opinion provide evidence of the sort of in-depth independent coverage necessary to meet WP:CORPDEPTH criteria. I've tagged it for sourcing and notability, and left a comment on the talk page. AndyTheGrump (talk) 21:47, 12 January 2024 (UTC)

I was wondering if eventually that page would turn eventually into something similar to Maxim where the sourcing was so problematic that the page was flagged somehow. Given how notorious Maxim is as an entity, it strikes me as odd that these smaller self-promoting outlets can slip through the radar. You could argue Maxim is less deserving of their current page distinction as 'being made for payment' since anyone could have legitimately thought they were worthy of describing for free. Additionally, one of the creators of that Remezcla page seems to have an inhuman number of contributions, which suggest it handled by multiple people. Tough to discern whether for profit. Chicano Culture (talk) 05:17, 27 January 2024 (UTC)

You can see some of their own marketing services on LinkedIn and on own site, which makes sense since that is the professional network in which the average enterprise user would solicit their niche marketing. They do manage direct social media accounts for large companies, like Target, and these are involve creating content for the brands (likely not disclosed as paid content to the average reader) here. I think the overall disclosure has to be done, however, due to legal requirements on both the publicly traded company and likely FCC. Why split hairs when they make no bones about identifying themselves as marketers to the right clientele? They're just not a good source for news, oh well. Chicano Culture (talk) 15:59, 20 February 2024 (UTC)

@Chicano Culture, I don't understand this. The company says they're a magazine. You say they're a digital advertising agency. What is your belief based on?

The magazine has been acquired by an investment company that also owns El Diario La Prensa, a daily newspaper (proudly printed on cheap paper for over a century).[9] WhatamIdoing (talk) 00:35, 2 February 2024 (UTC)

Hello. It's not a 'belief'. The marketing agency aspect is in their promotional materials. Remezcla was first owned by Hemisphere Media, then MyCode and now Ariel Investments is a majority shareholder of the holding company 'MyCode'. You can consult a very specific detailing of their marketing work here. Additionally, prior to acquisition(s), Remezcla presented themselves as a marketing agency to investors via their usual PR Newswire articles. Hope that helps. Chicano Culture (talk) 03:14, 2 February 2024 (UTC)

That link tells me that they run sponsored content/advertorials in their own publication. It does not tell me that they buy and sell advertisements for a variety of clients in a variety of publications ("an agency").

(It should be "a belief"; otherwise, you would have been posting something you did not believe to be true. One generally believes things one perceives to be facts, for example.) WhatamIdoing (talk) 07:28, 2 February 2024 (UTC)

You can see some of their own marketing services on LinkedIn and on own site, which makes sense since that is the professional network in which the average enterprise user would solicit their niche marketing. They do manage direct social media accounts for large companies, like Target, and these are involve creating content for the brands (likely not disclosed as paid content to the average reader) here. I think the overall disclosure has to be done, however, due to legal requirements on both the publicly traded company and likely FCC. Why split hairs when they make no bones about identifying themselves as marketers to the right clientele? They're just not a good source for news, oh well. Chicano Culture (talk) 15:59, 20 February 2024 (UTC)

Do I need to do anything about a source that is widely cited in articles published in peer-reviewed journals but the evidence suggests misattribution?

For a transcription and translation of the orginal texts, see user:JMF/Chastelet.

I have found quite a few sources that attribute to the French mathematician Alexis Clairaut (1713–1765) the aphorism "The example of Hooke and that of Kepler [serves] to show what a distance there is between a truth that is glimpsed and a truth that is demonstrated". Invariably the source they cite is

• Ball, W W R (1893). An essay on Newton's "Principia". London: MacMillan. p. 69.

who explicitly attributed the aphorism to Clairaut. Ball says that he took it from

• Marquise du Chastellet (1759). "Exposition abregée du systême du monde, et explication des principaux phénomenes astronomiques tirée des Principes de M. Newton" [Abridged explanation of the world system and an explanation of the principal astronomical phenomena drawn from the Principia of Mr Newton]. Principes mathématiques de la philosophie naturelle. Vol. 2. Paris: Desaint et Saillant.

