Wikipedia talk:No original research

There's a sub-thread at RSN right now (approximately hereabouts) in which editors disagree over whether searching a database to find the relevant record is prohibited original research.

We have generally prohibited search engine results (e.g., https://www.google.com/search?q=blue-green+widgets) but not all databases have a deep-linking scheme that will give you a URL that takes you exactly to the bit of information you want.

Should we address this? Do we already address this somewhere else? I have a vague memory of a guideline saying that if you need to cite such a site, then you should put in the relevant directions, e.g., "http://example.com?database.asp, then search for id #123456 and filter for 2025", but I can't find it right now. WhatamIdoing (talk) 19:32, 5 December 2025 (UTC)

What I've seen in the past, and what I've suggested in other instances is to add the instructions of how to find the correct record to the reference. It certainly doesn't match my understanding of OR, but I can't find any guidance on the matter. Something in one of the help pages about citations would probably be the best place to explain it. I would say that there's a difference between checking a database record and looking up something more dynamic or ephemeral. Still not really OR, more a V issue.
It does raise some rot concerns as changes to the database may make it unverifiable, as it's unarchivable and may not exist anywhere else. -- LCU ActivelyDisinterested «@» °∆t° 00:27, 6 December 2025 (UTC)

We could put a note in WP:ORMEDIA:

• Source information does not need to be in prose form: Any form of information, such as maps, databases, 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. If it will be helpful, instructions for how to find the information can be added in the citation (e.g., "Search for id #12345 and filter for 2026", "Map sector 16B", or "Row 1236, column PU").

We could also add spreadsheets to the list while we're at it. WhatamIdoing (talk) 18:52, 10 December 2025 (UTC)

WP:PRIMARY seems relevant here 'A primary source may be used on Wikipedia only to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge. For example, an article about a musician may cite discographies and track listings published by the record label, and an article about a novel may cite passages to describe the plot, but any interpretation needs a secondary source.'

If the average person cannot verify the claim then I'd consider it an issue. Traumnovelle (talk) 20:21, 10 December 2025 (UTC)

Yes, of course. But I think an average person can use a search box or click a few filter buttons. An example given in the RSN discussion was "type in YUL in the departure airport, followed by TLS in the arrival airport" and then look on the left side of the page to find the answer. I think that's manageable for an average person. WhatamIdoing (talk) 05:59, 11 December 2025 (UTC)

Remember that Original Research on WP has less to do with the source material, and a LOT to do with how Wikipedians can misuse it. If extracting information from a database requires interpretation or analysis… then that’s OR. But it is not OR to simply search a database, see that it says X… and then write “the database says X”. Blueboar (talk) 01:45, 12 December 2025 (UTC)

Sure, but quoting an isolated fact might itself be misleading because it is possible that a secondary source would have concluded that the previously believed fact has been found to be mistaken or misrepresentative. Johnuniq (talk) 04:12, 12 December 2025 (UTC)

That's true for any source. There is nothing unique about a database that makes it more susceptible to containing outdated information than non-database sources. Also, most well-maintained databases get updated, which means that checking a reputable database is more likely to set you on the right path than diving into the literature directly. In the airline/airport discussions, one of the complaints is that the databases are too up-to-date ("Too many changes, dear Mozart, too many changes").

(Also, it's more likely to be a primary source that concluded the old fact was mistaken or not representative, at least initially.) WhatamIdoing (talk) 08:46, 12 December 2025 (UTC)

When an edit is problematic, it is sometimes difficult to identify which policy applies. I am not saying that repeating what a source says can’t be problematic… I am simply that the problem isn’t WP:OR. Blueboar (talk) 13:34, 12 December 2025 (UTC)

It's possible, but it's also common for reliable publishers to make demographic or economic data available in database or spreadsheet form. A lot of UK census data is eventually made available at unique location-specific URLs, but not all and often well after it's published in spreadsheet form. I've seen sound national economic data published by international agencies on websites that require the user to specify the country and maybe even the particular kind of data required or the period. I'd often like a note in the ref that would help me do that quickly when verifying, but I can't see that we should reject the data because the website design doesn't support unique citation. NebY (talk) 14:03, 12 December 2025 (UTC)

A larger problem with the airport example is that the data is ephemeral, and can be dependant on when you personally check the data. The verification issue isn't that it's a database, but that verification is likely to fail. As databases can't be archived the reference is effectively useless, and readers can't verify the information. -- LCU ActivelyDisinterested «@» °∆t° 20:40, 20 December 2025 (UTC)

It seems to me that a flight database would be an excellent source for determining whether the claim is still true. People can't use it to verify whether it was true at the time the information was added, but they should be able to determine whether it's true today.

By contrast, we can find newspaper articles from the 1930s talking about the new transatlantic flights between New York and Southampton. Those are great for showing that a flight used to exist, but worthless for verifying that it still does.

So: Which do you want to verify? That it used to (and might/might not still today), or that it is (and might/might not have done previously)? WhatamIdoing (talk) 21:28, 20 December 2025 (UTC)

The database shows that there is currently a flight between two places, but that could just be a flight that is happening today and never again. The problem is that there's nothing to show this entry will still exist tomorrow or even in a few hours, and no way to archive it's content at any point. -- LCU ActivelyDisinterested «@» °∆t° 22:24, 20 December 2025 (UTC)

That's true for the newspaper article, too, though. It shows that there was a flight between two places, but that could just be a flight that was cancelled the next week and never again. There's nothing in a newspaper article to show this entry will still exist at any point in the future, and archiving it still doesn't make the newspaper article more recent. WhatamIdoing (talk) 23:16, 20 December 2025 (UTC)

I would hope a newspaper article contained more context than "A flight from A to B happened". If a news sources was pumping out articles with nothing more than that I would start to suspect that it was a content farm. -- LCU ActivelyDisinterested «@» °∆t° 00:19, 21 December 2025 (UTC)

Sure, it could contain a lot more information than that. But what it can't do is predict the future. Here's a 1939 New York Times article on transatlantic flight. Can it tell you whether that flight is still happening today? Of course not. And the flight tracking website can tell you whether it's happening this week, but it can't tell you whether it was happening in 1939.

Maybe we need more than one source, if your goal is to cite both that an airline both was flying and also still is flying between two airports? One source to indicate the "was in the past" part, and a different source for checking the "still is" part? WhatamIdoing (talk) 00:30, 21 December 2025 (UTC)

I'm reminded of the referencing for climate data in New Zealand articles. Take a look at Morrinsville#Climate and tell me if you can figure out how to find the source. Traumnovelle (talk) 18:38, 12 December 2025 (UTC)

That's not an example that I'd characterize as someone having "put in the relevant directions." I figured a bit out from watching a brief video and playing around. If you start at https://data.niwa.co.nz/, click on Climate data, and then on the Climate station normals option. Then scroll down a bit and input 23908 in the Agent no search bar, click on the three temperature parameters, and click on Climatology period 1991-2020. Then click on Apply filters in the lower right-hand corner. Choose one of the results and then choose View full product details and it displays the monthly norms for those two decades. In theory, you should be able to find the rainfall data by instead choosing the rainfall parameter and Climatology period 1971-2000, but that didn't work for me. Even if someone had put in these directions, it's a longish sequence of steps and I could see disallowing it for that reason. FactOrOpinion (talk) 19:46, 12 December 2025 (UTC)

Thank you, I genuinely was not able to figure it out, but as you've explained its a ridiculous task, although it seems you can actually link data directly: which makes me wonder why the citation is provided in such a format to begin with. Traumnovelle (talk) 19:52, 12 December 2025 (UTC)

No idea. And now I'm feeling foolish for not having noticed that one could link directly to the output. FactOrOpinion (talk) 20:00, 12 December 2025 (UTC)

It looks like the direct URL doesn't have all the information in the table, so a more 'generic' or 'high-level' URL was chosen as encompassing all of them. I've changed that one to the direct URL. WhatamIdoing (talk) 20:44, 12 December 2025 (UTC)

Well if its citing multiple datasets it should have multiple citations, even if combined into a single footnote. I agree that some sort of guidance is required on how to cite such material after looking at this specific example. Traumnovelle (talk) 20:50, 12 December 2025 (UTC)

Mostly I don't expect this to be a problem, because you're usually going to have a general/summary entry (example) and it will be obvious to most people that if you want more than what's immediately displayed, you should click on relevant links/tabs/sections.

