Wikipedia talk:No original research

Rewrite 2004-2005: Draft • Talk "Primary v. secondary sources" subpages "Change needed": Proposal • Talk "Various examples": Examples • Talk "Transclusion example": Example • Talk

Tie-in books

Shouldn't e.g. works that build on a particular fiction franchise be regarded as primary sources unless they provide substantial discussion above the level of narration? User:Dorftrottel 13:07, January 30, 2008

Original calculations

In the article Tsar Bomba has an edit been deleted as "original calculation". Howerer, since the calculation was of an elementary kind, there is no doubt that the result is correct. Does the NOR policy really prohibit any kind of own calculations even if everyone can easily verify that they are correct? Does it further mean that no one may derive e.g. the formula for some geometrical relationship (e.g. the ellipse parameters, volume of geometric bodies, interpolation/regression formula (with reference to the method) of graph of published data etc.) by him/herself without citing a reference? If so, than it would be strange, really strange...:-/--SiriusB (talk) 20:59, 26 April 2008 (UTC)

Well, those things you've mentioned aren't really "elementary"; they're not the kind of thing that anyone with a basic education can check. Simple addition/subtraction/multiplication/division people can do themselves. However, the example in question is really a numerical analysis requiring nontrivial background knowledge and not readily checkable by the typical editor. There may be other problems as well... SamBC(talk) 21:10, 26 April 2008 (UTC)

I do not fully agree. Take the very first example (Tsar Bomba), where some "complex calculations" have been deleted with reference to this very discussion. These "complex calculations" consist of some elementary operations (in at least one case just multiplications!) that everyone who owns a calculator can reproduce within a few seconds (and students are usually expected to do this just by brain). And if the input values like the solar constant, TNT equivalent or the duration of the energy release are just linked with related Wikipedia articles then the "original research" would be reduced to these basic calculations. I cannot see why it should be illegal just to assume that, analogously, 1+1=2 without proving by citing the literature. And do you really believe that the equations in Articles like Volume or Kepler's laws of planetary motion are entirely taken from the literature, without any transformation or substitution done by WP authors? As long as these are checkable by applying standard mathmatics, I do not see any problem.

Furthermore, scientific literature has IMHO the same problem as the more complicated examples above since many peer-reviewed journal articles cannot be completely understood by readers with only basic education but require advanced specific knowledge. Even worse, most peoble (especially those who do not work in a university or scientific institute) do not have easy access to those sources since they require a subscription. The main difference between "original research" and curve fitting is IMHO that the former is probably connected with personal interpretation of the author while the latter is a more or less straightforward application of well-proven methods (however, I agree that the author must state which methods he actually used).

The main reason for my question is that I have, in my early Wikipedia days, contributed some content like this which is created e.g. by numerical integration while the underlying source data are taken from the literature. You may now argue that these contributions now have to be speedy-deleted, but is that really what the nro-policy aims at (in the case given above I do not know any free content that might be used as a replacement)?--SiriusB (talk) 08:35, 28 April 2008 (UTC)

I believe you are correct in the distinction you make between "original research" and "curve fitting". I also believe you correct that the assumptions need to be specified/cited, rather than a pure simple formula run. Vassyana (talk) 08:58, 28 April 2008 (UTC)

Original images

See Wikipedia:Reliable_sources/Noticeboard#Radical_Gun_Nuttery.21_website. While this policy is about original research, unreliable sourcing in images is a problem. This policy is the only source of information about restrictions or permissibility of original images. The text, [emphasis in original] "images generally do not propose unpublished ideas or arguments, the core reason behind the NOR policy", was added to policy to clarify the permissibility due to nonsense like this. I would suggest that a single sentence follow that statement:

“ This "exception" does not cover images and diagrams based on data and statistics from unreliable sources. ”

I believe this would suffice to prevent the abuse of WP:OI to promote images based on unreliable sources, while clearly distinguishing the OR exception (the nature of this policy) from editorial decisions about the reliability of source material. Vassyana (talk) 03:04, 28 April 2008 (UTC)

To me the issue with images comes down to this: Images should illustrate something stated in the text of the article. If the text that the image is illustrating is not OR, then the image based upon that text is not OR either (even if the image is an original creation). Blueboar (talk) 12:14, 28 April 2008 (UTC)

I would agree with that. However, editors are using WP:OI as a defense stating that it's OK to use unreliable sources and that OI encourages that. The link above is not at all an isolated incident of using the section in such a way. Considering the common misuse of the section, I felt we should clarify the point I raised above. Vassyana (talk) 12:29, 28 April 2008 (UTC)

Sorry if I was not clear... I agree completely on your proposed addition. I am actually thinking that we need to re-think the wording on the entire section. Make it clear that images and the text they illustrate are directly related. While user created images are allowed, they must illustrate something stated in the text... and that text must not be OR or based upon unreliable sources. An image that illustrates text that is deemed Original Research is considered Original Research by association... and an image illustrating text that is based upon unreliable sources is considered unreliable by association. Blueboar (talk) 13:35, 28 April 2008 (UTC)

