Wikipedia talk:Manual of Style/Lead section

Citizenship in the lede

We are having a debate on Talk:Stevie Wonder#Citizenship over his newly acquired dual citizenship status. It doesn't seem as though there is a standard to go by on how to refer to a person with dual citizenship; for instance Tina Turner lived in Switzerland for many years, yet she was not referred to as "American-Swiss". Also, Tom Hanks is American-Greek and this is not reflected in the lede. The infoboxes of these articles do point out additional citizenships, and I made sure this was the case on Stevie Wonder. However, several editors insist that his dual citizenship status must be mentioned in the lede despite him being an American citizen for almost the entirety of his life. dekema (Formerly Buffaboy) (talk) 05:24, 16 May 2024 (UTC)

MOS:LEADLANG

Thoughts on adding the proposed sentence (shown below in bold so editors can readily see the change):

Do not include foreign equivalents in the text of the lead sentence for alternative names or for particularly lengthy names, as this clutters the lead sentence and impairs readability. Do not include foreign equivalents in the lead sentence just to show etymology. Do not include in the text of the lead foreign equivalents written in non-Roman script, as this is unhelpful to the non-specialist reader. Foreign-language names should be moved to a footnote or elsewhere in the article if they would otherwise clutter the first sentence.

Some examples where non-Roman script provides clutter to all readers, but probably helps only a very tiny minority of readers:

• Districts (Sinhala: දිස්ත්‍රි‌ක්‌ක, romanized: Distrikka, Tamil: மாவட்டம், romanized: Māvaṭṭam) are the second level administrative divisions of Sri Lanka, preceded by provinces.
• The Pakistan Institute of Development Economics (also known as PIDE) ( ‏‎پاکستان دانشگاہِ ترقیاتی معاشیات) is a post-graduate research institute and a public policy think tank located in the vicinity of Islamabad, Pakistan.

These would be better as footnotes or moved elsewhere in the article, and not in the text of the lead. Thoughts? CUA 27 (talk) 20:02, 7 June 2024 (UTC)

Checking to see if anyone wants to weigh in? CUA 27 (talk) 12:46, 11 June 2024 (UTC)

I don't think this change is necessary. There is nothing wrong with having a fairly short expression in another script in the lead sentence. Excessive cases are to be avoided, but that's common sense and doesn't need a particular rule. Plus the rule against clutter is already there. Gawaon (talk) 12:55, 11 June 2024 (UTC)

I wonder if there's a way to put these in an infobox instead. There's nothing inherently wrong with giving the name of a city in all the official/practical local scripts, but I don't necessarily want to see more than one or two in the first sentence itself. WhatamIdoing (talk) 22:37, 25 August 2024 (UTC)

Worth implementing short description guidance, e.g. WP:SDNONE?

While the mobile apps continue to terrorize the world by not making this at all visible to anyone using them, WP:SDNONE is effectively treated like a MOS guideline: we should probably mention it briefly, right? Remsense诉 20:44, 14 July 2024 (UTC)

Guideline on citations in the lede

This guideline is a horrible idea. It radically decreases transparency and greatly weakens the verifiability requirement. Editors and readers are expected to either just trust the implicit anonymous claim that every single assertion found in the lede is sourced somewhere in the article, or else read the entire article, which may be very long, just to find where the claim in question and its sourcing could possibly be located. This makes verification far more difficult than under normal circumstances. Given how often unsourced information is added, incorrect conclusions are made from sources, sources are misrepresented (unconsciously or not) or are unreliable, making part of the article - and the most prominent and important one to boot! - extremely difficult to verify is a horrible idea. The benefit is unclear - adding a few citations in the lede, even if they are also found in the body of the article, is not difficult and does not harm readability significantly - and certainly can't compensate for the harm caused. 'Just trust us, it's there somewhere' is an unacceptable principle for an encyclopedia whose articles are produced collectively and anonymously.

Locating the source might be easier when there is an editor that monitors the article constantly, is very familiar with its content and structure and can point the user demanding a citation to the right place in the article body. This may be the true in some cases, but it normally isn't and our official guidelines cannot be based on the assumption that it always is. It also requires the user who placed the tag to take the extra steps of coming back to the article and having a conversation with the editor in question before they can even check whether the source is reliable and says what it is cited for, and the user cannot be reasonably expected to do that each time. Anonymous44 (talk) 12:52, 24 July 2024 (UTC)

I don't feel similarly about this guideline, but I do note that inexperienced editors will frequently add new information (sometimes cited) directly to the lead rather than under a more appropriate subheading, possibly because the "edit lead section" icon is so prominent in comparison to the rest. By consequence, in practice this guideline is rarely followed except at well-established and stable articles. Folly Mox (talk) 14:15, 24 July 2024 (UTC)

The lead is meant to be a summary of the rest of the article. If it has its own citations and referecing, the possibilities of original research and/or synthesis increase 10 fold. But as FM says above, this is a rather "late-stage" guideline, which (in practice) is really only applied to rather well written articles. – Aza24 (talk) 18:02, 26 July 2024 (UTC)

I disagree totally, echoing points above. Moreover, contrary to your vibe check, when this guideline is actually considered or enforced, it only ever makes articles better. To me, it's not really under question that summary paragraphs of cited material do not require their own inline citations. I will also insist that a requirement would mostly result in lazy, less helpful, distracting, and misleading citations. Moreover, it would result in worse prose, as editors would be more likely just to copy-paste to ensure their summary doesn't need to grab citations from anywhere else in the article.

Think about it like this: when this guideline is invoked, it has the effect of making editors look the lead over, and reconsider how well it represents the article as a whole. A much better article usually results from this. Remsense诉 07:36, 31 July 2024 (UTC)

I agree with Remsense here. Something that's in the lead without being expanded upon (and sourced) in the body can usually be challenged and removed on that reason alone. The lead is meant to be a summary of the main points of the body, therefore in theory it would never need any references of its own (direct quotes exempted). Gawaon (talk) 09:37, 31 July 2024 (UTC)

when this guideline is actually considered or enforced

I think this is exactly what's wrong with MOS:CITELEAD. This guideline has never even been read by anyone, nevermind actually being considered or enforced. If anyone did read it, they'd notice it says the exact opposite of what everyone linking to it claims:

The lead must conform to verifiability,[...] "all quotations, and any material whose verifiability has been challenged or is likely to be challenged, must include an inline citation to a reliable source that directly supports it". [...] There is no exception to citation requirements specific to leads.[...] The presence of citations in the lead is [not] prohibited in any article.

Here's what everyone linking to it claims it does:

1. Removes the need for inline citations
2. Makes exception to the citation requirement specific to leads
3. Prohibits the presence of citations in the lead

Frankly I'm fine with throwing out the current policy. I'm also cool with keeping the policy as written and telling people who think references are ugly that this plugin exists.

But after having one GAN rejected for uncited claims in the lead, and a renomination rejected for citing these claims, I'll go ahead and say the current policy is awfully-written and unclear. I've removed the weird parts of the policy that subtly imply you should remove references (without ever outright stating this). If someone wants to open an RfC to get consensus on changing this, please go ahead and be my guest. Closed Limelike Curves (talk) 01:52, 26 August 2024 (UTC)

I have no idea why you think that other people just don't see where it plainly says direct quotations et al. in the lead still require inline citation. Your GAN got rejected either because the review was incompetent or because the lead just didn't do its job of summarizing the body with nothing more. To proclaim based on that experience that nobody else understands what this guideline plainly says and does not say is specious and insulting. Remsense ‥ 论 02:03, 26 August 2024 (UTC)

The lead summarizes the body. The two reviewers (presumably) disagree about whether the material I linked is controversial or technical enough to require citations in the lead. Closed Limelike Curves (talk) 19:29, 26 August 2024 (UTC)

In re the (intentionally hyperbolic) statement that This guideline has never even been read by anyone: section-specific shortcuts are getting about 250 page views per month, which suggests some interest in its contents.

My own experience is that the line about Although the presence of citations in the lead is neither required in every article nor prohibited in any article is frequently followed by a paragraph or to that amounts to so you are required to do it my way.

The questions I asked below are relevant. Specifically: If the material in question "must include an inline citation", and that material appears in more than one place, is that material required to have an inline citation in each location? Or is once enough? And if once is enough, does it have to be the first?

For example: I might someday clean up the slightly self-contradictory information in Cancer about risk factors. The article should say (in some form) that most cancers and most cancer deaths are not actually preventable by individuals. That is: About 40% of the (significant) cancers being diagnosed today could have been prevented through lifestyle changes, and the first, second, and third most important lifestyle changes are "Stop smoking already", "No, really, smoking kills", and "You moron, I told you not to smoke". This one risk factor is half of the individually preventable 40% of cancers. So if you're a lifelong non-smoker, and you get cancer, there's a ~67% chance that there was nothing you could do to prevent it. Maybe others (e.g., pollution regulators) could have prevented it, but the odds are that there was nothing you could have done as an individual. Smoking is three times as bad as obesity, four times as bad as moderate alcohol consumption, etc. If you are (or were) a smoker, and you get diagnosed with cancer, there's a 25% that the cancer was caused by smoking.