In this case, the high-level URL with search directions (even as small as "Click on 'Apply filters' to search") might be better than multiple citations. WhatamIdoing (talk) 20:57, 12 December 2025 (UTC)

One thing that might be useful here is that Evidence Explained method for citations, commonly used in genealogy to explain precisely how to access the same part of a database. Katzrockso (talk) 22:57, 20 December 2025 (UTC)

Evidence Explained is a red link. It is this reference book? https://www.evidenceexplained.com/ WhatamIdoing (talk) 23:19, 20 December 2025 (UTC)

This intersects with other policies like WP:NOT, WP:DUE, and WP:V. If no reliable third party sources have written about a fact, how did an editor come to evaluate it as an important fact for inclusion in the encyclopedia? This does risk violating WP:OR, and in doing so, violates a lot of other policies. I've never seen anyone cite something to a search engine, and maybe it's more common than I thought. But this is a bad way to build an encyclopedia. Shooterwalker (talk) 16:58, 17 December 2025 (UTC)

If no reliable third-party sources have written about a fact, how did editors decide to use {{cite twitter}} on >40,000 articles? If third-party sources is the only standard for inclusion of all facts, then why do we add birth dates cited to the subject's own social media posts, or alma maters cited to résumés?

I suggest that sometimes the answer is that encyclopedia articles should contain certain basic descriptive information about a subject even if the available sources . For example, an encyclopedia article about a person should include approximately when and where the person lived, even if you have to get that information from less-than-ideal (but still reliable) sources.

What I don't understand is how looking up information in a particular media format (e.g., a database) risks violating OR any more than using any other particular media format (e.g., a spreadsheet, or a table printed in a book). WhatamIdoing (talk) 22:58, 17 December 2025 (UTC)

It would be reasonable, even desirable to contest those citations. Yes, twitter, resumes, and social media often contains self-serving, false, or irrelevant statements. Editors might promote them recklessly, negligently, or because they are doing original research to WP:SYNTHesize a point in favor of their editorial biases. It remains a bad way to build an encyclopedia, and should be done with extreme caution. Shooterwalker (talk) 01:20, 18 December 2025 (UTC)

And yet the WP:ABOUTSELF and WP:BLPSELF policies, which are the opposite of using third-party sources, are widely accepted by the community. WhatamIdoing (talk) 01:30, 18 December 2025 (UTC)

Yes, we are looking at the same policy, which includes several cautionary notes and objections. If we begin to rely rely heavily on primary sources, it's removed. If an editor raises a reasonable doubt, it's removed. If it becomes self-serving, it's removed. If the claim is exceptional, it's removed. Using primary sources raises more risk than reliable secondary sources with a reputation for fact-checking and accuracy. I realize I'm repeating myself, but it's a bad way to build an encyclopedia and needs to be done judiciously and with caution. Shooterwalker (talk) 01:48, 18 December 2025 (UTC)

I don't think that using primary sources with a reputation for fact-checking and accuracy raises significantly more risk than reliable secondary sources with a reputation for fact-checking and accuracy, assuming that the claims being supported are something that any ordinary educated reader would find straightforward to verify in the source. WhatamIdoing (talk) 02:52, 18 December 2025 (UTC)

Well, for example, take City: you'll find a number of secondary and tertiary sources on City that state a number of demographic facts are relevant to understanding City. Taking those facts from a good database requires no stretch of editor imagination nor original thinking. Alanscottwalker (talk) 01:45, 18 December 2025 (UTC)

Who's suggesting we cite a search engine? And do you want to exclude census data published by, say, the UK's Office for National Statistics? NebY (talk) 20:47, 18 December 2025 (UTC)

Unlike @Shooterwalker, I have seen people cite search results (though not for a very long time).

Relevantly, there is a difference between actual search results like https://www.ons.gov.uk/search?q=inflation ("6592 results") and giving instructions on how to find a specific source on a specific website ("Search for the title of this report, and then set filters to Topics > Census > Housing".) WhatamIdoing (talk) 21:30, 18 December 2025 (UTC)

I admit I may not understand the use case. I've seen this kind of thing misused for WP:SYNTH and WP:NPOV, which is coloring my advice here. If there's a WP:CONSENSUS that one use of search results is factual and verifiable, that it isn't being overused compared to reliable third party sources, and isn't being used to spin some narrative, then WP:CONSENSUS often wins. I still stand by my advice to proceed with caution and care. Shooterwalker (talk) 17:29, 23 December 2025 (UTC)

Right, you definitely need a reliable source to begin with. This is only relevant when you've got a choice between "Authoritative Data Set #123 (good luck figuring out how to check that)" and "Authoritative Data Set #123, and now I'll give you the directions on how to actually find #123 on their website". WhatamIdoing (talk) 22:35, 23 December 2025 (UTC)

Thinking in how I see things, something about the hatnote redirect templates arranged as is doesn't seem to fit. So, I tried to fix it with the following: {{Redirect2|WP:OR|WP:NOR|text=They should not be confused with the WikiProjects of [[WP:OREGON|Oregon]] and [[WP:NORWAY|Norway]].}}, however when that happens the "{{hatnote group|" ends up becoming broken, and I don't know why that's the case, so if anyone knows what might be wrong with my code, that'd be greatly appreciated. I think merging the 2 templates fits in how I see it, but if it can't happen, that's totally understandable and fine by me. ѕιη¢єяєℓу ƒяσм, ᗰOᗪ ᑕᖇEᗩTOᖇ 🏡 🗨 📝 00:59, 21 December 2025 (UTC)

 You are invited to join the discussion at Wikipedia talk:Good article nominations § Extensive use of primary and pre-WWII sources. Rollinginhisgrave (talk | edits) 09:46, 24 December 2025 (UTC)

I'm trying to make sure I understand the Synth examples and possible versions of them that *would* be appropriate.

With the U.N. example, am I correct that that the two sentence would be appropriate one after another without the "but" or "and," if both are appropriately sourced?

The United Nations' stated objective is to maintain international peace and security. Since its creation there have been 160 wars throughout the world.

With the Smith/Jones example, would this version be appropriate of the first two acceptable sentences plus one sentence stating Harvard's definition without any further analysis, if all are appropriately sourced?

Smith stated that Jones committed plagiarism by copying references from another author's book. Jones responded that it is acceptable scholarly practice to use other people's books to find new references. Harvard's Writing with Sources manual defines plagiarism as using a source's information, ideas, words, or structure without citing them. newsjunkie (talk) 22:26, 1 January 2026 (UTC)

• I don't think it would be appropriate in either case. Synth isn't about phrasing and word choice. It's about a sequence of ideas that leads the reader to a novel conclusion that isn't stated in the sources.

In for example the UN case, the SYNTH is the juxtaposition of the UN's objective (to maintain peace and security) and the number of wars. It leads the reader to a conclusion that the UN has failed, or succeeded in the alternative UN example. Each sentence might be appropriately sourced, but you still can't put the ideas together in that way unless a reliable source does.

The rule came about because we needed to reduce the misuse of Wikipedia for political and marketing purposes, so it's interpreted fairly broadly.—S Marshall T/C 12:06, 2 January 2026 (UTC)

What exactly is meant by "This policy does not apply to talk pages and other pages which evaluate article content and sources, such as deletion discussions or policy noticeboards."? IOHANNVSVERVS (talk) 08:40, 3 January 2026 (UTC)

I'm coming here from observing a discussion on-wiki:

User 1: "WP:OR says This policy does not apply to talk pages and other pages which evaluate article content and sources, such as deletion discussions or policy noticeboards. So it does not apply."