I should have known we were on the same wavelength. :) How would you express that relation for things like user-taken photos of celebrities and landmarks? Vassyana (talk) 15:01, 28 April 2008 (UTC)

Although I too am on this wavelength, a word of caution about cyclic proofs: The completely unencyclopedic "she has brown hair" was "substantiated" by a -- also wonderfully photoshopped -- accompanying image. -- Fullstop (talk) 18:18, 28 April 2008 (UTC)ps: since this example is from an article on a porn-star, I should add that it was the hair on her head that was being being referred to :-)

I agree, but I think this should not be phrased as an "exception to an exception", which is too confusing. None of this, either "images themselves are usually good", or "images based on unreliable information are bad", is an exception anyway; this is all based on fundamental principles of verifiability. We can state this in a positive way, such as:

"An editor-produced photograph or drawing is not original research if (1) it has been was published in a reliable source, or (2) the information depicted in the image is verifiable and could be appropriately included in the article in text form. Any modifications to the contents of an image that have been introduced by the editor, such as "Photoshopping", should be disclosed either in the caption or on the Image page."

COGDEN 20:43, 28 April 2008 (UTC)

Images (including photographs) are primary sources, and as such should not be used as a source on Wikipedia, except for a statement about that particular image itself. Ideally, an image should be used as a source at all ... but instead should be used as an illustration of statements that are cited to reliable sources. ... you know, I think I will bounce that thought off of the folks at WP:V and WP:RS and see if a statement like that should be included on those pages. Blueboar (talk) 21:36, 28 April 2008 (UTC)

Yes, images should not be used as sources. Own interpretation == OR. This is what I think the most image OR comes from, i.e. the editors are not using images to demonstrate what text already says, but basing text on images. The issue of using images to further a point-of-view is not an OR matter; in the Radical Gun Nuttery! affair, the image is using "published" data that is (however whacky or unreliable) being faithfully reproduced, and hence not OR. -- Fullstop (talk) 03:02, 30 April 2008 (UTC)

What should happen when WP:SYN contradicts WP:NPOV?

Consider the following hypothetical scenario:

A claim arises that the Cuban govenment is directly involved in smuggling cocaine into the United States. Pulitzer award-winning journalist Jow Blow (a notable source) investigates, and publishes a story apparently confirming the claim and suppying specific details: the drugs are carried on Cuban submarines and transferred to small boats just off the Florida coast. Wikipedians are in general agreement that this story is notable enough to deserve its own article, "Allegations of Cuban involvement in Cocaine Smuggling".

However, there is a problem: Cuba has no submarines. At least, that's what Admiral Popeye (USN) is saying: and he's a notable and (normally) reliable source. But, even though he said this in response to a claim that Cuba does have submarines, he was NOT specifically referring to Blow's claim that Cuban submarines were used for drug smuggling.

This contradiction is all over the Internet: however, if there are any reliable sources pointing this out, nobody can find them among the 98,172 forum and blog postings regarding this issue. And fans of Blow (or enemies of Cuba) won't allow anyone to point out the lack-of-submarines problem without a valid reference that directly ties in to the article's subject, the alleged drug-smuggling (if this situation strains credibility, imagine a less high-profile variant: John Smith is accused of smuggling marijuana in his truck, but Smith has no truck...).

So, does Wikipedia only present one side of this controversy, thereby violating NPOV? --Robert Stevens (talk) 12:10, 1 May 2008 (UTC)

NPOV is defined by Wikipedia as "representing fairly, and as far as possible without bias, all significant views that have been published by reliable sources." Hence, if there is a viewpoint that hasn't been published by a reliable source, then it should not be represented. Hence there is no WP:NPOV problem - in fact, WP:NOR is helping to prevent an WP:NPOV problem.

But realistically, if the story is indeed notable enough to deserve its own article, then there will be reliable secondary sources that comment on Blow's allegation. If the lack-of-submarines issue is truly significant, you can pretty much guarantee that at least one source will mention it. So if such a situation were to arise, it would probably indicate that either a) the story fails notability, or b) there isn't a verifiable controversy (expressed in reliable sources) about the issue that needs to be represented. Jakew (talk) 12:23, 1 May 2008 (UTC)

Well, OK, but in this case I'm assuming that Popeye's claim has been published in a reliable source. So, it could be argued that his claim regarding the existence of Cuban submarines qualifies as "a significant view that has been published by reliable sources". The problem arises with the interpretaion of the phrase "significant view" (significant to what, and who decides?). Would I be correct in assuming that if there was a second Wikipedia article entitled "Controversy regarding the Existence of Cuban Submarines", a "See also" link to that page would be admissible? This seems to be a common pattern, and I haven't yet seen anyone comment that the linked page doesn't mention the subject of the previous one (though I suppose someone could raise that objection...). --Robert Stevens (talk) 13:08, 1 May 2008 (UTC)

What many wikipedians who edit the non-scientific wiki articles get wrong is that so-called "reliable sources" can write nonsense (e.g. the Wall Street Journal editorials on Global Warming are mostly garbage and not acceptable for the wiki global warming page). The best thing to do is to demand peer reviewed sources to back up information. This is not always posible, so the next best thing is to see if the claims in the source are verifiable. Does the published story cite references, does it cite witnesses, does it say that the author him/herself witnessed it him/herself? Count Iblis (talk) 13:26, 1 May 2008 (UTC)