But here is the problem: the French edition of 1759 (MDCCLIX) of Newton's Philosophiæ Naturalis Principia Mathematica was written by Madame la Marquise du Chastellet, who gives Books 1 and 2 of the Principia in volume 1. In volume 2, she provides Book 3 then (a) her explanation (Exposition) for the mathematically challenged, (b) Clairaut's "Treatise on the Figure of the Earth" and (c) "a more learned one" by Bernoulli. In her the introduction (Avertissement) in Volume 1, she wrote that the Exposition is drawn in the main from the works of Clairaut or from the notebooks that he had previously given in the form of lessons to her. I see nothing to suggest that the Exposition is anything other that her original work. She gives explicit credit to Clairaut and Bernoulli for their contributions. Which leaves the uncomfortable conclusion that Ball couldn't give the credit for the aphorism to a woman?

Having found the issue, do we have to do anything about it? Is it a wp:OR violation to even identify it? How far is wp:verifiability not truth to be taken? (The context is that I'd like to work Robert Hooke up to wp:FA standard and this is a hulking iceberg in the way.)--𝕁𝕄𝔽 (talk) 12:37, 18 February 2024 (UTC)

I'm not sure what the problem is that you see. Could you explain exactly why you see a problem with it? She says most of the stuff is due to Clairaut but you are saying this particular bit is probably due to her? Why is that? Verifiability does trump the truth mostly on Wikipedia but if you can make a very good case for a mistake then it might be possible to put that in using a WP:RfC at the article but really it would be best if someone outside Wikipedia wrote it first. NadVolum (talk) 13:12, 18 February 2024 (UTC)

The problem is that we say in a number of articles that it was Clairaut who said it, because Ball declares that he did. But the source that Ball cites says no such thing. Since the Marquise attributes every other part of her book except this one to the original authors, we must assume that it was she herself who wrote it. She says that she used the teaching of Clairaut to help her write it, not that he wrote it for her, especially as it expresses an opinion rather than a mathematical analysis. (Elsewhere she writes that Clairaut reviewed her translation of Newton, so omission is significant.) The reason we use the remark is that it is attributed to a noted mathematical physicist; if it is attributed to a student, then its inclusion is WP:UNDUE. I rather suspect that Ball made the same judgement. (See also Newton-Hooke priority controversy for the inverse square law.)

We could just regard this as an advanced case of I personally know that this information is false. Isn't that good enough to remove it? but it is not obviously "personal belief or knowledge", but neither is it beyond reasonable doubt.

Yes, an external source would resolve the issue. Chance would be a fine thing!™

My inclination is to let it stand but not with equanimity. --𝕁𝕄𝔽 (talk) 17:06, 18 February 2024 (UTC)

Getting away from the specifics of this case to the more general issue: if we discover that the RS we have been citing has misrepresented one of its own sources, do we have a responsibility to identify that fact or do we blithely go on using it knowing it to be invalid? --𝕁𝕄𝔽 (talk) 17:42, 18 February 2024 (UTC)

Supposing you actually very much do want to include something and you have good proof that the citation is wrong you'd really need to flag that you are going against policy in saying there is a problem even though no source says so, and raising an RfC would be best to say how to phrase it. There's always room for WP:COMMONSENSE but you really do need straightforward convincing evidence, and it might still fail for some reason. Unfortunately it is altogether easy for people to make mistakes and have nobody bothering to point them out in a reliable source and yes they can be a pain on Wikipedia if someone has got it into their head there is nothing wrong with them and cites them. I had a case recently where someone cited something and there was even a comment after it pointing out it was completely silly but that counted as opinion and not a reliable source. Unfortunately it required a little technical expertise to evaluate and so there was no hope of an RfC passing and I just had to live with it. NadVolum (talk) 18:39, 18 February 2024 (UTC)

In this specific case, my inclination would be to delete it, but that would only make sense if deletion applied to all the articles where the phrase is used:

• Newton-Hooke priority controversy for the inverse square law
• Philosophiæ Naturalis Principia Mathematica
• De motu corporum in gyrum
• Robert Hooke

I really couldn't see an RfC succeeding for the obvious reason that no RS says that Ball faked it and the FAQ I quoted above would be cited.