Something about risk factors should be in the lead, and it should actually say that most cancers can't be prevented through lifestyle changes. This is, in my experience, very surprising to people. My friends are all convinced that not eating enough fruits and vegetables is the #1 cause of cancer, and it's just not true. So I consider this Wikipedia:Likely to be challenged, and it will be cited. But: If the lead says "Most cancers can't be prevented through lifestyle changes", and the article has a whole section, with "twenty-seven 8-by-10 color glossy pictures with circles and arrows and a paragraph on the back of each one explaining what each one was, to be used as evidence", does the article also need to have some of those citations duplicated up in the lead?

LEADCITE doesn't answer that question either way. LEADCITE says stuff has to be cited, but it doesn't say that the stuff has to be cited repeatedly. WhatamIdoing (talk) 03:07, 26 August 2024 (UTC)

Perhaps these questions would help:

1. As a general rule, do you personally believe that if the same facts appear multiple times in the same article, they should be cited every time (e.g., copy/paste the same ref to the end of each sentence), or do you personally believe that once is usually enough?
2. WP:BLP requires an inline citation for contentious matter about living people. If contentious matter about a BLP appears in both the lead and also in the body (a very common situation), do you believe that it should be cited in both places, or is once usually enough?
3. WP:V requires require an inline citation for direct quotations. If the quotation appears both in the lead and also in the body (a rare situation), do you believe that it should be cited in both places, or is once usually enough?
4. If once is usually enough, should that once be the "first" mention (e.g., lead) or the "biggest" mention (e.g., a whole section)?

WhatamIdoing (talk) 22:35, 25 August 2024 (UTC)

I don't care as long as we pick one and stick to it consistently, or just clearly and explicitly state that this is up to whoever creates the article and can't be changed (just like current WP:CITEVAR policy).

Closed Limelike Curves (talk) 01:54, 26 August 2024 (UTC)

You might be looking for WP:STYLEVAR, which is the generic version for style questions. CITEVAR allows changes (you just need to have a quick chat on the talk page, to make sure nobody will be mad about it). I haven't looked at STYLEVAR in years, but I assume it's a similar approach. WhatamIdoing (talk) 03:10, 26 August 2024 (UTC)

In practice, if there are already citations in the lead, other editors are unlikely to remove them – why should they bother? In that sense, the first person to write a proper lead (not necessary the first article author) might have a good say in actually getting them added to the existing lead text. However, this cannot extend as far as making a requirement that newly added or significantly rewritten lead paragraphs must be referenced as well. Consider this case (fairly typical in my experience): There's a long and detailed section with 37 references. In the lead it's summarized in just one sentence, or at most a paragraph. So how should we reference that one? Repeating all 37 refs would clearly be excessive and make the lead much less readable, but cutting it down to just 2 or 3, while still getting everything mentioned properly referenced, would at least require detailed knowledge of all the references, and might in many cases be simply impossible. So "here's the summary, see below for details and references" is the most feasible solution in such cases. It's a good thing that CITELEAD allows this, and no first or other editor should be able to forbid it. Gawaon (talk) 06:16, 26 August 2024 (UTC)

I usually would remove them if they're clearly tacked on by a third-party editor who thought they were needed, or are of clearly lower quality than those actually citing the material in the body being summarized. Remsense ‥ 论 06:43, 26 August 2024 (UTC)

Same here. I routinely remove them, especially when they are of poorer quality than the sources in the article body. If it isn't in the body then it cannot be in the lead. Hawkeye7 (discuss) 07:21, 26 August 2024 (UTC)

That rule doesn't apply to sources, and it doesn't apply absolutely to lead content anyway. Sometimes a simple classification, or a statement about related/confuse-able subjects needs to be in the lead, but there isn't anything else to be said, and redundancy is both bad writing and also inconsistent with encyclopedic style. Simvastatin needs to begin with a sentence like "Simvastatin, sold under the brand name Zocor among others, is a statin, a type of lipid-lowering medication", but the article does not need a ==Classification== section that repeats the fact that it's classified as a statin. Once is enough for that kind of super basic fact. WhatamIdoing (talk) 15:34, 26 August 2024 (UTC)

I think the comments above show this is common. It's also routinely brought up at WP:FAC and WP:GAC, where editors demand you mechanically remove all sources from the lead, regardless of objections or past controversy on a topic.

I agree with you that newly-added or substantially-rewritten leads shouldn't be required to stick to the old citation style. Changing the style as part of a substantial rewrite is allowed by WP:VAR, so long as you provide notice on the talk page. Closed Limelike Curves (talk) 16:43, 26 August 2024 (UTC)

Please note that this is not a question of style. A good lead is for the most part a summary of the body. As such, it should not require citations, because they can be found in more detail in the body. Gawaon (talk) 17:58, 26 August 2024 (UTC)

More strongly, when a lead does require citations, it is a hint that some of its content may really belong in the body of the article instead. So omitting the lead citations serves a useful purpose as a flag to editors for that sort of issue.

Additionally, when a lead does properly summarize sourced and expanded body content, it may be very difficult to find a single source that summarizes the same material in the same way, so if we were to cite leads we might be faced with citation overkill when we repeated all the sources needed for the material that the lead summarized. —David Eppstein (talk) 18:20, 26 August 2024 (UTC)

So:

• If the lead requires citations, that might indicate a problem.
• Due to the summative nature of leads (and our rules that we have to Wikipedia:Use our own words), it may be difficult to find a single source that Wikipedia:Directly supports the summary (e.g., "She was an educated woman", even if you have sources that spend pages and pages on the details of her education and how unusual that was for a woman in her culture).

But perhaps the question is closer to:

• If the lead doesn't require citations, and you want to add them anyway, should a GA or FA reviewer be able to force their exclusion?

WhatamIdoing (talk) 18:42, 26 August 2024 (UTC)

We're talking about the WP:Manual of Style/Lead section page, so I think it is a question of style. WhatamIdoing (talk) 18:37, 26 August 2024 (UTC)

@Gawaon: I'm not making radical changes to the policy, or any changes at all. I'm keeping the policy exactly the same but wording it clearly. Right now, the policy says "Citations are permitted but not required in the lead", but is worded so confusingly that everyone linking it claims the opposite.

My edits are just clarifying the current policy:

1. The manual of style discusses citations in leads, but makes no recommendation for or against including them.
2. Therefore, WP:STYLEVAR currently applies, because there's no established rule about citations in leads. You can change the style if you have a good reason, but not just because you feel like it.

Now, if you want to change the policy to enforce one way or the other, go ahead. You can rewrite this to say "Citations are forbidden in leads, except if X, Y, Z" or to say "Citations are always required in leads", or you can rewrite it to say "you can include or exclude citations in leads whenever you want, but don't change it without a good reason because edit warring over this is stupid". I personally won't revert regardless of what you do, since I don't care. Closed Limelike Curves (talk) 19:21, 26 August 2024 (UTC)

I suggest simply leaving the policy as is. Nobody except you so far has even expressed a desire for changes, as far as I can see. So if you're happy with its content too, what would be the point of changing anything? Gawaon (talk) 19:55, 26 August 2024 (UTC)

I'm OK with leaving the policy, i.e. its content, as is. I think the original commenter is the only person interested in a change (they want to require citations in leads). But the phrasing on this page is confusing and easy to misunderstand.

In fact, so far as I can tell, 2 editors commenting here have explicitly said they want to enforce a policy that does not allow citations in leads, outside exceptional cases (e.g. BLPs). 2 others have implied the same. I think that's fine and dandy, but it's a completely different policy from the one actually written on this page, which waffles between requiring citations and allowing but not requiring them. – Closed Limelike Curves (talk) 20:11, 26 August 2024 (UTC)

BTW, I looked at 10 current FACs, and 90% of them have no citations in the lead. 20% had an Wikipedia:Explanatory footnote. There may be a perception that FAC "expects" uncited leads, even if there is no actual rule. WhatamIdoing (talk) 18:38, 26 August 2024 (UTC)

The guideline waffles and ultimately ducks the issue by saying "The necessity for citations in a lead should be determined on a case-by-case basis by editorial consensus." So, there's clearly no hard rule and the OP's point is moot.

Myself, I usually start articles as small stubs and don't create a summary lead until the size of the article seems to warrant one. Citations will therefore appear throughout.

As a related point, note that Wikipedia's main page, which is its primary showcase, never has any citations at all. They are obviously not essential. What matters more is that the statements are accurate and correct. But even, so there's still a disclaimer on that and every page which emphatically advises the reader that Wikipedia cannot guarantee the validity of the information found here.