I suppose another related issue is this: supposing a Blow-supporter (blowhard?) admits on the article talkpage that he fully accepts both the notability AND the factual accuracy of Popeye's account (i.e. he fully accepts that Cuba does indeed have no submarines): so there's really little room for doubt that his refusal to admit Popeye's testimony is POV-pushing (after all, we do have WP:IAR and WP:COMMONSENSE, and even WP:TE might be applicable). In this case, the "nonsense" would appear to be Blow's report itself, which is probably the main reference for the article! --Robert Stevens (talk) 13:45, 1 May 2008 (UTC)

I disagree, Robert. The standard for Wikipedia, after all, is verifiability, not truth. As such, it makes little difference whether the editor in question thinks that Popeye's account is correct. Inclusion of Popeye would constitute synthesis of sources to advance a position about the subject that hasn't been made in reliable sources, and that needs to be avoided. We shouldn't even try to determine whether OR is correct; we should just report on what reliable sources say.

Consider the following (slightly silly) hypothetical scenario. Suppose we did include Popeye's quote in order to contest Blow's claim. Another journalist, Moe, reads Wikipedia's coverage, and decides to interview Popeye. Popeye says "oh no, my words were taken way out of context. What I actually said was that our intelligence services determined that Cuba had no submarines as of the late 1980s. Wikipedia got it wrong." Now we're in a tricky situation, because instead of being a neutral observer, Wikipedia has actually entered the debate. We have no option but to include self-referential material. Jakew (talk) 14:05, 1 May 2008 (UTC)

To include Admiral Popeye's comment that Cuba does not have any submarines, we need to cite a reliable source that reports that Popeye says this. However, if there is such a source, I see no problem with including the fact that Popeye said it in the article on Cuban drug smuggling. More to the point, there are likely to be numerous other reliable sources that would back up the fact that Cuba has no submarines (such as Janes "Fighting Ships" which lists all ships in every Navy). Blueboar (talk) 14:26, 1 May 2008 (UTC)

If the subject of the article were "Cuban naval forces", I would entirely agree with you. However, unless the Popeye source (or Janes) makes these comments in the same context, I think it would constitute synthesis of sources to advance a position (namely, that Blow is incorrect). Consider what WP:SYN says:

• This entire paragraph is original research, because it expresses the editor's opinion that, given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it. To make the paragraph consistent with this policy, a reliable source is needed that specifically comments on the Smith and Jones dispute and makes the same point about the Chicago Manual of Style and plagiarism. In other words, that precise analysis must have been published by a reliable source in relation to the topic before it can be published in Wikipedia by a contributor.

Bearing this in mind, I think that we would need a source that specifically comments on the non-existence of Cuban submarines in the context of drug smuggling. Jakew (talk) 14:42, 1 May 2008 (UTC)

OK, let's toss in another curveball (this is rather fun, isn't it?): Jow Blow mentioned a specific class of submarine that was based in Cuba during the Cold War (on loan from the then Soviet Union). Editor Blowhard (the pro-Blow editor) has fleshed this out with various technical details of these subs (which, unsurprisingly, imply that they could have been used for drug-smuggling). In doing so, he has used naval sources which make no mention of drug-smuggling (or even, in some cases, the loan of these subs to Cuba). He considers this to be "background information": but the pro-Popeye editor (OliveOyl?) objects, accusing Blowhard of a WP:SYN violation. Is she correct? Should Blowhard and OliveOyl then haggle over whether discussion of these submarines is sufficiently "on-topic", and does OliveOyl have the right to demand removal of the info about these subs if Blowhard does not agree to relax his restrictive interpretation of what is "on-topic"? --Robert Stevens (talk) 15:10, 1 May 2008 (UTC)

From your description, I'd say that OliveOyl is indeed correct. The article should simply reflect what reliable sources have said about the subject, not try to build the case for either viewpoint using unrelated material. Jakew (talk) 15:17, 1 May 2008 (UTC)

I agree that a statement along the lines of "The allegation that Cuba is using submarines to smuggle drugs is false, because Cuba has no submarines." would need a source that states this. ... but I disagree with the contention that you can not include the simple blunt statement: "Cuba has no submarines." The policy states: Synthesizing material occurs when an editor tries to demonstrate the validity of his or her own conclusions by citing sources that when put together serve to advance the editor's position. In this case there is no conclusion or position being advanced. All that is being stated is a blunt fact. I think you are being overly strict in your application of the policy. One has to look at the intent of the Policy as well as its exact wording. And if you wish to use an overly strict interpretation of the policy, then I would say that we are dealing with a situation where IAR is eminently applicable. Blueboar (talk) 15:22, 1 May 2008 (UTC)

I seem to recall seeing something like that on one of these policy pages: "Let the facts speak for themselves" (i.e. if the editor doesn't embellish them, that's OK). However, I can't find it now. A change of policy, or have I just not looked in the right place? --Robert Stevens (talk) 15:34, 1 May 2008 (UTC)

