Wikipedia talk:Verifiability/Archive 39

SPS with sources

Right now there's a discussion [1] going on about navweaps.com, a SPS source that uses verifiable sources. It is essentially one guy's compilation of naval weapons data compiled from a variety of published sources, not OR or simple opinion.

As written WP:SPS doesn't address the issue of sources at all, which, in my mind, is a major flaw in the policy. Given that the site meets the overarching standards of WP:V and WP:RS, I would argue that its self-published status is irrelevant to its overall reliability. But people are arguing that it is inherently unreliable because it is self-published. I think that they are focusing on the trees rather than the forest because I believe that the SPS policy was established to deal with sites that contain unsupported opinions/allegations/facts, which is not the case here.

So I propose to amend the SPS policy to state that a self-published source is reliable if it doesn't contain OR and its sources are themselves reliable and verifiable, matching the overall policies on OR, RS and V.--Sturmvogel 66 (talk) 18:46, 6 June 2010 (UTC)

The fact that an SPS contains reliable sources is immaterial to its own reliability. If there are links to reliable sources on such a site, those sources can be used directly, with no need for an unreliable intermediary. Crum375 (talk) 18:58, 6 June 2010 (UTC)

(Removed excess indention which messed up browser:Jc3s5h) Unfortunately not all of those source are available in English, so that's no substitute.

No... don't change the policy. The solution to the debate about navweaps.com is whether the author can be considered an established expert or not. I don't know enough about the topic to opine on that, but there are others who can. I think the issue should be settled by consensus at the relevant project pages. Blueboar (talk) 19:14, 6 June 2010 (UTC)

I agree that it would be a mistake to change the policy, and I don't see why we need to. SPS seems to me to allow this source: "work in the relevant field has previously been published by reliable third-party publications". Ed gave a list at the RSN link of 23 books, reliable and even authoritative, that reproduce, cite and rely on the words and data from this source, so this author's work has been widely published, and he's been credited. - Dank (push to talk) 19:18, 6 June 2010 (UTC)

(ec) Yes, but bear in mind that the WP:SPS exclusion is not for any "expert", but "an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications." Crum375 (talk) 19:20, 6 June 2010 (UTC)

In concept I agree... but I also think "expertness" can be established though broad consensus of editors. If the editors at the relevant project all agree that the author is an expert in the field, I don't think we should bar the source because of technicalities (especially we can always invoke IAR to allow it anyway). Our policies are meant to be somewhat flexible. The intent is to allow SPSs that are written by an expert ... publication is the best way to ensure that the person is an expert, but there are other ways to determine that fact. Blueboar (talk) 19:37, 6 June 2010 (UTC)

I disagree. You could have a bunch of editors on a page who decide among themselves that someone is an expert, but that doesn't make him an expert for SPS exclusion purposes. WP:SPS specifically says the author must be "an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications." This means that some of this author's work has already been published reliably elsewhere, and this is more of the same (or similar) work, except he is publishing it on his own. Crum375 (talk) 19:46, 6 June 2010 (UTC)

I am not talking about a "bunch of editors on a page" ... I am talking about a broad consensus of editors who participate in the relevant WikiProject. In the specific case in question, it would be WikiProject: Military history. That project has over 600 members who could potentially opine, which I think would give you a very clearly demonstrated consensus as to whether the source is considered an expert or not. Heck, we sometimes change policy with less of a demonstrated consensus than that. On the other hand... if only six of them opine, you obviously can't say there is a broad consensus. Blueboar (talk) 20:35, 6 June 2010 (UTC)

We've had this situation several times, where someone sets himself up in an obscure intellectual area, and turns himself into an expert by self-publishing a lot of material about it. As with all our sourcing issues, this is fine until it is challenged. By definition, if someone is challenging it (assuming it's not someone being a DICK), there's a question mark over the expertise. At that point you look to see whether this person is acknowledged as an expert by others, and the way the sourcing policy recommends we do that is to see whether his work in the field has been published by anyone but himself. The problem with depending on broad consensus of editors is that POV issues might come into play. SlimVirgin ^{talk contribs} 20:42, 6 June 2010 (UTC)

BB, I think that's just inviting discord. This is why we have content policies, to provide us key critical elements, around which we can build consensus. In this case, the key provision of the SPS expert exclusion is "previously published by reliable third party publications." The editors can debate the meaning of "published" or "reliable", but not the key concept that this expert's work must have been published before by a reliable third party publisher. If there is a broad consensus, then it can be used to modify the policy, not to override it. Crum375 (talk) 20:48, 6 June 2010 (UTC)

But they can override it anyway... by invoking IAR. If you need to, think of an appeal to project wide consensus as being the same as asking: "This does not technically qualify as a reliable source under WP:SPS ... but I think using it is justified due to the reputation of the source... shall we invoke IAR and allow it?" Blueboar (talk) 21:10, 6 June 2010 (UTC)

IAR can't be invoked when someone is legitimately challenging a source. That's where the policy kicks in. SlimVirgin ^{talk contribs} 21:18, 6 June 2010 (UTC)

(ec) WP:IAR is intended to cover special or unusual cases we have not previously foreseen or covered in our policies, not as a tool to undermine the content policies. In this case, there is nothing special or unusual about a group of editors, from a project or otherwise, wanting to push a source as "expert" and therefore declare it reliable. This is why the WP:SPS policy specifically requires such expert to have been previously published by reliable third party publishers. To say that a group of editors can decide by IAR that their opinion trumps the SPS policy would undermine the concept of the content policies, and we'd end up with anarchy. Crum375 (talk) 21:19, 6 June 2010 (UTC)

As part of regularly assisting at the RS/N page, I participated in that discussion. As is quite clear, based on both policy and practice, "expert" means "expert according to reliable sources". Not "expert according to often-anonymous Wikipedia editors". One of the main values of policy is to impose the broader Wikipedia consensus on small groups of editors with narrow interests and views, editing on specific pages, who may often "agree" to "local policies" that contradict Wikipedia policies, and, in general, damage the quality of the articles themselves (and the credibility of the project). Jayjg ^(talk) 21:43, 6 June 2010 (UTC)

Well, I am not going to push it (especially since I don't know the specific details of the case in question). I still think the rest of you are being overly pedantic by sticking to the letter of the policy, and I disagree with saying that we can never determine expertize by consensus, but there is no point in discussing it further. And just to make it clear, I am definitely not suggesting any change in the wording of the policy. Only a consideration of what I see as a very limited interpretation of it's intent, applicable in rare situations. I am obviously in a minority of one. Blueboar (talk) 21:57, 6 June 2010 (UTC)

BB, you can determine expertise by consensus of knowledgeable editors. But you can't override policy. And the WP:SPS policy says that the only exception to SPS is a published expert, i.e. one who is adding a self-published work to others he has already published on this topic via reliable third party publishers. So the criterion is not just being an expert, but a published one, and then we allow you to bypass SPS. Crum375 (talk) 22:22, 6 June 2010 (UTC)

Can we close this thread out and continue as needed at RSN? The best I can tell, only Sturmvogel was in favor of changing the wording on this page; it looks like we're not going to get consensus for the change. - Dank (push to talk) 23:28, 6 June 2010 (UTC)

All I can says is that it seems crazy to me to call something unreliable simply because it's self-published when it follows WP:V and WP:RS and doesn't contain OR or unsupported opinion. Any lack of editorial control, as is so often mentioned herein, or whatever is irrelevant when the guy is compiling data from a variety of sources, all given, in a variety of languages. Languages, I'll remind y'all that not all editors can read and replace.--Sturmvogel 66 (talk) 02:35, 7 June 2010 (UTC)

Must references be in-line?

A while back, this page was changed from a nuanced discussion of the reasons why in-line citations are preferred to an absolute statement that citations must be in-line. While I agree that there is widespread preference for the style choice that cites be in-line and that they work for the vast majority of situations, I can find no consensus for the statement that only in-line cites are acceptable. Can someone please point me to where/when this was decided? Rossami (talk) 23:56, 2 June 2010 (UTC)

• I want to say that I much prefer the version Rossami advocates.—S Marshall T/C 00:01, 3 June 2010 (UTC)

If some material is challenged or is quoted, the reference must be included inline. Otherwise, someone could pile a bunch of links to generic sources at the bottom, and say when challenged, "It's in there!". The crucial point of WP:BURDEN is that if you want to add or restore material which is challenged, the burden is on you to provide the source and clearly prove that it supports the material in question. Including it inline, next to the specific challenged (or quoted) material, is part of that proof. Crum375 (talk) 00:36, 3 June 2010 (UTC)

• I see SlimVirgin's reverted on the basis of a "widespread consensus", and await a link to that discussion with interest.—S Marshall T/C 01:22, 3 June 2010 (UTC)

Yes, of course they must be inline, exactly as Crum375 states. Jayjg ^(talk) 01:48, 3 June 2010 (UTC)

• And I see Crum375's reverted again. Still no link to the discussion supporting that consensus has appeared. Should I take it that the current version isn't based on a discussion at all, but on WP:SILENCE?—S Marshall T/C 12:32, 3 June 2010 (UTC)

• Generally, it is better not to use absolutist language. Here, although SV's and Crum's preferred language feels absolute, there is wiggle room in that I think it is assumed that challenges have to be credible, not facetious, and Wikipedia:Common sense should be exercised.
• If inline citations are only preferable, what are the acceptable alternatives; I think one or two would have to be explicitly suggested. What is the downside of always requiring an inline citation for a contentious or surprising statement? --SmokeyJoe (talk) 12:45, 3 June 2010 (UTC)

We have required inline citations for several years now... it was a consensus that developed gradually on multiple policy and guideline pages, and not just here. This means that there is no single discussion we can point to. You would have to look through the archives of all of these pages (and some discussions at the pump) to find where and how the consensus arose. If you think the current consensus has changed (or never really existed in the first place), feel free to run an RfC ... I am confident that doing so would establish that there is still consensus. Blueboar (talk) 12:48, 3 June 2010 (UTC)

• I would characterise most of these responses as repeated assertions without evidence, combined with misguided attempts to place the onus on me to find the relevant discussion, when the burden of proof properly lies with the other side. But SmokeyJoe's remark is one I can engage with. I think that there is and should be wiggle room on a commonsense basis, mainly because the current wording gives too much licence for vexatious requests from tendentious editors.—S Marshall T/C 12:58, 3 June 2010 (UTC)

I would say the same about your comments. And it is pointless to engage in a mutual "you don't have consensus for that" battle. Several editors have told you that requiring inline citations reflects consensus, and has for quite a while. Don't believe us?... fine... run an RfC and determine what the current consensus actually is. Blueboar (talk) 13:24, 3 June 2010 (UTC)

Going back to Rossami: Are there any situations where in-line citations don't work? Are there examples where something else is working better? --SmokeyJoe (talk) 13:32, 3 June 2010 (UTC)

• Blueboar, if you say there's consensus for the current version, then why is it so unreasonable of me to ask you for a link to the discussion where that consensus was established? And if five other editors say the same thing, but none of them can link to such a discussion, why is it unreasonable for me to assume that there is no such discussion? This is a sort of meta-version of WP:BURDEN, if you like:- if you say there's consensus then the onus is on you to prove it.

  What we do if there's no consensus is separate matter. At the moment we're establishing whether the current version of this page was discussed at all.—S Marshall T/C 13:54, 3 June 2010 (UTC)

This discussion itself is one example of the consensus, with which I agree. You don't need to look any further. --SteveMcCluskey (talk) 15:57, 3 June 2010 (UTC)

The reason why it is unreasonable to ask us to point to a discussion is that no single "discussion" took place for us to point to... as I said, it was a consensus that developed gradually, over a broad period of time, and involved multiple discussions that took place on multiple policy and guideline pages. Consensus is not always reflected by !votes and polls that we can point to. Sometimes consensus develops incrementally, through a progression of small changes.

But ultimately, it does not matter how a past consensus developed (or even whether it did develop). Since consensus can and does change, what matters is whether there is currently a consensus or not. So, let's find out. Run an RfC, and see what the current consensus on inline citations is. Then there will be a single defined thread we can point at to say: "OK, as of this point in time, the consensus is/is not to require inline citations". Blueboar (talk) 16:36, 3 June 2010 (UTC)

I don't think we need an RfC. Let's just demonstrate the consensus (or lack) here. Rossami (talk)

Crum375 clearly articulates the argument for absolutist language in his/her response above. While very well stated, I'm not buying it. To require in-line citations in the absense of an explicitly disputed statement is to violate the spirit of WP:AGF. It also imposes an unnecessary burden on new participants, raising the threshold of knowledge that a new user must master before being a full participant in the project. From that regard, the absolutist language comes across as elitist and even anti-wiki. I completely support the in-line style choice as the strong preference but do not agree that the wording should be absolute. Rossami (talk) (splitting the comment to allow separate replies)

I also think that there are some articles where the in-line citation style becomes a distraction to the reader rather than a help. The thread above about citations for "common knowledge" stem-articles might be one example. Frankly, I'm still struggling with the proposition that any citation is an actual improvement to the reader in such situations but assuming for the moment that they are, you have the mechanical question of implementation. How, for example, would you incorporate the requested citations into the pulley article in a way that actually improves the article? Are you going to cite the same (randomly chosen) textbook each time, changing only the pagenumber? That's hardly efficient or helpful. Pulley is a synopsis article, not a synthesis article. Assuming there is a valid need for any sourcing, there is certainly no need for multiple sources. A single cite with a chapter reference would cover the entire page. Which sentence will you hang that cite on? Whatever you pick falsely implies that the cite does not apply to the rest of the page. And why would you bother when a single end-note achieves the same end while more clearly implying that it applies to the entire article?
Maybe I'm wrong. That's why I posed the question above. But I've been working this question off and on for about three years on the pulley article without success. Someone please demonstrate for me by using in-line cites to make that article better from the reader's perspective. Rossami (talk) 17:20, 3 June 2010 (UTC)

There's no single discussion anywhere agreeing that words should be spelled correctly, but it's nevertheless the case that if you add poor spelling to a page, someone is likely to fix it. There's been similar agreement for several years that anything challenged or likely to be challenged, including quotations, needs an inline citation, to the point nowadays where it goes without saying among experienced editors.

On this page, we have S Marshall wanting to remove the requirement for inline citations, while we have Gavin collins coming from the opposite perspective, wanting to insist that every single tiny point be sourced. Both are saying there may be consensus for their positions. Both need to draft clear proposals and put those proposals to the community. I can almost guarantee both will fail, because there has been strong consensus for years for the middle ground. SlimVirgin ^{talk contribs} 17:51, 3 June 2010 (UTC)

I want to say that correct spelling is an excellent example of this point. WhatamIdoing (talk) 22:17, 3 June 2010 (UTC)

What I think I'm coming to is a multiple suggestion designed to reduce the amount of time content authors must spend defending material they've already written to allow them to write more content.

1) A clarification that WP:V is a licence for good faith requests for proof, but not a licence for frivolous, vexatious or tendentious demands. In cases where a neutral, third party editor feels a request is frivolous or vexatious, the third party should legitimately be able to reinstate the material in question or remove a {{fact}} tag.

2) A rule that for mathematically deducible statements, a proof is as good as a citation, provided the proof can be derived using high school mathematics. (Possible exception: in technical articles within the scope of mathematics, formal logic, or hard science, undergraduate-level proofs might be acceptable.)

3) A rule that where evidence is unusually difficult to supply—such as a foreign-language source, where a demand for evidence would necessitate a long translation on the talk page—there's also an onus on the challenger to seek help in resolving the dispute by, for example, asking an editor who speaks the language to confirm what the source says.

4) A rule that until there are editing tools that let inexperienced or un-technically-minded editors format inline citations more easily, sources may be cited in any way that would be acceptable in an academic publication.—S Marshall T/C 19:01, 3 June 2010 (UTC)

• The policy already says don't be too quick to remove uncited material. But bear in mind that it's usually faster to add a reference than to engage in protracted discussion about it—though of course WP:DICK always applies.
• With non-English material, as with everything else, the burden falls on the person wishing to add or restore the material. I don't think we'll ever gain consensus to change that.
• Inline citations are very easy to add—e.g.<ref>Smith, John. ''Name of Book''. Routledge, 2010, p. 1.</ref>—so the technical argument isn't one that will gain any traction, in my view.

SlimVirgin ^{talk contribs} 19:13, 3 June 2010 (UTC)

Inline citations are even easier than that in some articles: You can type (Smith 2010), exactly like you would have done thirty years ago, using a typewriter. "Inline" is not a code word for <ref> tags (clickable footnotes). Any system that directly associates the ref with the text complies with this policy. WhatamIdoing (talk) 22:17, 3 June 2010 (UTC)

I think that the statement (which I agree with) "Any system that directly associates the ref with the text complies with this policy." needs to be very clear within policy. I have on more than one occasion been told that I was using the wrong citation format. I know there is a guideline that says stay consistent in citation styles, but I have seen editors come through and change all of the citations to the format they prefered and then argue that I was not using the right format. 66.102.198.60 (talk) 00:54, 4 June 2010 (UTC)

Well, you should try not to mix different citation formats in an article, because that can confuse the reader. Jayjg ^(talk) 03:42, 4 June 2010 (UTC)

Break

The argument appears to be about this footnote, which had been part of the policy for a long time, several years, but was removed in this edit about 6 months ago, without discussion. I think the removal was a very bad idea which makes this policy too rigid, and would have reverted it immediately had I seen it. As an example of a bad effect, I (and another editor) argued some time ago with a very well respected editor, a steward, at the time of the mass BLP deletions. Based on the inflexible current language, he felt it was OK to delete obviously notable stub BLP's that did not have inline citations, even though they had external links under references, which supported the scant one or two sentences of the article. Making them into inline refs was a purely mechanical, mindless process. We should not privilege form over content this way, and WP:V should stick to its original purpose of helping people understand principles of verifiability rather than enforce a particular style.

Again, since the footnote was boldly removed without discussion, I think there is a good argument that the removal is a change which needs to attain consensus, not vice versa, so I will restore it. Six months is on the long side, but I've seen restoration of unnoticed changes reverted here and accepted after about that long. (I intended to do so a few months ago but never got around to it.)John Z (talk) 01:10, 5 June 2010 (UTC)

John and S Marshall if you require a consensus then take the consensus reached here by Jay, Slim, Blueboar, Crum, and myself. This is ridiculous that long-standing changes are called "unsupported" because you personally disagree with it. If multiple editors say there's been a consensus your answer shouldnt be "show me the consensus" because the very fact that multiple editors are telling you the same thing shows a consensus. Plus the fact that the inline citation requirement is common sense around here already encoded in our way of doing things, policies normally lag behind common procedure, it is time to make sure that the policy matches reality. Which was done six months ago, and with the number of editors that watchlist this page the idea that it would have been reverted back then if you and others had noticed it is a big stretch. WP:SILENCE isnt always right but it does have a kernel of truth to it when its such a commonly watched page and its been six months.Camelbinky (talk) 01:24, 5 June 2010 (UTC)

Wait a minute! You can't allege consensus by cherry-picking just the people who agree with you and ignoring the people who disagree. The current discussion here is robust and I hope will reach consensus soon but there clearly is no consensus yet. And to rebut Blueboar's claims above, I have now read through hundreds of archived discussion pages all across the project. While I have found an absolute consensus in favor of citations and a very strong preference for the in-line style, I have yet to find anything approaching consensus that they must be inline. Rossami (talk) 03:16, 5 June 2010 (UTC)

That we require inline citations was added to this policy here without qualification in July 2007. It was added to reflect WP:CITE and the widespread use of inline citations by all experienced editors, as well as at GA and FA. This was bottom-up evolution, not something imposed by a small number of editors. SlimVirgin ^{talk contribs} 01:55, 5 June 2010 (UTC)

Yes, and the footnote was added the next day here, quoting the edit summary "attempt at compromise sufficient to meet acceptable contrary viewpoints and remove the "dubious" tagging [on the inline citation rule]" In-line citation without the footnote did not attain consensus in July 2007. Inline citation with the footnote did.John Z (talk) 02:19, 5 June 2010 (UTC)

If you read the foonote it doesn't say inline citations aren't required:

When content in Wikipedia requires direct substantiation, the established convention is to provide an inline citation to the supporting references. The rationale is this provides the most direct means to verify whether the content is consistent with the references. Alternative conventions exist, but inline citations are considered "best practices" under this rationale. For more details, please consult Wikipedia:Citing_sources#How_to_cite_sources.

What "alternative convention" would provide direct substantation? SlimVirgin ^{talk contribs} 02:27, 5 June 2010 (UTC)

I agree, "Alternative conventions exist" are weasel words, and for it to be considered for inclusion, such conventions much be referenced. Wikipedia:Citing_sources#How_to_cite_sources doesn't clarify anything about this for me. --SmokeyJoe (talk) 02:34, 5 June 2010 (UTC)

The July 2007 addition about in-line citations says "Material ... should be cited using an inline citation" (emphasis added). Even without the added footnote, that is nothing like the current version which explicitly requires cites to be in-line.
Alternative conventions at the time included citation in the edit summary, documentation of sources on the Talk page, documentation at the bottom of the page and manual footnotes among others. Some of those are now deprecated but the currently acceptable alternatives are clearly listed in Wikipedia:Citing sources#How to present citations (so yes, the section-link needed updating). Rossami (talk) 03:16, 5 June 2010 (UTC)

Rossami, can you give me an example of a method of citation that isn't inline, but that nevertheless makes clear which source supports which part of the text? I am worried here that we are talking about non-existent alternatives. SlimVirgin ^{talk contribs} 03:19, 5 June 2010 (UTC)

Slim, I find the statement "If you read the foonote it doesn't say inline citations aren't required" incomprehensible. It very clearly says inline citations aren't absolutely required, in so many words. Everybody else who has commented agrees, and that is how it was understood (including, it seems, by you) when it was written. I implicitly gave one example of an alternative convention above - general references at the end of very short articles. The original discussion that led to the adoption of in-line citation plus the imho essential footnote is at Wikipedia talk:Verifiability/Archive 21 and has more, like prose attributions. Experience is the best teacher. The footnote was there for the great majority of the time we have required inline citations, and may have been necessary for the adoption of this requirement. What harm did it do? What good did removing it do? I pointed above to a clear, concrete case where removing the footnote damaged the encyclopedia. I think the case for restoration in some form is very strong.John Z (talk) 03:43, 5 June 2010 (UTC)

(edit conflict) A general reference might be part of the way around the problem we're having on pulley where no definitive source exists and where section-specific sources are unnecessary because the entire article can be cleanly referenced as a unit. (If nothing else, I'm glad this discussion has led me to that page.)
Parenthetical referencing is less intimidating and perhaps more comfortable for some new editors. And in very long articles with a more academic tone, parenthetical referencing might even improve readability since the source is immediately visible rather than having to jump down the page then try to find your way back up. Now, you might consider that and embedded links to be "in-line citations" but all our help documentation treats them differently from the preferred <ref>-based style.
The point is that it's not always necessary to make clear connections to a specific part of the text. Sometimes a more general reference can be clearer by telling the reader that the reference applies to most of the article and that our article is a synopsis of it. Rossami (talk) 03:56, 5 June 2010 (UTC)

Looking at Wikipedia:Citing_sources#How_to_present_citations, points 2 to 5 sound like in-line citations. #5, mid-text off-site links, I think is a issue better covered by some other guidance. I think the point here is that #1, "General reference", is not good enough where there is a credible and substantive challenge to the material. If we are talking about "must" vs "should", then if you advocate "should", then are you saying that there can be a case where a credible and substantive challenge to a particular sentence is not best answered with an in-line citation? --SmokeyJoe (talk) 04:38, 5 June 2010 (UTC)

Rossami, a parenthetical reference is an example of an inline citation. A general reference (e.g. the name of a book at the bottom of the page) is useless for showing which part of the text is sourced to what, so that's not an option. SlimVirgin ^{talk contribs} 05:35, 5 June 2010 (UTC)

General refs can be links to webpages, or short papers, and provide sufficient specificity; if the article text is short enough it can be perfectly clear what part is being sourced. I recall an academic AfD a couple years ago where a substantial numbers of sitting Arbiters weighed in against demanding excessive formalism in referencing, (maybe even citing the footnote) I'll try to dig it up.John Z (talk) 07:13, 5 June 2010 (UTC)

SV is absolutely right that parenthetical references are inline. So are embedded citations.

Do editors think that specifying that ref tags aren't the sole method of producing an inline citation would help address the fundamental issues? I'm thinking of something like "...inline citations (such as parenthetical references or footnotes)..."

Also, looking at Pulley, I'd like to say that {{Refimprove}} does not actually constitute a challenge of any particular information, and therefore general references are an acceptable response to that tag (NB that I'm not actually recommending them; I'm just telling you that they are permitted). WhatamIdoing (talk) 16:53, 5 June 2010 (UTC)

Agree. (1) SV is right. (2) the footnote instead referring to different methods of in-line citations might be to the solution to this under-defined dispute; (3) There is no challenge to anything specific in Pulley, and in-line citations are not necessarily the way to improve the article. Perhaps the whole article needs to be restructured to follow the structure some comprehensive secondary sources. --SmokeyJoe (talk) 23:39, 5 June 2010 (UTC)

Consensus

I don't know any experienced editor who doesn't use and support the use of inline citations, and they've been in widespread use since around 2005. If editors here seriously want to question whether they're required for anything challenged or likely to be challenged (incl quotations), you're going to need a strong and clear wiki-wide consensus to change it. So please post on the pump and elsewhere, or post an RfC, to try to obtain that. Two or three editors trying to change the wording of this policy isn't appropriate in the face of such widespread acceptance. SlimVirgin ^{talk contribs} 05:43, 5 June 2010 (UTC)

There were sufficiently many experienced editors that opposed something like the current text so that the footnote was inserted back in July 2007, immediately after the inline citation requirement was added. And that is what is wanted - just a footnote! And now merely a discussion pointer - surely that is not too much to ask for temporarily? I think that there are a great many editors who support the general philosophy of the pro-footnoters back in Archive 21, who saw this, not even stated as rigidly as it has been for the last few months, as representative of a dangerous tendency in policy-writing, of "'strictly construed edicts' making their way into policy".