User 2: "It does if you want to amend or edit an article based on that same OR, but it's a confusing line to tread."

I think this needs clarification in the article. IOHANNVSVERVS (talk) 09:44, 3 January 2026 (UTC)

Is it considered Original Research or Synthesis to use the birthday from one source and the birth year from another source to come up with a full DOB?

Example:

Public figure stated on Instagram that their birthday is January 4. A reliable source elsewhere stated that the public figure was born in 2002.

Are editors allowed to say then that the public figure was "(born January 4, 2002)" in the first sentence of the BLP article? Or would this be considered original research/synthesis?

(This is assuming that the public figure is telling the truth and that the reliable source is accurate.) Some1 (talk) 18:52, 4 January 2026 (UTC)

Courtesy ping to: Sangdeboeuf. Some1 (talk) 21:17, 4 January 2026 (UTC)

• As I said in the other thread, as long as the other concerns from BLP are satisfied (the BLP is public with their birthday and has not wished for its removal/tried to hide it) then this is not at all an OR issue. We are drily stating a fact they supplied. If this is OR then any form of constructing an article based off of multiple sources also is and we are limited to one source per article. PARAKANYAA (talk) 18:58, 4 January 2026 (UTC)

  That is self-evidently not what WP:OR means. —Sangdeboeuf (talk) 01:36, 5 January 2026 (UTC)

• when we include 2 facts from 2 different sources into one sentence ("he was born on January 4, 2002"), we are not creating a new synthetic statement. there is zero original material. Rjensen (talk) 19:48, 4 January 2026 (UTC)
• It is not OR/SYNTH to have parts of a date from different sources. My reasoning is here. JFHJr (㊟) 20:35, 4 January 2026 (UTC)
• Given that we could cut this into two sentences… each supported by a reliable source… I see no problem combining the facts into one sentence. This isn’t what we mean when we talk about SYNTH. Blueboar (talk) 20:57, 4 January 2026 (UTC)

  Would the same logic "individual sentences each supported by a reliable source" apply to the sentences in this example (assuming they stay as individual sentences and are not combined in anyway) or is there a difference and if so could you explain it? newsjunkie (talk) 22:24, 4 January 2026 (UTC)

  If the following is improper synthesis:

  N The United Nations' stated objective is to maintain international peace and security, but since its creation there have been 160 wars throughout the world.

  Are we really saying that splitting the same material into two sentences is somehow acceptable, like so:

  Y The United Nations' stated objective is to maintain international peace and security. Since its creation there have been 160 wars throughout the world.

   ? —Sangdeboeuf (talk) 01:32, 5 January 2026 (UTC) edited 04:14, 5 January 2026 (UTC)

  This is what I was asking in the discussion above. in a talk discussion on another related page a few years ago somebody seemed to suggest it was appropriate. My interpretation was that it's the addition of the and or but that was making it a new claim. newsjunkie (talk) 01:44, 5 January 2026 (UTC)

  So the second sentence doesn't imply anything about the first? Seems like willful ignorance to me. —Sangdeboeuf (talk) 04:18, 5 January 2026 (UTC)

  I don't think it implies anything more than any other sentences do when put together in a paragraph, especially since per the example there are multiple different ways a reader could interpret it based on their own predispositions. And there is no value judgement based on the use of "but" or "only." newsjunkie (talk) 11:26, 5 January 2026 (UTC)

  If you think about it, you should. As sourced, "The UN promotes peace. The UN promotes security." is just not written very well, or is somehow literary writing. The two sentences together don't imply anything. Now, add a third sentence, "There was war." "War" has obvious implications for "peace", "security" and the implied topic of third sentence, "the UN". (Compare, "He was born on June 4. He was born in 1982." and now replace ". He was born in" with a comma. Now add a third sentence "There was war." OMG, his birth!) -- Alanscottwalker (talk) 12:14, 5 January 2026 (UTC)

  But if it's sourced that there was a war, can't one also argue that's a relevant fact to mention in the context of an organization trying to promote peace to illustrate challenges for example in terms of NPOV? The three statements don't say what the connection is. And the other example could just mean he was born/grew up during a war. So when is it just stating chronology? How does it depend on what the specific sources mention or don't mention or the context or the source background? Is it OR or just a due/undue argument? newsjunkie (talk) 14:46, 5 January 2026 (UTC)

  Certainly not if that connection (between the UN's stated goal and the existence of wars after its formation) has not been made in RS. That's a straightforward violation of SYNTH regardless of whether there's a comma or period since "implication" is not predicated on type of punctuation. Even with well-sourced material, if one uses it out of context, or to state or imply a conclusion not directly and explicitly supported by the source, one is engaging in original research. JoelleJay (talk) 19:45, 5 January 2026 (UTC)

  If someone was born in 1942 in Germany or the U.K, is it it OR to say he was born during World War 2? newsjunkie (talk) 14:49, 5 January 2026 (UTC)

  That there may be other ways not to be OR, does not make a specific text not OR. Alanscottwalker (talk) 16:37, 5 January 2026 (UTC)

  It would be SYNTH if there are no sources directly commenting on the subject's relation to WWII. JoelleJay (talk) 19:48, 5 January 2026 (UTC)

  What comes after the "but" doesn't inherently have to do with the UN, it's a fact about the state of peace in the world, but the sentence uses that fact to comment on the UN. That's editorializing. Largoplazo (talk) 15:10, 5 January 2026 (UTC)

  Can't *not* characterizing the "state of peace" in the world also be editorializing if it's an organization explicitly related to "peace"? newsjunkie (talk) 15:23, 5 January 2026 (UTC)

  No more so than failing after "Mother's Against Drunk Driving is dedicated to eliminating traffic fatalities from intoxicated driving" to add that traffic fatalities still occur is editorializing. Largoplazo (talk) 16:21, 5 January 2026 (UTC)

• Under WP:DOB, it seems rather obscure for a living person and not sufficiently sourced but it is not OR per se. Alanscottwalker (talk) 21:27, 4 January 2026 (UTC)

• Given that the question specifies a BLP article, the presumption in favor of privacy applies. If a celebrity has deliberately not posted their full date of birth because of concerns about stalking and/or harassment, then extrapolating the full DOB is pretty clearly a case of using material out of context, inconsistent with the intention of the source. Yes, it's original research. —Sangdeboeuf (talk) 01:20, 5 January 2026 (UTC)

  No, it isn't, if someone says in one interview that they are born in 2001 and in another they are born on September 27 then they clearly have no objection to it. The distinction you are suggesting is not meaningful. If that is OR then everything is, then constructing an article with more than one source is OR because it was not the intention of that one source to say the things the other sources in the article do. PARAKANYAA (talk) 05:27, 5 January 2026 (UTC)

  It's not clear that the subject in your example has no objection. If they wanted their full date of birth published, they would have stated their full date of birth. Maybe they forgot what they said in one interview when giving the other. It may be true that the internet never forgets, but that's not a reason to trap people into revealing more about themselves than they intend at any given time.

  Your contention that I'm proposing some kind of slippery slope that would lead to everything being considered original research is also off-base. For example, one can state that cars are wheeled transportation based on Source 1 saying cars have wheels and Source 2 saying cars are used for transportation. In that case no new information is introduced. That is different from extrapolating someone's personal identifying info from multiple, partial sources when the context indicates that neither source intends for that information to be made public. —Sangdeboeuf (talk) 10:20, 5 January 2026 (UTC)

  In that case no new information is introduced precisely what new information do you think is introduced in the claim "John Doe was born on 22 September[1] 1982[2]"? You can argue that there's a WP:BLPPRIVACY reason for not including full dates of birth in particular cases, but I don't see WP:OR being at all relevant. Caeciliusinhorto-public (talk) 10:59, 5 January 2026 (UTC)

  A person's full date of birth is useful information to cyberstalkers and identity thieves, not so much their year of birth alone. Both WP:OR and WP:BLP are relevant when publicizing the info is inconsistent with the intention of the source[s]. —Sangdeboeuf (talk) 02:04, 6 January 2026 (UTC)

  The fact that something is useful to identity thieves has precisely zero bearing on whether it is original research to state it. That's a BLPPRIVACY/DOB question.