Andrew🐉(talk) 09:37, 26 August 2024 (UTC)

• So it's quite common to see citations in the lead during the early stages of article development, especially in the absence of experienced editor participation. If and when the citations are pulled from the lead as the article progresses, we get a chance to see if this creates any issues. Is there something highly controversial in the lead that warrants a citation? Do random editors repeatedly challenge a claim in this section, despite the lead qualifying as a legitimate summary of the article?
  If you find yourself constantly reminding others that, "Hey, it's in the body right here, complete with inline citations", well, maybe that's when you know you need to make an exception. Add a citation or {{Efn}} to the lead where there are issues, or even consider a rewrite. Exceptions should generally be a rare occurrence. Not sure there's a good way to codify that in the MoS, but MOS:CITELEAD isn't likely to settle the dispute for you. --GoneIn60 (talk) 19:20, 26 August 2024 (UTC)
• It radically decreases transparency and greatly weakens the verifiability requirement.
  Honestly, I disagree, A lead is meant to be a summary of content that's cited elsewhere on the page. While it might be nice to have citations in the lead (especially for more contentious articles), we shouldn't be requiring citations in the lead. For some articles, it may be helpful; for others, it defeats the point of having the lead be a concise summary. This is why WP:CITELEAD, as currently written, neither requires nor bans citations in the lead—editors may put as many or as few citations in the lead as they like.
  I'm just speaking from personal experience here, but for well-written articles, I can typically Ctrl+F the claim and find the citation within 3-5 seconds. If editors and readers can't do that, then it's not a well-written lead. – Epicgenius (talk) 21:34, 26 August 2024 (UTC)
• The lead should never have citations in it (barring quotations, which should probably not be in the lead anyway). The article is the source for the lead and it is right there for anyone willing to actually read. --User:Khajidha (talk) (contributions) 12:00, 27 August 2024 (UTC)

  @Khajidha, in the RFC below, does your view align with Wikipedia talk:Manual of Style/Lead section#Citations should be removed from article leads, except for BLP or direct quotations? WhatamIdoing (talk) 17:02, 27 August 2024 (UTC)

  My view is exactly what I typed. No more, no less. --User:Khajidha (talk) (contributions) 12:14, 3 September 2024 (UTC)

  It sounds to me like "The lead should never have citations in it barring quotations" is the same as "Citations should be removed from article leads, except for BLP or direct quotations", but I wasn't sure if you thought contentious matter about BLPs should be uncited in the lead. WhatamIdoing (talk) 16:21, 3 September 2024 (UTC)

• I think citations should be avoided in the lead and only put in if contentious. The body of the article is the source. The lead should summarize the article, putting in citations seems to stop people doing that and can make the lead say things which aren't in the article. A citation in the lead indicates to me that probably someone has stuck stuff in the lead without checking the article and making it fit in properly. NadVolum (talk) 21:44, 28 August 2024 (UTC)

MOS:CITELEAD

Should citations be included for most claims in an article's lead, just like in the body? 19:51, 26 August 2024 (UTC)

Citations should be included in article leads just like in the body

1. Oppose. Most featured articles do not have citations in leads and I do not want to be the one to correct that.– Closed Limelike Curves (talk) 19:51, 26 August 2024 (UTC)
2. Oppose; at best, this will clutter the lead, and at worst, it forces editors into an unpalatable WP:CITEBOMB situation. This is also far beyond what WP:V requires—the policy does not say that a claim must be cited every time it is mentioned, only that claims be cited somewhere in the article. (See footnote [a], which says: The location of any citation—including whether one is present in the article at all—is unrelated to whether a source directly supports the material.) If a claim is not sourced in the lead claim in the lead is not sourced anywhere in the article, it should be tagged with {{citation needed lead}} or removed. Epicgenius (talk) 21:15, 26 August 2024 (UTC)
3. Oppose As argued above, this oversteps policy and would invite overloading intros with little blue clicky linky numbers that don't actually help. XOR'easter (talk) 03:32, 27 August 2024 (UTC)
4. Support - keep it simple and use citations on every claim worth fact-checking The strength of Wikipedia is matching claims to citations and everything about Wikipedia's processes for fact-checking and review would improve if all surprising claims, including in the lead, had fact-checking. The original rationale from decades ago for not including citations in the lead was that typical readers would not comprehend footnotes, because Wikipedia was the first and is still the only mainstream publication to present clickable citations to non-scholarly audiences. A generation later, I think society has arrived at a point where it can understand the concept of citations, hyperlinks to a reference section, the need to cite authoritative sources, and the distinction between sources with transparent fact-checking and sources without. The major problem caused by absence of lead citations is increased difficulty in identifying incorrect claims in the lead, and that causes the leads to be lower quality than the rest of the article and for the leads to include claims which do not appear elsewhere. It is also common for someone to put a sourced claim in the article body with an unsourced version of the claim in the lead, then sometime later another editor removes the claim and citation in the article body as articles develop. I also feel that English Wikipedia should recognize and react with responsibility to how influential it is among speakers of English as a second language, and to other language versions of Wikipedia where editors have scarce resources to only translate leads of articles. In both cases, non-English speakers have a burdensome experience in trying to match unsourced claims in the lead with the sources elsewhere in the article, which may use different phrasing. Finally, Wikipedia's reputation would improve immensely if we could all just explain it with the simple rule, "all fact-checked claims in Wikipedia have citations". Wikipedia has experienced confusion for decades as readers notice that sometimes there are citations, and sometimes not, but it is inconsistent. If a claim is worth fact-checking at all then it is worth having a citation attached to it. Bluerasberry (talk) 13:44, 29 August 2024 (UTC)

   Can certainly understand this POV, but on the flipside, a summary may combine several claims from the body into one sentence. While this may be a relatively clean presentation in the body (usually 1 or 2 citations at the end of a sentence), now you're faced with making sure each part of that sentence combining multiple claims in the lead is clearly marked with the reference that supports it. This can get messy and convoluted in a hurry, and for some editors, may require additional skills or expertise to pull off. Maybe that's overthinking it, but it's a point worth taking into consideration. The rule may not be so simple as it may seem on paper. Having it as a requirement vs. an option? I'm not so sure about that. What might make sense in a complex article may make a lot less sense in a simple, short article. -- GoneIn60 (talk) 15:47, 29 August 2024 (UTC)

   @GoneIn60: Yes, if we mandate citations in the lead then in 1% of cases there will be valid problems of the sort you described, but without citations in the lead, 100% of articles have leads with no accessible evidence of fact-checking. We can work through the editorial challenges of combining claims with multiple citations, but besides changing the policy for citations in the lead, we have no plan for resolving the serious problems caused by making the most accessible and popular content in Wikipedia to be the most difficult content for users to verify. A generation ago our priority was to avoid presenting superscript hyperlinks to readers because we needed to maximize readability at a time when both the idea of footnotes and clicking links were surprising new concepts for the general public. My view is that all of this is about readability versus reliability, and with our current priorities about addressing misinformation in the media ecosystem, I would like to sacrifice some readability to try to gain a more robust fact-checking process. Bluerasberry (talk) 20:42, 29 August 2024 (UTC)

   I see your point, but I don't think we should be sacrificing readability so people could save a few minutes verifying the facts in the lead. As mentioned above, a single sentence may combine several claims from the body, which might possibly lead to severe WP:CITEKILL issues. I would rather that we not have to say something like "The team has won many championships over the years." if it's already cited in the body. – Epicgenius (talk) 15:24, 31 August 2024 (UTC)

   Wouldn't WP:BUNDLING solve this particular problem? – Closed Limelike Curves (talk) 01:33, 1 September 2024 (UTC)

   No; I don't see how it is even relevant. The issue is whether citation footnotes in the lead are desirable or even necessary. -- Shmuel (Seymour J.) Metz Username:Chatul (talk) 10:53, 1 September 2024 (UTC)

   all fact-checked claims in Wikipedia have citations
   

   Isn't this already the case if the citation is in the body rather than the lead (in other words, what WP:CITELEAD currently allows)? – Epicgenius (talk) 18:43, 31 August 2024 (UTC)