There was no consensus or discussion when it was removed in back in November 2009. It is not unreasonable to say that that is the change that needs consensus. Would just saying "consensus", against explicitly cited discussions, have been enough back in November? Why should it be now? In any case, a few anti-footnoters (who I hope to convert) vs a few pro-footnoters, are hardly a consensus. If the wikiwide consensus for the precise wording of the anti-footnoters is/was so strong why wasn't it strong enough in July 2007?John Z (talk) 07:37, 5 June 2010 (UTC)

Policy must reflect what is actually done in Wikipedia. John you are assuming that policy dictates by fiat what is done on Wikipedia, it is the reverse, the community has already decided multiple times that inline citation is required for a professional looking article, policy must now be changed. It already IS standard practice that if a fact or sentence is questioned it is removed if there is no inline-citation, you must have an inline-citation. If policy is erroneously stating otherwise then policy is what must change. Wikipedians change policy by doing what is best for the encyclopedia and THEN the policy is changed to reflect what is done, policy lags behind common practice. John Z, THAT is why you may be confused regarding what happened 3 years ago and what happened six months ago. Bring this to the village pump, or better yet take an article with no inline citations and nominate it for FA or even just GA and see the responses you get. Inline citations are a must, alternatives, if they exist, are simply not seen as professional for Wikipedia.Camelbinky (talk) 11:23, 5 June 2010 (UTC)

Going back and forth saying "there was no consensus for this" and "yes there was consensus" is pointless and disruptive. It does not matter whether a change that occurred back in 2007, or even in 2009, reflected consensus or not... since consensus can change, the only thing that matters is whether the language reflects consensus now ... and the only way to determine that is inquire beyond this page... to ask at the pump, or to run an RfC. Blueboar (talk) 13:14, 5 June 2010 (UTC)

I'm afraid I simply don't accept that four or five people supporting a stealth edit to a policy, which was undiscussed and arguably misdescribed in the edit summary, can now pretend they have a consensus that it requires an RfC to overturn.—S Marshall T/C 15:02, 5 June 2010 (UTC)

And I can not accept one or two people challenging long standing policy by claiming there was no consensus. So... we have two choices ... either we continue to edit war over it (in which case the "inline citations are required" viewpoint "wins", simply because there are more of us and we can out revert you until you get banned for 3rrr)... or (by far the preferred method) we run an RfC and settle the matter. Your choice. Blueboar (talk) 15:12, 5 June 2010 (UTC)

Um, no. Continuing to edit war on a core policy page, whether or not 3RR is technically violated, is not a good option. I've protected the wrong version for 24 hours, presuming that you can all come to some sort of consensus (even if it's just consensus on an RFC wording) within the next 24 hours. Cheers, Jclemens (talk) 15:52, 5 June 2010 (UTC)

Thanks J... now that edit warring is off the table (darn... I love a good edit war!)... I guess that leaves running an RfC as the only way to settle this. For the language I suggest something short, simple and to the point... like "Are inline citations required when material is challenged for lack of sourcing?" I am happy to post it if no one else will. Blueboar (talk) 16:01, 5 June 2010 (UTC)

Slap a "disputed" notice on it, would you, Jclemens? That doesn't seem unreasonable when we've got three editors in good standing who don't like the version you've protected.

(interrupted comment) Except when six or seven editors do like the version he's protected. Please note that Jclemens did say (and I quote) "I've protected the wrong version for 24 hours," If you look behind the humor, this means he is fully aware that when locking a page to end an edit war, no matter which version it gets locked at, the other side will consider it to be the "wrong" version. Since the {{disputed}} tag is part of what we were editing warring over in the first place... you will just have to live with it being the "wrong version" for a while (The same would have applied to me if he had locked it with the {{disputed}} tag in place). Blueboar (talk) 19:05, 5 June 2010 (UTC)

suggested RfC language

(resumed comment) Given that we're decided to go the RfC route, I think the question would be better phrased like this:

Should WP:V read as:

a) The November 2009 version as last edited by User:Fences and windows, containing a nuanced discussion of why inline citations are best practice and the recommended way of dealing with disputed content; or

b) The current version as substantially revised by User:Slimvirgin, containing a flat and unqualified rule that inline citations are the only way of dealing with disputed content.

Happy with that?—S Marshall T/C 16:37, 5 June 2010 (UTC)

How about this: "When a source is provided for material which is being challenged, is an inline citation required or optional?" Crum375 (talk) 16:52, 5 June 2010 (UTC)

Crum's version; S Marshall's version is non-neutral, and poisons the well by calling one of the versions "nuanced" and the other a "unqualified rule".  --Joshua Scott (LiberalFascist) 17:36, 5 June 2010 (UTC)

Yes, if this is the only point at issue, then we don't need to talk about whole versions. But the question could be a bit more intelligent than that - I can think of at least two reasons why an inline citation might not be required or even desirable (1. the challenge is disruptive/bad-faithed; 2. the material is part of the lead or a summary-style paragraph and citations appear elsewhere), and there may be others, even though I generally support the idea that our style is to show verifiability by means of inline citations.--Kotniski (talk) 17:51, 5 June 2010 (UTC)

Agreed that we don't want to cover the article in inline citations, but after reading this whole discussion, I have yet to see a concrete suggestion that would replace inline citation for challenged statements. As for the disruptive/bad-faith requests: If a request is bad-faith, then arguing will accomplish nothing, since the challenge was designed to cause contention; arguing plays right into the troll's plan. Simply put: just cite the damn thing, it's faster than arguing.  --Joshua Scott (LiberalFascist) 18:09, 5 June 2010 (UTC)

OK... how about:

From 2007 to 13 Nov. 2009 WP:V#Burden of evidence read as follows:

• The burden of evidence lies with the editor who adds or restores material. All quotations and any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation.

followed by a footnote that read:

• When content in Wikipedia requires direct substantiation, the established convention is to provide an inline citation to the supporting references. The rationale is that this provides the most direct means to verify whether the content is consistent with the references. Alternative conventions exist, and are acceptable if they provide clear and precise attribution for the article's assertions, but inline citations are considered 'best practice' under this rationale. For more details, please consult Wikipedia:Citing sources#How to cite sources.

On 14 Nov. 2009 this footnote was removed (with this edit). The policy has stood without the footnote since then. An editor now challenges the removal. Other say it is better without the footnote. Both sides of the debate claim that their preferred versions reflect consensus. Greater community input is needed to settle the dispute and determine current consensus... Do you prefer the statement with the footnote or without the footnote?

This explains what occurred and when, and asks which version has consensus without any spin. Blueboar (talk) 19:36, 5 June 2010 (UTC)

I'm fine with that, but would like to tighten it a little, and remove "claim". I'd like to suggest:

---

From July 2007 to November 2009 Wikipedia:Verifiability#Burden of evidence read as follows:

• The burden of evidence lies with the editor who adds or restores material. All quotations and any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation.

followed by a footnote that read:

• When content in Wikipedia requires direct substantiation, the established convention is to provide an inline citation to the supporting references. The rationale is that this provides the most direct means to verify whether the content is consistent with the references. Alternative conventions exist, and are acceptable if they provide clear and precise attribution for the article's assertions, but inline citations are considered 'best practice' under this rationale. For more details, please consult Wikipedia:Citing sources#How to cite sources.

On 14 November 2009 the footnote was removed with this edit. The policy has stood without the footnote since then. Three editors now challenge the removal. Others say it is better without the footnote. Greater community input is needed to determine consensus. Do you prefer the statement with the footnote or without the footnote?

---

SlimVirgin ^{talk contribs} 19:50, 5 June 2010 (UTC)

Works for me Blueboar (talk) 19:57, 5 June 2010 (UTC)

However you word it, were you to ask me that question (required or optional), my answer (and I suspect most people's answer) would be very simple: required. And having just spent time (before the newest comments above were added) finding and considering the November 2009 version you refer to under (a) at the top, I am none the wiser about the point. That version too states, "All quotations and any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation." That statement has the footnote you now cite, which turns out to qualify it ("Alternative conventions exist, and are acceptable if they provide clear and precise attribution for the article's assertions"); following the daisy-chain further from the footnote, Wikipedia:Citing sources#How to cite sources appears not to exist, but another section, Wikipedia:Citing_sources#Inline_citations, says, "In most cases, an inline citation is required, either in addition to, or instead of, a full citation in the References section, depending on which citation method is being used (see below). Inline citations show which specific part of the article a citation is being applied to. They are required by Wikipedia's verifiability policy for statements that are challenged or likely to be challenged, including contentious material about living persons, and for all direct quotations. Inline citations are also mandated by Wikipedia's featured article criteria, where appropriate. An inline citation should appear next to the material it supports."

The point that I'm making is: if we (those who will consider your RfC if and when you raise it) are to have any chance of being persuaded that inline citations should be made optional, we will need to be shown an example where that is effective (i.e., provides "clear and precise attribution for the article's assertions"). Currently I am not able to find such from your suggested RfC wording. PL290 (talk) 20:28, 5 June 2010 (UTC)

RfC language isn't supposed to persuade... an RfC is supposed to present the issue neutrally and invite comments. The place for persuasion is in the subsequent comments. Blueboar (talk) 20:41, 5 June 2010 (UTC)

Who said anything about RfC language?!!! If an example exists demonstrating that an alternative to inline citations can be effective, but you don't make that clear in your RfC, then you haven't presented the issue neutrally. As I said, however you word it, that is the problem. (But of course if no such example exists, there's probably no point raising the RfC.) PL290 (talk) 20:59, 5 June 2010 (UTC)

I think you may be responding to the RfC before we have even posted it. Essentially all we are asking is the simple question: should the footnote should be returned to the policy or not? Any examples and persuasion have to come after we ask that basic question, as part of the comments we are requesting. Blueboar (talk) 21:23, 5 June 2010 (UTC)

I agree that we need to see an example of an alternative. No one has offered one yet that I can see. It would be good to give such an example in the RfC so that people know what they're commenting on. SlimVirgin ^{talk contribs} 21:07, 5 June 2010 (UTC)

Another request for an example

Would S Marshall, Rossami, or John Z please give an example of an alternative to inline citations? The footnote you want to restore says:

When content in Wikipedia requires direct substantiation, the established convention is to provide an inline citation to the supporting references. The rationale is that this provides the most direct means to verify whether the content is consistent with the references. Alternative conventions exist, and are acceptable if they provide clear and precise attribution for the article's assertions, but inline citations are considered 'best practice' under this rationale (my bold).

What are the "alternative conventions" that are acceptable "if they provide clear and precise attribution"? There is no point going through this discussion or a wider RfC if we don't know what the alternatives are. SlimVirgin ^{talk contribs} 21:13, 5 June 2010 (UTC)

The only alternative that comes to mind is one that I personally feel is a bit dicey. It's pretty common to see "[[wikilinked topic]] is an x" without an in-line citation, where the assertion is clearly supported by the wikilinked article. Probably the most common place I see it is in musical single infoboxes, where the previous and successor singles are indicated and wikilinked. Most people would consider it pedantic to require an in-line citation in the infobox if clicking the wikilinks takes you to articles that correctly and completely support the assertion that the singles in question are, in fact, the predecessor and successor singles.—Kww(talk) 21:24, 5 June 2010 (UTC)

I agree that it's dicey. I've never seen a discussion in which cross-wiki citation gained consensus (citations in one article sufficing in another). SlimVirgin ^{talk contribs} 21:29, 5 June 2010 (UTC)

True, but, as a hypothetical, what do you think would happen to an editor that began placing notices on every single and album infobox in Wikipedia for the "previous single" and "next single" fields? I suspect a consensus would rapidly develop that the editor was being disruptive, even if he was being technically correct.—Kww(talk) 21:35, 5 June 2010 (UTC)

It's true that the challenges must be reasonable, and that common sense and WP:DICK always apply (to this and all other policies). My concern throughout this discussion is that S Marshall, Rossami, and John Z are pushing for alternatives to be allowed, but without spelling out what those alternatives are. SlimVirgin ^{talk contribs} 21:42, 5 June 2010 (UTC)

Agree again with SV, but further, making repeated unreasonable challenges goes beyond WP:DICK to WP:POINT (disrupting editors, forcing excessive referencing) which leads to WP:BOCK.

Kww's example of a the cross-article link is not a citation, although it is frequently used as a pseudo-citation (especially on talk pages). Cross-article links are in-line "further reading" pointers. If a statement is reasonably challenged, it needs to be cited there, immediately, on that page. Wikipedia articles are supposed to stand as complete, stand-alone publications, and referencing to another page breaks the wise "The topic of Wikipedia articles should always look outward, not inward at the Wikipedia itself" creed. --SmokeyJoe (talk) 23:52, 5 June 2010 (UTC)

I would say that there is a clear consensus that when challenged, a wikilink to an article where the information is cited is not enough... when challenged in one article, the information needs to be cited in that article. I have seen this issue pop up in list articles especially, and every time the consensus has been to include a citation in the list article (all it usually takes is a quick cut and paste of the citation from the main article). Are there other alternatives besides this? Blueboar (talk) 22:02, 5 June 2010 (UTC)

Speaking for myself rather than for Rossami or John Z, I deliberately don't want to get tied down into detailed, creepy iterations of what the alternatives are because then we'll create a situation where those and only those alternatives are acceptable.

What WP:V ought to do is to ensure the encyclopaedia contains only verifiable content. In almost all cases, and certainly in 99.9999999% of cases involving a dispute between good faith editors who're acting in a collegial manner, the best way to achieve that will be through an inline citation to a reliable source. WP:V ought to recommend inline citations as best practice and the optimal way to resolve a content dispute of any kind.

But not all editors are in good faith, and at the moment, the policy as written is too favourable towards tendentious editors who seek to use it vexatiously. There are good reasons not to give examples of how the current wording might be misused, but I think you're all intelligent enough to see ways a bad faith editor could employ it. If I were a tendentious editor who wanted to annoy one of you, I could use it to destroy a substantial proportion of your article contributions in one weekend, and that's not okay.

Contrary to some of the criticisms above, I've actually given an example of other ways in which information might be verifiable, without being sourced. Later on this page, there's a long discussion about mathematically deducible statements. (If I can source that a famous person was born on 23 April 1564 and died on 23 April 1616, then I shouldn't need a source to say that he died on his 52nd birthday.)

In short, WP:V shouldn't say baldly that inline citations are the only way. It should explain why they're best practice, and why they're the strongly-preferred option.

If we reduce all this discussion to a one-line question that boils down to, "Ought people to be able to get away without using inline citations?" then the answer's certainly a foregone conclusion, but it's also a travesty of the much subtler positions Rossami, John Z and I have taken.—S Marshall T/C 22:48, 5 June 2010 (UTC)

You must offer an alternative if you're saying we should modify the policy to refer to alternatives. Otherwise you're causing a lot of trouble here in the name of something non-existent. SlimVirgin ^{talk contribs} 22:59, 5 June 2010 (UTC)

Very well. My example alternative is a formal mathematical proof using high school arithmetic.—S Marshall T/C 23:11, 5 June 2010 (UTC)

A formal mathematical proof would not fall under WP:CALC and would therefore constitute original research unless a reliable source is supplied. Crum375 (talk) 23:15, 5 June 2010 (UTC)

S Marshall, that's not what's being asked. I'm asking for an alternative method of citation, as the footnote you want to restore refers to. I'm not familiar with an "alternative convention ... [that] provide[s] clear and precise attribution for the article's assertions," as the footnote you want to restore says. If you know of one, please give an example. Otherwise you are just wasting time. SlimVirgin ^{talk contribs} 23:22, 5 June 2010 (UTC)

And if you are referring to an established and well known proof, as discussed under "Blueboar's Law" in the thread below, then you are referring to a "Paris is the capital of France" type situation where you do not need to provide a citation at all (because the information is such common knowledge, that could easily be cited to any basic high school math book, that it is unreasonable to challenge it), not an example of an alternative method of citation. Blueboar (talk) 00:12, 6 June 2010 (UTC)

Are we talking at complete cross-purposes here? I thought I'd been challenged to provide an example of a means of verifying something that doesn't involve an inline citation to a reliable source. Have I not done so?—S Marshall T/C 00:29, 6 June 2010 (UTC)

People are being very patient here. You want to restore an old footnote to a policy, a footnote that was removed because it made no sense. It said: "Alternative conventions [alternative to inline citations] exist, and are acceptable if they provide clear and precise attribution for the article's assertions ..." You are being asked for an example of an alternative citation method that provides clear and precise attribution, because we can't think of one, and that is why the footnote was removed. Therefore, please provide an example. SlimVirgin ^{talk contribs} 00:35, 6 June 2010 (UTC)

The foregoing concerns itself only with "ways in which information might be verifiable, without being sourced". Sorry, but that misses the point. The footnote you are asking for relates to content that "requires direct substantiation". It says that as well as inline citations as a means to verify whether the content is consistent with the references, "Alternative conventions exist, and are acceptable if they provide clear and precise attribution for the article's assertions". That assertion is being challenged. Unless alternative conventions can now be shown to exist, and to "provide clear and precise attribution for the article's assertions", then the question becomes academic and, frankly, a waste of time.PL290 (talk) 07:16, 6 June 2010 (UTC)

If the challenged material is a tiny stub, then a general reference might be sufficiently specific. Imagine an article whose sole (current) content says "_____ is a type of cancer." I don't believe any reader would have trouble figuring out what content the source supported/what source supported the sentence. WhatamIdoing (talk) 01:38, 6 June 2010 (UTC)

Right, but that's because the article is so under-developed that a general reference is right next to the claim that might be contested. We're talking here about regular articles with several points that might be contested. What kind of citation system would work in those cases, if not inline citations? The three editors wanting to change the policy must have something specific in mind; otherwise why bother? SlimVirgin ^{talk contribs} 01:48, 6 June 2010 (UTC)

I suppose the obvious example is in-text attribution in lieu of inline citation. It's a perfectly valid way of enabling the reader to verify an assertion. For articles about a published work (say, Macbeth) that make many references to the subject work it is frequently convenient to directly refer to the relevant part of the work rather than send the reader to footnotes best reserved for other sources:

The Porter's speech (Act II, scene III, lines 1–21), in particular, may contain allusions to the trial of the Jesuit Henry Garnet in spring, 1606; "equivocator" (line 8) may refer to Garnet's defence of "equivocation" [see: Doctrine of mental reservation], and "farmer" (4) to one of Garnet's aliases.[19]

LeadSongDog come howl! 02:47, 6 June 2010 (UTC)

In text attribution also includes an inline citation. Inline citation does not (necessarily) mean "footnote". Crum375 (talk) 02:53, 6 June 2010 (UTC)

Thanks, LSD. That's parenthetical referencing, which is included in what we mean by inline citation. SlimVirgin ^{talk contribs} 02:54, 6 June 2010 (UTC)

Just as we have certain scholarly standards that we (generally) hold our references to, we too must hold our own work up to high standards. Wikipedia has been around long enough, and most of us as individuals, that we've grown up and expect certain things to be professional. When I first started over three years ago (or four, I was an IP so I cant look it up right now) when I created my first article and I knew nothing about cite web or cite book or any other templates that could help with making things look "nice and neat" I simply put on the talk page "all info in this article is from..." Now, by the time I got the courage to create my second article I had done more exploring and work along more experienced editors at established articles and that time I created an article that quickly became a GA. So basically what Im saying is other than newbies we expect everyone to understand that we are professional, mature, and we have standards and practices that we expect to be common across Wikipedia. What S Marshall is supporting is dumbing down our standards and going to a previous way of doing things when Wikipedia itself was full of novices. Collectively we've become quite the experts on encyclopedia-building. If nothing else, this is a dispute over the direction of the encyclopedia-at-large; ie- professional and standardized or novice and amateurish.Camelbinky (talk) 03:15, 6 June 2010 (UTC)

Quoting from inline citation: "An inline citation is a citation next to the material it supports, rather than at the end of the article. Inline citations are used to directly associate a given claim with a specific source." In other words, when material is challenged, to provide the source "inline" simply means to place it near the challenged material, not at the bottom of the page. The actual citation format is flexible. Crum375 (talk) 03:10, 6 June 2010 (UTC)

As I added that particular text a few hours ago, I wouldn't count on the wording too much: someone else might find a better way of describing it. But I'm glad that you seem to have found it a useful description so far, and I hope that it will help editors figure out that ref tags are only one of several possible inline citation styles. WhatamIdoing (talk) 06:19, 6 June 2010 (UTC)

Citing sources says "Inline citations show which specific part of the article a citation is being applied to." That's the key. Crum375 (talk) 13:46, 6 June 2010 (UTC)

SV, I don't think that we should exclude the sub-stubs. There's absolutely nothing in BURDEN that limits the requirement for inline citations to "regular articles with several points that might be contested". If we're not going to state that particular limitation in the policy, then we shouldn't impose that particular limitation on potential alternatives. WhatamIdoing (talk) 06:19, 6 June 2010 (UTC)

WaId, As SV pointed out earlierelsewhere in this discussion:

"They [Citations] can be mid-sentence (not good form), at the end of the sentence (fine), or at the end of the paragraph (my preference), but there has to be text-source integrity of some kind. That's what's meant by inline citation.

Since a stub is usually just one paragraph (sometimes just one sentence) a general reference after that paragraph would be consistent with SV's description (which I agree with) of inline citations, thus there really is prectically no "exclusion" or exception for stubs. A good question to ask now is: Is there a policy or guideline that supports SV's description? Another good question is: Is there a policy or guideline that CLEARLY supports the various styles of inline citation formats including parenthetical references? 66.102.205.233 (talk) 02:09, 7 June 2010 (UTC)

Based on SV's response, she does not believe that a general reference at the end of a single-sentence (or single-paragraph) stub is an inline citation. I also don't believe that general references are inline citations, and I seriously doubt that any editor, seeing a ==References== heading between the body and the contested sentence, would actually claim that general references are inline citations.

As for pages clearly supporting different styles, see WP:CITE: "Editors are free to use any method; no method is preferred over another..." Parenthetical refs are clearly included in that (as is, BTW, any style that any editor invents), and they're specifically listed as an inline citation method. (They seem to be most popular in articles about non-science academic subjects, e.g., Irish phonology.) WhatamIdoing (talk) 03:10, 7 June 2010 (UTC)

You're quoting CITE selectively. It makes clear that inline citations are needed for anything challenged or likely to be challenged, and quotations, per V. This page is the policy. SlimVirgin ^{talk contribs}

I'm concerned that we may be getting into more detail now than is appropriate for this page. The point of this policy is that the information we present should be verifiable in reliable sources. Detail should focus on what it means to be "verifiable" and what it means to be "reliable". Matters of style (how and where we present citations) and procedure (how to deal with "challenges") shouldn't take over this core aspirational policy - they should be mentioned in passing, certainly, but this isn't the place to go into them in exhaustive detail. In particular I wouldn't want a situation to develop where compulsory inline citations become a kind of enforceable religious dogma - we know (or at least, those of us who haven't had our minds set through editing in contentious subject areas know) that too many [1]'s and [2]'s and [n]'s in the text make it harder to read, and where we can happily get away without them we should continue to do so.--Kotniski (talk) 09:22, 6 June 2010 (UTC)

As I pointed out above, the issue here is not citation format. The question is simple: when material is challenged, and a reliable source is supplied for it, do we place the reference near the challenged material it supports, or at the bottom of the page? This is the key issue that must be decided. Crum375 (talk) 11:57, 6 June 2010 (UTC)

Near the challenged material... that has been clear since 2007. Blueboar (talk) 12:08, 6 June 2010 (UTC)

Well, we place the (ref> tag near the material, and the reference itself appears at the bottom of the page. But this is very much a style matter, not a matter of fundamental importance that this policy needs to lay down in stone (and in practice there may well be exceptions).--Kotniski (talk) 18:43, 6 June 2010 (UTC)

K, the point is that refs need to be in the text, not at the end of the page, for anything challenged or likely to be challenged, and quotations. They can be Harvard refs (rarely used nowadays on WP) or footnotes. They can be mid-sentence (not good form), at the end of the sentence (fine), or at the end of the paragraph (my preference), but there has to be text-source integrity of some kind. That's what's meant by inline citation. We can no longer do what we did when I started editing in 2004, which was to have a several hundred or thousand word article followed by a list of references at the end, with no indication of which source supported which point.

This is an odd discussion to be having because there are no developed articles now that don't use them (or that aren't tagged as lacking them); there is no sourcing policy or guideline that doesn't say they're needed; and there is no FA, GA, or other peer review process that won't insist on them. Trying in 2010 to turn us back to the sourcing we used when the project first started isn't going to happen. SlimVirgin ^{talk contribs} 19:47, 6 June 2010 (UTC)

While the exact method of inline citation may vary, editors need to be able to easily find the source for specific claims in an article. That cannot be done effectively in any article that is more than a few words, unless some sort of inline citations are used. That is why all article quality-review process (FA, GA, etc.) insist on them. Jayjg ^(talk) 21:57, 6 June 2010 (UTC)

• Not sure where this is going or whether it is settled but agree with SV - except I think the standard is higher at peer-reviewed articles (ie. GA, A and FA), where pretty much all facts should be supported by in-line cites, not just ones "likely to be challenged". But that is another debate... hamiltonstone (talk) 01:25, 7 June 2010 (UTC)

Per Hamiltonstone, putting it next to the relevant snippet is the obvious thing YellowMonkey (vote in the Southern Stars and White Ferns supermodel photo poll) 01:39, 7 June 2010 (UTC)

• While I'm sure Hamiltonstone and YellowMonkey's contributions to this debate are in the best of faith, it might have been better to disclose the canvassing on the FAC talk page. A relevant grouping to consult, yes, but the only grouping to consult?