  I think your reading of the phrase "inconsistent with the intention of the source" is nonsense, frankly. Caeciliusinhorto-public (talk) 10:02, 6 January 2026 (UTC)

  There is no new information being introduced in a birth date case. Exactly as much "new information" is introduced in your cars argument as here. PARAKANYAA (talk) 17:30, 5 January 2026 (UTC)

  The relevant WP:OR question is whether the sources have intentionally omitted certain details. There's a much stronger argument for that in the case of a living person's date of birth than random technical information about cars. —Sangdeboeuf (talk) 02:09, 6 January 2026 (UTC)

  I don't think you have any support for your assertion: "they clearly have no objection to it": two sources that don"t say that and are only two, is not anything that supports you clearly can know what's in mind, nor can presume what's in their mind, based on just the little information given, here. (Also, it may not be doxxing without malice, but privacy still always matters.) Alanscottwalker (talk) 12:21, 5 January 2026 (UTC)

  "John Doe was born on a September 27 and he has said so publicly. He was also born in 2001 and has said so publicly. However, it is unclear whether he wants anyone to know that he was born on September 27, 2001."

  If John Doe has said in an interview that he was born on September 27, we can say he was born on September 27. We don't have to worry "Gee, what if he's regretted ever since then that he said that?" or "Gee, what if he doesn't remember that he said that?"

  The same goes for his having been born in 2001.

  Therefore, we could write "John Doe was born on a September 27, and also in 2001" and delude ourselves that we've kept secret a detail that John Doe would like to have kept secret, or we can write what's exactly the same thing, "John Doe was born on September 27, 2001." Largoplazo (talk) 15:18, 5 January 2026 (UTC)

  I don't agree that any of it meets the standards of DOB sourcing, based on the initial scenario. Alanscottwalker (talk) 16:20, 5 January 2026 (UTC)

  We don't have to worry about it, because we can simply list the year of birth alone per WP:DOB. In the above scenario we cannot reasonably infer that the subject wants their full DOB to be publicized. —Sangdeboeuf (talk) 02:16, 6 January 2026 (UTC)

  That also goes for someone saying "my birthday is x day x month x year", we cannot perfectly read someone's mind and know what they would like having on a wikipedia page, and people change their minds. It is not any better for them to say the full date on their twitter account then to have to combine the date and year, they are both the person being public with the pieces of required information. WP:DOB says "sources linked to the subject such that it may reasonably be inferred that the subject does not object to the details being made public", regardless of whether it is date of month/date or + year of birth. PARAKANYAA (talk) 17:34, 5 January 2026 (UTC)

  And that's the issue. What's reasonable. Is it reasonable to root around to find the DOB from two perhaps primary sources, private information that is not widely published anywhere, and apparently no one except some Wikipedians thinks is worth widely publishing. I say, no. Alanscottwalker (talk) 17:59, 5 January 2026 (UTC)

  DOB explicitly says yes if the primary sources are from the subject. Widely published applies to sources not affiliated with the subject. If the person says "i was born [Date]" on Twitter that is a better source per DOB than 1 reliable news source or journal or encyclopedia. PARAKANYAA (talk) 18:08, 5 January 2026 (UTC)

  I don't buy that cramped reading, given the stakes, set out. And I don't buy we must have a birthdate at all (I only link that article because I have been dealing with it presently). Alanscottwalker (talk) 18:14, 5 January 2026 (UTC)

  It isn't a cramped reading it is what DOB literally says.

  I don't buy we need that article at all. And that person ism't even alive; do you just not like dates? We don't need any article on any one person, we could simply not have it and all would be well with the world. But why make articles less informative when there is no reason to?PARAKANYAA (talk) 18:16, 5 January 2026 (UTC)

  I disagree, it is cramped reading.. Now, you get to it, with a BLP, we have to be reticent in how we treat BLP's. We have been given the reasons in the BLP policy. --Alanscottwalker (talk) 18:24, 5 January 2026 (UTC)

  No, it isn't "cramped reading". WP:DOB is there to read for all.

  Sure, but none of those reasons apply to the current cases we are discussing: people who have made their date of birth public and people who died decades ago. PARAKANYAA (talk) 18:28, 5 January 2026 (UTC)

  I plan to repeat what I said, again. (That.)

  The BLP policy applies to Living Persons, and the reasons do apply, no one else in the whole world has put this information together. Alanscottwalker (talk) 18:32, 5 January 2026 (UTC)

  No one else in the whole world has ever put this information together goes for every single article we have. It goes for the Garner article, no one in the whole world has put the information there together. So all articles are OR? Or what is the difference in that case, exactly? PARAKANYAA (talk) 18:34, 5 January 2026 (UTC)

  I think we should move on. Because 1) We are have been discussing BLP, not OR. 2) you moving to OR makes little sense in the context of this discussion we have been having. 3) as for the rest, we are not going to agree. Alanscottwalker (talk) 18:41, 5 January 2026 (UTC)

  We are on the No Original Research talk page. That is all we are supposed to be discussing. But I agree this conversation won't get anywhere. PARAKANYAA (talk) 18:43, 5 January 2026 (UTC)

  Multiple people have discussed BLP, as we would when dealing with a living person. You yourself have been discussing the BLP, the title of this section says DOB, part of the BLP policy. Alanscottwalker (talk) 18:47, 5 January 2026 (UTC)

  The question is whether it is original research to combine sources to get a full DOB. This applies whether that person is living or dead, OR is a universal rule, the question we are asking is not about BLP. There is a BLP debate about DOB but it is entirely orthogonal to the question we are discussing, which would apply whether the person was alive or dead for 200 years. PARAKANYAA (talk) 18:56, 5 January 2026 (UTC)

  Not at all orthogonal, intersectional. BLP and OR are both "universal" in every article or information they cover. Alanscottwalker (talk) 19:00, 5 January 2026 (UTC)

  BLP does not apply to material on the dead, so it is not universal. OR applies to literally everything we write in mainspace. If something is OR, you cannot do it, ever, while there are plenty of things you can do on non-BLP articles that you cannot do on BLP articles.

  The issue here would also apply to dead people, so BLP is not what we are discussing. PARAKANYAA (talk) 19:03, 5 January 2026 (UTC)

  Read again, universal "in every article or information they cover." The question is not about a dead person, it is about a living person. It is still the case, we can write without the birthdate in either the living or dead, what's also the case, is we have special considerations for the living. Alanscottwalker (talk) 19:11, 5 January 2026 (UTC)

  They are not both "universal", where one policy applies only to a subset of information and one applies to literally everything we can write. The question in this case is about both the living and the dead, it is about whether it is OR to "combine sources to get a full DOB".

  The special considerations for the living do not apply in a case when the consideration that sources from the person have volunteered this information have been satisfied. So, we are left with the (incorrect) argument that it is OR, which would apply to both the living and the dead. We can do many things; it does not mean we should. PARAKANYAA (talk) 19:23, 5 January 2026 (UTC)

  BLP always applies to the living, its universal to the living. OR is also universal to the living. You have not convinced me that BLP offers no consideration when Wikipedia are the first publication to put together private information about living people. And your unsupported inference from any one source is not enough. Alanscottwalker (talk) 19:36, 5 January 2026 (UTC)

  "Universal" only to a subset is not universal if you are comparing it to something that applies to everything, the class under consideration in this case is all articles.