5. Oppose The purpose of the lead is to summarize the body. The body is what is being cited. To check the lead you should check that what it says is in the body. If you think there's something wrong then it can mean either the summary does not summarize the body or the body is wrong. Essentially I see citations in the lead as being opposed to its purpose and liable to separate the lead from the body and make the article worse overall. If anybody is able to follow a citation they can check the artcle for what is said in the lead. That said I can see citations in the lead can be useful sometimes but in general if there isn't a good reason for duplicating them I see them more as something to satisfy the sort of PITAs who think the article title isn't correct unless it accurately desribes everythng in the article. NadVolum (talk) 22:05, 29 August 2024 (UTC)
6. Oppose per Epicgenius and XOR'easter. Draken Bowser (talk) 17:42, 1 September 2024 (UTC)
7. Oppose the purpose of the lead will still be to summarise the body, no? What's the point in requiring citations for a section where the defining policy is not WP:V but WP:NPOV. Badly thought out idea. ~~ AirshipJungleman29 (talk) 20:58, 1 September 2024 (UTC)
8. Oppose per the above. Leads summarise the body, which is where the citations should be in place. As many readers will only read the lead, let’s not clutter up their reading experience with cumbersome citebombing. Bad idea. - SchroCat (talk) 03:35, 2 September 2024 (UTC)
9. Oppose; for a substantial, carefully referenced article with a good lead, many references, either to entire books or to lists of pages, will be required. Mid-sentence references will often be requured for the reader to know which reference supports the particular claim they are wanting to check. A far easier way for the reader to check will be locate the claim in the body and use the reference there. Thincat (talk) 06:33, 2 September 2024 (UTC)
10. Oppose - The information should already be cited in the body of the article and it should be reasonably trivial for a reader to find the corresponding claim and citation later in the article. The lead essentially functions as an abstract of the article, and I don't believe citations are normally presented in abstracts. Exceptional claims in the lead can and should be cited as necessary, but this is certainly not necessary for most information. Hog Farm _Talk 15:32, 2 September 2024 (UTC)

Citations should be removed from article leads, except for BLP or direct quotations

1. Oppose. Most featured articles do not have citations in leads, but some do. A campaign to retroactively delete those citations will be a pain in the ass.– Closed Limelike Curves (talk) 19:51, 26 August 2024 (UTC)
2. Oppose. This goes too far in the other direction, and in some cases (e.g. basically any article about a controversial subject), there may be a good reason to put citations in the lead. Epicgenius (talk) 21:15, 26 August 2024 (UTC)
3. Oppose We don't need a blanket rule in favor of inclusion or in favor of removal. There are more occasions than just direct quotations and BLPs where citations are good. For example, in an article about a constant of nature (e.g., Planck constant), it makes sense to have the numerical value for that constant and a supporting citation to a standard reference work in the lede. Biographies of deceased but still controversial figures are another case where attaching citations directly to a description can back up a statement that partisans would want to dispute. XOR'easter (talk) 03:40, 27 August 2024 (UTC)
4. Oppose. I'm unconvinced that mandating citations in the lead is a good idea, but mandating that we remove them seems even sillier. How does mandating the removal of existing citations improve anything? Caeciliusinhorto (talk) 15:53, 1 September 2024 (UTC)
5. Oppose What does this improve? ~~ AirshipJungleman29 (talk) 20:58, 1 September 2024 (UTC)
6. Oppose but... if the tag is ambiguous in scope it could sensibly be removed. If a reader wants to check a quote's authorship they should refer to the body. All articles, not just BLPs should be carefully referenced but this is not adequately done in the lead. Thincat (talk) 06:34, 2 September 2024 (UTC)

Current policy (editors may do either, but should not switch without a good reason)

1. Support. This is by far the easiest approach. – Closed Limelike Curves (talk) 19:51, 26 August 2024 (UTC)
2. Weak support, although not sure on "should not switch without a good reason". We realistically should make it clear that citations are optional only if the information in question is already cited in the body. If a claim is not sourced or mentioned in the body, it should be sourced in the lead, added to the body with a source, tagged with {{citation needed lead}}, or removed. Epicgenius (talk) 21:15, 26 August 2024 (UTC)

   I'd definitely qualify that as falling under "good reason"! When I say "without a good reason" I'm just reiterating Wikipedia:STYLEVAR here, which is the current guidance and therefore policy. – Closed Limelike Curves (talk) 00:12, 27 August 2024 (UTC)

• Comment: That the "current policy" is "editors may do either, but should not switch without a good reason" is disputed, as MOS:CITELEAD nowhere actually says so. Gawaon (talk) 18:56, 27 August 2024 (UTC)

  The statement in bold above about current policy is false. SandyGeorgia (Talk) 05:42, 2 September 2024 (UTC)

• Comment also highly sceptical about the "should not switch without a good reason", because there is no "switching"—it's not like you can only have zero or all lead sentences cited. Cite what needs to be cited, and don't cite what doesn't. Simple. ~~ AirshipJungleman29 (talk) 20:58, 1 September 2024 (UTC)

  Yep. For example, in medical leads, I only cite hard data that might change with time (epidemiology and things like that). It's not either/or-- it's cite what needs to be cited. SandyGeorgia (Talk) 05:43, 2 September 2024 (UTC)

• Comment; nothing should ever be done without a good reason and we do not need to explain this frequently. Thincat (talk) 06:35, 2 September 2024 (UTC)

No change, leave MOS:CITELEAD as it is

1. Support. None of the other options makes sense under all circumstances, and no change is needed. Gawaon (talk) 21:17, 26 August 2024 (UTC)
2. Support. I think a clarification of the MOS:CITELEAD guideline might be warranted (see my comment under "Editors may do either, but should not switch without a good reason"), but I don't think either requiring or banning lead citations would be helpful at this time. Epicgenius (talk) 21:25, 26 August 2024 (UTC)
3. Support. Current wording provides adequate guidance. No change is needed. —David Eppstein (talk) 21:42, 26 August 2024 (UTC)
4. Support clearly this is the communal norm.Moxy🍁 21:55, 26 August 2024 (UTC)
5. Support It ain't broken. XOR'easter (talk) 03:43, 27 August 2024 (UTC)
6. Support But more obviously the primary instance I have seen of citations in leads becoming unwieldy is when there is contentious content. At one point I think it was Gamergate had 17 citations for one simple assertion that Gamergate "harassed" some people because of bad-faith arguments and persistent edit warring. Adding all the citations was overkill - but it made the arguments in defence of that sentences inclusion ironclad. As such I feel there is value to allowing the inclusion of some sources - but they should always be used and included within the article also. Orphan lead citations I disapprove of. Koncorde (talk) 23:25, 28 August 2024 (UTC)
7. Oppose This was never well established. This practice causes problems and it was a careless choice made so long ago. It is an ancient Wikipedia practice, but it is not dogma, and it is not like our other old practices that were well discussed, tested, and examined. If anyone can, prove me wrong by linking to the policy discussions which demonstrate that this practice had its origin in thoughtful decision. Bluerasberry (talk) 13:52, 29 August 2024 (UTC)

   That's a really terrible argument! What problems? Discussions are because editors think there is a problem with something. Therefore lack of discussions can indicate lack of problems. I believe there have been a few discussions on this - which indicates there may be a problem but nobody has come up with anything better. I don't think one can use that there has been discussions so there might be problems and there hasn't been discussion so there might be problems in the same sentence without a little tinge of something though. NadVolum (talk) 22:21, 29 August 2024 (UTC)

   It's hard to imagine that you may have already forgotten the considerable discussion about forcing personal preferences--mostly citations in leads, along with other issues--that led up to WP:ARBMED. I don't recommend rehashing those discussions on a guideline talk page, nor do I agree they were not based on "thoughtful" considerations about the problems with forcing citations into leads. SandyGeorgia (Talk) 05:58, 2 September 2024 (UTC)

8. Support no need to fix ~~ AirshipJungleman29 (talk) 20:58, 1 September 2024 (UTC)
9. Support, & vote as follows from that in the sections above. Johnbod (talk) 02:52, 2 September 2024 (UTC)
10. Support per the above. I’ve not seen any problems with the status quo and the sections above look like solutions in search of problems, rather than any attempt to deal with any existing issues. - SchroCat (talk) 03:38, 2 September 2024 (UTC)
11. Support and all other sections ignored accordingly. The guideline is as clear as it needs to be, and there is no problem to be fixed here. SandyGeorgia (Talk) 05:47, 2 September 2024 (UTC)
12. Support, and oppose other contradictory proposals accordingly. I would remind that the current MOS:CITELEAD allows (and in rare cases requires) citations in the lead where they are particularly valuable, such as for direct quotations and controversial statements, and indeed MOS:LEAD requires everything in the lead to be cited somewhere anyway. UndercoverClassicist ^T·C 05:54, 2 September 2024 (UTC)
13. Support Works fine as it is. Hawkeye7 (discuss) 06:09, 2 September 2024 (UTC)
14. Support until a useful change is proposed and agreed. Thincat (talk) 06:37, 2 September 2024 (UTC)
15. Support - already works; let's please not break it. Hog Farm _Talk 15:30, 2 September 2024 (UTC)

Editors may do either, but should not switch without a good reason, nor should they flag the lead as Citation needed

1. Support If the editors shouldn't gratuitously add citations to the lead then they shouldn't gratuitously add {{Citation needed}} to the lead. Of course, a lack of citations in the body would be a good reason. -- Shmuel (Seymour J.) Metz Username:Chatul (talk) 10:53, 1 September 2024 (UTC)
2. Am I supposed to be voting for or against this? What change does this mean? ~~ AirshipJungleman29 (talk) 21:00, 1 September 2024 (UTC)