I think you were looking for WP:PEACOCKTheRedPenOfDoom (talk) 17:07, 1 May 2008 (UTC)

Actually, there's another policy that could cause difficulties here: WP:UNDUE. It's quite possible that Blow's claims are considered to be so unlikely (perhaps because of the lack-of-submarines problem) that they represent a rather fringe viewpoint: but it's also quite likely that this fringe status is actually the reason why notable and reliable sources haven't bothered to address it specifically. I can think of many scenarios which are so far-out that no specific refutation is likely to be available. How should these be placed in perspective (i.e. given due weight) if WP:SYN is preventing a tailored response? --Robert Stevens (talk) 16:08, 1 May 2008 (UTC)

Interesting point. Looking at this from another point of view, if Blow's claims are so fringe that few secondary sources have even discussed them in any detail, then maybe Wikipedia shouldn't do so either. Maybe we haven't got sufficient secondary sources to justify having an article in the first place? Jakew (talk) 17:14, 1 May 2008 (UTC)

In the spirit of IAR, I'd say create a source. If it's this obvious, drop an email to a couple more accessible journalists pointing out the contradiction. It's probably not something you'd want to admit on-wiki, but realistically I think it's the best alternative. But I really don't see anything wrong with tipping off a journalist. Some people might see it as an attempt to skirt the rules, but quite frankly, the person you tip off is still the one making a call on the facts of the matter.

You might have more of a problem with a more obscure topic. Guettarda (talk) 16:27, 1 May 2008 (UTC)

(edit conflict) Blueboar, if this hypothetical article were to include the statement that "Cuba has no submarines" immediately after discussing Blow's claim, then the net effect would be to very strongly imply that Blow is incorrect. To my mind, that's advancing the editor's position, and I don't think that the intent of the policy is to say that OR is permitted as long as it isn't explicit. I can't see that it matters whether the argument is made explicitly or not; the point is that we should report on viewpoints expressed in reliable sources, rather than try to prove or disprove them ourselves.

Consider the "Chicago Manual of Style" example given in the policy. What we're discussing here is like saying: "Jones denies this, and says it's acceptable scholarly practice to use other people's books to find new references. The Chicago Manual of Style requires citation of the source actually consulted. The Chicago Manual of Style does not call violating this rule "plagiarism". Instead, plagiarism is defined as using a source's information, ideas, words, or structure without citing them." Sure, it doesn't explicitly make a link between Jones and the CMoS, but there is a strong implication there, and I think that the intent (as the policy basically states) is that editors should instead cite a reliable source making that specific analysis. Jakew (talk) 16:57, 1 May 2008 (UTC)

Guettarda's right. If an editor thinks unpublished "evidence" like this needs to be aired, the editor needs to take it to a reporter or find a publisher, but not to bring heretofore unpublished original research to wikipedia. It's worth asking if there is undue weight, for example could the story about the Cuban drug running be considered a WP:FRINGE issue? But editors can't "balance" the scale of a fringe topic with original research, there are much better ways to handle it. Professor marginalia (talk) 17:13, 1 May 2008 (UTC)

1. Indeed. Irrespective of how authoritative a source is, the topic is fringe until acknowledged (either positively or negatively) by multiple other reliable sources. As such -- and coinciding with the net result of Guettarda's suggestion -- there would need to be sources that discuss Blow's journalism in light of Popeye's remarks. Until that happens, WP editors shouldn't be adding two and two together.
2. OR (to include SYN) never contradicts NPOV. NPOV is established by providing both 'A' and a refutation of 'A'. The two must naturally be speaking of the same subject. When they are not speaking of the same thing, then the two sources are not comparable, and can't be used to balance each other. In this context, OR (to include SYN) occurs when an editor takes an only orthogonally-related source to refute 'A'. For example, taking Popeye's statement (which only notes Cuba's lack of submarines, but does not mention Blow or Cuban cocaine smuggling) as a refutation of Blow.
   To understand the distinction, imagine that the subject was Cuba's submarine fleet: Blow asserts it exists (and secondarily, that it is used it to smuggle cocaine). Popeye says it doesn't exist. Now the two are on the same wavelength, and can be used to establish NPOV. Get it?

-- Fullstop (talk) 21:00, 1 May 2008 (UTC)

It isn't actually that simple. From time to time I've come across things that I know are wrong, but that fairly reflect the sources provided. Back in the old days it was easy - we weren't too particular about sources. But as the quality improved, and as articles became more contentious, sourcing became a bigger issue.

Faced with the scenario that Robert outlined, it's entirely likely that I would have added the Popeye ref without giving much thought to the issue of SYNTH...after all, if you know the statement is wrong, and you can source it, common sense would dictate that you document the problem. Chances are, no one would ever notice until a Blow supporter (who was policy-savvy enough) came along and complained.