  You aren't putting together any new information, any more than using more than one source ever is. But we're going in circles and we will clearly never convince each other, PARAKANYAA (talk) 19:43, 5 January 2026 (UTC)

  We don't have to perfectly read someone's mind since we have a presumption in favor of privacy. In the example of someone self-publishing their birthday and year of birth separately, it cannot reasonably be inferred that the subject does not object to their full DOB being made public, especially when doing so more effectively enables cyberstalking, harassment, and/or identity theft. —Sangdeboeuf (talk) 01:56, 6 January 2026 (UTC)

  Yes, it can be reasonably inferred that someone who publicly states what their birth day is, is public about their birth day. PARAKANYAA (talk) 02:20, 6 January 2026 (UTC)

  But not their full date of birth, which is what we are actually discussing. —Sangdeboeuf (talk) 02:26, 6 January 2026 (UTC)

  The date of birth is the birth date and the year. It is not greater than the sum of its parts, it is simply those two facts. If the are open about those two facts then they are open about their birth date. PARAKANYAA (talk) 02:27, 6 January 2026 (UTC)

  You are confusing birthday and birth date. A full date of birth is actually more than the sum of its parts, since it more effectively enables cyberstalking and identity theft. Once again, the context in which a person self-publishes the information matters. —Sangdeboeuf (talk) 02:35, 6 January 2026 (UTC)

  No, I am just saying the year + month + day = year and month + day. If someone is public about those two facts, they are public about their birth date. We are sticking to the source, per the page you linked. PARAKANYAA (talk) 02:38, 6 January 2026 (UTC)

  No, you are suggesting using material out of context. You also said we cannot perfectly read someone's mind, and I agree. That's why we err on the side of caution and don't automatically include any and all trivia the subject may have posted over the course of their entire lives, because we don't know how much of that information they really want publicized. —Sangdeboeuf (talk) 02:49, 6 January 2026 (UTC)

I'm not sure that this is original research, but WP:DOB says dates are birth are included that have been widely published by reliable sources. If you have to find two different sources for different parts of the DOB I'm unsure how widely published it is. Maybe that's not so much an issue if both sources are ABOUTSELF. -- LCU ActivelyDisinterested «@» °∆t° 13:59, 5 January 2026 (UTC)

That's more or less where I fall on this as well. ScottishFinnishRadish (talk) 14:57, 5 January 2026 (UTC)

Agreed. JoelleJay (talk) 19:54, 5 January 2026 (UTC)

+1. Also, I like what @Alanscottwalker was saying about being reasonable. It's probably reasonable, in a WP:CALC-like way, to say that the person was born on Octember 32, 1987 (company was founded on, event happened on...), if you have one reliable source saying "Octember 32" and another saying "1987". But very specifically for WP:DOB, if you have to do that, it's not "widely published" and therefore moot/irrelevant. WhatamIdoing (talk) 20:35, 5 January 2026 (UTC)

I'm not sure that this is original research We could use another example involving a hypothetical deceased person:

Reliable sources A, B says "Anna Smith was born on May 4".

Reliable sources W, X, Y, Z says "Anna Smith was born in 1867."

Assuming that all of these RSes are accurate, is it OR/Synth to state that Anna Smith was born on "May 4, 1867" ? Some1 (talk) 23:34, 5 January 2026 (UTC)

That's a separate question to the one posed. If it's the DOB of a living person then it doesn't really matter, as it's not widely published. If its not a BLP concern, and so WP:DOB doesn't apply, then I don't think it's OR ("Anne Smith was born May 4,[A][B] 1867[W][X][Y][Z]"). Others may disagree. -- LCU ActivelyDisinterested «@» °∆t° 00:32, 6 January 2026 (UTC)

One has to consider the intention of the source(s). What are the likely sources for a person's birthday who was born 150 years ago? A more real-world example might help here. —Sangdeboeuf (talk) 02:23, 6 January 2026 (UTC)

What source would you consider acceptable with regard to the "intention of a source" for someone's birthday? And how are you deciding what the intention is? PARAKANYAA (talk) 02:28, 6 January 2026 (UTC)

That's impossible to answer without a specific, real-world example. Can you think of one? —Sangdeboeuf (talk) 02:57, 6 January 2026 (UTC)

Here's a more modern example (again, these are made up scenarios):

• Lisa, a public figure, responded to a fan's question on Twitter of "when is ur birthday??" with "June 7"
• Lisa posted images of her birthday party on Instagram on June 7 with the caption "Happy birthday to me!! :)"
• Lisa's BFF (who is also a public figure) said in an interview with Vogue that Lisa's birthday is June 7, which is a day before theirs
• 5 reliable sources reporting Lisa's birth year as 2001

Assuming that Lisa is telling the truth about her birth day, and that the RSes are accurate about her birth year: Is it Original research/Synthesis to state that Lisa's DOB is "June 7, 2001"? Some1 (talk) 03:02, 6 January 2026 (UTC)

In this made-up scenario, yes. (The second and third bullet points don't matter, only the first and fourth are important.) It's original research because it's taking the birthday out of context and using it in a way inconsistent with the intention of the source (Lisa). When Lisa posts her birthday on Twitter I don't think her intention is to enable cyberstalkers and identity thieves to make her life miserable, which is what publicizing her full date of birth can do. —Sangdeboeuf (talk) 03:24, 6 January 2026 (UTC)

• Well WP:OR states 'Do not combine material from multiple sources to state or imply a conclusion not explicitly stated by any of the sources.' so by the letter it is OR. Of course rules can ignored with good cause but I don't see how providing the exact date of birth of a living person meets that threshold. Traumnovelle (talk) 19:50, 6 January 2026 (UTC)

  Two adjacent facts aren't automatically "a conclusion". NOR is looking for things that a person with a decent education would agree/disagree is supported by the source. NOR is not asking us to cater to particularly stupid or rigid-minded people – just an ordinary person. That means, e.g., that if we have a source that says something like "Victoria, Princess Royal was born on 21 November 1840. On November 9th of the following year, Queen Victoria gave birth to the future monarch, Edward VII", then we don't have to pretend that we can't figure out which year Edward VII was born in. With a source like that, it's find to say "Edward VII was born on 9 November 1841". We would not have to restrict ourselves to saying "Edward VII was born on November 9th of the year after 1840". This is not OR because it's obvious to everyone from the source. Similarly, I don't think it's OR (including SYNTH sub-type of OR) to say that an event reported to have happened "on November 9th" and to have happened "in 1841" actually happened "on November 9, 1841". WhatamIdoing (talk) 23:18, 11 January 2026 (UTC)

  That isn't combining two different sources. Traumnovelle (talk) 20:20, 14 January 2026 (UTC)

  OR is OR. OR across two sources is the subtype called SYNTH, but if it's not OR to combine something in two parts of the same source, it's probably not OR to combine something in two sources, either. WhatamIdoing (talk) 20:47, 14 January 2026 (UTC)

• No. Not OR or synthesis since no new facts or inferences are being created. However, for BLPs, depending on context, it might be better to stick to the year per WP:DOB, but that's something to be determined on a case-by-case basis since the extent of publication can be subjective sometimes. -- Patar knight - ^chat/_{contributions} 01:30, 7 January 2026 (UTC)
• No, it is not WP:SYNTH to combine birthday and year of birth to arrive at DOB, despite neither source explicitly stating a DOB. The first line of SYNTH is confusing with its use of “combine”. It really should use “synthesize”. Is the problem the repetition of the word in the section title? “Combine” implies simple addition. Searching for words intermediate between “combine” and “synthesize” yields: these words are “semantically related”. The OED definition 1. “To put together or combine into a complex whole…” and contrasts with “analyse”. “Complex” seems a pertinent word. It is synthesis of it is a complex calculation? Adding day and month to a year is not a complex calculation. Google tells me that a complex calculation involved multiple steps and operations. I might suggest: a complex calculation might see different results obtained by different people. -SmokeyJoe (talk) 03:27, 12 January 2026 (UTC)

  “Do not synthesize meaning from multiple sources to state or imply something not explicitly stated by any of the sources”.