  I suspect any RfC is now bound to achieve the result one side desires, because that side is now in control of the question that's asked. I still don't agree with the current wording but I withdraw support for a RfC, since it's now pointless.—S Marshall T/C 11:19, 7 June 2010 (UTC)

Mathematically deducible statements

Is it necessary to provide a citation for something that can be deduced mathematically?

If it's trivial maths then I would think not; we wouldn't want to see "2+2=4" in an article. But I'm considering a more challenging example here, from Talk:HIP 56948. This article describes the star as "class G5V", and it was disputed whether that designation was correct. In practice, James McBride has come up with a source that says it's so, but I'd like to examine the hypothetical situation in which he hadn't. In that case the "G5V" classification would rely on a calculation of the star's absolute magnitude—and the maths is demonstrable, but non-trivial—followed by reading and understanding the Hertzsprung-Russell diagram.

My position is that when an intelligent non-expert could follow the steps and see it strictly proven that the description is accurate, then that ought to suffice for WP:V, but I'm uncomfortably conscious that there isn't a consensus supporting that, so I'd like to know whether editors here think I was wrong.—S Marshall T/C 00:31, 3 June 2010 (UTC)

• In discussion that follows, I envisage the possibility of a conversation with a difficult editor. I've realised that this could be taken the wrong way. The possibility of a difficult editor isn't connected with the example article I use; in other words, when I talk about "vexatious" and "tendentious" requests, I don't mean anything negative about JGHall, who's one of the most pleasant interlocutors I've found on Wikipedia talk pages.—S Marshall T/C 08:56, 3 June 2010 (UTC)

I think it depends on how trivial the computation is. If I find a source that says Utah and Iowa have something in common then it would be OK to summarize that with a comment about "two states". OTOH, a definition of obesity is a BMI of over 30. However calculating an individual's BMI and concluding that they are obese would probably be going too far, even if any reader could do the same calculation. .   Will Beback  talk  00:40, 3 June 2010 (UTC)

In general, we need a source to do the calculation or deduction for us. Otherwise we'd be flooded by editors making what to them seem "trivial" proofs for mathematical problems, and we are not staffed to reliably vet such proofs. Which means that we could have junk on articles because people would just assume it's correct. This is why WP:CALC only allows "routine calculations". Crum375 (talk) 00:48, 3 June 2010 (UTC)

• Well, I'm not keen on Will Beback's example because it introduces uncomfortable factors: there are differing definitions of obesity and "obese" is an emotionally-charged word that could have BLP implications. Here I'm talking about a matter that has no such differing definitions and no possible BLP subtext.

  To Crum375, I'd answer by analogy. I'm allowed to source an article to material in a foreign language, am I not? Or to material that's on my bookshelf but not yours?

  I think it's up to me to prove my case, but it's not up to me to make sure other editors can understand the proof.—S Marshall T/C 00:55, 3 June 2010 (UTC)

Any material which is challenged requires a reliable source. If the source is offline, you still need to provide all details necessary for other editors (or readers) to get it from their library. If it's in a foreign language, you need to provide a translation, per WP:NONENG. Crum375 (talk) 01:02, 3 June 2010 (UTC)

• Eek! That's a new rule (or at least, new since I last read WP:NONENG). I don't like that at all. It opens me up to all sorts of vexatious demands from tendentious editors, and it needs discussion. My position is that if there's a substantial volume of material in a foreign language, then the challenging editor needs to show that they've made reasonable efforts to verify the aspect they're challenging—such as asking a neutral editor who speaks the language for help—before they're allowed to demand that I type several hundred words of translation onto the talk page.—S Marshall T/C 01:22, 3 June 2010 (UTC)

The spirit and the letter of WP:BURDEN is that the burden of providing proof is on the person adding or restoring the material. If the source is in a foreign language, a valid "proof" requires a translation, since this is the English Wikipedia. The point of these rules is not to make life hard on editors adding the material or easy on the ones challenging it: the point is that we want Wikipedia readers to easily verify that anything we say can be directly attributed to a reliable source. We are not here to create new material, but to summarize what others have already published. Crum375 (talk) 01:40, 3 June 2010 (UTC)

The question is whether there should be a burden to provide the translation. We don't force the editor to put the book into the challenger's hands. We don't appear to force the editor who provides a library-based reference to type in the text. Should we force them to provide the translation? --John (User:Jwy/talk) 02:01, 3 June 2010 (UTC)

We "force" the editor adding the challenged material to provide detailed information about the book, including page number and quotes when required, so that other editors can judge on their own whether the what the book says actually supports the material. And that's the key: the person adding the challenged material must convince the others not just that a source exists, but that that source directly supports the material in question. For example, the source could support it only implicitly, if you read between the lines, but there is no way for us to decide that without seeing the actual quote and discussing it on the talk page. Similarly, if it's a translation, we need to see the translation so that editors can judge whether the source directly supports the material. Crum375 (talk) 02:10, 3 June 2010 (UTC)

IMO talking about "translations" is potentially confusing: Supporting a disputed translation of a direct quotation (e.g., Should we translate a German quotation containing "Torschlusspanik" as 'biological clock', or 'mid-life crisis' or something else?) requires a different response than supporting a standard statement with a source that happens to be in a non-English language. I assume that the issue here is only the second.

For example, this Spanish-language news source easily supports a claim that the unemployment rate recently dropped in the Charlotte region of North Carolina, and IMO an editor has done his duty by this policy if he has provided either a link to the article or the direct quotation, "En el área de Charlotte, Gastonia y Concord el desempleo disminuyó hasta ubicarse en 11.1%, comparado con el 12.1% de marzo, mientras que en todo el estado el índice fue de 10%." He shouldn't be required to provide a translation for monolingual editors, especially if it would require more than a sentence or two -- and if there's a dispute about whether the unemployment rate really did drop, then they probably shouldn't trust his translation, even if he offered. The editors who need the translation should be the ones who supply it, exactly like we would require them to get software necessary to read a pdf of a source, or to borrow a computer that can display Flash video, or otherwise supply whatever else they lack. WhatamIdoing (talk) 02:54, 3 June 2010 (UTC)

(e/c) The process you describe has some holes in it as far as verification is concerned. If the discussion is contentious, the level of good faith in the editor providing the information will likely be lower than one would like. At some point, a critical challenger will have to go to the book with an independent translator to truly verify. Providing the translation is a convenience, not a taut part of the verification process.

re last edit: I agree with the sentiment in general, but my first question would be "isn't there an English source for that?" --John (User:Jwy/talk) 02:59, 3 June 2010 (UTC)

• Going back to mathematically deducible things, my position is the same as with translations. I think it should be up to the challenging editor to show that they've made reasonable efforts to verify the mathematics involved—such as asking for help on the talk page of WikiProject Mathematics, say—before they're allowed to demand a step-by-step proof that they can follow on their pocket calculator (or maybe even a pencil and paper?) Because if we don't place any onus on the challenger at all, then the Wikipedia dispute process could well become a tyranny of the most ignorant, in which the least-educated editors "win" disputes where they don't understand the issues and those who do give up in defeat.—S Marshall T/C 07:05, 3 June 2010 (UTC)

No, trivial means very common place trivial even in maths articles. Dmcq (talk) 07:48, 3 June 2010 (UTC)

• Well, Wikipedia's maths articles seem to be broadly at undergraduate level, mathematically speaking, with a few of the more rarefied ones at postgraduate level. When you say "very common place trivial", do you mean undergraduate-level maths, or do you mean that only high school arithmetic's allowed?—S Marshall T/C 09:00, 3 June 2010 (UTC)

All articles need to cite sources... even those on basic math concepts. Obviously an article on a basic math concept can be easily cited to a basic math textbook... while articles on higher level concepts will require higher quality sources. The more complex and "higher level" the math concept is, the more we need to provide sources to support it. It isn't a question of the accuracy of the math involved... We need sources to establish notability, and we need sources to show that the concept being discussed is not OR (remember, Wikipedia is not the venue to debut a new mathematical proof ... even if the math involved is flawless). Blueboar (talk) 15:18, 3 June 2010 (UTC)

This is an example of an article including undergraduate-level maths. Notice that the fundamental equations were introduced with this edit of 18 May 2004, when WP:V looked like this. The equations are unsourced, and have never been sourced, though they've several times been corrected.

In 2005, an editor questioned the mathematics involved on the talk page, where Richard B explained clearly and patiently why the article is correct. His answer was a mathematical proof without reference to sources. 2006 was the first time that an editor asked for a reference supporting one of the equations. He went unanswered, and the equation still stands in the article.

In practice, can/should any of the mathematical content of that article be removed on grounds of being unsourced? Your answer seems to imply that I could, if I wished, remove them all, and I want to be sure that this is really what you intended.—S Marshall T/C 16:14, 3 June 2010 (UTC)

Yes, you may, if you think that it is in the best interest of the encyclopedia, remove the equations as being unsourced, just like any other unsourced information. You don't have to do so... there are other options which might be better (such as trying to find a source for the equations yourself, or tagging them to see if someone else can find a source... and you can even leave it as it is if you think that is best)... but you do have the right to challenge and remove them. Blueboar (talk) 16:45, 3 June 2010 (UTC)

Putting citation needed on them is probably best because you don't actually doubt them but would like to see where they came from. The question you talked about didn't ask for a citation - they were asking a reference desk type question about the equation trying to understand it. Dmcq (talk) 19:45, 3 June 2010 (UTC)

I think that we need to remove the "and its translation" from NONENG's line about citing sources. Nobody is going to demand that an editor spend hours typing up a translation, or even the original, for that matter. A news story in Spanish is going to be treated exactly like a news story in English: editors who have access (physical or linguistic) should provide a reasonable level of help to those that don't, and should refuse to provide an unreasonable level of help. We don't actually permit editors to demand that someone else type up ten pages out of an English-language source [even if it's in the public domain] so that they can verify the content without any trouble to themselves, and we shouldn't claim that editors can be forced to do this for a non-English-language source. WhatamIdoing (talk) 01:45, 6 June 2010 (UTC)

Examples

1) In a geographical article, I can prove by sources: 1) "X is a place with an area of 1,200 square kilometres"; and 2) "X is a place with a population of ~10,000." I wish to write: "The area has a population density of roughly 8.3 people per square kilometre." I think it is reasonable to write this, provided I cite both sources.

2) In an astronomy article, I can prove by a source that "X is a planet with a radius of 5,000km." I wish to write: "The planet's surface area is ~314,150,000 square kilometres." (Note that I have assumed without proof that everything called a "planet" is an oblate spheroid whose surface area can be approximated by the same formula that gives the surface of a sphere.) I think it is reasonable to write this, provided I cite the source.

3) In a technical maths article, an editor wishes to write:

"The push forward maps

$f_{\mbox{!}}=\sum (-1)^{i}R^{i}f_{*}\colon K_{0}(X)\to K_{0}(Y)$

and the pushforward

$f_{*}\colon A(X)\to A(Y),\,$

are related by the formula

${\mbox{ch}}(f_{\mbox{!}}{\mathcal {F}}^{\bullet }){\mbox{td}}(Y)=f_{*}({\mbox{ch}}({\mathcal {F}}^{\bullet }){\mbox{td}}(X)).$"

Even if the editor can prove this mathematically on the talk page, I think the third equation would need to be sourced.

Does this help illuminate what I'm saying?—S Marshall T/C 10:28, 3 June 2010 (UTC)

• 1, 2 & 3: Don't. Wikipedia covers what others have already covered. If population densities and planet surface areas are interesting, there will undoubtedly be published data already containing such things. We collect information that already exists. --SmokeyJoe (talk) 10:35, 3 June 2010 (UTC)
  • (2). You've taken 1 significant figure data and created five significant figure data. Are you trying to make us look foolish? --SmokeyJoe (talk) 10:37, 3 June 2010 (UTC)
    • You have assumed that none of those zeroes are significant. A "radius of 5,000km" could mean 5,000km, 5,000km, or 5,000km. WhatamIdoing (talk) 16:11, 3 June 2010 (UTC)

(1) - I would be wary about this because two different sources may not be referring to the same entity, even if they're using the same name. Many countries have territorial disputes (what is "China"?) and even without those, sources are often inconsistent on whether to count islands etc. (2) - the assumption needs justification, and I doubt it's true to that level of accuracy even for Earth. --GenericBob (talk) 12:06, 3 June 2010 (UTC)

Remember that we are writing for a general audience, not mathematicians. For an unsourced statement that is based on a mathematical formula to be verifiable, the average reader must be able to follow the formula... the average reader can be expected to know basic formulas ... Area, Volume, conversions from miles to killometers, etc. and do the simple math that is associated with them. A more complex formula (especially one containing all those funky Greek letters, arrows and symbols) will be beyond the ability of the average reader to understand. For a statement that is based on such a formula, the reader needs a citation to a source in order to verify that the information is accurate (or at least that someone reliable says it is accurate). Blueboar (talk) 13:13, 3 June 2010 (UTC)

• Yes, Blueboar's view in this exactly accords with mine. If I can prove it with high school maths, then it ought to pass WP:V without a citation, but if I need undergraduate-level maths, then it's not enough to show my working: I need a source. (Arguably, there might be an exception for highly technical articles on maths, physics or astronomy where we can expect a reader who's able to follow the article will also be able to follow an undergraduate-level proof, but I don't see the need to push that point.)—S Marshall T/C 13:18, 3 June 2010 (UTC)

The problem with doing that is it is original research to start generating figures that other people haven't though worthy of generating. A small obvious bit is okay but there are lots of pitfalls as others have pointed out above. I've seen too many people trying to push their original research with things that are obviously wrong on the basis that they are doing a trivial calculation. Dmcq (talk) 19:52, 3 June 2010 (UTC)

Well said. Crum375 (talk) 20:06, 3 June 2010 (UTC)

Yes, OR and V are directly intertwined here. I think the distinction is whether the basic calculation in question is commonplace or not. Consider the following... a newly discovered planet, where the diameter has been given by a reliable source but nothing else. A Wikipedian wants to add the approximate surface area of the planet, but no source actually gives it. Technically it would be OR for the editor to calculate the surface area himself. However, WP:OR allows "routine calculation" ... and in this case the calculation is not only simple, but is nothing but an application of a commonplace and well known formula that itself is not OR. Adding the aprox. surface area is allowed.

Now let us take a different situation... say I were to invent my own formula for figuring out the optimum consumption of beer by sports fans at a baseball game, based on body weight, temperature and the score (a measure of nervous energy expended)... I can not add that to an article, as it would be OR. The calculation might involve nothing more than basic high school math... but it is not routine... the formula itself is my invention and is not commonplace.

Now we take a third situation... an editor wishes to write an article on a complex mathematical formula that is known, understood and accepted by most post-graduate physicists and mathematicians, but not to the average reader. Here we need to cite reliable sources to verify that the formula exists, what its constants and variables are, and how they work. But, having verified that through citation it would not be OR for us to plug in some numbers and calculate a result using the formula. We are not performing original research... we are merely applying specific numbers to the cited and existing formula. Blueboar (talk) 21:27, 3 June 2010 (UTC)

An excample could show a completely hypothetical situation or a documented case, but it shouldn't be used to present original research. Putting some famous actor's statistics into an example in the body mass index article and coming up with a figure 'as an example of calculation' would be original research and very wrong. Dmcq (talk) 22:30, 3 June 2010 (UTC)

I am not sure I agree with you... If the article on Body mass has established (through citation) that a set, known formula exists for calculating body mass, and if the actor's statistics are established (again through citation), the act of plugging his numbers into the formula and calculating the actor's body mass is not OR. It is merely an illustration... an application of the known formula using known numbers. From a WP:NOR or WP:V stand point, this is no different than if you used a hypothetical set of body statistics. I could certainly agree with objecting to using a real person's statistics on WP:BLP grounds (as publication of a living person's body mass could potentially cause harm or embarrassment to that living person), but I don't think it is a WP:NOR or WP:V violation. Blueboar (talk) 00:05, 4 June 2010 (UTC)

Oooh! Could we illustrate the body mass of Ginger versus Mary Ann? :D 66.102.205.233 (talk) 02:31, 7 June 2010 (UTC)

• I'd like to suggest summarising and closing this discussion with a principle henceforth to be known as Blueboar's Law: "If Euclid proved it in 300 BC, then it's not original research".—S Marshall T/C 23:36, 3 June 2010 (UTC)

I would fully agree with that, as long as we amend it to "If Euclid proved it in 300 BC, and we have a reliable source for it, then it's not original research". Crum375 (talk) 23:50, 3 June 2010 (UTC)

Blueboar's Law is actually between those two... ""If Euclid proved it in 300 BC, and we know there is a reliable source that says he did so (whether cited or not), then it's not original research to use it now." Blueboar (talk) 02:29, 4 June 2010 (UTC)

The antecedent of the parenthetical comment may be misinterpreted. I believe that you meant "If Euclid proved it in 300 BC, and we know of the existence of a reliable source that says Euclid proved it in 300 BC (even if we have not named this reliable source in the article), then it's not original research to use it now."

We already get plenty of people saying that reliable sources aren't reliable if the source itself doesn't provide a bibliography/is "uncited". WhatamIdoing (talk) 18:41, 4 June 2010 (UTC)

Yes... that is what I intended. Maybe we should call it WhatamIdoing's Law.  :>) Blueboar (talk) 02:25, 5 June 2010 (UTC)

I got around this in one article where the audience were not mathematicians by stating in a reference "The proof of this is a standard undergraduate maths problem". Is this appropriate? Martinvl (talk) 10:54, 8 June 2010 (UTC)

Reliance on Published Works (revived discussion)

In an earlier thread I criticised Wikipedia's reliance on published works as being (in effect) censorship of politically incorrect beliefs -- of Men's Rights, in particular.

I was refereed to WP:Righting Great Wrongs. TFD (talk) 09:50, 15 May 2010 (UTC).

However, that policy advice merely restates that the issue is the concept of "reliable secondary sources". I am not interested in (although I can see why someone might think I was) using Wikipedia to "to set the record straight and Right Great Wrongs". What I want to be able to do is to have articles on Wikipedia which give information of Masculism and Feminism on an impartial basis.

Having a definition of "reliable secondary sources" which privileges universities and publishing houses is arbitrary and discriminatory. That is because Feminist have been around for about two centuries, have appreciated where the centres of power are, and have proceeded to infiltrate and dominate them to the extent possible. The institutions which dominate our brains -- publishing, the education system and the media -- are obvious centres of power, and mere political office-holders are downstream of those institutions, since the politicians can hardly hope to operate effectively outside the information world created by the information institutions.

Wikipedia is a non-published entity par excellence. It claims (explicitly and implicitly) to offer something that other, published entities cannot themselves provide. It is therefore contradictory to the very spirit of Wikipedia that it gives credibility value principally to sources which are published in a conventional, old-fashioned, pre-Web, pre-Wikipedia manner.

I agree that "what matters is not truth but verifiability." I am not arguing, here, about whether Masculism or Feminism is "true." What I am arguing about is the sources which Wikipedia chooses, arbitrarily, to view as "reliable." PeterZohrab (talk) 16:12, 8 June 2010 (UTC)

How would you suggest we define "reliable secondary sources" less "arbitrarily." It appears to me you believe that the culture's dependence on the publishing industry, the education system and the media is "A Great Wrong" and you are hoping to "correct" it. --John (User:Jwy/talk) 19:19, 8 June 2010 (UTC)

John raises two issues, which I will deal with out of order.

The second one is whether I am actually trying to correct a Great Wrong. I am perfectly entitled to want to correct Great Wrongs within the Wikipedia editing guidelines. Since Wikipedia is part of Western culture, it is likely to reflect Great Wrongs which exist in the wider culture. Righting this Great Wrong within Wikipedia does not automatically correct the Great Wrong in the wider culture, although it may be one small step along that road. The point here is not what happens in the wider culture -- the point is what Wikipedia ought to be doing.

The first issue is how I would suggest we define "reliable secondary sources" less "arbitrarily". The only answer to this is that there is no such thing as a "reliable secondary source." I have taught at universities, I have worked a a journalist, I have published a book, and I know someone who works in a publishing house. The notion that one could regard myself, my colleagues or my friends as "reliable" merely because I/they work(ed) in institutions that pull themselves up by their own bootstraps and want to project an image of "reliability" is ludicrous. That is largely spin and public relations hype. The realm of ideas and knowledge is truly anarchic. People may like to cling to the notion that these institutions have systems in place for fact-checking, or that peer-review is some kind of safeguard, but what actually happens in practice is minimal -- in all sorts of different ways. For example, facts are only one small part of journalism. If I interviewed someone and quoted him/her at length in a journalistic article, there was no need to check facts, since I was attributing the claims to the interviewee. So all a Feminist journalist, for example, has to do is to keep interviewing Feminists and avoid interviewing Masculists, which gives Feminists credibility and Masculists invisibility (the actual situation as of now). As another example, any university can set up a department of Phrenology, Astrology, Voodoo Science or Women's Studies and anyone can found a peer-reviewed journal of Phrenology, Astrology, Voodoo Science or Women's Studies.

Wikipedia should have a system for attributing all articles and edits to sources, and let the reader decide how much to trust the sources.210.246.20.174 (talk) 17:15, 14 June 2010 (UTC)

Conflicting clauses

Addition of SlimVirgin at 19:37, 9 April 2010 - All material in Wikipedia articles must be attributable to a reliable published source to show that it is not a Wikipedian's original research, but in practice not everything need actually be attributed. I can't understand it. There is wide consensus about this addition to the rules? X-romix (talk) 09:47, 27 May 2010 (UTC)

Makes sense to me. It must be theoretically possible to attribute it, but in practice there's often no need to actually attribute it.--Kotniski (talk) 09:51, 27 May 2010 (UTC)

Not true. If an article's coverage is not attributed with an in line citation, how can you tell if it meets WP:BURDEN if that is the only coverage there is? --Gavin Collins (talk|contribs) 09:56, 27 May 2010 (UTC)

Depends who "you" is. Being attributed with a citation is only one (possibly fairly minor) factor in whether a particular reader is able to check a particular statement. If it's something for which 1000 sources can readily be found just by Googling, for example, then it really doesn't help much if we provide a specific citation. (I'm not saying we shouldn't, just that it's not an absolute priority.)--Kotniski (talk) 11:29, 27 May 2010 (UTC)

"In practice there's often no need to actually attribute it" sounds like WP:ATA#Crystal to me. --Gavin Collins (talk|contribs) 11:55, 27 May 2010 (UTC)

Not sure what you mean, but the fact that in practice not everything needs to be attributed has been WP's core policy from day one. Essays, which are random thoughts by random anonymous individuals, without consensus, cannot override policies or guidelines, and I am not even sure that the essay you cite is conflicting with this policy. Crum375 (talk) 12:02, 27 May 2010 (UTC)

While it is possible to find and cite a source that says "2+2=4", we all know that it would not contribute much to the accuracy or reliability of WP to cite every detail that trivial. Of course editors considering the base 3 POV may slap on a {{cn}} and argue on the talkpage "That's wrong, 2+2=11 in base 3!", leading to further refinement of the article (although in general human readers work in decimal arithmetic and our anthropocentric POV reflects that). We often don't cite things that are so trivial unless someone actively challenges them. Every citation takes time and effort away from other work and there is a modicum of editorial judgement involved. Slapping {{cn}} on every assertion in WP would not be constructive. So far, this approach has worked out pretty well as most editors actually do want to get things (small r) right.LeadSongDog come howl! 13:32, 27 May 2010 (UTC)

The key is that things need to be cited "when challenged or likely to be challenged"... This concept has a long history on Wikipedia and has firm consensus. What causes all the commotion, confusion and angst are statements that were originally added with the idea that they were "unlikely to be challenged" (and so were not cited)... but which are now being challenged.

I think we all understand what a challenge is... although we often disagree as to whether a specific challenge is legitimate (or frivolous). But whether legitimate or not, I think it has been established that the ultimate solution to any challenge is to provide a reliable source. We may feel that it is a pointless hassle to do so, but doing so will resolve the challenge.

We don't have a clear statement as to what "likely to be challenged" means. Nor can we. This is because "likely to be challenged" if often a judgment call... and often depends on how controversial the topic is. A good rule of thumb is... "when in doubt, cite it". But even then, editors will frequently disagree as to whether a statement is "likely to be challenged". That is in the nature of Wikipedia. Leaving aside the addition of deliberate OR and POV pushing... we are always going to have editors who add material that they assume does not need a citation (because they are sure that it is correct and non-controversial), but sometimes they will be wrong in that assumption... On the other side, we are also always going to have editors who take an overly legalistic, literal approach to challenging (and challenge things that realistically don't have to be cited). We can not legislate either out of existence, no matter how we word this policy. What we can do is encourage our editors to do good research, to not take challenges personally, and ask them not to make challenge frivolously. Blueboar (talk) 14:24, 27 May 2010 (UTC)

I agree with X-romix: this amendement cannot stand. I think it is a very personal interpretation of WP:BURDEN that "in practice not everything need actually be attributed". I think this policy can only be silent on this issue, because I read this as a licence to stall any possible challenge. In practise, the one statement that is not attributed is going to be the one challenged and disputed. --Gavin Collins (talk|contribs) 16:39, 27 May 2010 (UTC)

It's not an "amendment", since it's always been part of the core policy, from the day WP was founded in its current format. And it works very well in practice, since all WP content is based on it: anything challenged, likely to be challenged, or quoted must be attributed, anything else can be unattributed until challenged. All material, without exception, must be attributable to a reliable source. Crum375 (talk) 16:53, 27 May 2010 (UTC)

It's not any kind of "personal interpretation" either, just an obvious statement of fact - just click around Wikipedia for a few seconds and see what a large proportion of statements are not actually attributed, and consider what a pale shadow of its globally successful self Wikipedia would be if all those statements were to suddenly disappear.--Kotniski (talk) 17:05, 27 May 2010 (UTC)

Yes, this is correct: The policy is, and AFAICT has always been, that material must be verifiable, not already verified. Entirely unreferenced articles (although bad practice) are actually 'legal' on Wikipedia, so long as it is possible to verify the information (e.g., by asking your favorite web search engine about it).