   It would mean no change in the policy. Yes, it's confusing to have three separate options for the same thing. Unfortunately, someone else added an additional choice without clarification, because they think MOS:STYLEVAR doesn't apply to whether claims can be referenced every time they're included. – Closed Limelike Curves (talk) 03:01, 2 September 2024 (UTC)

   Please understand the difference between policy and guideline. SandyGeorgia (Talk) 05:44, 2 September 2024 (UTC)

   I've been seeing a lot of editors, especially newer (e.g., last five years) use the word policy in its everyday, unofficial meaning of "usual practice" or "what I understand to be standard operating procedure", encompassing not only the contents of Category:Wikipedia policies, but also guidelines, help pages, information, essays, what someone told them once at the Teahouse, etc. It's not necessarily bad, but I think it might cause confusion sometimes. WhatamIdoing (talk) 06:41, 2 September 2024 (UTC)

3. Comment, nothing should be done without a good reason. I think a {{cn}} in a lead is very rarely helpful and so there will very rarely be a good reason for putting one there. Thincat (talk) 06:39, 2 September 2024 (UTC)
4. Comment - I'm a bit confused on how this differs from the status quo proposal above. {{Citation needed/doc}} already says that the template should not be added to lead sections that are clearly summarizing the article ({{Not verified in body}} should be used instead). As for changing stuff without good reason, I would say that every good-faith edit ostensibly should have a good reason behind it. Epicgenius (talk) 17:12, 2 September 2024 (UTC)
5. Support While it already says this in {{Citation needed/doc}}, putting it here elevates it to guideline status. Hawkeye7 (discuss) 20:56, 2 September 2024 (UTC)
6. Partial support per Hawkeye7 and Epicgenius; that is, elevate the content of the {{citation needed}} to guideline, without the "editors may do either, but should not switch" WP:CREEP. SandyGeorgia (Talk) 00:01, 3 September 2024 (UTC)

Discussion

How about SNOW-closing this? An RfC whose OP opposes all suggested options obviously won't go anywhere (leaving aside the fact that none of the options makes much sense). Gawaon (talk) 20:20, 26 August 2024 (UTC)

If you want to propose some other option, feel free to add it. I initially opposed all because I don't care that much, but I've edited my comment to support the option I think makes most sense. – Closed Limelike Curves (talk) 20:25, 26 August 2024 (UTC)

@Gawaon I have no objections to including additional options, but duplicating an existing option just confuses people. I've merged both options, but I don't like the confusion this might lead to. – Closed Limelike Curves (talk) 22:38, 26 August 2024 (UTC)

I don't really see this leading directly to a change in this guideline, but it might be interesting to see what the community thinks overall and see whether some adjustments should be considered. There are editors who believe that "one sentence, one citation" is the ideal, and they might support a change. There are also editors who think that the fewer rules we have, the better.

In case the current state of citation density interests anyone, based on some work done recently by BilledMammal, the middle half of our articles, sorted by length, have about 5–30 sentences, and the middle half when sorted by number of refs have about 2–10 inline citations (about 35% of articles meet both of these conditions). We're running around one ref per four sentences for all articles, and one ref per 3.5 sentences for the middle-ish articles. Stubs (when defined as ≤250 words [mean of 110 words] or as ≤10 sentences) tend to have one ref per two sentences. WhatamIdoing (talk) 21:17, 26 August 2024 (UTC)

As if the RFC were not already dubious and compromised enough, in Special:Diff/1242462087, Closed Limelike Curves attempted to refactor it into three questions instead of four, merging two groups of questions that had already received separate answers. CLC: Do not change the questions in active RFCs. That is not how RFCs work. —David Eppstein (talk) 22:53, 26 August 2024 (UTC)

As a note, I tried to clarify choices 3 and 4 are the same, with both referring to current policy (MOS specifies both styles are permissible). I'm of the opinion the confusion caused by duplicate options is more dubious than just combining them, but if you disagree, I'm fine with leaving it. – Closed Limelike Curves (talk) 00:30, 27 August 2024 (UTC)

That is not how RFCs work. ...but it might be how RFCs happen all the time. @Gawaon boldly changed the third one after two people had replied there, seemingly to draw a distinction between "current policy" and "no change", or perhaps to indicate a belief that the current rules aren't that editors may do either (or perhaps that they shouldn't switch without good reason?). The practical difference between those two is unclear to me. WhatamIdoing (talk) 01:11, 27 August 2024 (UTC)

Woah... I thought something was off with my question title, but I figured I must have been misremembering what I wrote somehow.

@Gawaon If you object to the characterization of this as current policy that's fine. However, in the future, please provide a ping and explanation of the change (like I did when I merged the two), so someone can dispute this. I'd be happy to see a proposal from you clarifying exactly what you think the current policy says; given both I and @WhatamIdoing seem to think the description I gave of current policy is accurate, but you think it's inaccurate, I think it's pretty clear there's some kind of ambiguity here. – Closed Limelike Curves (talk) 02:32, 27 August 2024 (UTC)

You had explicitly written "If you want to propose some other option, feel free to add it", so that's what I did. MOS:CITELEAD says "editors should balance the desire to avoid redundant citations in the lead with the desire to aid readers in locating sources for challengeable material", which is far more nuanced than "Editors may do either, but should not switch without a good reason". Hence I don't think your third option is an adequate summary of the status quo. If you meant it as "don't change MOS:CITELEAD at all", well ... in that case my apologies to you, but you certainly didn't write that, and I didn't read it as such. Hence I added "No change" as a fourth option since it's the only one I can really support. And from the comments this RfC has attracted so far, other editors seem to think the same. Gawaon (talk) 06:36, 27 August 2024 (UTC)

I think the concern is that you not only added "some other option", but you also "changed an option people had already voted on", potentially changing how other people would interpret their votes. WhatamIdoing (talk) 17:05, 27 August 2024 (UTC)

I merely changed "Current policy (editors may do either, but should not switch without a good reason)" to "Editors may do either, but should not switch without a good reason" and considering that "Editors may do either, but should not switch without a good reason" is very clearly NOT the current policy (which doesn't address the question of switching at all), I think that change was indeed called for to prevent confusion. The only person, besides Closed Limelike Curves themselves, that had supported that option at that time was Epicgenius. They later change their position to "Weak support" and added their support for the "No change" option, indicating that "Current policy" rather than "Editors may do either, but should not switch without a good reason" had supposedly been their preferred position all along. Gawaon (talk) 18:10, 27 August 2024 (UTC)

I absolutely wrote that, as @WhatamIdoing showed. My original title explicitly described the option clearly and unambiguously as maintaining current policy. It included a description of what this policy is as well, unlike the newly-inserted option, which does not adequately summarize current policy (contra WP:RfC guidelines, which state each choice should be adequately summarized and explained).

From the comments this RfC has attracted so far, other editors seem just as confused by this. The question just asks:

Should citations be included for most claims in an article's lead, just like in the body?

In other words, both options are the same with regards to the question being asked. Both involve maintaining the status quo that citations are optional, and there are no situations where they are prohibited (which by MOS:STYLEVAR and other guidelines means users should not attempt to remove them without consensus). – Closed Limelike Curves (talk) 18:05, 27 August 2024 (UTC)

If you really believe that "Editors may do either, but should not switch without a good reason" is the current policy, then I challenge you to prove it by citing the exact sentence(s) in MOS:CITELEAD that say so. Gawaon (talk) 18:12, 27 August 2024 (UTC)

Although the presence of citations in the lead is neither required in every article nor prohibited in any article, there is no exception to citation requirements specific to leads. The necessity for citations in a lead should be determined on a case-by-case basis by editorial consensus.

– Closed Limelike Curves (talk) 18:26, 27 August 2024 (UTC)

If you'd like to propose a change exempting this one specific guideline from the declaration at MOS: that:

Where more than one style or format is acceptable under the MoS, one should be used consistently within an article and should not be changed without good reason.

Then you're free to propose doing so as an additional option. – Closed Limelike Curves (talk) 18:51, 27 August 2024 (UTC)

Ah, so you admit that MOS:CITELEAD doesn't say so itself, but you take it from elsewhere in the MOS. However, note that the presence or absence of citations is not a "stylistic choice", so that rule doesn't apply here. Gawaon (talk) 19:01, 27 August 2024 (UTC)

I did not say that. If the Manual of Style explicitly states "you can do either this or that", that's clearly a stylistic choice. I cited the exact lines clearly stating that this is a stylistic choice.

Although the presence of citations in the lead is neither required in every article nor prohibited in any article, there is no exception to citation requirements specific to leads. The necessity for citations in a lead should be determined on a case-by-case basis by editorial consensus.