  - SmokeyJoe (talk) 11:12, 14 January 2026 (UTC)

  Yes? That is the definition you just changed the page to. Traumnovelle (talk) 20:18, 14 January 2026 (UTC)

  “Do not synthesize meaning from multiple sources to state or imply a conclusion not explicitly stated by any of the sources” looks good as the lead sentence. SmokeyJoe (talk) 11:40, 17 January 2026 (UTC)

• The specific example is tightly defined and caveated so I would agree with most above that it is not OR. However, extending this to all general instances doesn't seem right. It seems highly unusual that we would have to combine sources like this, and that suggests there might be an unusual reason behind the lack of a single comprehensive (the loosest possible use of the term comprehensive) source. The sources and caveats should probably be looked into for each unique case. CMD (talk) 12:27, 17 January 2026 (UTC)
• Much of the above does not really grapple with why this particular situation is likely unique. OR is concerned with what is put into articles, it is not concerned with inferences editors make in their heads to justify their edit, as long as that inference does not make it into the article. WP:DOB is concerned with editors making an inference in their head ('the person being written about wants the DOB to be public') and the inference is not so much from what's written in the source, rather its from what the editor thinks about the mode of publication of the source. And that inference does not show up in the article, only letters and numbers do: still, we could never write in the article: "Mary does not mind that the world knows their birthdate", or anything like it, based on these sources - that would be OR. It remains, that given the artificial situation presented, I remain convinced there is not sufficient support to make the needed BLP inference (even in the editors head) under the BLP policy. Alanscottwalker (talk) 15:14, 17 January 2026 (UTC)

  In this case, if reliable source1 state a birth DAY (September 2nd)… and reliable source2 state a birth YEAR (1992)… we could write two sentences:

  • She was born on September 2nd(cite1). She was born in 1992(cite2).

  We can SUMMARIZE these two sentences as:

  • She was born on September 2nd(cite1), 1992(cite2).

  or

  • She was born September 2nd, 1992(cite1)(cite2)

  or even

  • She was born 9/2/92 (cite1)(cite2)

  Again… this is Summarization… not Synthesis. We are not drawing or implying any conclusions not stated in the sources, merely rewriting the information in a briefer form.

  However, we can not draw a “rule” from this example. In some different situation, combining the facts and citations into one sentence may well result in an implied conclusion. Blueboar (talk) 17:09, 17 January 2026 (UTC)

  I agree, and I also agree with CMD that each situation has its own unique facts. WhatamIdoing (talk) 18:16, 17 January 2026 (UTC)

  Yes , but the point I was making here is that, "She was born September 2nd, 1992", does not imply anything about what the person wants or thinks about making widely public the information. And it would still be entirely OR to write in the article "She thinks this date should be widely public.", based on those sources that don't say that. Alanscottwalker (talk) 18:19, 17 January 2026 (UTC)

  Oh… sure. Then again, why would we ever want to write that in an article? I don’t see that as being encyclopedic information even if we had a reliable source explicitly quoting her as saying it. Blueboar (talk) 18:39, 17 January 2026 (UTC)

  We don't want to write that into the article, but it is an inference that we are required to make with given sources per WP:DOB policy, before we write in the article, "born, September 2nd, 1992". Alanscottwalker (talk) 18:46, 17 January 2026 (UTC)

  You are WAY overthinking this. Blueboar (talk) 19:12, 17 January 2026 (UTC)

  Really? Do you not understand that the "inference" required by BLP:DOB policy is not an inference made from article text, it is an inference made from circumstances of publication, whereas OR is concerned with the "inference" arising from article text. Alanscottwalker (talk) 19:18, 17 January 2026 (UTC)

I was advised by a user on the Discord group to post something about this at Template talk:Notelist and a second user recommended to correct the note, so I am posting this here. I spotted a formatting issue in the notes section involving the Notelist template. As can be seen in this edit, the notes section appears to be formatted for three columns, but with a completely blank center column and weird spacing for the fifth note. The fifth note in particular is an Efn template that has a bulleted list template inside it that uses references. I "fixed" the formatting by setting the em spacing to 39em, which is more just masking the issue and moved the problem to a different resolution per a second user. However, my understanding of the second user's response was that the note should be edited to be similar in size to the other notes --Super Goku V (talk) 13:55, 11 January 2026 (UTC)

Yeah 39em looks very funky for me. It might be helpful to upload an image of what you're seeing, as the edit you linked formats differently dependent on skin and screen resolution. The issue is how very long notes are handled. They're not allowed to roll over into another column, which means if there's one very long note it can cause weird column formatting. -- LCU ActivelyDisinterested «@» °∆t° 15:46, 11 January 2026 (UTC)

Sure, here is a screenshot of what it looks like. (Anomie suggested it might be a browser bug.) --Super Goku V (talk) 11:30, 13 January 2026 (UTC)

That's is a weird one. Are you using a odd browser / system, e.g. not Chrome on Windows, (which I'm guessing makes up most editors) or Safari on iPhone / Chrome on Android (which makes up most of the readership)? If not I would doubt it's a browser issue, as it would have been reported previously. Are you able to view it on a different browser to see if it stays the same? -- LCU ActivelyDisinterested «@» °∆t° 11:47, 13 January 2026 (UTC)

Firefox on Windows 10 with Zoom size at 100%. About Firefox says I am on 146.0.1 (64-bit) and up to date. Window Night Light was on if that matters.

Chrome on Android only shows one single column with no issues for me. --Super Goku V (talk) 12:12, 13 January 2026 (UTC)

I would agree with Anomie then, this sounds like a browser bug rather than an issue with notelist. There is a large random whitespace in the middle of note 'e' in your screenshot, has that disappeared with the change to 39em? It's not something I see before or after your edit, and there's no reason for it I can see in the sourcecode. -- LCU ActivelyDisinterested «@» °∆t° 12:27, 13 January 2026 (UTC)

FWIW, by resizing my browser window I can repeat it in Firefox 146.0.1 on Windows 10 with 100% zoom and Vector legacy (2010), both in your diff above and as it is now. It's finicky, happening only in a small range of window sizes, but right now it's much as your screenshot: three columns in the See also section and two in the Notes, with a column's worth of white space between them, plus a lot of whitespace above the first and third bullets of note e. NebY (talk) 13:26, 13 January 2026 (UTC)

Mayne this would be better off at WP:VPT, they're more likely to know what's going on. -- LCU ActivelyDisinterested «@» °∆t° 13:55, 13 January 2026 (UTC)

Huh. I just came back here and after undoing my earlier edit that I forgot to deal with, everything looks normal. --Super Goku V (talk) 09:15, 24 January 2026 (UTC)

Regarding his revert.

. Late 2009, User:SlimVirgin introduced fontcolor for “Our policy” subheadings in WP:PSTS. For me, it worked. Much of what PSTS contains is educational, introductory, for the actual “policy”, and the color coding achieved clarity. It was originally red, then maroon, now crimson. I preferred maroon. The “Our” got dropped. I guess it is unambiguously implied, and first person plural seems not quite right, if necessary I might prefer “Wikipedia policy”, but the one word “Policy” works. It also saddens me to toss away Sarah’s styling. SmokeyJoe (talk) 11:38, 17 January 2026 (UTC)

I think "Our policy" would be better. I don't feel strongly about the colors. WhatamIdoing (talk) 18:17, 17 January 2026 (UTC)

It was maroon throughout the 2010s. I prefer maroon to crimson, crimson being too bright against white. SmokeyJoe (talk) 08:57, 20 January 2026 (UTC)

• Visually, it looks a bit like "Policy" is a redlink.—S Marshall T/C 09:16, 20 January 2026 (UTC)
  • Less than when it was red. The crimson produces a dazzling shimmer that occurs when the contrast goes low. Maroon is the same colour as your signature. —SmokeyJoe (talk) 10:48, 20 January 2026 (UTC)

Hi MichaelMaggs, re this, is there a wording you think would be more accurate? I think a sentence like this is much needed, I doubt newbies read and comprehend this whole article Kowal2701 (talk) 18:59, 19 January 2026 (UTC)

It should be something that summarises the body of the text, and that has consensus approval. It's not something that can be added as a new statement directly into the lead. MichaelMaggs (talk) 21:07, 19 January 2026 (UTC)