Additionally, there is, and has never been, any requirement to use inline citations (so long as there are no direct quotations and no challenges). See WP:CITE#General_reference: This section wouldn't exist if Wikipedia actually required inline citations in every article. WhatamIdoing (talk) 17:07, 27 May 2010 (UTC)

I tried to say that with this edit, but was instantly reverted. I couldn't be bothered to bring the matter to the talk page at that time so I left it. Right now, the policy says inline citations are a "must", which is fairly typical: policies document how the kind of editor who's active on policy pages thinks things ought to be done, not how the kind of editor who writes content thinks things ought to be done.—S Marshall T/C 18:47, 27 May 2010 (UTC)

Inline citations are a must for anything challenged or likely to be challenged, and for all quotations. That doesn't mean that every single thing needs a citation. SlimVirgin ^{talk contribs} 18:52, 27 May 2010 (UTC)

I don't agree with you. Jimbo wrote: "If it is true, it should be easy to supply a reference". [2]. I think that article with lacks of sources - is a wide gate for mass (hundreds in one article) "mistakes", original researches, conflicts of interests, hoaxes, nonsence, non-quality and non-neutral articles. Falsificators, propagandists and original researchers do not wants to show their sources. Conscientious users always can supply references in all paragraphs of their text, or in the bibliography section of their article. X-romix (talk) 23:03, 27 May 2010 (UTC)

Certainly: The absence of sources can cause all sorts of problems. But the fact remains that unless and until some editor actually gets around to challenging a given statement, Wikipedia does not actually require editors to support the statement with an inline citation (except for direct quotations). WhatamIdoing (talk) 23:17, 27 May 2010 (UTC)

The absence of sources can cause all sorts of problems. The presence of sources makes article better and more helpful for readers and students, allows to check and clean up article with some errors and falsifications, allows to evaluate the quality of used sources. So, WHY NOT provide all sources in Wikipedia? I don't understand WHY I need not to read and write articles, but talking and challenging with falsificators, clowns and political propagandists who can't find sources for their own texts. X-romix (talk) 13:39, 29 May 2010 (UTC)

S Marshall, I see where you're coming from, but BURDEN really only applies when a challenge exists. Once that challenge exists, an inline citation is required. Consequently, in the context of that particular section, an inline citation is actually required. WhatamIdoing (talk) 22:53, 27 May 2010 (UTC)

Gavin collins is wrong. Not all material on Wikipedia must be verified. The "verifiable not verified" soundbite is longstanding. Often facts are such common knowledge that to verify them would be ridiculous, or at least of little practical importance. Where the boundary is crossed is an editorial judgment on a case-by-case basis. If an editor wants individual facts or opinions verified or attributed, then they should: 1. do it themselves; 2. tag them; 3. remove them if dubious or potentially violating BLP. Fences&Windows 17:18, 27 May 2010 (UTC)

...or 4. discuss the issue on the article talk page. or 5. discuss the issue at WP:RSN... there are lots of ways to deal with material that you think should be sourced but isn't. Which one is the best way to deal with it depends on the specifics of the situation. But Fences is correct... you must be able to cite even the most obvious statements... but you are not necessarily required to actually cite them. (unless they are challenged or likely to be challenged.) Blueboar (talk) 22:51, 27 May 2010 (UTC)

• I agree with Blueboar and F&W. Inline citations are not required except for specific disputed facts. Requiring them is way beyond the standard of any normal academic tertiary source, of even the highest standards. Things must be referenced, certainly, but to in a sufficiently exact manner to permit verification,and nothing more is generally necessary. The only sorts of writing I know which actually use inline references for everything are legal opinions and medical textbooks, two very specialized genres, notyorious for their lack of general readability. DGG ( talk ) 05:03, 28 May 2010 (UTC)
• After my attempt at a FA and long experience at AfD, I've learned that if it can be challenged, someone will challenge it. There are Wikipedians who assume everyone's either a marketer, a spammer or a vandal—and if they've done a lot of new pages patrol, that suspicion's understandable. I'm finding myself giving one or two citations a sentence. (I actually started two yesterday, which is very unusual for me; they're Hannah Monyer and Dominique Reiniche, both BLPs. You can see I've ended up writing to a pattern: first sentence, name, date of birth, nationality, profession, with inline citation. Second sentence, assertion of notability, with inline citation.) Experience with suspicious patrollers has taught me that if you want to get past NPP and through the AfD process, one citation per sentence is the only safe way. But the encyclopaedia shouldn't be like that.

  Constant inline citations that appear next to the material they support should not be necessary. In a logical world, I should be able to write, "The following three paragraphs are based on Thompson 2002, chapter III," and provided Thompson 2002's in the bibliography, that should suffice. As a longstanding Wikipedian and autoreviewer, I should arguably be able to source something to "Mr D Smith (with his credentials as an expert), personal correspondence" and editors should trust me.

  But there's a vociferous group of editors who believe that inline citations are there so that they can conveniently check (and quibble) every sentence. I think we've all seen the editors at AfD who will challenge a source because it's not in English. ("I can't check it!"—Well, mate, that's not my fault. I can prove I've done the research, but I'm not going to teach you German, or drive to your house with half the contents of my bookshelf so you can read the offline sources. Whether or not you can check it is not my problem.)

  If I were writing for de.wiki, or fr.wiki, or for any print encyclopaedia, it would be enough to provide a bibliography. Only en.wiki has the inline citation rule and it makes it unnecessarily hard to write.

  Another problem is that when I say these things, editors come back to me with alphabet soup links to policy, combined with strongly-worded sentences in the emphatic declarative. They say to me, "We do things like this, because of WP:THISRULE. It's supported by a longstanding consensus", apparently without understanding that this is a logical fallacy. I'm not talking about how we do things now. I'm talking about how, in a logical world, things ought to be done.—S Marshall T/C 08:39, 28 May 2010 (UTC)

Simply put, the idea that "All material in Wikipedia articles must be attributable to a reliable published source... but in practice not" is rubbish policy. --Gavin Collins (talk|contribs) 09:03, 28 May 2010 (UTC)

That addresses one small part of my argument rather than the whole. Can I take it you agree with everything else I say?—S Marshall T/C 09:08, 28 May 2010 (UTC)

I sympathise, but like you I have found that adding citations is the only way to avoid disputes, and all of Wikipedia policies hinge on being able to review the sources, so attribution is key to resolving these disputes. If there is a fallback position that attribution is not necessary in every case, then these disputes can't be resolved. I have been in a long mediation case where a group of editors were adamant that just because content was unsourced, it could not be labled as original research, but once the case got underway it was shown that all the unsourced content was original research. This is why I object strongly to this "get out of jail free" clause. --Gavin Collins (talk|contribs) 09:24, 28 May 2010 (UTC)

You seem to think that all of Wikipedia is the subject of dispute. Maybe you work in more controversial subject areas, but in many areas people don't go around disputing things all the time, so there really isn't any compulsion (whatever that would mean) to add sources for every statement. It would be quite nice if we did have a citation for every statement, but we don't want to give people the impression that they're somehow breaking Wikipedia's rules any time they add statements without citations - that would drive away even more editors than we currently are doing.--Kotniski (talk) 11:11, 28 May 2010 (UTC)

There's a nub of substance in S Marshall's comment above. When stubbing a new article we do need to have a suitable citation for the assertion that establishes the subject's notability. Otherwise it simply should not survive. OTOH there are indeed entire classes of articles, such as discographies, that are very sparsely cited without really compromising accuracy. Sometimes things are not "likely to be challenged" simply because it is too easy to check. LeadSongDog come howl! 12:56, 28 May 2010 (UTC)

I disagree: Subjects are notable if the sources exist, not if the sources are already cited in the article. Lion, for example, was a notable subject in its very first, utterly unsourced stub, not a while later when someone started adding sources. If one of those aggressive NPPers tags an unsourced article for deletion because the NPPer is ignorant of the existence of sources, then the NPPer has screwed up — an understandable error, but still a mistake.

Yes, if we want to avoid disputes, we'll all do the equivalent of defensive medicine on Wikipedia — but that doesn't mean that we must write and cite to defend against potential challenges. WhatamIdoing (talk) 16:12, 28 May 2010 (UTC)

(edit conflict) Well, I write/translate a lot of biographies, and Wikipedia's BLP policy presently treats anything biographical as if it had anthrax, leprosy and bird flu, so yes, I probably do work in more "controversial" subject areas.

In response to Gavin Collins: I've written a lot of material that the average editor couldn't check—because the average editor doesn't speak the language the source is written in, or because he doesn't have the same books on his bookshelf. Things don't have to be checkable by everyone.—S Marshall T/C 12:59, 28 May 2010 (UTC)

I agree that standalone articles do need to have a suitable citation to establish the article's compliance with WP:BURDEN, which is why I think saying that its "not necessary in practise" shold be dropped as it is a licence for original research. However, I don't agree with LeadSongDog that there exist facts that "too easy to check"; my experience with fictional topics is that a lot crap is contained in unsourced articles such as Gaius Baltar that seems plausible, but in reality is actually madeup and because there are no citations it is impossible to check whether it is original research or not. I propose that we drop the "everything need actually be attributed clause althogether" clause because it encourages sloppy editing. --Gavin Collins (talk|contribs) 13:40, 28 May 2010 (UTC)

Stand-alone articles do not need a suitable citation until their contents are challenged. BURDEN does not say "The burden of evidence lies with the editor who adds or restores material and he has to provide that source the very second he adds or restores material." BURDEN does not apply until some other editor comes along and says, "Really? Then WP:PROVEIT." WhatamIdoing (talk) 16:17, 28 May 2010 (UTC)

Leadsongdog, discographies are not an example of articles with few sources, because each disc (record or CD) mentioned is a citation (unless, of course, the label, cover, or case does not mention that the person who is the subject of the discography was involved with the disc.)

Gavin.collins, fiction is a poor example to mention when trying to give examples of unacceptable uncited statements, because unlike the real world, anything is possible in fiction, so there is know way to know if a statement is true except by reading/viewing the work of fiction.

For a more realistic example, consider this example from an article Gavin.collins has worked on, Moneylender:

Moneylenders who are unregulated, engage in predatory lending or seek to enforce loan agreements by illegal means such as extortion are commonly referred to as loan sharks.

Anyone familiar with the English language knows this is true. Anyone unfamiliar with English can easily look it up in almost any dictionary. No citation is necessary. Gavin.collins, are you willing to say that anyone could possibly start a good-faith dispute about this statement? Jc3s5h (talk) 14:14, 28 May 2010 (UTC)

(edit conflict)Well, fictional or pop culture topics are perhaps a different animal. They tend to be attractive to fans, and fans are, by and large, people with strong opinions who believe themselves knowledgeable about their subject. Only some are correct, and that's a place where verifiability standards need to set the bar very high.

Personally I think Wikipedia's coverage of fictional topics is beyond excessive in proportion to our coverage of factual ones, and it blows my mind that we've got people writing FAs about The Halo Graphic Novel and the Metallica Discography when the biographies of three quarters of the Gottfried Wilhelm Leibniz Prizewinning scientists, and about ninety percent of the national-level politicians of Europe, are still redlinks.

But producing an article that's unchallengeably-referenced in every respect results in proseline. My last attempt at that—History of Hertfordshire—has 170 separate references, 14 footnotes and 27 volumes in the bibliography. It's totally neutral in point of view and it's also virtually unreadable, partly because I could've organised it better, but mostly because there's nothing to it but lede followed by a seven thousand word list of facts that are pretending to be prose.—S Marshall T/C 14:19, 28 May 2010 (UTC)

• ... and in response to Jc3s5h, you were the person who reverted my change to inline citations. The problem with the line you're taking is that things that ought to be uncontroversial, aren't. I ought to be able to write:

  A human is a hairless ape ultimately descended from tree-dwelling primates.

  But, there's the Christian right who will insist on describing that as "theory" and wanting what they would call "legitimate alternative views" to be represented in the article. Many of them have long since departed for safer climes like Conservapedia, but not all, and the truth is that no matter what you write, there's usually some raving lunatic or someone with a point to make who'll go around challenging what ought to be uncontroversial.—S Marshall T/C 14:27, 28 May 2010 (UTC)

The statement S Marshall puts forward as something that ought to be uncontroversial is anything but. Many religious people believe that humans have immortal souls and will experience an afterlife, but other animals do not. If they are right, there is a huge difference between humans and other animals, and there is justification for saying that humans are not apes.

The approach of the cite-every-sentence crowd in this discussion, of giving examples of statements that seem uncontroversial, but upon closer examination are controversial after all, is the wrong approach. This crowd should demonstrate that all statements are controversial, and thus need a citation. I don't think they can prove any such thing, so I don't think their argument can prevail. Jc3s5h (talk) 14:40, 28 May 2010 (UTC)

In response to Jc3s5h, the quotation from the article Moneylender was written by me, and to be frank, its not very good, as it contains a form of mass attribution ("they are are commonly referred to") that could easily be challenged. I don't think we should encourage unsourced content, even if I have been guilty of adding some myself.

The biggest objection I have to unsourced content is that it invites what I call "goldfish editing", namely lack of attribution encourages other editors to nibble away at unsourced coverage, changing a word or a sentence here and there, but with no memory of what went before. Eventually the content may be changed hundreds of times, without any substantial change being made. For example, until citations were added the lead paragraph of the article Accountancy, the unsourced content was the subject of tiny goldfish edits virtually every day. We need to encourage editors to add citations, not unsourced content which is why the "not everything" clause needs to go. --Gavin Collins (talk|contribs) 14:45, 28 May 2010 (UTC)

Wikipedia is built on unsourced content. Adding unsourced content (if you know that you could source it if you wanted to) probably adds far more to the value of the encyclopedia that spending the same amount of time adding citations. So no, I don't agree that "we need to encourage editors to add citations, not unsourced content". At least, not everywhere - it would probably be true in areas that are known to be controversial, but most areas aren't.--Kotniski (talk) 14:53, 28 May 2010 (UTC)

Perhaps the discography example was poorly chosen. I've stubbed quite a few articles on academic journals. These I usually start out by populating a few fields in template:infobox journal, particularly the Title, url, ISSN, and OCLC fields. The latter two are in effect citations to the respective WorldCat entries. They do not meet a normal form of citation, but they still provide sceptical editors with an easy means of verification that I'm not just engaging in WP:OR. The link to the publisher's website is helpful for populating things for which SPS are normally accepted as reliable, e.g. the editor, publisher, frequency parameters. If I'll assert that it is the official journal of a certain society then I populate the infobox with a link to that society's website. Then I'll write the first line of the article with information from those sources, tag it with {{journal-stub}} and hit "Save page". I have not, at that point, made any formal citation, but I have still linked to sources that support the accuracy of what I say. I've yet to have a problem with new page patrollers using this approach. It isn't the form of citation that really matters, but that the information can be readily verified by anyone who cares to. Only a few articles about journals grow to any great size or accumulate large numbers of citations, but even at this stub stage, they have utility to readers. I then add redirects from the various abbreviate titles (with and without periods). Suddenly, all over WP, dozens of redlinks turn blue.LeadSongDog come howl! 15:37, 28 May 2010 (UTC)

Sure, unsourced additions to articles are not great and I always work from and cite sources myself. I would never encourage anyone to add unverified information into Wikipedia. But disallowing unsourced content in the way Gavin.collins proposes would be a major policy change. It would in effect allow removal of all unsourced content and speedy deletion of all unsourced articles, which would be massively disruptive. This is why the "if challenged or likely to be challenged" bit is sensible - it allows the sourcing or removal of unsourced material to proceed at a measured pace rather than in a blitz of deletion. Fences&Windows 15:40, 28 May 2010 (UTC)

I am not saying we this policy should state unsourced content is not allowed, but that the statement "in practice not everything need actually be attributed" is rubbish guidance, and should be removed. --Gavin Collins (talk|contribs) 21:17, 28 May 2010 (UTC)

Sounds like excellent guidance to me, correct and well stated. Why do you think it's "rubbish"? Crum375 (talk) 21:23, 28 May 2010 (UTC)

Maybe I am the only one who sees the conflict between say "All articles must be ...but in practise not". Mayb we should change this sentence to "all articles should be attributable to a reliable published source to show that it is not a Wikipedian's original research" and drop the get out exemption at the end? --Gavin Collins (talk|contribs) 16:16, 29 May 2010 (UTC)

As we said right at the start of this thread, there is no conflict. "Attributable" is not the same as "attributed", any more than "payable" is the same as "paid" or "breakable" the same as "broken". --Kotniski (talk) 16:20, 29 May 2010 (UTC)

You are splitting hairs. If a sign says "You must never touch the overhead power line...but in practise you can", we would say that is a rubbish warning. It does not matter if the overhead power line is never to be "touched", or should not be considered "touchable"; its the get out clause at the end that is nonsense. --Gavin Collins (talk|contribs) 16:37, 29 May 2010 (UTC)

Yes. But what does that have to do with the sentence we're discussing here, which isn't in anything like that form? --Kotniski (talk) 16:49, 29 May 2010 (UTC)

I might be wrong on this, but its a good analogy. On the one hand, this policy says that articles must be attributable to show that they are original research, but on the other it says not everything in practice need actually be attributed. I see conflict between the two statements. Surely we can bring these two conflicting statements together to write one clear statement of guidance: articles should be attributable to show that they are not original research. --Gavin Collins (talk|contribs) 21:01, 29 May 2010 (UTC)

Which policy says that "articles must be attributable to show that they are original research"? Crum375 (talk) 21:22, 29 May 2010 (UTC)

I see no conflict at all and I truly wonder if many here are talking right past each other because those questioning the language added do not understand what the words ending "able" and "ed" mean in this context. The complete missing of the mark of the power line analogy above makes me think I'm right.

• Attributable means capable of being attributed. It means that it is possible to find reliable sources for the information.
• Attributed means sources are present in the article; the subject that was always attributable has had the reliable sources added to become attributed.

There is no a "get of of jail free" clause, and the statement can only be seen in that way if the subject statement's meaning is misunderstood. So, following on my defining of the words, let me translate the text we are here about:

"All material in Wikipedia articles must be attributable to a reliable published source to show that it is not a Wikipedian's original research, but in practice not everything need actually be attributed"

It means:

All material in Wikipedia articles must be capable of being verified with reliable published source (it must be possible to place sources), because Wikipedia does not announce new things; information that has not already been published outside of Wikipedia is original research. Because we require the ability of information to be sourced, as noted by "attributable", but do not require actual sourcing for every statement to appear in article text, a corollary of the verifiability policy is that if someone questions whether something is capable of being sourced (whether is attributable), by challenging it, then showing the capacity to be sourced must be proved by producing and citing the actual source(s). Once that is done, the material has been attributed.--70.107.78.246 (talk) 22:16, 29 May 2010 (UTC)

How many articles without sources you have already written? In some cases any other user can find source, but in other cases it is too difficult (especially if author used some propaganda or fiction). Jimbo wrote: "I heard it somewhere" pseudo information must be aggressively removed.[3] - now it is a part of rule about verifiability. There is many factual inaccuracy (we found more than 100 factual errors in one "featured" article, and more than 50 errors in other "featured" article, written by arbiters of Wikipedia) and political propaganda without any sources. Propagandists and falsificators do not like to show their sources and prefer to hide sources and names. X-romix (talk) 09:12, 30 May 2010 (UTC)

So name and shame if necessary. That's part of why we have a full history of edits. Wikiblame can help, too. But for most articles, where propaganda isn't an issue, a {{cn}} tag serves the purpose. LeadSongDog come howl! 15:23, 31 May 2010 (UTC)

We don't need to shame and blame; we already have policies which say that this amendment conflicts with Wikipedia policy. Lets put the relevant statements side by side:

• WP:V: "All material in Wikipedia articles must be attributable to a reliable published source to show that it is not original research, but in practice not everything need actually be attributed";
• WP:OR: "Citing sources and avoiding original research are inextricably linked: to demonstrate that you are not presenting original research, you must cite reliable sources that provide information directly related to the topic of the article, and that directly support the information as it is presented".

My experience in the Kender mediation is that any content that is not the subject of an inline citation is defintely original research. No if's, no buts, unsourced content is crap - it can't be checked, and more than likely it has been madeup. The example given that "Paris is the capital of France" is entirely misleading. A better example would be given by "Paris is the cultural capital of Europe" or "Paris is the city of lights". Who said this, where to these statements come from? These are the questions that only attribution can provide answers to. --Gavin Collins (talk|contribs) 21:16, 31 May 2010 (UTC)

Attributed and Attributable

Gavin, you're confusing unattributed with unattributable. The former is okay in certain circumstances, the latter not allowed. SlimVirgin ^{talk contribs} 21:36, 31 May 2010 (UTC)

Gavin, I admit I'm new to this discussion, but can you explain/expand on your point that "Paris is the city of lights" is "crap" if it isnt attributed. From where I stand reading your post you are saying the sentence, if in an article unattributed it is crap; but once it has a citation it becomes legit and "Truth"... Either the sentence is good or not does not depend on all of us knowing it is good or not through observation, if the citation was in French and I cant speak French then I must assume AGF that someone who speaks French will speak up if the citation does not say that; until then I must assume it does. How is this any different than assuming that the sentence without a source is in fact true? If an unsourced sentence is so dubious that I believe a citation is needed I can go to the search engine of my choice and find one, if I dont find one THEN it becomes crap. A "factoid" in an article is not Schrodinger's cat in limbo between truth and crap until observed by an outsider. It seems your entire philosophy is based upon an entire lack of good faith on any Wikipedian's contributions. This is entirely against the philosophy on which Wikipedia is based upon. Once you go down the slippery slope of saying EVERYTHING must be attributed not just attributable you then will see that the next step is that EVERYTHING must be verified by everyone, if one person cant verify something themselves then it is "crap". We simply can not throw away AGF and become paranoid on everyone's motive for an addition.Camelbinky (talk) 23:17, 31 May 2010 (UTC)

"I can go to the search engine of my choice and find one" - it is not so simple in real articles. Some special information can't be quickly found. There is vicious circle - search engines find sites whose authors takes information from Wikipedia. "Good faith" is not allows to verify articles for non-reliable sources, "errors", political propaganda and original researches. X-romix (talk) 12:36, 1 June 2010 (UTC)

X-romix is bang on the money. I would go further and say that good faith is a sharp sword that cuts both ways. In order to demonstrate good faith that you are not presenting original research, you must cite reliable sources that provide information directly related to the topic of the article, and that directly support the information as it is presented. This policy should not then say that in practise its not necessary to attribute sources - that is a good example of doublethink. The wording should be rephrased to say "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research", and by doing so, the ambiguity of the current version can be elimated (with the added benefit of brevity thrown in). Surely this is "win-win" proposal? --Gavin Collins (talk|contribs) 15:27, 1 June 2010 (UTC)

Gavin, several people have pointed out here and elsewhere that you're confusing "unattributed" with "unattributable." This policy and NOR already make clear that all material in WP must be attributable. But it has never been the case that every single thing must actually be attributed. That has never been policy and I don't think it ever will be. SlimVirgin ^{talk contribs} 22:23, 1 June 2010 (UTC)

Amen and ditto Slim.Camelbinky (talk) 23:37, 1 June 2010 (UTC)

The distinction between "attributable" and "unattributed" is meaningless in the absence of inline citations, which is why attribution should be provided for all statements of opinion or fact to verify their source and to demonstrate that they are not original research. Show me an article without attribution, such as Gaius Baltar, and you can guarantee it to be filled with unverifiable statements. I would love to see an example of an article without attribution that does not contain original research. Can anyone show me an example? --Gavin Collins (talk|contribs) 06:50, 2 June 2010 (UTC)

I'm agree with Gavin Collins - "attributable" concept is a doublethink and loophole for falsifiers, original researchers and political propagandists. X-romix (talk) 10:22, 2 June 2010 (UTC)

No-one said that whole articles should be without attribution (though I'm sure Wikipedia contains many such, particularly stubs, that are not in any way original research). But there are many, many statements in WP articles that are not attributed but which are not OR (because they are attributable). This is really only an obvious practical point, not in any way disputable - you may think it not worth stating at this point in the policy, but because sometimes people see "policy" pages and interpret them as commandments to wage holy war, I think it needs to be made clear right from the start that we don't go around gratuitiously removing (or even challenging) statements purely on the grounds that they are not currently attributed.--Kotniski (talk) 10:35, 2 June 2010 (UTC)

Many statements is not attributed? What a pity! These statements is not verifiable and must be deleted. X-romix (talk) 10:54, 2 June 2010 (UTC)

I am still not following Kotnsiski position. The conflict of statements is apparent to me:

1. quotations and any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation...
2. ...but in practice not everything need actually be attributed.