If you believe this policy is not clear in this way, you may propose a change in the phrasing that would clarify this specific section is exempt from MOS:STYLEVAR. Otherwise, I'm done replying to WP:ICANTHEARYOU. – Closed Limelike Curves (talk) 21:47, 27 August 2024 (UTC)

MOS:STYLEVAR says: "When either of two styles is acceptable it is generally considered inappropriate for a Wikipedia editor to change from one style to another unless there is some substantial reason for the change." Having citations or not is not a question of style. I rest my case. Gawaon (talk) 22:00, 27 August 2024 (UTC)

If you believe that the location of citations is not a question of style, then perhaps you will consider making a separate proposal to remove CITELEAD from WP:Manual of Style/Lead section. Any time a time a rule is presented in a page that actually says "Manual of Style" in its title, I think it's fair for editors to assume that it's "a question of style".

Additionally I'm not sure whether you object to:

• Editors may do either, or
• but should not switch without a good reason, or
• both.

Which one(s) of these do you think is not the community's current rule and practice (regardless of where that rule may or may not be documented)? WhatamIdoing (talk) 22:55, 27 August 2024 (UTC)

I don't think any change is needed, because I don't think the MOS is only about what it refers to as "stylistic choices" in its lead, while STYLEVAR is precisely about such merely stylistic choices – such as the use of DMY vs. MDY dates. The MOS also has a chapter on images, but that doesn't mean that STYLEVAR also applies to the use or non-use of images. Consider: "Editors may or may not add images to articles, but should not switch from one style to the other without a good reason, hence they normally shouldn't add images to articles that don't yet have them." Would you consider that a good summary of our image policy? I for one, would not, despite that fact that that chapter lives indeed in the MOS. Gawaon (talk) 06:12, 28 August 2024 (UTC)

Has anyone made a decision to not include any images at all in any article, even if a good image became available? If not, then we can ignore that as a strawman.

I've seen editors decide to not include an image in the lead, and I'd usually classify that as a stylistic choice.

I've also seen editors object when someone switching the image from one to another similar image without a good reason. I think that's also a stylistic choice. WhatamIdoing (talk) 06:28, 28 August 2024 (UTC)

@Gawaon your comments are conflating the inclusion/non-inclusion of citations in an article as a whole (not a stylistic choice) and where exactly they should be located within the text (a stylistic choice). – Closed Limelike Curves (talk) 20:29, 28 August 2024 (UTC)

It's not merely a stylistic choice, as myself and others have already said. Like, when summarizing a long section with 37 references in a single lead sentence and you add just 3 references to that sentence, then you're potentially misleading your readers, since those 3 references might be insufficient to back everything in the summary. But if you tackle all 37 references to that sentence, unreadability ensures. So the best choice, and not at all stylistic, in such cases is to omit the references in the lead summary altogether, implicitly pointing to where they can actually be found, namely the body. Gawaon (talk) 20:39, 28 August 2024 (UTC)

I don't know if you knew this, but readers basically don't read the refs (per doi:10.1145/3366423.3380300), so there's very little you can do with a ref that would actually mislead any reader.

"Unreadability" is a stylistic choice. Style choices include what things look like. Thirty-seven little blue clicky numbers in a row looks like a mess. Preferring or avoiding that mess is therefore at least partially a stylistic choice. WhatamIdoing (talk) 23:51, 28 August 2024 (UTC)

"We should not have more than n references in a row" is a stylistic judgment. n=37 is a situation where it's an obviously-correct stylistic judgment, which would qualify as a good reason for the change. That said, it's still a stylistic question. For example, say some editor tried to delete all references after the first 2 in every sentence of an article, because they think anything with 3 citations in a row is "unreadable". That would be an attempt to impose a particular style and should be avoided. – Closed Limelike Curves (talk) 02:30, 29 August 2024 (UTC)

@WhatamIdoing: Would it be helpful to get the statistics limited to the lede of articles? BilledMammal (talk) 13:30, 27 August 2024 (UTC)

@BilledMammal, I think it would be difficult to reliably differentiate a stub from a lede, but if you want to have a go at it, I'd love to see the numbers. Perhaps it would be possible to filter for articles that contain a ==Section== with at least two non-list/non-template sentences in it? (Or some roughly equivalent number of words, if that's easier to calculate?)

If you were to run the first set on just, say, 50 or 100 articles, I'm willing to manually look at them all to see how the filtering is doing. WhatamIdoing (talk) 16:10, 27 August 2024 (UTC)

Max length guideline similar to the total length guideline

We're often getting instances where a lead is kept to 4 paras, but is almost 700 words e.g. [1]. It's worth strengthening the lead guidance to give a suggested max, like we have for total length: 15,000 words. We state most featured articles have up to 400 words, but this isn't keeping leads to a reasonable length. Surely adding a guideline very similar in wordage to the total length guideline, e.g. "More than 500 words means it almost certainly should be trimmed," would help. What do people think? Tom B (talk) 10:47, 28 August 2024 (UTC)

I agree such a recommendation would be good to have. My own rule of thumb is to strive for 225–450 words in the lead. Gawaon (talk) 12:49, 28 August 2024 (UTC)

@Tpbradbury, see Wikipedia talk:Manual of Style/Lead section/Archive 23#FA numbers and Wikipedia talk:Manual of Style/Lead section/Archive 22#Seeking consensus for table modification and probably some others.

In general, I think editors are willing to remove the paragraph count from the table and replace it with a word count. Something approximately in the 200–500 range is suitable for a longer article. Few leads, even on a Start-class article, should be less than 100 words.

However, I think there are two other statements they'd agree with:

• Most articles should have four or fewer paragraphs.
• Changing the number of paragraphs by adding/removing line breaks is not the point of this guideline. Having 700 words across five paragraphs is not automatically worse than 700 words across four longer paragraphs. What's usually needed is to not have 700 words.

WhatamIdoing (talk) 17:27, 28 August 2024 (UTC)

Also, I don't know if this has been discussed before, but I'm not sure that article length is actually the best indicator of lead length. It makes some intuitive sense, but a short article about (e.g.,) a disease needs the same information in the lead (e.g., definition, symptoms, treatment, epidemiology, prognosis) as a long article about the same subject. A long list might need a short lead ("This is the list of widget styles. There are ___ significant variations in widget styles. Blue-green widgets are widely agreed to be the most commercially important category" – a mere three sentences/27 words, but not much else is needed, regardless of whether the number in the blank is 15 or 5,000). Subject complexity might be a better indicator of lead length than the number of words it takes to write the article. Bangladesh is given as an example above, and it seems like an article about a country is a complex subject that would need a lot of room to fully summarize. WhatamIdoing (talk) 18:08, 28 August 2024 (UTC)

Table of contents is a better guide than overall length. If topics are important enough to gain a dedicated section or subsection, that topic is likely part of the article's "most important contents", and we further state weight in the body and lead should be roughly proportional. CMD (talk) 09:03, 29 August 2024 (UTC)

thanks, yes, I'm name-checked in the FA numbers one from archive! I've put forward one proposal, "More than 500 words means it almost certainly should be trimmed." Alternatively we could add or replace a column in the table, with something like:

Article length	Lead length	Lead length
Fewer than 2,500 words	One or two paragraphs	100 - 200 words
2,500–5,000 words	Two or three paragraphs	200 - 300 words
More than 5,000 words	Three or four paragraphs	300 - 500 words

What changes do people think we should make? Tom B (talk) 08:37, 29 August 2024 (UTC)

I'm in favour of just specifying a general recommended upper limit (500 words or similar). I don't think smaller articles necessarily need a short lead. Gawaon (talk) 09:03, 29 August 2024 (UTC)

How's this for a replacement section?

The appropriate length of the lead depends on the total length of the article complexity of the subject and development of the article.As a guideline, the lead should usually be no longer than four paragraphs. A lead that is too short leaves the reader unsatisfied; too long is intimidating, difficult to read, and may cause the reader to lose interest halfway. These suggestions are useful:

• The length should conform to readers' expectations of a short, but useful and complete, summary of the topic.
• Few well-written leads will be shorter than 100 words. Most featured articles' leads are about three paragraphs, containing 10 to 18 sentences, or 250 to 400 words.
• The Table of Contents may be a useful guide to how complex a subject is.: if it needs many sections, then the lead will likely need to be longer.
• If the lead is longer than 500 words, it almost certainly needs to be shortened.
• Most articles should have four or fewer paragraphs. However, it is better to have five well-organized paragraphs than to combine multiple topics to artificially force a smaller number of paragraphs, or to avoid a single-sentence paragraph.

Lead sections that reflect or expand on sections in other articles are discussed at Summary style. Journalistic conventions for lead sections are discussed at News style.