The issue here is that this isn't really what SYNTH and NOR were meant to protect against. The issue here is one of deriving novel conclusions. We should not do our own analysis, we should not draw new conclusions. Drawing a line between Blow's book and Popeye's prior statement falls somewhere in the middle continuum of drawing connections. If it were common knowledge that the Cuban navy lacked submarines, then pointing out the obvious error, while technically problematic, is balanced by common sense. One reliable source saying so...a little more problematic. A reliable source which lists all the vessels in the Cuban navy (from which one could conclude that the Cuban navy lacks submarines) - now we're getting into the type of stuff that these policies are meant to combat. The best way to avoid making the wrong call is to create a source. Guettarda (talk) 05:26, 2 May 2008 (UTC)

Another factor to consider is that what seemed to be an obvious error may later turn out to be correct, but we didn't have enough information at the time to realise. Similarly, what seemed to be obviously correct may later turn out to be incorrect. True, reliable sources can and do make mistakes, but we're not infallible either. Jakew (talk) 22:30, 2 May 2008 (UTC)

I recognize the dilemma, but the solution is simple even if the issue isn't:

As long as no one else has critically responded to Blow's book (at which point Popeye's comment would presumably have been noted), then Blow's book remains non-notable per Wikipedia:NOTABLE#General notability guidelines.

Inversely, once Blow's book has become notable there will no longer be a dilemma because someone else will have said something about Popeye.

But that regulation aside, when the dimwits "inclusionists" insist that Blow's hypothesis deserves an article on WP, and your perfectly-justified AfD has failed (i.e. IAR is in effect), you have no choice but to apply IAR as well. -- Fullstop (talk) 23:16, 2 May 2008 (UTC)
(Because lack of recognition equals 'fringe', I would have instinctively said 'fringe' instead of 'non-notable'. Unfortunately, wp:fringe is now worthless)

I agree that in this example, the claim's notability is the key to judging the dispute. But SYNTH should never be sacrificed to provide balance which is otherwise missing. I love wikipedia, but we all know that some of our fellow editors tend to get a bit full of themselves sometimes. In this case given to us, we have a Pulitzer prize winning journalist--and even he couldn't get his conclusions published without editorial review. Has a self-published reporter ever been awarded a Pulitzer for his report? I highly doubt it. There's a reason we don't allow unpublished or self-published claims in wikipedia, from anybody practically speaking, and it's not to leave more room for some wikipedian's claims instead.

Returning again to this example. We have a ref written by a pulitzer prize winner making Claim A, and several refs quoting Adm. Popeye making Claim B. Claim B perhaps contradicts Claim A, or so thinks a wikipedian and some bloggers. Everybody else is asleep on this big story but bloggers. Not even the pulitzer prize winning investigative reporter has the smarts to figure out the Adm's Claim B out there undermines his entire report. But we do have a wikipedian who's thinking, "If any human being alive has the stuff it takes to straighten this out and expose such a fantastic hoax, it's gonna be me!" Well, I think they need to also be thinking hard on this too: "but the encyclopedia is not the place for me to be do it in".

New conspiracy theories are literally born every day from ambiguous or seemingly contradictory juxtapositions just like what's given in this example. What needs to happen next is a reporter's task, not an encyclopedian's task. It's a reporters job to look further, because many conclusions are possible other than the one hypothetical wikipedian has synthesized. The only rationale proposed for including Popeye in the article about Cuban drug running is an assumption that Popeye's claim should jive with the Pulitzer prize winning reporter's, but that's not necessarily true at all. And if nobody knowledgeable who is watching the story thinks the two claims should jive, then it's no surprise at all they're not addressing the "controversy". If, for example, Admiral Popeye is giving a report on Cuban military subs does that mean he is also providing assurance Cuba has no drug trafficking subs? Conversely, would Popeye's report on the US naval fleet inventory also include the inventory of air and marine craft used in US drug enforcement? If not, would that mean there is a "controversy" over the discrepancy? Spend 5 minutes at 911 truth and know most of it is built from SYNTH drawn from disparate facts not all that much different than those posed in the hypothetical posed here.Professor marginalia (talk) 00:37, 3 May 2008 (UTC)

Yep, its precisely because we are not authorities (or -- if the ego says otherwise -- pretend that we are not) that we should abstain from leaping to conclusions. Either way you slice it, publishing novel conclusions on WP would make WP a source of never-before-published information. And that is what OR policy is really there to protect against. We can't undermine it. Ever. Or all the work honest people have put in will be worthless. -- Fullstop (talk) 01:38, 3 May 2008 (UTC)

The problem with "directly related"

I've run into a problem quite a number of times regarding the words "directly related". The problem, in a nutshell, is that people insist that whatever sources they use are "directly related" to the topic at hand, based on the argument they've constructed, regardless of what the actual topic of the article is. So, for example, if an article on Mr. X states that he was acquitted of murder, someone would then bring legal sources to argue that the judge in the case erred in his decision, based on the decisions in cases Y,Z, and W. When you point out that the sources do not actually mention the case of Mr. X, they insist that since the charges were identical, they are "directly related", and can be used to prove that the judge erred. Is there a wording that can more explicitly cover this problem? I'm thinking instead of

"if the sources cited are not directly related to the subject of the article"

stating

"if the sources cited do not refer directly to the subject of the article"