Effectively 1 (taken from WP:BURDEN) is negatated by 2. I think for me the "in practice" is a major sticking point. This is a very broad, sweeping statement that gives the impression that not providing citations is genrally acceptable, when we know that in practise there is not a single statement in Wikipedia safe from challenge. Since content is challenged all the time, even cited content, I think the statement "but in practice not everything need actually be attributed" makes the false presumption that some types of content are somehow exempt from challenge, as if there existed content that is super-truth, or above being sourced. --Gavin Collins (talk|contribs) 11:14, 2 June 2010 (UTC)

Well yes, I think that's the point. There are things that no-one would reasonably challenge, and for which it would be unnecessary in practice to provide a source. (Not that it would matter if somebody did provide a source; but it would matter if people went round removing all such statements as X-romix seems to think should happen, or even wasting people's time by "challenging" such statements just to make a point, and that's why we benefit from the qualification about not everything having to be attributed.) --Kotniski (talk) 12:00, 2 June 2010 (UTC)

I don't see any false presumption. It is simple logic...

• All A are B (All statements must be attributable).
• Some B are C (Some attributable statements - those that are "challenged or likely to be challenged" - are required to be actually be attributed).
• Thus, some B are not C. (Thus some attributable statements - those that are not "challenged or likely to be challenged" - are not required to actually be attributed.)

We can (and often do) debate where to draw the line between "C" and "not C" ... but "not C" still exists. Blueboar (talk) 12:12, 2 June 2010 (UTC)

I think you are missing the point I made earlier that the distinction between "attributable" and "unattributed" is meaningless in the absence of inline citations, because there only editorial opinion can make unsupported distinctions between the two. Therfore if in practice not everything need actually be attributed, then both attributable and unattributable statements are allowable. In answer to Blueboar

• All A are B or D (All unattributed statements can be attributable or unattributable).
• Some B are C (Some attributable statements - those that are "challenged or likely to be challenged" - are required to be actually be attributed).
• Some B are not C (Thus some attributable statements - those that are not "challenged or likely to be challenged" - are not required to actually be attributed.)
• Some D are not C (Thus some unattributable statements - those that are not "challenged or likely to be challenged" - get a free pass as they are not required to actually be attributed "in practise".)

Its clear to me why we need to change the wording to "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research". The idea that unattributable statements are allowable " in practice" is rubbish policy. --Gavin Collins (talk|contribs) 12:43, 2 June 2010 (UTC)

You are still confused, Gavin. Unattributable material is never allowed, but some material may be unattributed, unless it is challenged, likely to be challenged, or quoted. Crum375 (talk) 12:56, 2 June 2010 (UTC)

The distinction is simple, Gavin... Unattributable means a source does not exist. Unattributed means that a source does exist, but is not provided. Unattributable material is never allowed... unattributed material is sometimes allowed. Blueboar (talk) 13:17, 2 June 2010 (UTC)

I take your point that this is what the policy should say. As I have said, there is no discernable difference between unattributed and unattributable content in the real world as you can't tell them apart. The whole idea of providing citations is that you can, so that is why I object to this policy saying that "not everything need actually be attributed" because that is licence to cheat, and in any case, its as clear as mud. If we shorten the offending sentence to "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research", then we are being both brief and clear - a win win situation, no? --Gavin Collins (talk|contribs) 16:19, 2 June 2010 (UTC)

Except that as you yourself have shown, a lot of people don't realize there's a difference between "-able" and "-ed", and are liable to read this sentence as "...should be attributed...", and then (possibly) start kicking up a fuss about the vast number of valuable statements in Wikipedia that don't carry a citation. I agree the clarification isn't vitally necessary, but I think it's worth having it there, and it doesn't do any harm, since there is plenty more in the policy you can quote at people if they start trying to use this as some kind of get-out clause.--Kotniski (talk) 16:33, 2 June 2010 (UTC)

I believe strongly it should be removed, as it is harmful and I have been through a long dispute over this issue. Once editors start claiming that "in practise, not everything need actually be attributed" then we have an entirely new layer of policy that is licence for original research and plagiarism. Lets agree to get rid of it. There must be some wording we can agree upon. --Gavin Collins (talk|contribs) 16:42, 2 June 2010 (UTC)

What do you suggest? (But editors are not starting claiming that "in practise, not everything need actually be attributed" - this is true, and has been so for ages - if you're proposing that it shouldn't be, then you should make a specific proposal and try to convince the community to change its practices.)--Kotniski (talk) 17:07, 2 June 2010 (UTC)

Gavin, please listen to all the editors who are telling you that we have never required that everything be sourced. If you want to change that, please make a solid proposal here, on the village pump, and elsewhere. SlimVirgin ^{talk contribs} 19:42, 2 June 2010 (UTC)

Sorry guys, I am going against the grain here. Just for the record, I am not saying that everything has to be sourced. I am saying that should be attributable, period. Its a standard we should aim for, not must aim for. Afterall, even I am made of only flesh and bones. --Gavin Collins (talk|contribs) 20:48, 2 June 2010 (UTC)

Okay, but I think this discussion has reached the end of its useful life. If you want to make a concrete proposal to change the policy, please suggest new wording and let people know about it. SlimVirgin ^{talk contribs} 20:53, 2 June 2010 (UTC)

Late to this discussion, too, but I'll add my opinion that the standard is and always has been attributable, not necessarily attributed. The statements above that in-line citations are essential and that edits without them are, as one editor put it "all crap" fly in the face of Wikipedia:Assume good faith. Are citations good? Obviously. Are in-line style citations preferred? Generally. Are they required to make a good encyclopedia? Not always. Rossami (talk) 21:49, 2 June 2010 (UTC)

Agreed. That is why the proposal to change the wording is as follows: ""All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research". I think this meets everyone's concerns. --Gavin Collins (talk|contribs) 07:38, 3 June 2010 (UTC)

>SlimVirgin wrote: "we have never required that everything be sourced" ???? What about "Zero information is preferred to misleading or false information" and (WikiEN-l) insist on sources letters? "If it is true, it should be easy to supply a reference. If it is not true, it should be removed. I really want to encourage a much stronger culture which says: it is better to have no information, than to have information like this, with no sources. Any editor who removes such things, and refuses to allow it back without an actual and appropriate source, should be the recipient of a barnstar. --Jimbo"? X-romix (talk) 14:25, 3 June 2010 (UTC)

Gavin Collins wrote: "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research" - Yes, I'm agree with it. X-romix (talk) 14:25, 3 June 2010 (UTC)

Being "attributable" doesn't "show" anything on its own, since attributability is only a principle. This is why we mandate attributability (all material must be attributable to a reliable source), and if it's challenged, likely to be challenged, or quoted, it should be attributed to a reliable source, or else it may be removed. Crum375 (talk) 14:44, 3 June 2010 (UTC)

X-romix: the Jimbo quote was in relation to a not-very-plausible statement that had no source, so has no application to this discussion, which is about whether unchallenged statements can remain in an article without an explicit source. But that is no longer a problem, since your latest statement, "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research" [emphasis added] shows you agree with the current policy. Jc3s5h (talk) 14:45, 3 June 2010 (UTC)

We are agreeing with the current policy, no doubt. This form of wording is clearer, shorter and does not contain any ambivalent statements. --Gavin Collins (talk|contribs) 16:16, 3 June 2010 (UTC)

Gavin.collins: I have reverted your edit because I believe it is an effort to move the policy, against consensus, toward your personal view that every statement must be attributed with an inline citation, even when there is neither a quotation nor a likelihood of a challenge. You have deliberately removed a clarification/reminder of what "attributable" means. Jc3s5h (talk) 11:53, 7 June 2010 (UTC)

In what sense is it "against consensus" or "personal"? I think you are misrepresenting my position entirely. You have already stated above that you agree with this position. --Gavin Collins (talk|contribs) 12:00, 7 June 2010 (UTC)

The consensus is that although all claims must be attributable, the attribution need not appear in the article if the claim is neither a direct quote nor likely to be challenged. Your edit removed wording which helped make this meaning clearer to readers. Your change makes it easier for readers to be confused, and mistakenly think the policy conforms with your earlier statement, "The distinction between "attributable" and "unattributed" is meaningless in the absence of inline citations, which is why attribution should be provided for all statements of opinion or fact to verify their source and to demonstrate that they are not original research." Do you now agree that your older statement, which I quoted, no longer reflects your opinion? Jc3s5h (talk) 12:18, 7 June 2010 (UTC)

I don't follow you, Jc3s5h. You have gone on the record to say that this "shows you agree with the current policy", and now you say it is confused and mistaken. Lets make a side by side comparison:

Current version: "All material in Wikipedia articles must be attributable to a reliable published source to show that it is not original research, but in practice not everything need actually be attributed"

Proposed version: "All material in Wikipedia articles should be be attributable to a reliable published source to show that it is not original research".

The reading of this is for all intents and purposes is the same: a "must...but" statement is replaced with a "should" statement. Compare this with the version which you yourself edited back in Janurary[4]:

Older version: "Editors should provide a reliable source for quotations, and for any material that is challenged or likely to be challenged, or the material may be removed".

I might have missed a discussion or two in the interim, but its seems to me that this is pretty close to a version which you have endorsed in the past. As your rightly point out, my views have not changed: the "must...but" statement is confusing, and should be shortened. However, I am not understanding your position, because if the short wording is clearer and eliminates any ambivalence, I don't understand what material objection you have to the proposed wording. --Gavin Collins (talk|contribs) 13:30, 7 June 2010 (UTC)

Are you deliberately refusing to understand what people keep explaining over and over again, in the hope that eventually everyone will get so frustrated they'll go away and let you write what you want? It's perfectly simple and not at all confusing to people who can hold two different ideas in their heads at one time: everything "must" be attributable (no need to weaken that to a "should"); but in practice not everything needs to be attributed (attributed is NOT the same as attributable, so there is NO CONTRADICTION; and pointing this out potentially reduces confusion, not increases it, among reasonably intelligent readers).--Kotniski (talk) 13:51, 7 June 2010 (UTC)

You can drop the ad hominem attacks, they cast no light on these discussions. In fairness, it is not just me who objects to the current wording, and I am not objecting on purely spurious grounds either. The "must...but" statement is inherently contradictory to start with, so lets agree to eliminate it. --Gavin Collins (talk|contribs) 13:58, 7 June 2010 (UTC)

I just don't believe this. How many times do we have to tell you that the two words mean two different things, so whatever else this statement may be, IT IS NOT CONTRADICTORY. --Kotniski (talk) 14:11, 7 June 2010 (UTC)

I agree with you that the words "attributable" and "attributed" mean differnt things, but the statement is contradictory. So could we agree on alternative wording? I believe we were agreed that my alternative says the same thing, but because it does not resort to a "must...but" statement, I think it is both clearer as well as shorter. Although you have not said so, I think you would agree that there is a problem with the statement such as "You must not put your hand on a power line, but in practise its OK". I have a similar objection to the current wording, and so do other editors. --Gavin Collins (talk|contribs) 14:31, 7 June 2010 (UTC)

All we agree on is that your proposed version is shorter. It is not clearer except in that it ducks out of trying to explain what the longer version tried to explain, so for people who are not intelligent enough to understand the concept, I suppose it could be called "clearer", but hopefully they are not our audience. The power line statement is nothing to do with the issue - it's of a different form. I really think we've been through this long enough. (A better example would be "All power lines must be disconnectable, but not all of them need actually be disconnected." As you see, there is nothing even remotely contradictory about this.)--Kotniski (talk) 14:40, 7 June 2010 (UTC)

No, I have a fair point. It is just not consistent guidance to say that you (absolutely) must do this, or must not do that, and then add a qualification that turns the "must" into a "should". If you want to make clear and explicit execeptions, that is another matter, but a "must...but" statment is going to be challenged because it is ambivalent, even if you don't agree it is contradictory. Lets agree to change the wording now on the grounds that it will be challenged in the future and be done with it. --Gavin Collins (talk|contribs) 14:47, 7 June 2010 (UTC)

Gavin, here is another example. Imagine we say, "All cars must have doors, but not all car doors are open." Do you see anything contradictory in that? Crum375 (talk) 15:10, 7 June 2010 (UTC)

Nice try, but no go. I think what you meant to say is "All cars must have doors, but in practise, some don't." Lets face it, the fact that some open top cars don't have doors negates the first part of the sentence. Also, there is no regulatory requirement for cars to have doors, so it is not a good analogy. A better analogy is "All car drivers must have third party insurance, but in practise, some don't." We need to change it to "All car drivers should have third party insurance" just to make it clear that the uninsured should be insured. --Gavin Collins (talk|contribs) 15:29, 7 June 2010 (UTC)

Actually, WP:BURDEN is more along the lines of... "To drive a car the law says you must have a driving license ... if stopped by the police you must prove that you have a license, by showing it to the police. If you drive without a license your car may be impounded." Blueboar (talk) 15:44, 7 June 2010 (UTC)

Exactly. There should not be any qualification along the lines of "but in practice not everybody needs actually to be licenced". I am sure we can agree on wording to get rid the statement "but in practice not everything need actually be attributed", either as I have suggested or perhaps going back to an older version used in this paragraph a few months ago. --Gavin Collins (talk|contribs) 16:11, 7 June 2010 (UTC)

Rather than the car analogy, debt would be a better analogy. Something like "Lenders have the right to demand payment in currency, but in practice many allow payment with checks." Jc3s5h (talk) 16:31, 7 June 2010 (UTC)

Again, that is a misleading analogy, but this time because there is no "must" (absolutely) clause which is being qualified or watered down by a "but..." statement. --Gavin Collins (talk|contribs) 16:45, 7 June 2010 (UTC)

AH... now I see your confusion. I don't think WP:Burden is a "must...but" situation. It is more of a "must...and" situation. All information must be attriutable, and it must be attributed if challenged.

To belabor the drivers license analogy... Adding unattributable material to Wikipedia is like driving without a license... it is possible to do so, and you might get away with it if you are careful... but it is against the rules. Adding attributable material is like driving with your license in your wallet. You are following the rules as every good citizen should. Actually attributing material is equivalent to proving to the police that you have drivers license. In practice, you do not need to actually attribute unless challenged... just as you do not need to show drivers license to every cop you pass on the highway. You only are required to do so if pulled over (challenged). The difference is that our police (editors) can pull you over simply because they want to check that you have a license. Blueboar (talk) 17:41, 7 June 2010 (UTC)

Nearly there. Adding attributable material is not like driving with your license in your wallet, for there is no requirement to drive with your licence in your wallet - you cold keep it somewhere else. It is the licence itself that you must (absolutely) have, but a qualification along the lines of "but in practice not everybody needs actually be licensed" negates or diminishes the (absolute) requirement.

It is clear that a "must...but" statement appears to conflict with itself, because it sets up an absolute requirement on the one hand, but negates it with a generalisation on the other. What this policy should say is that "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research" without any qualification, or ambivalence. I propose we make this change to this form of wording. --Gavin Collins (talk|contribs) 21:23, 7 June 2010 (UTC)

As I understand it, the discussion is about whether or not the current wording says "You don't need a license" or "You need not have your license constantly on view." I've always understood it to be the second. --John (User:Jwy/talk) 22:04, 7 June 2010 (UTC)

That's exactly right. The analogy is that we must all own a driving licence. We needn't carry it around with us, but if someone asks to see it, we have to show it at the nearest police station asap. :) SlimVirgin ^{talk contribs} 22:18, 7 June 2010 (UTC)

No, because the "must" infers an absolute requirement. I am sure why the use of "must" was introduced in the first place, as Wikipedia policy is not law, which is another reason to change the wording. --Gavin Collins (talk|contribs) 08:33, 8 June 2010 (UTC)

The absolute requirement is for attributability, not for attribution. Do you not see it from Slim's example with the driving licence? The fact that we don't have to have our licence with us (or display it on the windscreen - depends on your local laws) in no way compromises the absolute requirement that we must hold a licence.--Kotniski (talk) 09:00, 8 June 2010 (UTC)

I think you are taking the analogy too far. In the real world, if you don't have your licence with you, you must present it to a police station if challenged, because the term "must have" is an absolute/legal requirement. In Wikipedia, there are no absolute requirements - hence the "must" statement has no teeth, in which case the "but" statement renders it totally useless. As a result, the qualificiation "but in practice not everything need actually be attributed" could provde a justification for keeping unattributed content that is original research. This is the point in my early analogy: some unattributable statements get a free pass as they are not required to actually be attributed "in practise".

Believe me, I have seen how this works in practice: see Kender OR Discussion for an example how one or more editors can deny that original research need not be attributed using the arguement that "in practice not everything need actually be attributed". Effectively this "but" statement can used as an excuse to prevent original rearch from being challenged or removed, which is why it must be changed. --Gavin Collins (talk|contribs) 10:03, 8 June 2010 (UTC)

I don't see what that discussion has to do with this one (except that then as now, you seem to be on your own). No-one in that discussion quoted the statement you're objecting to, and you didn't take the opportunity to quote any of the large number of maxims that this and other policies provide to those wishing to get dodgy claims cited or removed. Whatever the problem was in that discussion, I don't see how it was related to any problems with this policy's wording. --Kotniski (talk) 10:14, 8 June 2010 (UTC)

The point of the discussion was that "in practice not everything need actually be attributed" was being used as an excuse to retain unattributed content, even though it was challenged. When the case went to mediation, all of the unattributed content was was removed. If this point is lost on you, I am happy to bring other uninvolved editors into the discussion to consider the issues I have raised if you wish. --Gavin Collins (talk|contribs) 10:43, 8 June 2010 (UTC)

The Kender article given by Gavin.collins is an example of misuse of this policy on both sides of the argument. The usual crowd of editors at that article thought that since the article was about the Kender franchise of fiction games, there was an implicit citation to those games. While that might be good enough for a single book, an implicit citation to an entire franchise is a stretch. Gavin thought he could just slap a "Citations needed" tag on the whole article with no requirement to discuss which parts needed citations. A perfect example of a "discussion" where people just talked past each other without communication. Jc3s5h (talk) 10:42, 8 June 2010 (UTC)

The phrase "but in practice not everything need actually be attributed" that Gavin objects to was not in the policy as of April 2008 at the time of the dispute that Gavin uses to support his claim that the phrase should be removed. (Or at least, it was not there in the versions I checked, at the beginning, middle, and end of April.) Jc3s5h (talk) 11:04, 8 June 2010 (UTC)

This is true, but you can see how similar arguements were being used to frustrate cleanup. Since it is impossible to prove that unattributed content is original research, any challenge that is made that will flounder if policy says "in practice not everything need actually be attributed". In this case, policy could be used to deny the existence of original research. --Gavin Collins (talk|contribs) 11:33, 8 June 2010 (UTC)

The results of this discussion are:

• Gavin.collins' proposal to remove the phrase "but in practice not everything need actually be attributed" has been rejected
• Gavin.collins has demonstrated he will never stop discussing this
• Any removal of the phrase from the policy by Gavin.collins should be reverted without further discussion or explanation. Jc3s5h (talk) 12:25, 8 June 2010 (UTC)

Simple suggestion

May I make a simple suggestion? We're starting with this:

All material in Wikipedia articles must be attributable to a reliable published source to show that it is not a Wikipedian's original research, but in practice not everything need actually be attributed. This policy requires that anything challenged or likely to be challenged,...

If we apply some simple copyedits, we can get this:

All material in Wikipedia articles must be attributable to a reliable published source to show that it is not a Wikipedian's original research. All material does not need to be attributed, but this policy requires that anything challenged or likely to be challenged,...

Simply by splitting up the phrases, you remove the "must .. but" approach without weakening the policy, but not connecting it with the next line emphases the "show your license" concept to avoid OR inclusion. --MASEM (t) 12:33, 8 June 2010 (UTC)

Sounds OK to me (but "All material does not need" -> "In practice not all material needs").--Kotniski (talk) 12:40, 8 June 2010 (UTC)

It still looks odd becasue the statement "All material does not need to be "attributed" still sticks out like a sore thumb in a sentence that goes on to require in line citation. Here how the full text would read:

All material does not need to be attributed, but this policy requires that anything challenged or likely to be challenged including all quotations, be attributed to a reliable source in the form of an inline citation, and that the source directly support the material in question".

I do see conflict, in the sense that why would you say that all material does not need to be attributed on the one hand, and then say all quotations are required to be subject of inline citation? If anything, this amendment highlights the absurdity of saying "All material does not need to be attributed". --Gavin Collins (talk|contribs) 14:19, 8 June 2010 (UTC)

In this form it's saying: "Some X are not Z, but all X that are Y must be Z" where X is "statement", Y is "statement that may/has been challenged" and Z is "attributed". It's a logically clear statement. --MASEM (t) 14:25, 8 June 2010 (UTC)

Now you're equating "quotations" with "all material", which is even more absurd than equating "attributable" with "attributed". Do you know what "quotations" means?--Kotniski (talk) 14:35, 8 June 2010 (UTC)

Lets say for arguments sake that you are both right, and that it is a logical statement. Taking that aside, it still looks odd in the sense the requirement is first relaxed and then tightened in the same sentence. Surely you can see that they don't sit comfortably together, at least in terms of presentation? --Gavin Collins (talk|contribs) 14:44, 8 June 2010 (UTC)

OK... what about:

• While there are circumstances in which material does not need to actually be attributed, all material needs to be attributable to a reliable published source to show that it is not a Wikipedian's original research. Note that this policy requires that anything challenged or likely to be challenged,...

Is this acceptable? Blueboar (talk) 15:02, 8 June 2010 (UTC)

To be honest, I would much rather see the statement "All material does not need to be attributed" removed altogher because it is counter-intuative with the context of a policy dedicated to verifiability. It simply does not sit well within this policy, and I would go further bay saying it actually conflicts with the requirement for inline citation for quotations. If we can't agree on my proposal, then could we go back to the version without this statement? The originator of the current version is SlimVirgin in April of the year[5]--Gavin Collins (talk|contribs) 15:16, 8 June 2010 (UTC)

I think if people reading this policy have the kind of brains that can't cope with the idea of "A but B" - that the things we say sometimes need exceptions and qualifications to make them fully accurate - then we might as well give up. I see absolutely nothing wrong with any of the suggested wordings - even yours wouldn't be disastrous - but I don't see why we should make our attempt at explaining things a little worse just because one editor keeps haranguing everyone about it with obviously spurious arguments. All the time we've wasted discussing this we could actually have spent making this policy (or other worthy things) better.--Kotniski (talk) 16:33, 8 June 2010 (UTC)

Despite your complaints, there are valid points in my arguments, and your unwillingness to acknowlege them seems to be the source of your annnoyance. I think you could do more to assume good faith, in fairness. --Gavin Collins (talk|contribs) 08:23, 9 June 2010 (UTC)

(hollers out) what about a footnote like below? I must say I would much rather a footnote explaining obvious cases - I personally am uneasy with saying anything need not be attributed, as that runs counter to policy. But I would say that examples where inline referencing is left off are obvious ones which all editors agree could be referenced very easily. Casliber (talk · contribs) 00:28, 10 June 2010 (UTC)

It does run counter to policy, and this has not been acknowledged so far. The only reason I can think that is resistance to changing the current wording may be a (legitimate) concern that in doing so, we reach a point where fundamentalist editors (if there are such a group) will demand that everything be attributed with inline citation. If this is the case, then we need to bring this out into the open now. In the meantime, I have asked SlimVirgin why he made this change in the first place[6]. --Gavin Collins (talk|contribs) 08:26, 10 June 2010 (UTC)

How about a footnote?

I've come late to the party and am a bit goggle-eyed after trawling through the above. Maybe having:

All material in Wikipedia articles (should/must) be attributable to a reliable published source to show that it is not original research.

BUT with a footnote after the "research" - in this way the general flow and caveat for referencing is preserved, and the footnote can discuss common-knowledge statements (eg "sky is blue") where we needn't be so proactive about referencing. Then one can expound upon details there. Casliber (talk · contribs) 03:59, 9 June 2010 (UTC)

I would be happy with "All material in Wikipedia articles should be attributable to a reliable published source to show that it is not original research". The "should" relaces the "must" statement, as "must" is far to perscriptive, in which case there is no need to qualify the statement. --Gavin Collins (talk|contribs) 10:32, 14 June 2010 (UTC)

That completely weakens WP policy. All material must be attributable to sources; that's a basic principle of WP. --MASEM (t) 12:44, 14 June 2010 (UTC)

It does not in fairness; the current wording is ambigious because of the "but" qualification. The "must....but" statement has only crept in since April; before that policy has always been "should". --Gavin Collins (talk|contribs) 06:52, 15 June 2010 (UTC)

I chose a random revision (presuming no vandalism) before April [7] and clearly there's no "should". Reading what's there certainly implies "must", as well as what's stated in WP:5; the present chance only helps to clarify between attributable and attributed. Plus it is plain common sense given every other policy that we have - we aren't here to publish falsehoods and lies or stuff made up by an editor; information must be verifiable across the board.