This isn't a mainspace article, MOS doesn't apply. If it helps people understand then idk why we wouldn't add it, it isn't something prescriptive that I'm trying to change about policy, but just describe current practice, hardly needs high-level consensus, just some implicit/explicit agreement on wording and utility Kowal2701 (talk) 21:20, 19 January 2026 (UTC)

It's a fundamental English Wikpedia policy that requires the highest level of consensus. WP:PGCHANGE. Although there was nothing stopping you making a bold change, which you did, you now need to seek consensus as reasons have been given in the edit summary for not accepting your suggestion. MichaelMaggs (talk) 21:54, 19 January 2026 (UTC)

That’s what I did with my first comment. But happy to leave this alone Kowal2701 (talk) 22:04, 19 January 2026 (UTC)

I am not sure that User:Luka Maglc‘s new lead to the section is an improvement to the policy page. SmokeyJoe (talk) 10:43, 20 January 2026 (UTC)

@SmokeyJoe: Thanks for raising this question. My intent with the new lead is not to weaken or redefine WP:SYNTH, but to reduce a recurring misunderstanding: people read “synthesis” in the everyday sense (encyclopedias summarize multiple sources) and then get confused about why WP calls SYNTH a problem.

The added lead is meant to clarify that WP uses “synthesis” in a "policy-specific" sense: editors must not combine sources (or separate parts of one source) to imply a conclusion not explicitly made by any source. That distinction is already explained in Wikipedia:What SYNTH is not, so the change is explanatory and aligns with existing guidance rather than introducing a new concept.

If you think the wording over-emphasizes “Wikipedia is inherently a synthesis” or risks sounding like it grants editors leeway, I’m happy to tighten it so it more clearly signals “no change to requirements” and keeps the focus on “A and B, therefore C” being the prohibited case. But I do think having a short lead that separates the common-language meaning from the policy meaning will prevent misreadings and bad-faith wikilawyering. Luka Maglc (talk) 10:53, 20 January 2026 (UTC)

I think it's longer than what we need. Maybe something as short as "Wikipedia articles are meant to include different sources and different views, but not to combine them in ways that create novel conclusions" would be sufficient.

If your main point is that the real world has multiple meanings of the word synthesis, then that might be better handled as an essay. WhatamIdoing (talk) 17:23, 20 January 2026 (UTC)

Fair point, on length. I can definitely trim it down. I also take your point about not foregrounding “multiple meanings” on a policy page, since that can invite wikilawyering.

If the community thinks any lead sentence is useful here, I’d support something close to your proposed wording, but phrased in the same “A and B, therefore C” / “not explicitly stated” language the section already uses. For example:

"Wikipedia articles may summarize multiple reliable sources, but editors must not combine sources (or separate parts of one source) to state or imply a conclusion that no source explicitly makes."

That keeps the focus on the prohibition without implying any “acceptable synthesis” carve-out. If even that is redundant, I’m fine dropping the idea entirely. Luka Maglc (talk) 07:03, 21 January 2026 (UTC)

• I'm strenuously opposed to it or anything resembling it, and have removed it for now. My concern is that its hair-splitting invites people to say "ah, yes, this is synthesis, but not the special Wikipedia way we use the term." All that this does is create potential confusion, so I'm utterly opposed to anything on any policy pages that implies that "synthesis", by any definition of the word, is acceptable or present on any Wikipedia article; or which makes any effort whatsoever to introduce or discuss the idea of multiple definitions of synthesis. Our policies should focus on defining things the way we define them, not on vague, confusing, and unhelpful hand-waving like this. --Aquillion (talk) 20:24, 20 January 2026 (UTC)

  It sounds like you're foreseeing another round of "Yes, the subject is notable, but I'm saying that the subject isn't WP:Notable" (only this time with "synthesis" and WP:SYNTHESIS). I think you're right, and that we should always avoid anything like that. WhatamIdoing (talk) 21:36, 20 January 2026 (UTC)

I'm opposed to this addition as well. There is a strong distinction between summarizing sources versus synthesizing them. Those two terms may have some fuzzy overlap in the middle. But a summary states things other sources have already said, while a synthesis tries to assemble new conclusions or ideas, either implicitly or explicitly. I strongly disagree with muddying the waters to say "it's all synthesis actually", and we need to discourage synthesis more broadly. Shooterwalker (talk) 21:22, 20 January 2026 (UTC)

I hear y'alls concern, and I agree the last thing WP:SYNTH needs is wording that can be misread as “some synthesis is fine”.

My aim was purely reader-clarity, because I see newer editors routinely treat “synthesis” as “any use of more than one source” and then get confused by the policy examples. However, I accept the point Aquillion/WhatamIdoing/Shooterwalker are making: explicitly talking about “multiple meanings of synthesis” on a policy page risks creating exactly the loophole we’re trying to prevent.

So I’m not wedded to that framing, and I’m fine with the revert. If there’s willingness for a lead sentence at all, I think a shorter version that stays entirely within policy-language (and never says “Wikipedia is synthesis”) would address the misunderstanding without inviting wikilawyering. For example:

"Wikipedia articles summarize what reliable sources say, but editors must not combine sources (or separate parts of one source) to reach or imply a conclusion that no source explicitly makes."

That’s basically the “A and B, therefore C” point in one line, and it aligns with the existing examples and with Wikipedia:What SYNTH is not / Wikipedia:No original research without bringing in definitional debates. If people still feel even that sentence is unnecessary (since the section already states it), I’m happy to drop the whole idea and leave the section as-is. Luka Maglc (talk) 06:58, 21 January 2026 (UTC)

What you just stated is my understanding: we summarize, but we take care to avoid combining sources (through bundling or sequencing) to imply original conclusions. Do we need this somewhere, and if so where would it go? I'm not sure, but I'll let other editors chime in. Shooterwalker (talk) 18:21, 21 January 2026 (UTC)

Correct, that was what I originally meant. Luka Maglc (talk) 18:31, 21 January 2026 (UTC)

The "summarize" part is the second paragraph of WP:STICKTOSOURCES. The "bundling or sequencing" part is WP:SYNTH. WhatamIdoing (talk) 18:32, 21 January 2026 (UTC)

Look, I'm trying to clear some confusion for editors having a hard time understanding that Wikipedia is technically a synthesis. All encyclopedias are technically synthesis. And the fact that synthesizing is not allowed on Wikipedia, when it already technically is one. The guideline is more about advancing a conclusion through synthesis from different sources, instead of synthesizing stated facts directly from the source themselves. The second is what Wikipedia already does. This is the confusion that I am trying to clear. Luka Maglc (talk) 18:37, 21 January 2026 (UTC)

Why do you need editors to agree that encyclopedias are "technically synthesis"? WhatamIdoing (talk) 18:48, 21 January 2026 (UTC)

I don't. I was just trying to clear some confusion on the WP:SYNTH policy, because it lead me to originally misunderstand it at first and I'm sure others are also misunderstanding it. Luka Maglc (talk) 18:56, 21 January 2026 (UTC)

Can you point to evidence of others having misunderstood it? You misunderstood it once, what was the consequence of that misunderstanding? SmokeyJoe (talk) 10:49, 24 January 2026 (UTC)

I don’t like it, not put here for sure, at the very least, it derails the flow. A major policy section, like WP:SYNTH is, should get the point immediately. I thought we had just tweaked its lead sentence to be better: “ Do not synthesize meaning from multiple sources to state or imply a conclusion not explicitly stated by any of the sources”. This lead sentence states in one sentence what the section means. The suggested paragraph to go in front is introductory, telling the reader things as if they didn’t know, which is good if they didn’t, but at the expense of delaying getting to the point.

WP:NOR already has a fair bit of introductory material, and I don’t think more belongs right there.

If someone feels that editors don’t get something, or miss a nuance, I think someone should right an essay before writing their explanations into policy.