Thoughts? Jayjg ^(talk) 02:38, 17 April 2008 (UTC)

That makes sense to me. "Directly related" lends itself to abuse. The proposed change removes some of the ambiguity. Briangotts (Talk) (Contrib) 02:47, 17 April 2008 (UTC)

I could have really used this change of wording in the past...and what do you know, the only times I could think of this actually being useful were over law-related issues. A lot of editors unfortunately, but in good faith, feel the need to describe the strength of a legal argument that is presented in an article about the subject/origin of that specific argument. And of course they find reliable sources describing the strength of a very similar argument made by/at a completely different subject, or even more unfortunately resort to the laws or court decisions themselves. There are of course similar examples in non-legal situations, and they are similarly inappropriate. Someguy1221 (talk) 03:00, 17 April 2008 (UTC)

That seems like a reasonable wording change that more clearly expresses what is intended. Vassyana (talk) 03:16, 17 April 2008 (UTC)

The proposed change is an improvement. While the original "directly related" may have meant sources that refer to the subject, it can be loosely interpreted. The tighter "refer directly" closes that loophole. ·:· Will Beback ·:· 06:25, 17 April 2008 (UTC)

This is an excellent suggestion that should help to avoid an all-too-familiar problem. In a nutshell, editors need to be able to demonstrate that material is directly related to the subject of the article without creating a logical "loop" by relying on original research to do so. A possible alternative is "are not directly and verifiably related", but I think Jayjg's suggestion has slightly more clarity. Jakew (talk) 11:12, 17 April 2008 (UTC)

Referring to the subject of an article is too narrow. If Einstein is the subject of an article, and I want to include something about James Clerk Maxwell's theory of electromagnitism, which inspired general relativity, I can't use any sources that don't mention Einstein. In other words, the relevance may not be demonstrated within a single source; source A might say that Maxwell inspired Einstein, and source B might say something interesting about Maxwell's theory, without mentioning Einstein. Of course, this process could be carried too far, and result in original research, but I don't think this process should be prohibited altogether. --Gerry Ashton (talk) 12:19, 17 April 2008 (UTC)

I like Jayjg's proposal, but I think Gerry raises an interesting point. I have mixed feelings. I think the solution will not lie in the introduction to the policy but in how we explain the appropriate use of primary and secondary sources, but maybe not. In effect, i think Gerry is saying that some articles need background or contextual information that is not directly related to the topic. I agree. I have two points. First, a question: Gerry are there secondary sources that make the connection between Maxwell's equations and Einstein's? I would think so, and if the answer is yes, then your example really is not an objection to Jayjg's proposal. If the connection is made by a reliable secondary source, it is not a problem. If the answer is no - if we agree that an article may need contextual or background information despite the fact that no secondary source says it is related, we move to my second point, which is that I think we need to shift our attention from the introduction to the section on primary and secondary sources, and synthesis. The question now is, does background information simply help someone appreciate or understand something in the article, or is it being used to make an argument? If the former, there may be no need for a secondary source, even if it is not directly related. If the latter, there is definitely a need for a secondary source. I always see this policy in terms of how added information is being used. If it is being used to forward any kind of argument, it needs a secondary source for the argument i.e. that argues for/explains how one thing is connected to another. Slrubenstein | Talk 12:34, 17 April 2008 (UTC)

To answer Slrubenstein's question, there was a reliable TV documentary (I think it was Nova) that says Maxwell inspired Einstein; I'm not going to look up a full reference, but assume for the sake of discussion there is a secondary source. My example does not really illustrate advancing a position, so I suppose this clause wouldn't limit adding this kind of background information. But I still think there could be a case where secondary source A says "Jones believed in Smith's theory of X" and source B explains Smith's theory, which advances a position. --Gerry Ashton (talk) 12:50, 17 April 2008 (UTC)

Forgive me if I misunderstand, Gerry, but I don't understand why this is a problem. To consider your example, if Nova stated that Maxwell inspired Einstein, then Nova "refer[s] directly to the subject of the article", and we can document whatever they say. In all probability, Nova also gave some background information about Maxwell, which could be used without performing any synthesis.

The only thing we couldn't do is to use an independent source about Maxwell, which didn't mention Einstein. (Of course, in practise this wouldn't be a severe problem even if it did advance a position, since there are numerous highly reliable sources which no doubt discuss both in detail.) But we couldn't cite that independent source anyway, because according to the lead, "you must cite reliable sources that provide information directly related to the topic of the article". The only change is to clarify that "directly related" doesn't mean "I think it's directly related", but instead means something closer to "the source states that it is directly related".