Let's put it this way. If it was the case that "information should be attributable...", what are the cases or examples or the like where we would allow unverifiable/unattributable information on WP? --MASEM (t) 14:23, 15 June 2010 (UTC)

Good name for proposal please

Just a quick question, not a debate on the topic, but does 'Summary statements' sound like a good description of what was talked about in a recent discussion here #Two copies of a statement must each have their own inline citation? about a statement in an article that is based on statements or a subsection further down in the same article? Or can you think of a better name please? I'd like to get the title right before trying to write something up about it and proposing it properly at the Village Pump thanks.. Dmcq (talk) 13:54, 15 June 2010 (UTC)

Citations in the lead

In WT:Manual_of_Style_(lead_section)#Citations_in_the_lead (discussion now centralized here) I am disputing the requirement to stick in-line citations into the lead of articles where they are summarizing a part of the article. The statements would still be cited and verifiable if the summary is accurate so I'm saying one should instead have a template saying a summary statement is not an accurate summary rather than immediately trying to drill down a level to sources. Dmcq (talk) 10:44, 10 June 2010 (UTC)

Indeed we have little if any infrastructure for the support of accurate summary-style ledes. Very often eager new editors will "start at the top" of an article. Given that nothing says "this article uses summary style", they are quite justified in inserting material in the lede (with cites) irrespective of whether there is a more extensive version in the body. We really should establish firmer guidance on this.LeadSongDog come howl! 12:17, 10 June 2010 (UTC)

My understanding is that cites in the lede are optional:

• If you do use them, you need to be as consistent with them as you are using for body cites.
• If you don't use them (which is not a bad thing), you are still required to cite direct quotes. You also may be challenged if your lede includes statements and OR not directly spelled out in the summary. --MASEM (t) 12:21, 10 June 2010 (UTC)

"Summary style" refers to an article that consists of summaries of other Wikipedia articles, with links to those articles. It has been argued that summary style articles don't need their own citations, and can rely on the citations of the other articles. But that is not what we are discussing.

The lede is a special case of the principle that a claim need not be cited every time it appears in an article; once is enough. A special template that says something like "This claim in an article lede is not supported by properly sourced statements later in the article" would be useful. Jc3s5h (talk) 12:40, 10 June 2010 (UTC)

I can't see the lead paragraph being in any way different from the rest of an article. The distiction between one part of an article and another is the subject of style (WP:LEAD), not by content policy. In some articles, it is not possible to distinguish where the lead ends and body begins. We can't exempt the lead from inline citation, because like all content in Wikipedia mainspace, if it is going to be challenged, it need to be attributed in accordance with WP:CITE. --Gavin Collins (talk|contribs) 14:00, 10 June 2010 (UTC)

The current state of this discussion, combined with existing policy, is that statements in the lead that are neither quotations nor likely to be challenged need not have a citation, although it must be possible to find a citation if a challenge is issued. Also statements in the lead that are likely to be challenged must have a citation, but if the statement is substantially repeated later in the article, the citation may be located at the later instance of the statement. Finally, all quotations require a citation at the point where the quotation occurs. So, Gavin.collins, is there anything in my summary of the state of the discussion that you would regard as an unacceptable exemption? Jc3s5h (talk) 14:16, 10 June 2010 (UTC)

Why should a statement be challenged in a lead if it is properly cited in the article and the lead is a good summary? 'Likely to be challenged' should include some willingness on the part of the challenger to have at least a superficial look at the article. As to stub articles which don't have a body one would of course have to stick the citations on the statements because they would not be summarizing anything. Dmcq (talk) 14:20, 10 June 2010 (UTC)

When I wrote "statements in the lead that are likely to be challenged" I meant that the substance of the statement is likely to be challenged, so of course, the challenge would apply to every instance of the statement, whether it appears once, twice, or a dozen times in the article. The requirement, as I understand it, is that at least one instance of the statement requires an inline citation, but the remaining instances do not require a citation. If the challenge were sustained and the statement found to be invalid, every instance of the statement would have to be removed from the article. Jc3s5h (talk) 14:28, 10 June 2010 (UTC)

(edit conflict)Gavin has hit the nub of the issue. WP:V is a core WP:POLICY while WP:LEAD is only part of a guideline. Without actual citations in the lede, there is no effective tracking mechanism in place (short of an intensely commited reviewer) for mapping lede assertions to the corresponding assertion in the body in order to check how completely (or whether) the lede's assertions are supported by sources or merely disguised WP:SYNTH. Particularly for lengthy articles that are extensively cited in the text, an uncited lede is nearly impossible to review. Given the difficulty we have getting and keeping reviewers, we need to do what we can to make that job more straightforward. Removing this poorly-justified practice of uncited ledes (at least for the better classes of article) would remove a significant hurdle. LeadSongDog come howl! 14:37, 10 June 2010 (UTC)

Surely it's easier for a reviewer to check whether something is in the article than to check whether it's in some (quite possible inaccessible) external publication? --Kotniski (talk) 14:53, 10 June 2010 (UTC)

The Verifiability policy does not require that every instance of a statement (other than a quote) have an inline citation; if a citation is required at all, providing it for one instance satisfies the policy. The lead sentence of Common Era, "Common Era, abbreviated as CE, is one of the designations for the world's most commonly used year-numbering system" has been challenged and thus has a citation. If we decide that every instance of a challenged statement must have a citation, then we must go through the encyclopedia and provide a citation for every single instance of the abbreviation "CE".

If LeadSongDog wants to advocate for a system to make it easier to find where a statement in the lede is discussed in further detail later in the article, that's fine, it could be a useful readability or editing enhancement. But it is not required by policy. Jc3s5h (talk) 15:14, 10 June 2010 (UTC)

Is it not possible to have this discussion only in one place?--Kotniski (talk) 14:30, 10 June 2010 (UTC)

I see nor reason to partially exempt leading paragraphs from WP:CITE, which is what this proposal is essentially advocating. The source for a lead paragraph must come from somewhere, so it is not unreasonable to provide citations for the those sources. I think Jc3s5h is arguing that citation is not necessary on the basis that the lead is a summary of the main body of the article. However, I think he may have forgotten that writing articles involves researching the most reliable sources on the topic and summarizing what they say (unless the citation is a verbatim), and the lead paragraphs do not differ in this respect. Mainbody or lead, it is good practice to cite the source from which the summary has been derived. --Gavin Collins (talk|contribs) 16:38, 10 June 2010 (UTC)

Is this not the same argument of "attributable" vs "attributed" above? Moreso, citing everything is treating our readers as lazy who need to be pointed exactly where information came from. An encyclopedia is a research tool, it is not doing your work for you. Just as "attributable" is required but "attributed" only necessary when there could be question of where the info comes from, the lede can remain free of citations as long as no new information is needed. If someone is looking to research a topic and find more sources, they aren't just going to read the lede and call it good, they're going to read the whole article. --MASEM (t) 16:44, 10 June 2010 (UTC)

Gavin.collins, you have misrepresented what I advocate, and have failed to address the question of whether two instances of the same statement both require an inline citation (that is, two copies of the same citation), so your position is incomprehensible. I advocate writing a lead that consists exclusively of statements that

• are quotations and have an inline citation
• are not quotations and are not likely to be challenged, so they don't need any citation
• are likely to be challenged, and are clearly supported by citations, but the citations are placed in the area of the article that discusses the statement at greater length.

Jc3s5h (talk) 17:54, 10 June 2010 (UTC)

It looks like the most discussion is here instead of where I was intending so I'll try to centralize here. I'll copy over just a couple of things from the other talk page and put a note there about here. Dmcq (talk) 16:50, 10 June 2010 (UTC)

The main points I had in the other talk at WT:Manual_of_Style_(lead_section)#Citations_in_the_lead were:

There are a number of problems with the current scheme.

• It divorces the lead from the article in as much as instead of describing a part of the article it is describing what is in a source instead.
• Citations get duplicated meaning people must learn about named citations or have a mess at the end.
• The leader tends to have a long list of citations after individual statements since they summarize a number of sources.
• It makes the leader stilted and less approachable instead of being an easy introduction into the article.

These problems are exacerbated by the number of drive-by editors going round sticking on every statement that doesn't have a citation without doing even minor reality checks like looking at the content list and seeing if the statement is describing a well supported section of the article. Since the lead is the first thing they see they stick them in there.

There was an anti from Gavin.collins saying there was problems with Accountancy till they ensured citations in the lead and a pro from Kotniski saying the lead should be form the text and what we do now is wrong headed. More text and some responses from me there Dmcq (talk) 17:04, 10 June 2010 (UTC)

There's another related idea for supporting summaries properly which perhaps might be better put here too. This would support both summary style splitting or articles and within article summaries like in the lead or at the top of some subsections.

I think I'd want two templates

• {{summary disputed}} where someone thinks the summary is plain wrong or misleading
• {{summary improve}} where it just doesn't summarize reasonably well

Plus a type of summary citation which is stuck at the end of a summary sentence and refers to what's summarized. For summary style that is done already but within articles there's no real indication that a statement summarizes something further down. One can of course put in an ordinary link but I think something better is needed. Then a lead could have something like 'Monsoons in Mojimbe regularly kill thousands of people.^[Monsoons] ' which referred to the section 'Monsoons' instead of listing various citations from that section. Dmcq (talk) 17:22, 10 June 2010 (UTC)

Those are interesting objections, but I think they are somewhat ill-conceived. Citing an assertion is not "describing what is in a source". Rather, we use some of what is in a source and give credit where it is due. In the process we allow readers and other editors to verify (if they care to) that we haven't somehow misrepresented or cherry-picked the source. It is not necessary that every editor understand named citations or list defined references, just that some editor does. To a significant extent automation could help with the consolidation of duplicated refs. The "long list of citations after individual statements" is a hallmark by which we can easily recognize WP:SYNTH when it happens. "Stilted" is the only argument that seems to me to have any real currency. Yet if an idea is important enough to be asserted in the lede, it should be supported in one of the sources used. WP:Brilliant prose is still possible. With regard to drive-by tagging, yes, it can be momentarily frustrating, but in practice it does represent a (small) positive contribution that leads us to improve our work. LeadSongDog come howl! 17:50, 10 June 2010 (UTC)

As a curious statement, if we're using sources to give proper credit for a statement, then a statement that's made across several sources where no single source can be credited (eg reporting of news events, etc.) should not be inline-cited directly and considered immediately as non-contentious. This of course gels with the idea of quotes and contentious statements - those presenting novel OR from secondary sources or the like - should be inline cited. --MASEM (t) 18:23, 10 June 2010 (UTC)

Err, if I'm reading you correctly that's not what wikipedians usually mean when we capitalize "OR", especially in this venue. Did you perhaps intend to emphasize the conjunction "or", as in "those publishing novel research (primary sources) or reviews (secondary sources)"? LeadSongDog come howl! 19:40, 10 June 2010 (UTC)

I meant "novel original research" but abbreviated that to OR. --MASEM (t) 22:04, 10 June 2010 (UTC)

Not sure what that last statement was about so I'll just address LeadSongDog's point about synthesis. A summary in a leader very often has to be a little bit of synthesis. Otherwise we can't say things like 'A number of studies have shown smoking causes lung cancer' in a leader. Such statements currently have to be followed by a long list of citations because of someone sticking citation needed after them. It would be far better to point to the section of the article that lists the various studies and be able to say summary disputed on it if it wasn't a reasonable summary.

By the way the Accountancy article which was given as an example where the leader has been improved by requiring the statements in it be all cited is in my view a very good example of where the lead has been divorced from the article - the first item on my problem list. If the complaints about the lead had said it was not accurately summarizing then perhaps it would by now conform to WP:LEAD instead of going its own sweet way and ignoring the body of the article. Dmcq (talk) 19:50, 10 June 2010 (UTC)

Using a Accountancy, which is a C-class article instead of an FA for your example doesn't really exemplify what we think our best work should be. Still as I said above, WP:LEAD is only part of a stylistic guideline, while WP:V is policy that goes back to the origin of wikipedia. The former must be made compatible with the latter, not the other way round. LeadSongDog come howl! 20:15, 10 June 2010 (UTC)

I just looked at a 10 featured articles at random and I'm pleased to say that 6 of them had no citations at all in the lead and they were at most 8 in the rest and that was in a fairly long lead. Perhaps I've been lucky but it looks on that basis that featured articles are escaping this plague of {{cn}}s in the lead. Your point about WP:V is why there is this discussion here. The purpose of Wikipedia is to produce a free reliable encyclopaedia that anyone can edit. I believe that by dealing with summaries better we can improve the encyclopaedia bit whilst not impacting the reliability bit. Dmcq (talk) 20:45, 10 June 2010 (UTC)

The topics inspected by Dmcq must have been very boring or uncontroversial. Where a topic is the subject of competing points of view (e.g. Creationism), it is very hard to write a lede devoid of opinion that don't need to be backed up by citations. My view is that a lead has to contain statements of opinion if it is to provide wider context to the reader. For instance, in the article Accountancy, there has been some debate as to whether the article title is the correct one, as the term Accounting is more widely used in everyday speach. This debate is far from over, but it has since subdued since well sourced definitions have been found for both terms. It seems to me that to establish the "credentials" of any topic (i.e. a definition and/or explanation about its significance), a well written article will have to cite some form of expert opinion in its lede in order to provide context to the reader. In my experience, any opinion expressed in the leading paragraph will almost certainly be challenged, so I don't think is good guidance to say that in line citation is not required.

If it were possible to lead articles with simple bland statements of fact on their own, such as "Paris is the capital of France", then this proposal might work. However, I think it is naive to think that any decent article can be written without statements of opinion, and hence I don't think it possible to provide guidance that says that in line citations are not required in the opening paragraphs, unless lede are going to be stripped of opinions. --Gavin Collins (talk|contribs) 21:36, 10 June 2010 (UTC)

As long as the statements of opinion necessary to include the lead are summary of cited statements of opinion later in the article, there's no issue, because we're not introducing any new original research by summarizing existing content in the lede. If someone is putting a completely new opinion from a secondary source in the lede, then yes, but that's probably bad form more than anything else. --MASEM (t) 22:03, 10 June 2010 (UTC)

• How would you know that? Without in line citation, you have no idea that the lead is summarizing a particular source. --Gavin Collins (talk|contribs) 04:46, 11 June 2010 (UTC)

Example: no citations in lede

The article Humpback whale contains, in the lede, the statement, "During the winter, humpbacks fast and live off their fat reserves." I don't think that is obvious or well-known, so a citation is required. Later in the article, we find

Humpbacks feed primarily in summer and live off fat reserves during winter. They feed only rarely and opportunistically in their wintering waters...^[21][22][23]

I don't have copies of the sources, but suppose for the sake of discussion that they adequately back up the longer statement. Clearly the shorter statement in the lede is a reasonable summary of the longer statement. So does anyone contend that the presence of the longer statement in the lead, without an inline citation in immediate proximity, is a violation of the verifiability policy? Jc3s5h (talk) 22:26, 10 June 2010 (UTC)

I'm sorry, I haven't read this entire section, but in brief there is no exception for citations in the lead, as WP:LEAD makes clear. Anything challenged or likely to be challenged, and all quotations, needs a citation, no matter whether in the lead, in the body, or in captions. SlimVirgin ^{talk|contribs} 22:59, 10 June 2010 (UTC)

So you contend that if the same, or substantially the same statements, appear in more than one spot in an article, there must be an inline citation at every spot? If so, your recent revert to the policy is now revealed as surreptitious and major policy change which must be challenged. — Preceding unsigned comment added by Jc3s5h (talk • contribs)

Yes, if the lead says "Scotland is X miles in diameter," it needs a reference. The ref can be repeated the next time it's mentioned, if it's mentioned again, using ref name=. I can't see the problem with that. A lead is not an article abstract.

If people's concern is citation clutter, I wholeheartedly agree, but that's a separate issue of where to place footnotes: placing them at the end of paragraphs and avoiding the use of citation templates reduces clutter for both readers and editors. See WP:CLUTTER. SlimVirgin ^{talk|contribs}

Jc, I don't see the connection. SlimVirgin ^{talk|contribs} 23:35, 10 June 2010 (UTC)

The connection is that the version of the policy before your changes acknowledged "Alternative conventions exist, and are acceptable if they provide clear and precise attribution for the article's assertions". The presence of substantially the same statement in an article, which has a proper inline citation, would constitute such an alternative convention to a person such as yourself, who seems to think that every copy of a statement in an article is a separate entity which needs to be argued about, cited, or deleted separately from any other copies. Of course, I never held that view, and always thought all copies stand or fall together. Jc3s5h (talk) 23:55, 10 June 2010 (UTC)

My concern is more than just the clutter. I think that putting citations in the lead detracts from and interferes with its purpose to summarize the article. I want to see more infrastructure to support that a statement is a summary so people will check that a lead section or text at the top of a number of subsections is actually a reasonable summary rather than just a verifiable statement in itself with no relation to anything else. Dmcq (talk) 23:40, 10 June 2010 (UTC)

A lead is not only a summary. It can include material not repeated in the body, and generally acts as an introduction to the article. Unlike an abstract, a lead also has to justify the subject's notability, and create an interest so a reader would be enticed to read the rest of the article. Therefore, although a lead does include a rough summary, it is not exclusively so, and it is not exempt from the sourcing requirement. Crum375 (talk) 23:50, 10 June 2010 (UTC)

Dmcq, it's a misunderstanding of a lead that it's like an article abstract in an academic journal. It's not intended to be that. SlimVirgin ^{talk|contribs} 23:53, 10 June 2010 (UTC)

I agree with Crum375 to a certain degree. The first sentence in Tropical year (which I wrote in its present form) is such an elementary definition of tropical year that it need not be repeated later in the article, and yet, since there are so many definitions to choose from, the inline citation at the end of the paragraph is mandatory in order to demonstrate the validity of the definition. However, some of the other statements in the paragraph, such as "...currently define it as the time required for the mean Sun's tropical longitude (longitudinal position along the ecliptic relative to its position at the vernal equinox) to increase by 360 degrees" are substantially duplicated with citations later in the article; if all the statements were like that, no citation would have been needed in the lede. Jc3s5h (talk) 00:04, 11 June 2010 (UTC)

Ledes (on developed articles, natch) are not like abstracts - which from an academic POV set the stage for what the reader should expect in the article - but instead they should act like executive summaries, the 60 second blurb that if I need to understand a topic quickly I can read and access it and know that if I had more time, what I could expect to find in the body of the article. It needs to be simple and avoid introduction of new material (particularly contentious) that gets in the way of quick comprehension. Citations can get in the way of comprehension as well if they are used as frequently as is generally demanded in bodies of articles. --MASEM (t) 00:20, 11 June 2010 (UTC)

Leads do often act as a kind of mini-article, and therefore do require their own references. Many people, myself included when acting as a reader, use Wikipedia primarily to get a quick overview of a topic and its best sources. Yes, we can go to Google, but hopefully the WP sources have been vetted a bit. If we send such readers to the body to search for the sources, we lose much of the advantage of the lead. Bottom line: if we present material, it should be well sourced, esp. if challenged or likely to be challenged, and the sourcing has to be inline so that a reader would have an easy job retrieving it. Sending the reader to find the sources by forcing him to dig through a complex body is not reader-friendly. Crum375 (talk) 00:27, 11 June 2010 (UTC)

I agree that Wikipedia ledes should be executive summaries, but at least in electronics, abstracts are executive summaries rather than stage-setters. Indeed, if you send in a stage-setter abstract to a conference, you will not be speaking at that conference. Jc3s5h (talk) 00:29, 11 June 2010 (UTC)

Crum375, your argument that it is user-unfriendly to write ledes that require a reader to search through the article body to find a similar statement, and then consult the citation for that statement has some merit, from the point of view of good writing. However policy does not forbid it. Compare two cases. A, I find a surprising statement in a lede, spend three minutes finding the equivalent statement in the article body, and then 5 minutes checking the online source. B, I find a surprising statement in a lede, accompanied by an inline citation to a book. I drive to the library (20 minutes), fill out an interlibrary loan request (5 minutes), wait six weeks for the book to arrive, drive back to the library and check out the book, verify the citation, and return the book. B is clearly in full compliance with policy, but is much more difficult than A. So A might be a little more difficult than really necessary, and so be sub-optimal writing, but it isn't so difficult that policy should forbid it. Jc3s5h (talk) 00:39, 11 June 2010 (UTC)

Abstracts and ExecSums serve very different needs. Abstracts are written to actually draw you into reading the article/attending the presentation, and while usually establish what will be talked about, don't give anything away; I wouldn't call it "Advertizing" but it is meant to entire further reading. ExecSums are designed to be a replacement for reading the article in the interest of time, and thus need to describe what's contained, including results and conclusions, as a replacement for reading for, well, "execs", while still being a useful determinate for others to read further. Our ledes should be closer to the later, summaring the major structure of the article and hitting the major points for the fast read. --MASEM (t) 03:52, 11 June 2010 (UTC)

Many sources are online, just a click away. So if I read a lead and see something of interest, I like the convenience of being able to click and start reading the source. For me in many cases WP is merely a gateway to the real sources, so after getting a quickie intro to the subject I click and read the real stuff. But if the lead is bereft of sources, I need to wade through a sometimes complex body to find the justification for that interesting statement I saw in the lead. Sometimes I'll find it quickly, while other times it may be buried inside other stuff, or phrased differently in the body, making it hard to find. Why torture the poor reader? Because we are too lazy to add <ref name=X/> in the lead? Sometimes we forget our customer is the reader, not the editor. Crum375 (talk) 00:53, 11 June 2010 (UTC)

OK, but suppose I come across an article with some statements in the lede with no inline citations, and the statements are indeed supported by citations later in the article. If I have the time to add citations in the lede, I could do so, although opinion is mixed about whether that is good writing. If I lack to the time to do that, I certainly should not apply {{cn}}, because that would incorrectly indicate the article contains no citation to support the statement. So my choices would be to do a complete fix (at the risk of getting some flack), or do nothing. Jc3s5h (talk) 01:27, 11 June 2010 (UTC)

I think we need to focus on the reader, not the editor. And the reader would clearly benefit from not needing to wade through the body to find the source. If the editor doesn't have the time to add <ref name=X/> in the lead, he can do it later, when he finds the time, or when someone challenges it. You are under no obligation to add {{cn}} to your own material. What I normally do is to add references for the most important or critical points in the lead, and then if other people challenge more points, I can add their references too. Crum375 (talk) 01:37, 11 June 2010 (UTC)

I feel there's been a Verifiability Full Moon in the sky the last couple of weeks. We have people insisting that every single tiny thing be sourced (though there's never been consensus for that) and others arriving to claim there's no consensus for inline citations (though it's self-evident to anyone who edits a lot that the consensus for them is strong). Guys, the sourcing policy needs to be stable. The concepts need to be stable. The writing needs to be clear. The footnotes needs to be non-contradictory. The advice needs to reflect what happens on the ground in a best-practice sort of way without being horribly idealized.