(By the way, Madame la Marquise du Chastelet (aka Émilie du Châtelet) was quite a character!) --𝕁𝕄𝔽 (talk) 20:20, 18 February 2024 (UTC)

It would be OR to say that an RS is wrong about the attribution. It would probably be consistent with policy to attribute the attribution explicitly to Ball in the article, rather than to state Ball's attribution in Wikivoice. And document the problem on the article's talk page in case reliable sources pick it up in the future. An RfC will not circumvent the no original research policy. Geogene (talk) 21:39, 18 February 2024 (UTC)

Sounds about the best bet if you're fairly sure it is wrong. NadVolum (talk) 22:16, 18 February 2024 (UTC)

I was thinking it was quiet around here, and just realized that this is the talk page for the WP:RS guideline. JMF might want to move this to the WP:NORN No Original Research Noticeboard instead. Geogene (talk) 00:07, 19 February 2024 (UTC)

Yes, I agree it would be OR for me to research the primary source and decide unilaterally that I am right and that the secondary source [and all those academic authors who have cited it in the past 130 years] is wrong. I had and have no intention of doing that.

Perhaps I should explain that I was doing citation verification. The article had the full quote from Mme du Chastelet but the Ball citation was only good for half of it. So the obvious solution was to go back to Ball's source, which indeed has the full quote. [Of course that would raise the wp:PRIMARY v WP:SECONDARY sourcing issue, but let's park that for the time being.] But that just opened the can of worms re attribution that I've already described.

The reason I raised it here at wp:RS is really to ask what we should do when an RS turns out not to be reliable after all? It seems that the conclusion we have all come to is that we have to grin and bear it until a new source comes along and puts matters back on an even keel. In the meantime, we can only document it back stage. --𝕁𝕄𝔽 (talk) 00:40, 19 February 2024 (UTC)

FYI, I have revised the citation given at the concluding sentence of Robert Hooke#Gravitation so that it cites Ball without reader-evident comment. I have however included a hidden note that refers future editors to talk:Robert Hooke#A truth that is glimpsed and a truth that is demonstrated, which briefly explains the issue. --𝕁𝕄𝔽 (talk) 01:25, 19 February 2024‎ (UTC)

If you are fairly certain that the source is wrong, then you should consider using some Wikipedia:Editorial discretion and removing it from the article completely. You cannot say "They're wrong" without a source that says this, but you can stop saying anything at all about it. WhatamIdoing (talk) 00:00, 21 February 2024 (UTC)

The remark expresses a mathematician's perspective and has significant value in the article. (It explains why Newton is credited with the 'enlightenment' model of universal gravitation, because it was he who produced the mathematical proof – and why Hooke is not, despite the fact that it was he who identified that "all the heavenly bodies have not only a gravitation of their parts to their own proper centre, but that they also mutually attract each other within their spheres of action" when colleagues (notably Newton) were still explaining their motion as being due to "vortices in the aether".) So I don't think it can be discarded when the only problem is attribution. Ball says that it was Clairaut who said it, perhaps Ball knew something that is not obvious to the modern reader: per the FAQ point above, I have to accept it at face value.

But I did, however, delete that part of the quote ["We must not believe that this idea thrown at random into Hook's Book diminishes the glory of Mr. Newton, ..."] which an earlier editor contributed but is not in Ball. I confess that I bridled at her his remark that "this idea thrown at random into Hook's Book", since Hoook's "idea" was as significant in the development of planetary physics as was Copernicus's idea that the Earth and planets orbit the Sun. --𝕁𝕄𝔽 (talk) 10:48, 21 February 2024 (UTC)

I'm not sure I understand the specifics of this example, but I don't think policy is standing in the way of exclusion. First, no source is inherently reliable (except for itself). A clear-cut reason that something in Ball is wrong is evidence that for this item it fails WP:RS (but getting talk page consensus on that point would be wise). Second, per WP:VNOT even if something is verifiable there is no obligation to include it. About the only thing you are not allowed to do is to insert your own analysis about Ball being wrong. The best option would be to find a good published source that agrees with you. Zero^talk 12:37, 21 February 2024 (UTC)

I agree and (as I noted above, 𝕁𝕄𝔽 (talk) 01:25, 19 February 2024 (UTC)), I have removed all reader-visible questions about reliability of Ball from the article and created a talk page item about it. Hopefully in time a History of Women in Science journal will re-evaluate this aspect of the work of Émilie du Châtelet (Madame la Marquise du Chastelet) but in the meantime we have to accept the source at face value because it is widely accepted as reliable. No RS says that Ball is wrong: it is only my analysis of the primary source that differs from his and the FAQ at the top of this page says clearly that such questions are not for Wikipedia editors to decide.