WhatamIdoing (talk) 18:07, 29 August 2024 (UTC)

Reads good to me, except that I fear that "the lead will likely need to be longer" could be misinterpreted as "longer than (the) 500 words" mentioned in the previous item, which of course would not be the intent. Gawaon (talk) 19:40, 29 August 2024 (UTC)

yes i thought that too independently, before reading your comment, Tom B (talk) 21:43, 29 August 2024 (UTC)

thanks very much. i've suggested removing nearly all the underlining, we should maybe remove all of it, and the bit about 5 paras. I've never seen a situation where 5 paras is better. If you get the word count down, you don't need more than 4 paragraphs. I've reordered, as it said 500 words max, then the next bullet said if the TOC is long, then it needs to be longer, perhaps implying longer than 500 words, which isn't what we mean! i've brought up the featured bit, next to well written leads. I'm sure the TOC point will be used to justify more than 500 words, so i've suggested striking it, particularly as you've added the complexity and development point already, thanks again, Tom B (talk) 21:42, 29 August 2024 (UTC)

I think the current wording saying that 500 is a hard limit and anything over that "almost certainly needs to be shortened" is far too strongly worded. 500 words can mean merely a solid four paragraphs of text. For instance the current lead of Emmy Noether (a Featured Article) appears to be 591 words long. A 1000-word lead almost certainly needs to be shortened. A 500-word lead might be far too long for some subjects, just right for others, and not quite enough to adequately summarize a few such as maybe Noether's. —David Eppstein (talk) 21:51, 29 August 2024 (UTC)

it's 574, and significantly above average, as someone did analysis linked above in FA numbers [2] It was 527 when it got promoted [3] and now undergoing review. This is the issue; everyone thinks it's ok if a few are over a limit, and then we've got lots of 600 word leads! Tom B (talk) 22:23, 29 August 2024 (UTC)

Saying it's "now undergoing review" is kind of euphemistic. All identified problems were cleared months ago and it has been sitting in FAR limbo for months waiting for someone to do something. —David Eppstein (talk) 23:06, 29 August 2024 (UTC)

We could tone down the wording for that, e.g., "probably" instead of "almost certainly".

If we had more information about FAs (@BilledMammal has been thinking about it; I suspect that a comprehensive list of oldids for FAs at the time of their promotion would be a useful contribution to this goal), then we could phrase it another way, e.g., "Fewer than 1% of FAs have a lead length in excess of 500 words". WhatamIdoing (talk) 22:58, 29 August 2024 (UTC)

I'd leave the "better to have five well-organized paragraphs" sentence in, I think it's reasonable to have it. Nothing wrong with 5-paragraph leads now and then. Gawaon (talk) 22:02, 29 August 2024 (UTC)

Paragraphs are supposed to be unified around a topic, at least according to teachers of grammar and composition. That means that if you're writing an article about, e.g., a country, and you decide that the important things are history, geography, demographics, and culture, then you could end up with an introductory paragraph followed by four paragraphs (one each for history, geography, demographics, and culture), even if that's technically "five" instead of "four".

My main goal, though, is to stop people from claiming that these five paragraphs (numbered for your convenience):

1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
2. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
3. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
4. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
5. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

are magically better and shorter than the exact same 345 words with one fewer line break:

1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
2. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
3. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
4. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

Paragraphs should be divided according to the content of the paragraphs. If you need all 345 words here, then put in the number of line breaks that their content suggests. If you don't need all 345 words, then take some words out, rather than taking some line breaks out. Changing the number of line breaks to comply with a "suggestion" in a guideline is a bad idea. WhatamIdoing (talk) 22:52, 29 August 2024 (UTC)

I agree. Gawaon (talk) 06:38, 30 August 2024 (UTC)

Tom, the TOC idea came from a comment by @Chipmunkdavis. I've heard this recommended as a best practice for writing a lead from scratch. If some point was important enough to write a whole section about it, then it should be considered for at least a brief mention in the lead. It might not get much attention (e.g., epidemiology might be condensed to "is a rare disease"), but it should be considered. WhatamIdoing (talk) 23:02, 29 August 2024 (UTC)

I don't think it's really necessary. As people summarize sections for the lead, an article with multiple sections will automatically lead to a longer lead. If something is worth pointing out in that regard, it might be something like "Short articles with few sections will often have fairly short leads". Gawaon (talk) 06:41, 30 August 2024 (UTC)

I think it's also true that "Medium-sized articles with few sections will often have shorter leads" and even that "Long articles with few sections will often have shorter leads". WhatamIdoing (talk) 16:49, 30 August 2024 (UTC)

thanks, whatever improves things. the issue for me is we get too long leads being tagged, and i'm strongly assuming anything that gives people a hook to justify a too long lead will be used. it'd be great if we could get consensus on what too long is, but i'm always amazed how long some others want things. I don't know why we can't just set a simple hard limit, people always want to leave some complex living room, when there must some reasonable limit, most can agree on. one measure is when, on average, do people add too long tags. i've been working on this a lot this year and it's often about 500 words, i've seen tags added at only 400, which is wrong, Tom B (talk) 21:33, 30 August 2024 (UTC)

We can't set a firm rule against having >500 words in the lead because we have a fundamental policy that says we have WP:No firm rules.

OTOH, if we could get a reliable list of all leads >500 words, I could imagine someone thinking it would be helpful to add {{lead too long}} to all of them, but I don't think that we could guarantee that 510 or 550 words actually would be too long in every single case. WhatamIdoing (talk) 22:18, 30 August 2024 (UTC)

I would think a bean-counting gnome of the type you describe to be very tendentious, but YMMV. —David Eppstein (talk) 22:29, 30 August 2024 (UTC)

there's a list here that might be of use: https://pageviews.wmcloud.org/massviews/?platform=all-access&agent=user&source=category&start=2021-03-03&end=2024-03-23&subjectpage=1&subcategories=1&sort=views&direction=1&view=list&target=https://en.wikipedia.org/wiki/Category:Wikipedia%20introduction%20cleanup Tom B (talk) 14:52, 31 August 2024 (UTC)

I think we should instead set the lead's length as a percentage of the article's length. Based on the table presented by Tom B above, it seems like the optimal lead length should be "around 4-10% of the total article length":

Article length	Lead length (paragraphs)	Lead length (words)	Ratio of lead to article length
Fewer than 2,500 words	One or two paragraphs	100 - 200 words	>4-6%
2,500–5,000 words	Two or three paragraphs	200 - 300 words	4-12%
More than 5,000 words	Three or four paragraphs	300 - 500 words	<10%

This 5-10% recommendation would work even for larger articles. For example, the Cleopatra article (which I picked because it's among the most-viewed FAs) has a 690-word lead, out of a total readable prose size of about 13,000 words, so the lead is just a little over 5% of the total prose size. This falls right into the 4-10% range mentioned above. However, if we recommended that the lead be no longer than 500 words, then the lead of Cleopatra would be less than 4% of the total prose size, which may not be enough. My concern is that someone might misinterpret the 500-word recommendation as a hard limit, then add {{lead too long}} where it isn't warranted. – Epicgenius (talk) 15:19, 31 August 2024 (UTC)

thanks, it would be better to increase the word limit above 500, rather than have 10%, otherwise people will say a 14,000 word article means i can have a 1,400 word lead! Tom B (talk) 15:26, 31 August 2024 (UTC)

That's a good point. In that case, maybe we can set an upper limit of 600-700 words. Most articles don't go about 12,000-14,000 words, and if an article reaches that point, it might need to be split anyway. – Epicgenius (talk) 15:27, 31 August 2024 (UTC)

thank you, that's going quite a way above 400 which is suggested as best practice, which is why i put 500. whatever guideline you put down, people go past it, particularly as some don't like simple, firm guidelines. so it's worth putting it lower than you want it and then people will go past it, but only slightly, Tom B (talk) 15:42, 31 August 2024 (UTC)

I strongly disagree. Whatever guideline you put down, some gnome will go around putting attention-wasting screaming cleanup banners at the top of the article saying THIS LEAD IS TOO LONG! SOMEONE MUST IMMEDIATELY CLEAN IT UP! for any article that goes even a single word above the threshold even when the lead is perfectly appropriate. So any hard limit should be set at the point where almost all articles exceeding it really do have a too-long lead. —David Eppstein (talk) 18:23, 31 August 2024 (UTC)

This is one reason why I am uncomfortable with all numerical thresholds for this purpose, whether they are expressed as word counts or percentages. I don't know of any good and reliable way of evaluating a lede other than reading it, reading the article, and then judging whether the former does the job it's supposed to. Trying to put a number on that seems like an attempt to make a hard problem look easy. If we have to bikeshed, maybe the best course is to minimize the harm. Draw the line in a place that reduces the chance of drive-by banner-tagging busybodies starting arguments and taking time away from article improvement. XOR'easter (talk) 19:34, 31 August 2024 (UTC)

thank you, that's the situation we've got at the moment. the guidance says 4 paras, so people are tagging articles with 5/6. Or we're getting 4 paragraph leads almost 700 words long, when featured usually have 400 max, so people are tagging those too. Switching to, or adding a word guideline like we have for article length, would at least reduce the second problem, Tom B (talk) 13:08, 1 September 2024 (UTC)

The page views for Cleopatra are artificially inflated by an example used by Google Assistant.[4][5]

I don't think that the lead content is necessarily proportional to the size of the article. Consider an FA like Influenza. It's currently 9,731 words (per Prosesize gadget). The lead has 472 words (4.85%).