I happen to think we have the balance more-or-less right, which doesn't mean the policy can't be improved, but improvement needs to happen in a calm and considered way. SlimVirgin ^{talk|contribs} 00:13, 11 June 2010 (UTC)

Perhaps you should read the discussion more fully. Elsewhere it was explained that many featured articles do not have any citations in the lede, so the consensus among good working editors seems to be that when a longer statement with an inline cite appears in an article, a summary of that statement, without a separate inline citation, may appear elsewhere in the article. Your claim to the contrary tends to disrupt the stability of the policy. Jc3s5h (talk) 00:29, 11 June 2010 (UTC)

I also write FAs, and I have citations in my leads when they're needed, as with any other part of the article. The person who wrote the advice about citations for WP:LEAD was one of the featured article review delegates. The page that reflects best practice for sourcing at FAs is Wikipedia: When to cite, and it repeats the advice in WP:LEAD. SlimVirgin ^{talk|contribs} 01:56, 11 June 2010 (UTC)

• I strongly support a rule discouraging citations in the lead. Citations in the lead are a making the lead unreadable, especially in controversial articles. The lead should be verifiable against the article itself, without need for any additional citations. We should not allow the lead to contain any information not present in the article itself, and therefore citations in the lead will always be redundant. Marokwitz (talk) 10:45, 16 June 2010 (UTC)

But if the sources in the lede are different from those in the body of the article? This line of thinking is too generalised. In practice, a lede can contain many facts or statements of opinion about a topic in general that are not repeated in the more detialed sections that follow. --Gavin Collins (talk|contribs) 10:56, 16 June 2010 (UTC)

I think that we should discourage that. The lead should summarize the article, meaning that all sentences in the article should be verifiable as if they were cited to the article itself. I believe that in any article larger than a stub, everything written in the lead should be later repeated in the article. Akin to an academic abstract. "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. It is even more important here than for the rest of the article that the text be accessible. Consideration should be given to creating interest in reading the whole article..... Where the article is a stub, a lead may not be necessary at all. ". I see no contradiction between the requirement to make an interesting and enticing lead, that demonstrates the notability of the subject, and the proposed style guide modification , that is, in any article longer than a stub, everything in the lead should be repeated in the article, and no citations should be used in the lead. I challenge you to find such an article where implementing the above proposal will result in a worse article. Marokwitz (talk) 11:23, 16 June 2010 (UTC)

How can you discourage what is best practice? If the lede discusses a topic at a high level, and the body of the article contains more detailed analysis, then it would actually be better to use different sources, e.g. introductory level sources in the lede to provide a summary, and more advanced sources that provide coverage in the body of the article at a detailed level. A good example is the article Enron Scandal, where the leding sources are entirely different from the main body for this reason. --Gavin Collins (talk|contribs) 11:37, 16 June 2010 (UTC)

That's not true, for example, the same book cited in the lead (Enron and the Dark Side of Shareholder Value) is cited many times across the entire article. Same for ""The Quality of Corporate Financial Statements and Their Auditors before and after Enron" which is used in 2 other places in the article. Same for the citation "A Market Proposal for Auditing the Financial Statements of Public Companies" which is also used later. The only citation in the lead that isn't used in the main body is #4. I disgree with you that this is best practice. Marokwitz (talk) 11:50, 16 June 2010 (UTC)

I think what you mean by source is the publication; what I mean by source is the facts and opinions from the publication. The lede contains facts and opinions not expressed elsehwere in the article. This is why inline citation is so important: in order to verifiy a source, we can't expect the reader to sift through the lede and try and match it with the source. This is an encyclopedia, not a game of hide and seek. --Gavin Collins (talk|contribs) 12:34, 16 June 2010 (UTC)

WP:SPS

RE this revert by Kww: Please see the discussion at WT:IRS After some discussion it became clear, to me at least, that the insistence in WP:SPS on separate third-party-published material in order to establish expertise (so as to allow an exception to the prohibition of SPSs) is somewhat arbitrary, and sometimes contradictory to the fundamental definition of a reliable source. There exist self-published authors and websites, two examples of which are given at WT:IRS, for which it can be clearly demonstrated that they have a "reputation for fact-checking and accuracy". So the key to establishing valid exceptions to the general prohibition of the use of SPSs is in demonstrating expertise and more centrally "a reputation for fact-checking and accuracy". Of course, if a self-published author has other third-party-published works to her or his credit, all the better in terms of establishing reliability. ... Kenosis (talk) 00:06, 15 June 2010 (UTC)

It's the criterion that's been used for a long time, and I've seen it often be decisive in resolving a dispute. The idea is that if a person can't find anyone but themselves to publish their material it might signal a problem, or at least a lack of widely acknowledged expertise. Therefore we require that a self-published source at least be someone whose work in the field has previously been published by an independent reliable source. I know this does catch some sources in areas like pop culture, so editors have to engage in a bit of IAR sometimes, if those sources are the only ones. But in mainstream areas where other sources are available, it would be hard to argue that we should use someone who has only ever been published by himself. SlimVirgin ^{talk|contribs} 01:03, 15 June 2010 (UTC)

By your chosen example, it would appear you didn't read the discussion to which I linked. As it happens, one of the examples at WT:IRS (drawn from the noticeboard) involves an anonymously published website relied upon by the CIA and others who are heavily dependent on reliability of their sources; the other involves an author who has been acknowledged to be an expert in other RSs but hasn't published anything but the website material. Inevitably we're going to run into more instances of this kind as time moves forward. ... Kenosis (talk) 01:43, 15 June 2010 (UTC)

Could you give those examples here, so that people don't have to read the other discussions? (And, by the way, if the CIA is relying on a website e.g. one known to be operated by al-Qaeda operatives, the source is the CIA, or whichever news organization tells us the CIA uses it, not the website). SlimVirgin ^{talk|contribs} 02:11, 15 June 2010 (UTC)

SlimVirgin: I think that ultimately what matters most is whether the source has a reputation for accuracy and fact-checking. Occasionally, there are sources which are SPS and have earned such a reputation. I think the quintessential example of such a source is Snopes.com. It's an SPS and has an excellent reputation for accuracy and fact-checking of urban legends. Of course, urban legends falls under pop culture, but I don't consider using Snopes.com as an example of IAR. A Quest For Knowledge (talk) 02:17, 15 June 2010 (UTC)

I think WP will be in big trouble if we base our sourcing policy on pop culture as the lowest common denominator. Crum375 (talk) 02:30, 15 June 2010 (UTC)

I don't agree that we are lowering are standards at all. I see it as a re-enforcement. At its most fundamental core, what matters most is whether a source has a reputation for accuracy and fact-checking. Everything else is secondary. It doesn't matter what your editorial policy is or if you have professional staff if you have a poor reputation for accuracy and fact-checking. A Quest For Knowledge (talk) 02:57, 15 June 2010 (UTC)

Not exactly. What matters most for reliability is that a source has a reputation for accuracy and fact-checking as judged by third party reliable sources. Third parties do this by publishing the material from this source, or if the source is a single person who is a published expert in the field, we accept his self-published material because his previous publication by others counts as third party verification. But Wikipedia editors alone don't count as third parties by definition. Crum375 (talk) 03:11, 15 June 2010 (UTC)

Actually, I don't disagree with what you said. In my test example of Snopes.com, it does have an excellent reputation for accuracy and fact-checking as judged by third party reliable sources. A Quest For Knowledge (talk) 03:21, 15 June 2010 (UTC)

I agree with that too. If there is a reputable resource (such as a website) which is routinely cited and/or linked to by highly reliable sources for general information (not for information related to that source itself), I consider that a form of "publishing" of the source's material by reliable third parties. For example, in aviation accidents we have Aviation Safety Network which is our most valuable database resource, and is run by (as far as I know) a couple of guys. But it has proven itself over time to be highly accurate (though not perfect), and it is used as a reference by many mainstream publications. So I consider it a very reliable source, because the fact that reliable sources are relying on it for general information (not about itself) is akin to "publishing" its material, thereby making its authors published experts in the field. Crum375 (talk) 03:46, 15 June 2010 (UTC)

Snopes cites its sources, does it not, so we would never have to rely on it, and we'd never rely on it for anything contentious. SlimVirgin ^{talk|contribs} 06:56, 15 June 2010 (UTC)

This conversation keeps moving around like crazy, it's been on atleast 6 talkpages after its incarnation in a FAC. Anyway, Slimvirgin asked for the relevant examples, so here they are:

www.navweaps.com is mostly a convenience issue as it is a source collecting technical information on naval weaponry, currently linked on over a 1000 wikipages. Some of the sources he uses are non-english and/or very rare, so while theoretically possible it would be a pain to replace the site. It is a one man run SPS, but cited (and sometimes recommended) in 23 naval warships and weaponry books, including some important reference works. At least one the sources calls the author (Tony Digiulian) a "gun control expert."

www.Sinodefence.com is the other example, even thought the people behind the website are anonymous the site seems to be THE source for information on the Chinese military, cited by (among others): the US government, Reuters, Bloomberg, Al Jhazeera and practically every academic paper on the topic.

Both sites fail SPS because the author does not seem to have made additional publications in the relevant field. For more information I recommend you read the relevant discussions on the noticeboard Wikipedia:Reliable sources/Noticeboard#navweaps.com & Wikipedia:Reliable sources/Noticeboard#sinodefence.com Yoenit (talk) 08:03, 15 June 2010 (UTC)

The key question is whether someone has objected to these as sources. If not, there's no problem. But if someone does object, there has to be a test that establishes reliability, and the one we have chosen, which works very well, is whether someone other than the author has chosen to publish his work in that field. Why would an author prefer to publish without payment, if other publications are offering him payment and a platform to write the same material? That's the question you have to ask yourselves. SlimVirgin ^{talk|contribs} 10:00, 15 June 2010 (UTC)

By that argument all people who contribute to wikipedia must have tried to get them selfs published and failed, why else would they be here? Yoenit (talk) 10:22, 15 June 2010 (UTC)

But we're not offered payment to write the same material for a different publication. And anyway WP is a bit of an exception because it's fun and we're all nuts. It's just that it's difficult to judge expertise if no third-party has ever published that person's material. It doesn't mean the expertise isn't real; just that it's unconfirmed. SlimVirgin ^{talk|contribs} 11:34, 15 June 2010 (UTC)

Yes, there are some quasi-sane people out there who devote their lives to presenting reliably sourced information to the public for no pay, and whose work product is fairly reliable. In the case of Aviation Safety Network (ASN), when their sources are easily available to us (e.g. online NTSB reports), those sources are obviously preferred. Often ASN are a good starting point for an article, because they give us a road map and at least some pointers. We can then try to find more sources, until the ASN "scaffolding" is mostly superfluous. The fact that they are routinely cited by the major media and other reliable sources (typically for historical data) means that WP is not alone in relying on them, which means that in the modern age such highly-cited reputable sources are effectively a "published expert". Crum375 (talk) 11:28, 15 June 2010 (UTC)

I'm rather disappointed that assertions are being made here that have already been corrected elsewhere. It has again been claimed that At least one the sources calls the author (Tony Digiulian) a "gun control expert." Yes, but what source calls him that? It turns out that the source is an autobiography by D. Gossman, published by iUniverse, a "self-publishing company" (i.e. vanity press)! This approbation carries no weight, and this kind of misleading argumentation only weakens the case it attempts to bolster. Jayjg ^(talk) 01:15, 16 June 2010 (UTC)

My apologies, I must have missed the earlier correction. Yoenit (talk) 11:53, 16 June 2010 (UTC)

Two copies of a statement must each have their own inline citation?

The Verifiability policy requires inline citations for statements that are likely to be challenged, but if substantially the same statement appears more than once in the article, does each and every instance of the statement require a duplicate inline citation? (Of course, quotes require inline citations for each copy of the quote.)--Jc3s5h (talk) 00:50, 11 June 2010 (UTC)

Response to RfC

• No, a statement in a proper lead that summarizes content sourced in the body of the article does not need a separate inline citation. I suppose that a strict constructionist reading of the words of the policy, without regard for the spirit of the policy, can be used to support the double citation notion, but that crumbles in the face of common sense analysis. Material that is challenged or likely to be challenged must have an inline citation, yes, but is it ever reasonable to take the position that a challenge to material in the lead is not met by citation for that material in the body? I think not. Then the challenge is no longer about whether it can be verified—it has been verified—the challenge becomes about proximity. When we "challenge" we are saying "prove its verifiable." Just as it would be unreasonable to say a citation in one article is enough for the same statement in another article (too attenuated, there's no direct nexus between them), it's just as unreasonable, once a source has been provided in one article, to say that that same material, in that same article, is not verified because it's a few paragraphs away. In a world where articles were twenty-seven pages long, and a standard lead was 15 to 20 paragraphs of summary, I might have a different take.--Fuhghettaboutit (talk) 01:58, 11 June 2010 (UTC)

• See WP:LEADCITE. That guideline suggests that the question of citing the intro section should be handled on a case-by-case basis. Material outside the intro should certainly be cited if there's any question of it being contested.   Will Beback  talk  02:09, 11 June 2010 (UTC)

• Oppose changing the policy or guidelines on this point. WP:LEADCITE has the balance right at the moment, and I see no reason to change it. It says:

The lead must conform to verifiability and other policies. The verifiability policy advises that material that is challenged or likely to be challenged, and quotations, should be cited. Because the lead will usually repeat information also in the body, editors should balance the desire to avoid redundant citations in the lead with the desire to aid readers in locating sources for challengeable material. Leads are usually written at a greater level of generality than the body, and information in the lead section of non-controversial subjects is less likely to be challenged and less likely to require a source; there is not, however, an 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. Complex, current, or controversial subjects may require many citations; others, few or none. Contentious material about living persons must be cited every time, regardless of the level of generality.

The page that advises on sourcing in FAs, Wikipedia:When to cite, repeats the advice in WP:LEADCITE, and adds: "The distance between material and its source is a matter of editorial judgment. The source of the material should always be clear. If you write a multi-sentence paragraph that draws on material from one source, the source need not be cited after every single sentence unless the material is particularly contentious. Editors should exercise caution when rearranging cited material to ensure that the text-source relationship isn't broken." Again, I think this is the right balance. Some decisions have to be left to editorial judgment. SlimVirgin ^{talk|contribs} 02:22, 11 June 2010 (UTC)

• The issue will depend on the level of detail of the information provided in the lead. For example, in the case of a dispute, if the lead explain that X says A and Y says B, with more detailed explanations later, it would still need a reference. However, if the lead merely aknowledges the existence of a dispute, withoutspecific detail (for example, "whenever ovnis are alien spaceships or not is disputed"), then it would be less urgent to add a footnote for it. MBelgrano (talk) 02:42, 11 June 2010 (UTC)

• I think that statements in the lead which clearly summarize part of the article usually fall outside the 'likely to be challenged' category. However, we of course would need to cite an inline source whenever there is a good-faith challenge, and we should WP:AGF in those cases. I support the version of the policy without the theoretical footnote.  --Joshua Scott (LiberalFascist) 02:57, 11 June 2010 (UTC)
• Oppose How would you know if a statement appears more than once in the article if there is no source to identify it? It seems to me to run counter to WP:BURDEN if the reader has to play a game of "hide and seek" and search for a citation, and then match it with a particular statement. The whole idea of inline citations is that the reader can check a statement directly. I think this proposal would a step backwards to the middle ages. --Gavin Collins (talk|contribs) 04:53, 11 June 2010 (UTC)
  • Our readers are not idiots. --MASEM (t) 20:44, 11 June 2010 (UTC)
    • They aren't clairvoyants either! --Gavin Collins (talk|contribs) 07:21, 12 June 2010 (UTC)
      • We're not asking them to determine something that is not already included with the article. --MASEM (t) 08:59, 12 June 2010 (UTC)
        • Even if the readers are clairvoyant (disputed), and the lede already includes coverage within the article (disputed), then what is to stop editors from removing attributable content in the lede, and replacing it with unattributable content? It just does not work. --Gavin Collins (talk|contribs) 20:58, 12 June 2010 (UTC)
          • Because we assume good faith and deal with vandals when they do swap out false info. Starting from the assumption that editors are acting in bad faith and crafting policy around that is a dangerous starting point. --MASEM (t) 21:16, 12 June 2010 (UTC)
            • How can you tell that an edit is not vandalism if there are no inline citations? Unless you are clairvoyant, I don't see how you could tell which edits where made in good faith or not. --Gavin Collins (talk|contribs) 06:21, 17 June 2010 (UTC)
              • Very simple: We're still talking about duplicate information in the same article. One source we've assured stated will be sourced, so the question is if you suddenly find "X is Y" and "X is not Y" in the same article, you find the source that's attached to one of those statements and review it. It is actually less difficult in finding the problem than if someone maliciously changed information that's only sourced and cited ones, as it would be difficult to catch as time went on and would require a recheck of the sources. --MASEM (t) 06:41, 17 June 2010 (UTC)
                • Not so simple: How do you know if information is duplicate if there are no inline citations? Unless Masem is clairvoyant, I don't see how he could know. --Gavin Collins (talk|contribs) 06:53, 17 June 2010 (UTC)

No need for clairvoyance, just read the article. Of course, if the new claim is obviously suspect and one has either read the article in the past, or is familiar with the article subject matter, one will see the need to check out the new claim. If one is familiar with the subject matter but not the article, one can find the equivalent statement either by observing the section headings, or by using the browser's search feature to look for key words. Jc3s5h (talk) 09:58, 17 June 2010 (UTC)

For those of us not familiar with the article subject, we still won't know if a statement appears more than one is sourced, and we certainly won't know if a statement that appears only in the lede is sourced at all. Jc3s5h suggestion for those of us who are not clairvoyant is still problematical; not only will we have to read the article, check the revision history as well as search the internet in order to verify it. Forgive my extreme views on this issue, but I still think inline citation is still the most direct and simple way of verifying sources, and this policy should in any way suggest that it is not required.--Gavin Collins (talk|contribs) 10:55, 17 June 2010 (UTC)

Given that there are just as many of the same problems with assurance of reliability of a statement only made in the body of an article with a following in-line cite (such as if a vandal changed the statement, modified the reference, inserted a new statement between statement and inline site, or deleted either part), problems we can't get away from as long as we're an open-wiki, there is little issue about having a statement repeated/paraphrased twice but only sourced once. No one is asking the reader to by any more "clairvoyant" with dealing with duplicate statements and sourcing as they have to be with statements in general. --MASEM (t) 13:36, 17 June 2010 (UTC)

Even if inline citations can be corrupted, they do facilitate cross checking, and this this policy should in no way suggest that that they should not required. No one may be asking the reader to be "clairvoyant" in order to identify duplicate statements, but some form of supernatural assistance is required. Presuming that the reader will be able match facts and opinions with the source by reading body of article, the revision history as well as conducting an internet seach even before he can check the ultimate source is just unreasonable. This may not involve clairvoyance, but since effort required is like looking for a pin in a haystack within a haystack, this proposal is just not practical. --Gavin Collins (talk|contribs) 14:41, 17 June 2010 (UTC)

It is a reasonable assumption that a reader, who needs to make sure what he's reading is verifiable, will read the entire article comprehensively. Given that, it is further reasonable they are intelligent enough to recognize a duplicated statement, and can trace and track references to figure out if there's an issue. We don't need to do any more work then that. One can if they want, but up to that people, that is part of the assumptions we need to make to be consistent with our general disclaimers. Treating readers as any less intelligent than this is harmful and assumes bad faith. --MASEM (t) 16:55, 17 June 2010 (UTC)

You are asking to much. The readers are entitled to check the citations as they read the article; they are not clairvoyant, nor are they detectives. The lede has to stand on its own feet, just like the rest of the article. Treating the reader as if they are participants of a game of "hide and seek" is not in good faith. --Gavin Collins (talk|contribs) 17:13, 17 June 2010 (UTC)

If they are reading for comprehension, they are not going to stop at the lede and thus will be reading the entire article; if they are only reading the lede, they are only trying to get an idea of what a topic is and are not needed references. Given that duplicating sources once a statement is made in one place is rarely done in scientific journals and other encyclopedias, I don't see the need for it here. --MASEM (t) 17:28, 17 June 2010 (UTC)

LOL, Masem. If they are reading for comprehension and stop at the lead and they need references before reading further, what then? Assume good faith: the reader might want to see where the sources come in respect of the lede. QED --Gavin Collins (talk|contribs) 18:48, 17 June 2010 (UTC)

You are not going to stop at the lede if your reading for comprehension. We are not Cliff notes for topics. --MASEM (t) 19:01, 17 June 2010 (UTC)

We can't determine what the reader reads. That is why the lede must be subject to the same treatment as the main body of the article. We can't have "one rule for you, one rule for me". This policy has to consistent: the lede has to stand on its own two feet.--Gavin Collins (talk|contribs) 22:06, 17 June 2010 (UTC)

Masem, Wikipedia:LEDE#Citation clarifies that a bit. Wtmitchell (talk) (earlier Boracay Bill) 07:32, 18 June 2010 (UTC)

(←)We're actually here about LEDE#Citation, but it is useful to point out that we're still calling for quotes to be sourced in the lead, and I wouldn't hesitate to re-cite a highly charged statement ("John Q Smith is considered to be the worst person alive by some"). But as I've noted below, a proper lede is summarizing the article and should be avoiding going in these nitty details that we would require citation on; it should be summarizing the article in a fairly mundane and neutral manner, and not include details that would be challenged. As such, being a factual breakdown of the article, very little in the lede should be contentious, and the structure of the lede (sitting right about the TOC) should make it obvious where readers can look further in the article for more information and references if they need them. --MASEM (t) 12:36, 18 June 2010 (UTC)

I arrived at this discussion late. The assertion being discussed appears to be, "a statement in a proper lead that summarizes content sourced in the body of the article does not need a separate inline citation." My understanding of WP:V is that any statement likely to be challenged in the article, no matter where it is located, must be attributed to a reliable, published source using an inline citation. From that I infer that if the content of the lede is not sufficiently general that it is not likely to be challenged, it must be attributed. That seems to me to be logical and reasonable. Wtmitchell (talk) (earlier Boracay Bill) 13:19, 18 June 2010 (UTC)

• The effort required to read an article and observe where similar statements appear, and observe whether at least one instance of the statement has a proper citation, is about on a par with the effort of reading an online source, finding the relevant part of the web page, and deciding if it really supports the statement. If the reader does not possess the source and must obtain it from a library, the effort of checking the source is much greater than the effort of reading the article. Thus any wording which allows a person who has read the whole article to determine which source, and in the case of large sources, which page(s) of the source, supports the statement, is within policy. This is distinguished from a list of general sources, because in that case, one would have to read several potentially large sources to see if the statement is supported; that would be an unreasonable burden on the reader. Jc3s5h (talk) 21:11, 11 June 2010 (UTC)
• Oppose - WP:LEADCITE is fine as it is. We should not require the lead to become encumbered thus. The WP:V requirement is met by the main article text. Verifiability means that the article must provide the possibility to verify, not that it must repeat the cite at every opportunity. PL290 (talk) 09:39, 12 June 2010 (UTC)

Threaded discussion

(moved from above)

• No, a statement in a proper lead that summarizes content sourced in the body of the article does not need a separate inline citation. I suppose that a strict constructionist reading of the words of the policy, without regard for the spirit of the policy, can be used to support the double citation notion, but that crumbles in the face of common sense analysis. Material that is challenged or likely to be challenged must have an inline citation, yes, but is it ever reasonable to take the position that a challenge to material in the lead is not met by citation for that material in the body? I think not. Then the challenge is no longer about whether it can be verified—it has been verified—the challenge becomes about proximity. When we "challenge" we are saying "prove its verifiable." Just as it would be unreasonable to say a citation in one article is enough for the same statement in another article (too attenuated, there's no direct nexus between them), it's just as unreasonable, once a source has been provided in one article, to say that that same material, in that same article, is not verified because it's a few paragraphs away. In a world where articles were twenty-seven pages long, and a standard lead was 15 to 20 paragraphs of summary, I might have a different take.--Fuhghettaboutit (talk) 01:58, 11 June 2010 (UTC)

No, that's wrong. When specific material is being challenged, the challenge is about that specific statement, not another one with the same concept appearing elsewhere. Therefore, we need to supply a source inline, i.e. in reasonable proximity to the challenged statement so as to clarify which source goes with which statement. A reader must be able to find a source for a challenged statement without wading through the rest of the article, or going to the bottom of the page. Crum375 (talk) 02:04, 11 June 2010 (UTC)

The act of challenging is specific to an editor, not to others. If you challenge a statement, it is because you want to check verifiability. Once I provide the proof ("see inline citation to Russell, third sentence of second section," you know longer need to have that material verified. The idea that "A reader must be able to find a source for a challenged statement without wading through the rest of the article" misses, because it is directed at verifiability pre-challenge. Once challenged, and provided the location of the source for the material in the same article, they know longer can possibly still regard the material as "possibly unverifiable".--Fuhghettaboutit (talk) 02:14, 11 June 2010 (UTC)

No, if I challenge a statement, I want to see the supporting source right next to it, in a way that makes it clear that this source supports this statement, per WP:BURDEN. If I am told to go somewhere else for it, like some other section of the bottom of the page, it's no longer compliant with WP:BURDEN or WP:V, since the source is not inline with the specific statement I am challenging. Crum375 (talk) 02:18, 11 June 2010 (UTC)

And that is exactly what I meant about strict constructionism, verses the spirit of the policy. Once you learn the material is verified by an inline citation in the article, the underlying reason for your challenge has been rendered moot. Your challenge was not for the purpose of getting an inline citation into the article at that exact spot in the text—that's mechanics—your challenge was to have proved that the statement was verifiable. There is a balancing act. Like I said above, we can't just tell the person to look at a different article (or email them a citation), but an inline citation for the same material in the same article, where the lead summarizes the content, it's quite enough.--Fuhghettaboutit (talk) 02:26, 11 June 2010 (UTC)

No, you are inventing your own policy. What our WP:V and WP:BURDEN policy says is that when material is challenged a source must be provided inline for it, and the clear rationale is to identify which source goes with which statement. If the only motivation were, as you say, to satisfy the challenging editor that there exists a valid source for the challenged statement, we could supply the source at the bottom of the article, or even on the talk page. But the intent of the policy is not only to provide a source somewhere, but to clearly link it to the challenged statement in the article. Crum375 (talk) 02:36, 11 June 2010 (UTC)

No you are missing the fact that I am saying that an inline citation, not "somewhere", but present in the same article's body, for the same material, and where the lead is a summary of the body, IS a "source provided inline for the material" and requiring a double citation under those facts, is to ignore the spirit of the policy. You are interpreting the "material" as being only applicable to the precise textual location in the article. I brought up the issue of attenuation myself (predicting the reductio ad absurdum argument of "somewhere"), and this doesn't violate that (very valid) concern.--Fuhghettaboutit (talk) 02:55, 11 June 2010 (UTC)

No, if the intent were to provide the source somewhere, we'd say so. For all we know, maybe the same statement is made on another article, or on the talk page? The point is that if this statement is challenged, it needs to be sourced inline, i.e. in close proximity to it. This way if a reader reads this statement, he can immediately find the source for it, without needing to wade through the entire article to see where there is another place the statement is made, perhaps paraphrased, where he would hopefully find a source. The key is that when a specific statement is challenged, it gets a source close to it, not some other place near a perhaps similar statement. Crum375 (talk) 03:03, 11 June 2010 (UTC)

You've ignored what I've said about this being not just "somewhere". Person A: "This statement might be unverifiable; I challenge; prove it" Person B: "Look two paragraphs down, same material with an inline citation, and note that leads are supposed to summarize material appearing later in an article." My response if I was Person A: "Oh, you've addressed my concern." Your response if you were person A: "I'm going to ignore that you just addressed my concern over verifiability, prove it again right where I said to." Person B to you: "Sorry, but it's not as if the inline citation is just somewhere; it's in this very same article nearby and it verifies the very material you challenged." We have different tolerances for proximity. Unfortunately, I must go to bed now.--Fuhghettaboutit (talk) 03:24, 11 June 2010 (UTC)

Again, you are ignoring the point that all challenged material must be sourced inline, so that we can connect each challenged statement to its supporting source. By your logic, we might as well include all sources on the talk page. The lead is not only a summary, it is also an introduction, and may have material presented in a different way than in the body. Once a statement is challenged, it requires a source which can be clearly linked to that statement, not its paraphrased equivalent somewhere in the bowels of the article. Crum375 (talk) 03:37, 11 June 2010 (UTC)

Again you're treating an inline citation in the same article for the same material as being the equivalent of a citation in my basement in a lock safe that I'll provide upon request. You can say "bowels of the article" ten times; we do not agree that an inline citation in the same article for the same material does not meet the standards of "directly support the material in question" and "clearly support[ing] the material as presented in the article."--Fuhghettaboutit (talk) 12:33, 11 June 2010 (UTC)

WP:V requires the citation for a challenged statement to be "inline", which means in reasonable proximity to the specific statement in question, so a reader wouldn't have to work too hard to find it. Telling the reader to go find the source "somewhere" in the article does not provide it "inline" for that specific challenged statement. Crum375 (talk) 12:47, 11 June 2010 (UTC)

And around we go. It is inline, and it is for the specific material; the challenge has been met with that challenger's actual concern addressed, in a manner that is within the keeping of the policy.--Fuhghettaboutit (talk) 13:08, 11 June 2010 (UTC)

No, the challenger wants this statement to be sourced inline, not another statement in another place. The idea behind inline citation is that a reader should easily see which source supports a specific statement, and not have to search for something similar, perhaps paraphrased, in some other section. Crum375 (talk) 13:17, 11 June 2010 (UTC)

You've made me late for work:-( We are going in circles I think. Everything you say I am hearing, and I have told you why I don't agree with your view.--Fuhghettaboutit (talk) 13:22, 11 June 2010 (UTC)

Ditto, except I believe that the need for any statement that is challenged to get its own inline reference has a wide consensus. Crum375 (talk) 13:32, 11 June 2010 (UTC)

Ok, Crum375, since the abbreviion "CE" has been challenged in the past for the article Common Era, why don't you cruise on over there and provide a citation for every single occurrence of "CE". If you think that is taking the concept too far, then where would you draw the line between what is prohibited and what is allowed? Jc3s5h (talk) 02:15, 11 June 2010 (UTC)

Sorry, not following. I would include an inline source each time a statement is being challenged. If somebody goes around challenging every single word in an article, for example, he'd be violating WP:POINT. Obviously common sense is needed, and our goal is to make life easy for a reader encountering a surprising or non-obvious statement in an article, without having to wade through many sections to find the applicable source. Crum375 (talk) 02:23, 11 June 2010 (UTC)

(moved from above)

• The issue will depend on the level of detail of the information provided in the lead. For example, in the case of a dispute, if the lead explain that X says A and Y says B, with more detailed explanations later, it would still need a reference. However, if the lead merely aknowledges the existence of a dispute, withoutspecific detail (for example, "whenever ovnis are alien spaceships or not is disputed"), then it would be less urgent to add a footnote for it. MBelgrano (talk) 02:42, 11 June 2010 (UTC)

"Urgency" in this case translates into "likelihood to be challenged". If we think it's unlikely, we don't add a ref, and wait to be challenged, at which point adding a <ref name=X/> takes two seconds and we move on to the next problem. Crum375 (talk) 02:46, 11 June 2010 (UTC)

Summaries, not duplicate citations

The discussion seems to be concentrating on the the business of removing citations rather than the more important requirement as I see it of supporting summary statements. A summary statement summarizes a part of the article - you'd mostly see them in leads. Currently there is no support for summary statements so they have to all stand as separate entities. Supporting summaries would I believe help greatly in improving the consistency and readability of articles.