My reason to raise the question here was what in general we should do if we believe that a reliable source has made an error of detail. The consensus is (a) find another (preferably modern) source that corrects it or (b) don't use it or (c) let it stand but create a talk page item for a future editor who may be able to find a correction. --𝕁𝕄𝔽 (talk) 13:34, 21 February 2024 (UTC)

I think that pretty much sums up the options. I would put an option such as saying something like "a famous quotation in mathematics" (i.e., instead of "Hooke said...") in the category of "don't use it" (as it doesn't use the probable mistake). WhatamIdoing (talk) 19:25, 21 February 2024 (UTC)

I'll add to this the essay Wikipedia:When sources are wrong, which goes into more detail on the options available. Arcorann (talk) 12:51, 24 February 2024 (UTC)

Notes

2BHK Tamil Movie Director

2BHK tamil movie (sc-fi psychological thriller)directed by Bala palaniyappan, 2BHK movie was released on August 4 in ott platforms like Airtelxstream, hungama,vi movies,watcho,moviewood,etc. Bala palaniyappan (talk) 09:39, 5 March 2024 (UTC)

Upcoming LLM pseudo-sources

It looks like (archive) we're going to soon start having a bunch of LLM pseudo-sources: Google Is Paying Publishers to Test an Unreleased Gen AI Platform. In exchange for a five-figure sum, publishers must use the tool to publish 3 stories per day. ... As part of the agreement, the publishers are expected to use the suite of tools to produce a fixed volume of content for 12 months. ... These sources of original material are not asked for their consent to have their content scraped or notified of their participation in the process—a potentially troubling precedent, said Kint. When any of these indexed websites produce a new article, it appears on the platform dashboard. The publisher can then apply the gen AI tool to summarize the article, altering the language and style of the report to read like a news story. ... The program does not require that these AI-assisted articles be labeled. ... 'I think this calls into question the mission of GNI,' Kint said. 'It's hard to argue that stealing people’s work supports the mission of the news. This is not adding any new information to the mix.'

These sources will neither be based on journalists' investigations nor on fact-checking by humans who understand the meaning of things. Chances are they'll also merge a bit of Wikipedia-based content into them, creating a self-confirming loop.

A risk is that if the original sources are good enough, the pseudo-sources may look convincing enough that regular WP:RSP discussions will not be enough to classify them as unreliable. If they mostly say that 1+1=2 and that the world is round, and just occasionally talk about pi being 3, then chances are that they could be classified as mostly reliable, despite doing no journalism at all.

This problem will tend to exacerbate the bias introduced by filter #3 in the standard model of Western mainstream media.

Just brainstorming on possible ideas for handling this:

• create a list of the specific news media involved in this particular case (might need some sleuthing to to identify them); this would make it easier to debate their reliability
• be more sceptical about rich-country news sources that are suspected of involvement
  • in particular small "independent" (Google's term) news media
• create a Wikipedia:AI/Noticeboard (maybe there already is one?)

Boud (talk) 15:27, 9 March 2024 (UTC)

How do you assess reliability for Russian state media?

You have a group like RIA Novosti, for example. They are an officially approved Russian state media broadcaster. But if I were to cite them to support, say, a claim which Western publications would deny, how would their reliability then be assessed? What if all the other Russian publications then support RIA Novosti's information? This is not very specifically about them, or at least not only them. I use them only as an example.106.69.220.173 (talk) 17:10, 25 March 2024 (UTC)

When it comes to political matters I think all Russian media needs to be attributed. There are some which are better than others but there has been a general crackdown since the Ukraine war. I think you'd need to bring up a specific instance to get anything much better. See WP:RSP for some of them including RIA Novosti. NadVolum (talk) 19:05, 25 March 2024 (UTC)

By the way questions like this shoud be on WP:RSN. This page is about possible changes to the general guidelines. NadVolum (talk) 19:14, 25 March 2024 (UTC)

Don Spencer

Please see the discussion at Wikipedia:Teahouse#DOB issue - Don Spencer OAM. This seems to be a genuine, good faith request from the person involved, and a classic example of where our policies are getting in the way of improving an article. Is there any way we can solve this problem? HiLo48 (talk) 01:20, 26 March 2024 (UTC)

@HiLo48: Wouldn't this be acceptable under Wikipedia:Verifiability#Self-published_or_questionable_sources_as_sources_on_themselves? I'm going to post this at the teahouse thread. --David Tornheim (talk) 01:37, 26 March 2024 (UTC)

Yes. That looks reasonable to me. HiLo48 (talk) 01:42, 26 March 2024 (UTC)