The lead has four paragraphs which say:

• what it is, what the symptoms are, and what the complications are;
• how it's classified and how it's transmitted;
• how it's prevented, how it's diagnosed, and how it's treated; and
• how many people get it, including famous epidemics.

The thing is: If you halved or doubled the number of words in the body, the lead would still want this same information, and it would still take approximately the same amount of space to say it. Compare this against some other common airborne infectious diseases. All of these have four similarly structured paragraphs:

• Our article about the Common cold (GA) is only 30% as long as Influenza, but it still needs 70% as many words in the lead to cover the same information (2843 words total, 334 in the lead; 11.7%). The parts of the lead that are shorter in Common cold are shorter because of the facts (e.g., no vaccines, no diagnostic tests, no famous epidemics), not because of the article length.
• Measles (B-class) is 60% as long as Influenza, but the lead takes 15% more words to cover the same information (5,904 words total, 540 in the lead; 9.15%). The parts of the lead that are longer are the symptoms and transmission, and these are longer because of the facts, not because of the article length. The description of the symptoms in the body of the two articles use almost the same number of words (249 for Measles and 283 for Influenza).
• Tuberculosis (B-class) is as 77% long as Influenza, but the lead needs 85% as many words to cover the same information (7,486 words total, 400 in the lead; 5.34%). This lead is shorter because the two middle paragraphs have fewer facts to relate (e.g., no classification). This is partially balanced by the fourth paragraph having more facts to relate.

According to the proposed table, Common cold needs to have one fewer paragraph and at least 10% fewer words – even though it's the shortest of all of these, even the contents of each paragraph are necessary for a comprehensive summary of the article and there isn't a single sentence that I think is unnecessary. That doesn't make sense.

I can understand someone saying "Look, the lead in Measles is just a bit too long, so people won't read it", but I can't understand someone saying "This is the shortest of the bunch, but it needs to be even shorter. It's perfectly fine, though, if you want to add another 20% to Tuberculosis, though". WhatamIdoing (talk) 18:02, 31 August 2024 (UTC)

To be fair, the common cold article's lead is only about 12% of the overall length, which I think is fine. According to the table, the optimal lead ranges from 4% to 12% the overall length, and all the examples you cited fall into that range. I only mentioned 4-10% as an approximate range, and this could, of course, be adjusted later on.

However, Tom B made a good point above that a ratio-based lead length might not be optimal for much longer articles (e.g. a 14,000 word article doesn't need a lead that's 10% the overall length). – Epicgenius (talk) 19:17, 31 August 2024 (UTC)

I don't think we should have a suggested number of paragraphs. It is too easily gamed. WhatamIdoing (talk) 18:04, 31 August 2024 (UTC)

I agree that the length of a good lede isn't necessarily proportional to that of the article. The job of the lede, first and foremost, is to summarize the main text that follows. Sometimes, one can expand the main text with further details about whatever the topic is, details that belong in the article but don't demand mention in the summary. So, the main text grows more than the lede. That's just life. XOR'easter (talk) 19:16, 31 August 2024 (UTC)

Perhaps that suggests simplifying the suggestions to just these two points:

• The length should conform to readers' expectations of a short, but useful and complete, summary of the topic. Editors must use their judgment to determine whether the lead meets this goal.
• Few well-written leads will be shorter than 100 words. The leads in most Wikipedia:Featured articles contain about 250 to 400 words.

The first says that you need to include what you need to include. The second provides a suggested minimum and a fact about FAs that hint at a desirable length. No specific number is given for a maximum, but I suspect that in most articles, especially if they are not heavily contested articles, that will encourage conformity without encouraging mindlessness. WhatamIdoing (talk) 21:38, 31 August 2024 (UTC)

Sounds reasonable. Gawaon (talk) 22:06, 31 August 2024 (UTC)

thanks, i'm for making things simpler and it avoids the issue of people simply removing a paragraph break and therefore everything is fine, which it's not. maybe you've struck gold, or at least silver. Will removing the paragraph guidance lead to a general increase in lead size? Tom B (talk) 13:24, 1 September 2024 (UTC)

in fact i'd make it even simpler by removing the 'editors must' sentence as it's already implicit, and I can see a few using it to justify any length as simply their judgement. also people generally don't like 'must' directions if it's avoidable, Tom B (talk) 13:39, 1 September 2024 (UTC)

It's a bit surprising how allergic some people are to the word must. I'm equally opposed to saying should when we enforce a rule rigidly ("You should not add URLs to anything on the WP:Spam blacklist" – No, you must not, and you can not, because we instantly block that in software. It's not just a good suggestion about what you "should" do. It is an actual requirement).

But in this case, it's not a demand. It's a statement of fact: The only way to determine whether this goal is met is to use your best editorial judgment. Still, we can leave it out. If we really need it, someone will come around and ask how to do this ...and I doubt that will happen, honestly.

I'll go make these changes in a minute, so we can see what it looks like. WhatamIdoing (talk) 17:13, 1 September 2024 (UTC)

Here's a link to the new version. Please take a look. WhatamIdoing (talk) 17:35, 1 September 2024 (UTC)

It's ok. I might add that it's more important to prioritize concisely summarizing the main points of the article than it is to prioritize hitting numerical targets, but I'm not sure it's necessary to say that explicitly. —David Eppstein (talk) 19:26, 1 September 2024 (UTC)

thank you, we could remove "and complete"? as this might encourage people to try and cover everything when only the main points are need for the lead, Tom B (talk) 19:57, 1 September 2024 (UTC)

I can live with either outcome. "Complete" might encourage one editor to add too much, but it might stop a different editor from removing too much. WhatamIdoing (talk) 20:00, 1 September 2024 (UTC)

Generally speaking, I think this is a good change, but I'm surprised that there is now no explicit mention that the length of the lead will generally be proportional to the length of the article. It would be wise to spell this out, to avoid the impression that the numerical direction in this part of the guidance contradicts MOS:INTRO a little further up. UndercoverClassicist ^T·C 06:12, 3 September 2024 (UTC)

@UndercoverClassicist, I think we agreed here that the length of the lead generally shouldn't be proportional to the length of the article. See, e.g., my analysis of some medical articles, which all need approximately the same lead length, even though the articles vary 3x in length. Also, we tried coming up with some percentages, and they just weren't useful. In the table above, we considered a suggestion that a mid-length article of 2,500–5,000 words should have a lead that is 4% to 12% of the article length. That means somewhere between 100 and 600 words ...which is not actually useful advice, since 100-word leads are usually too short, and almost everyone thinks that a 600-word lead is usually too long.

I think that leads on complex subjects usually need longer leads, but a simple subject (e.g., most list articles) can be very long without needing a longer lead.

(I don't actually see anything in MOS:INTRO that says anything about lead length.) WhatamIdoing (talk) 06:44, 3 September 2024 (UTC)

My concern is that MOS:INTRO, as the overarching guideline, has The lead section should briefly summarize the most important points covered in an article in such a way that it can stand on its own as a concise version of the article. That's very much dependent on the length of the article to begin with, whereas all of the instructions currently in LEADLENGTH are agnostic to it. There's no contradiction as such, but there is now a difference of emphasis. UndercoverClassicist ^T·C 07:30, 3 September 2024 (UTC)

I don't think the quoted text is dependent on the length of the article. The most important points could be covered in the body as a single sentence each, or covered with a few paragraphs each. CMD (talk) 08:20, 3 September 2024 (UTC)

Are you concerned that it says the one section currently says "short, but useful and complete, summary of the topic" and the other says "concise version of the article"? WhatamIdoing (talk) 16:15, 3 September 2024 (UTC)

Yes: I think "article" is better, since elsewhere we say that facts stated in the lead should also appear in the body. Calling it "a summary of the topic" suggests, I think, that it's more independent of the body than it should be. UndercoverClassicist ^T·C 19:31, 3 September 2024 (UTC)

I think there are exceptions to the idea of "facts stated in the lead should also appear in the body". IMO the most important exception is that some leads also have to include information about the Wikipedia:Manual of Style/Lead section#Scope of article. WhatamIdoing (talk) 22:18, 3 September 2024 (UTC)

That is true, but this is also a true case of the exception proves the rule -- outlining that meta information is not required to be repeated in the body underlines that everything else, generally speaking, is meant to be repeated. UndercoverClassicist ^T·C 06:24, 4 September 2024 (UTC)