I believe summary statements should be checked to see if they are a reasonable summary and their citation check should involve checking the citations of the bit being summarized. Putting citations on the summary directly divorces it from the bit being summarized. Currently you get lead being developed almost independently of the body of articles. If there is something dubious in a summary it would be easier to find the relevant citation by going to the bit being summarized, there might be three citations currently on the summary statement but it would be easy to see the particular statement and its own citation in the bit being summarized. It also makes it easier to look after the citations as there would be far less duplication and reference names to worry about which makes it easier for people editing articles. This means the leads would reflect the contents of the articles better.

Of course any statement can be challenged but a reasonable response for a summary statement would be to put in an in-article cross reference to the bit being summarized. I would also add new tags saying that a statement wasn't a good summary or was wrong. These could be substituted for citation needed in many instances. Dmcq (talk) 11:08, 11 June 2010 (UTC)

Although leads typically include a general summary of the article, they are not exclusively so, since they also serve as an introduction, and may also have some unique material. Therefore enforcing a good summary there makes no sense, since it's up to editorial judgment. Any material on WP need to be sourced when challenged, so even if it's a supposed summary, there should still be a link provided to the source, not to another statement with a source. Crum375 (talk) 11:29, 11 June 2010 (UTC)

I never said a lead was exclusively composed of summary statements. Your other part is correct about the current situation. I'm saying if you could refer to other statements rather than copy the citations it would improve the articles and improve checking. Dmcq (talk) 11:49, 11 June 2010 (UTC)

If the point is style, i.e. where is best to place the source so that a reader would not have to work hard to find the support for a given statement, then we agree. If a statement is being challenged, the source (or a link to i butt) should be provided in a way that a reader can easily see how that specific statement is supported by the source. Crum375 (talk) 12:11, 11 June 2010 (UTC)

It isn't just style to find a citation, I believe the current system is actively harmful as I described above. I fully agree that it should be easy to find citations for statements, however for a summary style statement the source should be considered the statements being summarized. That is a tertiary source which then refers to the external citations because it is part of the same article, but there are benefits from an having an integrated article rather than a sequence of disconnected statements. Dmcq (talk) 13:00, 11 June 2010 (UTC)

I think the key point is simple: per WP:V, if a reader sees a contentious, surprising or non-obvious statement X, he should be able to find a reliable source which supports X "inline", i.e. with minimal effort. Searching inside some other section for the same or similar (possibly paraphrased) statement, where a source may hopefully be included, is not inline and not minimal effort. Crum375 (talk) 13:10, 11 June 2010 (UTC)

That is why we're having a discussion here. This is where changes to WP:V are agreed. You are saying minimal trumps everything i.e. that it means absolute minimum whatever bad effects might follow. And I've shown you how referring to the underlying statements the relevant citation can often be found with less effort. Dmcq (talk) 13:40, 11 June 2010 (UTC)

Unless of course, the lede contains facts or opinions not stated anywhere else in the article. The idea that it should be easy to find citations for the lede in the body of the article is a sweeping generalisation that does not applicable in every case. --Gavin Collins (talk|contribs) 06:26, 17 June 2010 (UTC)

Which is why I'll be proposing a form of within article citation for use where it isn't straightforwardly obvious from, say, the contents list. One would use this instead of putting in duplicate citations from the indicated section. Dmcq (talk) 10:20, 17 June 2010 (UTC)

The underlying problem with using summary style article ledes is that they presume one can summarize in a manner which is not WP:SYNTH. Although in rare cases this conceivably might be possible to do, we most certainly cannot do it in every article, or even for most articles, simply because we usually base an article on a specific set of sources, supporting a specific set of assertions that have never been combined before. If we are frank with ourselves, in practice we routinely ignore - or at least temper - WP:SYNTH for lede sections. We can do better. We should, at minimum, have a construct that distinguishes such synthetic summary from simple duplication of assertions in the lede. This could be as simple as inserting <ref>See under [[#About this]] and [[#About that]].</ref> by the synthetic summary statement, to make it clear to readers that the basis is not external, but just content of the article. LeadSongDog come howl! 15:42, 17 June 2010 (UTC)

Copies and citations: conclusion

Since there is no consensus that providing one inline citation to cover two instances of substantially the same statement, it would be improper to sanction any editor who does so, and any administrator who imposed such a sanction should be subject to disciplinary measures. Further, anyone who, knowing that a citation for a statement existed elsewhere within the article, deleted material claiming it was unsourced, is a vandal and should be treated accordingly. Jc3s5h (talk) 13:16, 11 June 2010 (UTC)

That is almost a perversion of WP:BURDEN: if any editor challenges meterial, then its down to them to find citation, not the editor who added or restored it. --Gavin Collins (talk|contribs) 13:32, 11 June 2010 (UTC)

I'm not sure how a discussion about the verifiability of summary statements ended up with sanctioning admins. Dmcq (talk) 13:36, 11 June 2010 (UTC)

I take offense at Gavin.collins' misstatement of what I wrote.

Dmcq: this discussion has conflated policy violations with less-than-ideal writing practices. Editors who violate policy are sanctioned. Editors who write in a less-than-ideal way are thanked for their contributions and others polish their writing. Editors are entitled to know in advance what conduct is subject to sanctions. Since there is no consensus about the policy on this matter, or even consensus on what constitutes good writing, it would be improper to impose sanctions on any editor who chooses to write a statement twice but only provide an inline citation for one of the copies. Jc3s5h (talk) 13:53, 11 June 2010 (UTC)

Jack, if you take offense at something, it helps to identify the something to which you refer. Traditionally we do this using WP:DIFFs. In some cases it also helps to describe what part of the diff you find offensive. LeadSongDog come howl! 19:10, 11 June 2010 (UTC)

The feeling I get from this is that someone with a chip on their shoulder from something that happened to them has jumped in here to mess up an unrelated discussion. It does not sound like it is trying to improve the associated policy. If so could the OP kindly just remove this section please. Dmcq (talk) 19:19, 11 June 2010 (UTC)

No. Jc3s5h (talk) 20:05, 11 June 2010 (UTC)

(edit conflict)

Your statement "Further, anyone who, ... deleted material claiming it was unsourced, is a vandal and should be treated accordingly." lacks a rationale for assuming that the actor knows and understands the concensus not to take that action. In that regard, the statement assumes bad faith. We don't do that. LeadSongDog come howl! 21:03, 11 June 2010 (UTC)

Well since verifiability only gets as far as admins when people try corrupting articles and acting disruptively I don't think sanctions are very relevant here.

My conclusion is that I should go ahead with a proper proposal to Village Pump proposals and the discussion has shown both what the main worries are and a few extra problems which such a proposal would also solve. Verifiability is a prime requirement and needs to be ensured, doing that will require a couple of extra templates and the whole business needs a good document explaining it. So I'll try writing an essay which might be suitable for sticking into WP:CITE as the next step and then put it to a vote as a proposal. Dmcq (talk) 20:59, 11 June 2010 (UTC)

Thinking it the other way - poorly written ledes

I don't think anyone questions that a quote duplicated in the lead needs a second citation. But it seems the focus is on contentious statements in the lede, and that makes me wonder : are we writing our ledes wrong that they include contentious statements?

A lede as I've explained above is doing the job of an Executive summary, which WP describes nicely as a summary that is presented "in such a way that readers can rapidly become acquainted with a large body of material without having to read it all". The only major change our ledes have over executive summaries is that our introduction statements in the lede are to establish what the topic is, while an exec summary will establish what the problem is that the longer report address; similar in nature, effectively. It is fairly intuitive that if we only We're also trying to write ledes as an introduction to the structure and content of the rest of the article, and thus it is important that the lede sticks true to how the article is constructed.

The lede should be written in the same structure that the rest of the article is in. For example, coming from video games, I know we have, in nearly every game article, "Plot", "Gameplay", "Development", and "Reception" sections, give or take a few. Thus, in the lede, after defining what the game is (genre, platforms, release) we will have one to two sectences on each of those points - plot, gameplay, development, and reception - in the order established in the article. Introducing anything else new in the lede that would likely need a citation and is otherwise not covered in the body is bad summation of the larger article; new information needs to be incorporated in the body before it can be stated in the lede.

So now lets consider contentious statements. I would argue that in most cases, significant contentious statements (ones that no one would ever question the need of a citation for) that are present in the body of the article never need duplication in the lede but instead should be summarized in a unbiased summary of any debatable issues. Take abortion for example. The lede does an excellent job of defusing the "Abortion debate" section into a few lines that say "it is a contested issue" and avoids stating any specific aspect that is contentious. When it is done this way, along with the logical progression that follows the outline of an article, a reader can quickly discover what section(s) of the article deal with the abortion debate and find all the sources in the world there. Thus, there is no need to really ever include a contentious statement in the lede since it usually can be summarized out into a unbiased statement that acknowledges the issues exist.

But even with all that, it may be necessary to include a statement that can be challenged. For a critically praised movie, we would probably include a line, summarizing from the Reception section, that "The movie was considered by many reviewers to be the best film of 20xx." (and for sake of argument, say we have at least 10 different reviewers that are sourced as saying that). It certainly would be a contentious statement without additional context, and it is also a statement you can't bury like with the abortion debate. But, and this is why structure of the lede is so important, if it follows this is part of the film's reception, and I wanted to validate that statement, I know exactly what section of the article to jump to to find this. If this was really challenged, and we had to source it, we'd need to include a link to every reviewer that stated that fact, which could lead to a major eyesore of inline cites. Instead, by structuring the lede (and the rest of the article) properly, validation is very easy to find.

But all of these are based on how the lede is written, and not about WP:V; that is, it is really a lack of what we have at WP:LEAD's "Introductory text". It skims over how to write a good lede in favor of explaining how to set it up. I think that much of the issues addressed here can actually be dealt with by explaining how to write a lede that includes no drastically contentious statements, and when necessary, providing an easy way to guide the reader to find sources to support any of the claims made in the lede. --MASEM (t) 21:41, 11 June 2010 (UTC)

Tell me about it. I've just gone and stuck another three duplicate citations into a lead after the statements were complained about. I pointed out that there were sections which were obviously on those topics in the contents list but no they wanted the citations copied over, quote the policy and tried deleting a bit because the citations weren't there. So now there's 21 citations in three medium size paragraphs. Dmcq (talk) 22:31, 11 June 2010 (UTC)

No sorry 24 duplicate citations in the lead but three of them are duplicated again in the lead itself! Dmcq (talk) 14:27, 12 June 2010 (UTC)

If the problem is visual citation clutter, it's easy to combine the references, so assuming these are repeated elsewhere, you'd end up with (say) one single <ref name=X/> per paragraph, which is barely visible. Crum375 (talk) 22:45, 11 June 2010 (UTC)

This is reminiscent of calendar-related articles which have calendar conversion formulas. The formulas are often proceeded by a statement that all divisions are integer divisions; remainders are dropped. Nevertheless, editors will mark the equations {{dubious}} or {{disputed}} and complain on the talk page, all because they didn't read the whole article. What are we supposed to do, make every other line say "these equations use integer division"? Now, if there is a way to prevent confusion without making the article look horrible, fine, but sometimes you just have to turn your back on readers who only read a few sentences of an article. Jc3s5h (talk) 22:42, 11 June 2010 (UTC)

At some point we have to stop holding the reader's hand. We need to assume things like a basic knowledge of English, a basic knowledge of grade-school science, geography, history, etc. etc. We also need to assure that if they are looking for references, they 1) know how to do research and 2) are capable of reading the whole article. This entire trend of late of citing every single line is irksome because no other real-world work would ever source that frequently. I can understand that we need to be a bit more frequent than most real-world sources due to the nature of an open wiki, the sabotaging by vandals, and the attempts to improve WP's reliability as a source. But the current trend is catering to people who cannot do a bit more reading on the same page to find the information they need, and instead want it NOW NOW NOW. This whole cites-in-lede is just an extension of it. That said, we can do something about it by making as clear as day of how to read and use the article effectively through a good lede that directs the reader right where they need to go for more info. --MASEM (t) 22:51, 11 June 2010 (UTC)

Masem, as several people have pointed out, a WP lead is not an executive summary. Therein lies the confusion. SlimVirgin ^{talk|contribs} 06:37, 12 June 2010 (UTC)

It may not be an executive summary by name, but it is serving the same purpose - summarizing the salient points of the article's body as to serve for brief but complete overview of the article. --MASEM (t) 08:57, 12 June 2010 (UTC)

Not really. It's an introduction to the article, not just a summary of it. It contains material that summarizes the key points, and it may contain other material. The mistake is in thinking of it as an executive summary, then insisting that it not have citations for that reason; you're begging the question, in other words. SlimVirgin ^{talk|contribs} 09:02, 12 June 2010 (UTC)

A quote in a lede should have an explicit, verbose citation. In 1999, so-and-so said "blah" [ref]. It must be a pretty significant quote to warrant placement in the lede in a good article, and then should probably be the subject of a section below.

In general, any contentious statement in the lede should be readily matched to an entire section. If it seems to hard for some, I suggest wiki-linking to the section.

I don't agree that a lede should necessarily reflect the structure of the rest of the article.

In the lede, citing the primary source for the article, and citing for notability, seems sensible. Much more citing would seem to mean that the lede material belongs below. --SmokeyJoe (talk) 09:24, 12 June 2010 (UTC)

Is lede the Wikipedia-preferred spelling over lead? --SmokeyJoe (talk) 09:25, 12 June 2010 (UTC)

I've only see it used to distinguish the verb "to lead" from the lede section. Either works, I'm just used to lede. --MASEM (t) 09:34, 12 June 2010 (UTC)

Our advice on the lede section says to start with a definition (where there is a likelihood of new info not in the body), followed by "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 an exec. summary style that also serves as an introduction. It needs to tell the reader what is in the article and what to expect in the article before reading it. Inclusion of new material in the lede is bad form outside of the definition; that info should be incorporated in the body somewhere. This is particularly true of new contentious statements which should never be in the lead and not repeated in the body. --MASEM (t) 09:34, 12 June 2010 (UTC)

There's a tendency all over Wikipedia to try to tell editors what to do at every point in article creation, from cradle to grave, as it were, leaving no room for editorial judgment, distinctive approaches, brilliant writing. But we're not bots. I don't want to read the same format over and over and over, and be told "you can't have this factoid in the lead because it's not repeated elsewhere," or "you can't use curly quotation marks, because someone has decided they're 'deprecated'." I do want to see well-researched, neutral, well-written articles with good sources, but I also want to see some humanity in them, even some quirkiness. SlimVirgin ^{talk|contribs} 09:47, 12 June 2010 (UTC)

Agree. It's great when a sense of a person behind it comes through. I see it achieved when a fresh editor comes in an copy-edits the whole article. It's even one of SV's strengths. --SmokeyJoe (talk) 10:45, 12 June 2010 (UTC)

To get this back on the point (citations)... I think there is a clear consensus that where something in the lede is expanded upon and cited later in the article, we do not need to do so in the lede. That does not mean we can't... just that we don't need to. That covers most of the material in the typical lede... however, we do have to allow for non-formulaic writing. Information may appear in the lede that is not expanded upon later. That information should be given a citation in the lede. As for quotes... My feeling is that quotes should always be cited the first time they appear, whether in the lede or elsewhere. To sum up... let's allow for some common sense and flexibility here. Blueboar (talk) 13:31, 12 June 2010 (UTC)

Agreed, I'd say these points still apply:

1. Quotes always need cites, lede or otherwise
2. A new piece of contentious information that is not repeated in the body should always be cited in the lede. (Whether the lede should introduce new info, that's a writing style issue)
3. Otherewise, the choice to use inline cites in the lede is up to the page editors (ala choice for format styles, US/UK spelling, etc.) It is not a style choice that should be forced either way and edit-warred over. If cites are used beyond the required cases above, they need to be used consistently as with the rest of the article body. --MASEM (t) 13:37, 12 June 2010 (UTC)

No, that's not right. If there's something contentious in the lead, it needs an inline citation, even if repeated elsewhere. The paragraph on this from WP:LEAD gets this balance right. It stresses the need to stick to policy, but also talks about leads perhaps being more general in tone.

The lead must conform to verifiability and other policies. The verifiability policy advises that material that is challenged or likely to be challenged, and quotations, should be cited. Because the lead will usually repeat information also in the body, editors should balance the desire to avoid redundant citations in the lead with the desire to aid readers in locating sources for challengeable material. Leads are usually written at a greater level of generality than the body, and information in the lead section of non-controversial subjects is less likely to be challenged and less likely to require a source; there is not, however, an 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. Complex, current, or controversial subjects may require many citations; others, few or none. Contentious material about living persons must be cited every time, regardless of the level of generality.

SlimVirgin ^{talk|contribs} 13:42, 12 June 2010 (UTC)

(ec)I think whenever some specific statement is actually challenged, a reference should be provided inline, per WP:V. Inline means in a way that makes it easy for a reader to find the supporting source for that statement, without having to wade through other sections and search for similar material with the hope of finding the source there. How exactly this is done should be left to editorial discretion. Crum375 (talk) 13:44, 12 June 2010 (UTC)

Ideally, I think, no sentence in a lede will be contentious. Unfortunately, most articles are not ideal. It is more important to get contentious statements cited than to idealise the ledes. --SmokeyJoe (talk) 14:03, 12 June 2010 (UTC)

(ec, but I also agree with what Joe just said) Not having been following this discussion in detail, I think the guideline as stated is not all that bad, although it should be strengthened, basically to make the (possibly obvious) point that if we want to give a cite for some piece of information that appears in both the lead and in the article body, then we prefer to cite it where it appears in the body rather than in the lead. (And in a mature article, almost everything in the lead should appear in the body, so that means almost no cites in the lead once the article is up to scratch.) Then we can mention the specific exceptions, like quotations and particularly contentious material.--Kotniski (talk) 14:09, 12 June 2010 (UTC)

I agree the focus of this discussion should be on what happens with mature (GA or better) articles. --MASEM (t) 14:18, 12 June 2010 (UTC)

To get GA approval, most reviewers require each paragraph (at a minimum) to be attributed, regardless of which section it's in. This has been my own experience, and I can't blame the reviewers, since I agree that forcing the reader to search through other sections for similar material in the hope of finding the right reference is unreasonable. The idea of a good article is that we make life easy on the reader, not ourselves as editors. Crum375 (talk) 14:42, 12 June 2010 (UTC)

Which in the case of leads, we do by leaving out the distracting footnote markers that make the text harder to read. That certainly seems to be the norm for most featured articles, which as a rule are heavily cited in the body, but usually contain few or no citations in the lead.--Kotniski (talk) 14:49, 12 June 2010 (UTC)

A tiny little a the end of the paragraph is hardly "distracting". If we combine the references, which is good practice for mature articles anyway, all that's needed is one tiny combined reference per paragraph. And even in edit mode, all one needs to see is <ref name=X/> if the references already exist below. Crum375 (talk) 15:07, 12 June 2010 (UTC)

If a ref is needed because the reader can't see the appropriate place easily I would much prefer for it to point to the section in article that deals with it in more detail. For a reader who has a worry about something this might clear up the problem immediately or point them to the appropriate citation to check rather than a list of citations. It also means the lead is checked against the text better. Duplicating the citation is an unhelpful and unfriendly way of dealing with the reader's queries. Dmcq (talk) 16:16, 12 June 2010 (UTC)

Combining citations will often create an unreasonable burden on editors. The same sources are likely to be used again later in the article, but it is unlikely that all the sources will support every subsequent point in the article, so the whole citation will have to be rewritten when the sources are reused. In the case of online sources, this will mean the same url will appear multiple times in the article. When the url needs to be changed because it has moved, there is a serious risk that some copies will be corrected and other copies will not. Jc3s5h (talk) 16:26, 12 June 2010 (UTC)

You're right that it can require slightly more maintenance, so it's best done on articles where editors are taking care, but the difference really is very slight. The benefit is that you don't have footnotes after every sentence, or even within sentences, which can be ugly when repeated a lot. SlimVirgin ^{talk|contribs} 16:30, 12 June 2010 (UTC)

That last bit's certainly true. But can you give an example of an article where this "combined referencing" style has been used? I somehow can't imagine it working in most cases (since as a rule different statements within a paragraph are supported by different sources - and that's particularly likely to be the case in the lead).--Kotniski (talk) 16:38, 12 June 2010 (UTC)

The era abbreviations CE, BCE, AD, CE, and the article Common Era are all subject to drive-by zealots who try to cleanse the encyclopedia of the version they don't like, and/or contest every word in the article. As a result, the article has some combined citations in the lead, because one dictionary just won't convince people, five are required just to keep the drive-by edits to a minimum. Jc3s5h (talk) 16:50, 12 June 2010 (UTC)

(ec, reply to Kotniski)

I use it a fair bit; for example: <ref>

• For Smith's opinion, see Jones 2010, p. 1.
• For his date of birth, see Simon 2009, p. 2.</ref>

You do often have to make clear which ref supports which point. You can see it in this version of the lead at David Icke, though someone has just changed the notes back to citations within sentences. If you look at it in edit mode, you'll see "for point X, see reference Y" in the footnotes. SlimVirgin ^{talk|contribs} 16:52, 12 June 2010 (UTC)

That's a really bad form that doesn't meet normal guidelines. If you are going to cite the lede (which no one is preventing) the citation style needs to be consistent with the body. Throwing all the cites into a single cite at the end doesn't gain any advantage and requires one to either have manually formatted all the cites or make sure that the formatting matches the rest of the cites. --MASEM (t) 17:11, 12 June 2010 (UTC)

Which normal guidelines doesn't it meet, Masem? The advantage is less clutter, and a lot of editors do use manual citations. You seem to be here to impose all kinds of stringent rules on us about what we're allowed to do. Editorial judgment and choice is still important on WP when it comes to writing, sourcing, and how to write citations. :) SlimVirgin ^{talk|contribs} 17:17, 12 June 2010 (UTC)

For one, WP:CITEHOW. That format would not fly at FA review neither per 2c WP:WIAFA. --MASEM (t) 17:43, 12 June 2010 (UTC)

CITE makes clear that editors can choose how to write citations. Discussions on the talk page show that combining references is acceptable; see here, for example. And it does indeed fly at FAC. I've had articles promoted using combined citations. SlimVirgin ^{talk|contribs} 17:46, 12 June 2010 (UTC)

Even if they do fly (I've seen FACs with awkward citations that fall apart because of that, so it may simply depend on other factors) it is nonstandard and also breaks the requirement that quotes be immediately inline-cited after they are given. If you do want to source the lede, you might as well go ahead and inline cite in the same, consistent manner as the body of the article to avoid any issue of having a cite too far away from where it needs to be. It is a matter of consistency on a single page, not so much across the rest of WP. --MASEM (t) 12:47, 14 June 2010 (UTC)

The citations are at the end of the paragraph, so there is sufficient source-text integrity. WP:CITE is very clear that editors may choose how to add inline citations. We can't dot every i and cross every t on other editors' behalfs. SlimVirgin ^{talk|contribs} 01:06, 15 June 2010 (UTC)