- SmokeyJoe (talk) 10:45, 24 January 2026 (UTC)

I'm in agreement that we don't need this at the top of the lead, and that it's already stated in so many words elsewhere in the guideline. I'm open to putting a clarification somewhere else, but I will wait to see if editors can form the beginnings of a WP:CONSENSUS. Shooterwalker (talk) 15:06, 24 January 2026 (UTC)

To me, "synthesis" does not naturally spring to mind "summarize", it more springs to mind "create", and while some of what we do is creative, it's not original or new creative thought (although it should generally be "original" writing in the cvio sense). Perhaps, it might help some to think of it as "original synthesis", rather than just "synthesis" but that seems already apparent in the context of this policy. Alanscottwalker (talk) 15:32, 24 January 2026 (UTC)

I think you're right about the lead having gotten long. I've just tried to reduce some of the redundancy in its second paragraph. Maybe that helps a little? WhatamIdoing (talk) 01:41, 25 January 2026 (UTC)

I've further shortened it. There is no reason to spend repretive time on "uncited material" in the first paragraph, the second paragraph and the notes are more than enough. Alanscottwalker (talk) 01:51, 25 January 2026 (UTC)

The notes were also redundant. I've merged a single copy all into the paragraph.

I don't think the notes were good enough in practice. The main difference between NOR and WP:V is that NOR doesn't care about citations. (NOR is about whether the material was made up by an editor vs could be cited; WP:V is mainly about what should be cited.) Maybe we need a redirect at Wikipedia:Uncited is not a synonym for original research. WhatamIdoing (talk) 17:22, 25 January 2026 (UTC)

Meh… Technically correct… however, anything that is questioned for WP:NOR falls into the “challenged or likely to be challenged” aspect of WP:V. The two policies address overlapping issues … and disputes are resolved in the same way (add citations). Blueboar (talk) 17:32, 25 January 2026 (UTC)

Well, I agree with you, but I see some editors claiming "OR" when they mean "you didn't cite it". Remember Wikipedia talk:No original research/Archive 63#"Anywhere in the world, in any language" and other matters with the new table, when a couple of very experienced editors asserted that WP:BURDEN failures were OR violations? WhatamIdoing (talk) 17:41, 25 January 2026 (UTC)

Of course NOR "cares" about citations, the sources are the only thing to judge OR on. And sources are identified by citation. But more importantly the droning on about uncited material is getting in the way of the point, which is that OR is what you make-up or missapply/mistake the cited or uncited sources that don't support the text. The second paragraph is already about uncited material, we should not have it in the first paragraph. -- Alanscottwalker (talk) 17:40, 25 January 2026 (UTC)

NOR doesn't "care" about citations per se. I agree that OR is what you make up (including when an editor believes RSes support the text but they actually don't, e.g., assuming that if the subject is "a doctor", then he's specifically a medical doctor, or believing that an Honours degree is the same as an Honorary degree or a degree with academic honors).

I think that an important point is that OR doesn't depend on citations. Instead, it depends on the real-world existence of the source. A citation is evidence against an OR accusation, but it's only the existence of the source that this policy cares about.

Blueboar is correct that anything that is questioned for WP:NOR falls into the “challenged or likely to be challenged” aspect of WP:V – but that's not a problem for this policy.

Consider:

• Source exists in the real world, none is cited in the article, editor challenges it: This was never an OR violation.
• Source doesn't exist in the real world, two are (improperly) cited in the article: This was always an OR violation.

WhatamIdoing (talk) 17:57, 25 January 2026 (UTC)

That's silly. How do you know the source exists, it can be identified; How do you identify, citation. (As an aside, misapplying/mistake is also when you use a primary source (or combine them) to make a secondary claim.)Alanscottwalker (talk) 18:00, 25 January 2026 (UTC)

How do you know that the capital of France is Paris? Did you actually need to identify a specific source?

To some extent, this is a technical distinction, but this kind of technical distinction needs to be made so that WP:V and WP:NOR policy pages can comply with the WP:REDUNDANTPOLICY policy requirement. If we'd adopted the WP:ATT proposal years ago, it wouldn't matter.

Also: How do you identify, citation should be How do you identify to other editors, citation. You normally identify the source to yourself before you have a citation. WhatamIdoing (talk) 18:08, 25 January 2026 (UTC)

Well, yes, I need to identify the source for Paris, in fact, I already have. Communication in a communication project, though is more that what I do in my head. Alanscottwalker (talk) 18:30, 25 January 2026 (UTC)

This policy says that "The capital of France is Paris" is not original research and that you do not need you to cite a source for that statement. Do you disagree with this policy? WhatamIdoing (talk) 20:01, 25 January 2026 (UTC)

Silly question. That information almost certainly original to almost all the 7, 000, 000 topics -- if you are going to write that in say, the hydrogen article, you better have a good source that directly connects that information to hydrogen and tell us what that source is and how. -- Alanscottwalker (talk) 21:12, 25 January 2026 (UTC)

Putting a sentence in the wrong article doesn't make the sentence itself original research. "The capital of France is Paris" is not original research no matter which article you put it in. WhatamIdoing (talk) 21:38, 25 January 2026 (UTC)

No, it likely makes the inference original, you have the article inferring for the whole world that Paris' status matters to understanding hydrogen. Alanscottwalker (talk) 21:49, 25 January 2026 (UTC)

The only inference I would make – and I think that any reasonable reader would make – is that someone screwed up when they stuck a sentence about Paris into the middle of an article about Hydrogen. WhatamIdoing (talk) 21:57, 25 January 2026 (UTC)

As long as its in the article, it is for the whole world part of the topic. Alanscottwalker (talk) 22:03, 25 January 2026 (UTC)

I don't think our readers are that stupid. WhatamIdoing (talk) 23:47, 25 January 2026 (UTC)

I don't think writers of original ideas really care what you think is stupid. Nor do writers who make mistakes and add original research. Nothing about this policy is based on intent, it's based on the work product. This is not a behavior policy, it is a content policy. -- Alanscottwalker (talk) 00:00, 26 January 2026 (UTC)

I think that if readers (NB: not "writers of original ideas") see this:

• Hydrogen is a chemical element; it has the symbol H and atomic number 1. Poop poop poop. It is the lightest and most abundant chemical element in the universe, constituting about 75% of all normal matter.

that they are smart enough to realize that's vandalism instead of a "part of the topic". I do not believe readers will see this as "the article inferring for the whole world that [poop] matters to understanding hydrogen".

I similarly think that if readers see this:

• Hydrogen is a chemical element; it has the symbol H and atomic number 1. The capital of France is Paris. It is the lightest and most abundant chemical element in the universe, constituting about 75% of all normal matter.

that they are smart enough to realize that's a mistake instead of a "part of the topic". I do not believe readers will see this as "the article inferring for the whole world that Paris's status matters to understanding hydrogen".

Do you actually disagree with me? WhatamIdoing (talk) 07:47, 26 January 2026 (UTC)

I think you are going for absurdism without considering the likely situation. Original Research is not invariably obvious to the reader. It's possible that on some occasions it is obvious but no one can bank on that. This hypothetical thing about hydrogen, which is only one of millions of articles that could have been mentioned, began with 'the editor' wanting to add Paris to the article -- editors get ideas about how things go together and sometimes they want to publish them -- and its point is, the editor (and Wikipedia and the reader) needs good sources that support that addition to the topic for it to be sustained in the article. Alanscottwalker (talk) 10:51, 26 January 2026 (UTC)

If there were some plausible connection between the two, then we might see it as being two things juxtaposed in a way to imply a conclusion that isn't supported by any source in the world. But that wouldn't make this sentence an OR violation; it would make the implied conclusion be an OR violation.

Also, in terms of "editors get ideas about how things go together", I think we have to allow a little bit of space for editors to Wikipedia:Build the web between articles. Nobody should be claiming an OR violation over something saying "Not to be confused with" a similarly named page, even though that implies a claim that some people might confuse two similarly named pages. WhatamIdoing (talk) 20:31, 26 January 2026 (UTC)