So I guess I don't understand the problem. Jakew (talk) 13:17, 17 April 2008 (UTC)

The attitude in this policy shouldn't be "let's muzzle everything that might be a problem". Information should be allowed unless there is a clear reason to exclude it. In an article about Einstein, everything that lead to his theories is relevant. While the claim that a certain earlier theory influenced Einstein must be sourced, not every source used in the article to describe the earlier theory has to mention Einstein. --Gerry Ashton (talk) 13:38, 17 April 2008 (UTC)

The example you bring up doesn't strike me as at all potentially problematic. Assuming for the moment that there weren't any sources discussing Maxwell's theories in reference to Einstein, except to point out that they influenced him, then it would seem to me so insignificant that it deserves no greater a mention in Einstein's own article. We can still satisfy the curious by linking to some appropriate articles on James Clerk Maxwell or Maxwell's equations. Someguy1221 (talk) 22:44, 17 April 2008 (UTC)

The key point is that it's really not up to Wikipedia editors to decide which (if any) of Maxwell's theories are relevant to Einstein's. Instead, we let reliable sources make those connections. And if reliable sources haven't made those specific connections, then we shouldn't be doing so either - that's the very essence of Original Research. Jayjg ^(talk) 01:58, 18 April 2008 (UTC)

Exactly. I find it interesting that, with examples of "reasonable synthesis" such as this, it is likely that a reliable source has already performed that synthesis. So instead of worrying about how to exclude crank syntheses while allowing sensible ones (which is practically an impossible problem), we substitute a much simpler test: can we cite a reliable source that has already performed the synthesis? It may seem like muzzling, but in practise it usually just encourages good sourcing. Jakew (talk) 12:32, 18 April 2008 (UTC)

I agree it has to be more than "directly related", but I don't think there has to be a direct reference, necessarily. All we need is that it be verifiably related--which usually means a direct reference, but there could be exceptions. For example, suppose we cite Jackson Pollock in an article about abstract expressionism, even though Pollock didn't actually use the term abstract expressionism in every paragraph he spoke. Everybody in the art community knows that Pollock is the quintessential abstract expressionist, so citing him for thoughts on the subject could very well be "verifiably related". COGDEN 22:21, 21 April 2008 (UTC)

Gerry, I think you are doing yourself a disservice when you declare that Jayjg's proposal = "let's muzzle everything that may be a problem." It sounds like you are complaining about being censored. But you know that Wikipedia is not a blog or chatroom or any kind of a space where you or any other editor has a right to express whatever they think. This is an encyclopedia and we need some form of accountability, some way to ensure the quality of the articles and the information they contain. The main way we ensure this accountability is that any editor can edit, e.g. delete, anything they think is wrong. Anyone can delete anything you, I, Jayjg, or anyone else adds to any article. That is what makes this a "wiki." But people need some kind of guideline as to what to delete. Surely you do not want people just deleting anything they happen not to agree with! Surely you do not want that? So we have a WP:V policy; all material that represents a notable point of view and comes from a reliable source can stay. Jayjg is correctly assuming that the claim that x is related to y is a verifiable view. If it is, then there is a source one can peg it to and it stays. But if there is no source that supports this view, i.e. no source that says that x is related to y, well, then, there is nothing to prevent an editor from deleting it. This is not "muzzling" you, this is editing an article to ensure quality and it is what Wikipedia is all about. Slrubenstein | Talk 20:50, 17 April 2008 (UTC)

This whole crowd wants to muzzle everybody! I am removing this policy from my watchlist; I'm done discussing with this crowd. I will follow what I consider to be the basic principle of no original research, but will ignore the specific wording of the policy, because I do not accept the thinking processes of those who shape the specific wording. --Gerry Ashton (talk) 12:07, 18 April 2008 (UTC)

"Crowd?" Gerry, WP:AGF!!! Slrubenstein | Talk 18:03, 18 April 2008 (UTC)

I'm with Gerry on this one. If anything we need to loosen up the language in that part of the policy. People are taking it too literally, and it's making it too difficult to include background information. For instance, if I'm editing an article about Mr. X who was accused of crime C in state X, I wouldn't be able to cite the laws of state X to explain the penalties of crime C. Squidfryerchef (talk) 17:30, 19 April 2008 (UTC)

Obviously not. Thankfully, that information will no doubt already be included in the available reliable sources, if relevant. Dlabtot (talk) 19:29, 19 April 2008 (UTC)

It happens quite often that the article on Mr. X won't include the background information. There is nothing wrong with citing the background information, as long as it isn't used to support statements like "Mr. X must be innoncent because the law says this..." Squidfryerchef (talk) 02:28, 20 April 2008 (UTC)

It seems to me that we still need to think about what "directly related" means. To my mind, it makes little sense for it to mean "the editor inserting the material thinks that it is directly related", because I'm fairly confident that all editors inserting material think that it is related. It makes more sense to understand it as meaning that others should be able to verify that the material is directly related. As such, I think that this proposal is largely a clarification rather than a change to existing policy. But I may be wrong: what do others think? Jakew (talk) 23:26, 19 April 2008 (UTC)

Very often, the problem is that editors see a connection between two things, but do not bother to establish that they are directly related. If you want to discuss X in an article about Y, try finding a source that connects X to Y... then you can go on to discuss X. Blueboar (talk) 00:37, 20 April 2008 (UTC)

I won't argue that some editors try to cram unrelated facts into articles where they don't belong, but I think the OR policy should be about WP:OR, and not try to also be WP:RS, WP:COATRACK, WP:RELEVANCE, WP:CRUFT, and so on. Squidfryerchef (talk) 02:28, 20 April 2008 (UTC)

Sure, but a lot of NOR violations involve synthesis - arguing a connection between two unconnected things. Jagz just wants to clarify a vague aspect of this policy. and of course any fix is likely to point to V or RS because the opposite of original research is research that uses sources appropriately; to explain what we mean by research that violates this policy, it might be very constructive to explain what kind of research would not violate this policy!! Slrubenstein | Talk 11:48, 20 April 2008 (UTC)

I support the wording change. Many people fervently believe that certain things are "directly related" because they've gone over it so much in their minds that to them, they are. Jayjg's proposal will help reduce OR.Windy Wanderer (talk) 11:44, 22 April 2008 (UTC)

I have to support Gerry's point of view on this. "Refer directly" is too strict, and policies should always be conservative. While connections between topics should always be sourced, once they have been sourced there's no reason that the source that made the connection should be preferred over a better reference for the related topic. For example, in an article about a song about the Titanic, we might say that it's about the Titanic, source this fact, link RMS Titanic, and briefly explain the parts of its history that are relevant to the song; but the overstrict wording "refer directly" would unintentionally exclude these last facts. Dcoetzee 21:35, 24 April 2008 (UTC)

I agree. Once another source has made the connection, then the connection is not "original", and information from a more detailed source should be admissible. The Wikipedian, in this situation, isn't performing the basic act of synthesis: someone else did that, our guy is just filling in more details. --Robert Stevens (talk) 00:14, 25 April 2008 (UTC)

So it is, but that doesn't diminish the value of Jayjg's clarification. As Rob Stevens notes, policy would not fetter the editor when someone else has made the connection. Consequently, Gerry's Maxwell/Einstein projection wouldn't actually happen. Moreover, no one is actually going to waste time arguing that X is not related to Y unless he/she had good reason to. But Gerry appears to think that throwing a wrench into the works is everyone's popular pastime, which in my experience is neither true, nor (given AGF) is it a good assumption.

On the other hand, in a dispute -- and this is where policy actually kicks in -- Jayjg's "if the sources cited do not refer directly to the subject" is simply a logical continuation of what policy already says: Don't yourself construct connections.

-- Fullstop (talk) 01:31, 25 April 2008 (UTC)

Well, no, the existing policy IS in fact fettering editors. And surely policies kick in when disputes occur? So, given the latest example: if someone else has their own strange theory regarding the "real meaning" of the lyrics of the song about the Titanic, they could (and will) insist that the second and more detailed Titanic reference be excluded. Out of spite? Yes, possibly, but thanks to AGF we can't say so: they're just "enforcing Wikipedia policy". --Robert Stevens (talk) 12:36, 25 April 2008 (UTC)

I've revived this section to see if there's a consensus for the the change, as I've just come across another example of the problem. For over a year now an editor has been inserting this "Criticism" paragraph into the article on Honest Reporting:

The issue of perceptions of media bias has been the subject of experimental studies. In 1985 one such study first demonstrated the existence of what has come to be known as the Hostile media effect , where partisans of a particular cause display a marked tendency to see news coverage as being biased against their position. Not only do they see news coverage as biased, but are also more likely to explain the perceived bias as due to malign intent. Numerous subsequent studies have confirmed the existence of the phenomenon.

When you repeatedly point out to him that the section is a synthesis and original research, and that the sources nowhere mention "Honest Reporting", he responds:

You are reading "directly related" as "directly referenced". The published research on HME is nothing but "directly related" to a group that deals exclusively with claims of media bias. Asking that it mentions Honest Reporting by name is demanding "referenced" not "related", which is not in keeping with WP:NOR.

This "nothing is Original Research if I say it's directly related to the article" is exactly the problem I am trying to avoid with the wording change. Jayjg ^(talk) 00:27, 6 May 2008 (UTC)

I will once aain state my support for Jayjg's proposal. Let me make it clear that my support of this idea is very strong. I just do not find the opposition convincing. (On the contrary, they reinforce my belief in the need for this policy as a whole and this change in particular.) Supporting and illustrative information is perfectly fine, but the allowance thereof should be carefully limited. Using the example about a song, it would be perfectly fine to have a few short statements about the specific correlation mentioned by reliable sources. It would be wandering into coatrack and synthesis territory to add "background" details that are not even referenced in passing in the reliable source(s) making the connection. Another way this comes into play is when a single source makes a claim, perhaps supporting one or two statements in article, and then paragraphs of "background" material are added (making it both a synthesis and due weight issue). Such approaches are a common sort of coatracking and synthesis, and so it is an important issue to address appropriately in policy.) Vassyana (talk) 00:46, 6 May 2008 (UTC)

Yes, I too remain convinced that Jayjg's suggestion is sound. In the described case, the other editor is quite obviously wikilawyering -- going by the letter and not by the spirit. The spirit of "directly related" is that the cited material be on the topic of the article, and not orthogonal to it. This is not to say that tangentially related content (with sources) cannot also be in the same article, but the off-topic source must be properly contextualized, and should not be framed as if it were co-eval with an on-topic source. -- Fullstop (talk) 01:02, 6 May 2008 (UTC)