Wikipedia talk:No original research/Archive 23

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 20

Archive 21

Archive 22

→

a proposal - just the intro

I have been trying to edit the intro section for clarity and brevity - I also changed the "nutshell" based on a comment by Tim. I want to propose changing the intro of the actual policy to the following, and I will not do it unless there is general agreement.

My approach has been conservative - to change content as little as possible, but to streamline.

Here goes:

This page in a nutshell:

Wikipedia is not a publisher of original thought, nor a forum for promoting one's own point of view; all material must be verifiable
Facts must be backed by citations to reliable sources that contain these facts
Interpretations and syntheses must be attributed to reliable sources that make these interpretations and syntheses

Original research (OR) is a term used in Wikipedia to refer to unpublished facts, arguments, concepts, statements, or theories. The term also applies to any unpublished analysis or synthesis of published material that appears to advance a position — or, in the words of Wikipedia's co-founder Jimmy Wales, would amount to a "novel narrative or historical interpretation."

Wikipedia is not a venue for publishing, publicizing or promoting original research in any way. no original research, or NOR, is a corollary to two other policies:
Our neutral point of view policy (NPOV) encourages editors to add undisputed facts, including unbiased accounts of various people's views. It forbids editors from inserting their own views into articles, and demands that Wikipedia balance the relative prominence of differing viewpoints based on their prominence in the relevant field.
Our verifiability policy (V) demands that information and notable views presented in articles be drawn from appropriate, reliable sources.
Compliance with our Verifiability Policy and our cite sources guideline is the best way to ensure that you do not violate our NOR policy. In short, the only way to demonstrate that you are not presenting original research is to cite reliable sources that provide information directly related to the topic of the article; the only way to demonstrate that you are not inserting your own POV is to represent these sources and the views they reflect accurately.
NPOV, V, and NOR are Wikipedia's three principal content policies. Since NPOV, V, and NOR complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all three.

Any strong objections? Slrubenstein | Talk 14:23, 7 September 2007 (UTC)

directly related bolded as original, .. dave souza, talk 09:39, 12 September 2007 (UTC)

Good work, Slrubenstein. No objections from me. ≈ jossi ≈ (talk) 15:03, 7 September 2007 (UTC)

I like it. The only part I wonder about (though minor), is the policy on living persons. It appears that it is also a content policy (as it's categorized), and that it is also a core content policy. Should it or should it not appear then "...four principal content policies..."? wbfergus 15:15, 7 September 2007 (UTC)

I would argue that BLP is a tightening of V and NPOV, and as such it is an extension of both. ≈ jossi ≈ (talk) 15:18, 7 September 2007 (UTC)

That's not what NPOV means. Jacob Haller 17:44, 7 September 2007 (UTC)

I like it. I think it's better than the current intro, although we might want to re-word the NPOV paragraph, possibly like this:

* Our policy forbidding OR is a corollary to two other policies: our verifiability policy (V), which demands that Wikipedia only present established information found in reliable sources, and our neutral point of view policy (NPOV), which demands that Wikipedia balance the relative prominence of differing viewpoints based on their prominence in the relevant field.

—Preceding unsigned comment added by COGDEN (talk • contribs) 18:48, 7 September 2007 (UTC)

Approve or disapprove?

Approve:

I approve of this. wbfergus^Talk 17:19, 9 September 2007 (UTC)

I approve, with the exception that the word "established" is adding something to verifiability that is not there. Verifiability requires sourcing and verifiability requires that reliable sources be used. In almost all cases what is added will be material that is "established" (which seems normal and right) and I agree with the concept that Wikipedia is not the place to introduce willy-nilly challenges to what is established. If the word "established" were removed the section would be beautiful. I'd prefer that to further discussion about whether "established" belongs. That could be discussed later, along with the discussion on source typing and the discussion on synthesis. This is an opportunity for forward motion. Forward motion would be wonderful. --Minasbeede 17:42, 9 September 2007 (UTC)

I added the word because I was incorporating COGDEN's suggestion. If COGDEN has no objection I can remove "established." However, I think Cogden was right - whereas it is important for us to provide not just majority but minority views - the threshold for inclusion of views is only that they be notable, not established - I think the threshold for "factual material" i.e. information should I think be higher i.e. established. Thoughts? Slrubenstein | Talk 17:53, 9 September 2007 (UTC)

I agree. COGDEN 00:01, 11 September 2007 (UTC)

Whether or not you both agree "Verifiability" doesn't say what the draft claims it says - or didn't when I looked. To be honest, I also do not even know what you're driving at with "established." I made a small change in the mastodon article. I've been looking at various early (pre-1810) mastodon references using Google books. I don't know which of the references I find would be praised as secondary or tarred as primary but if I found something notable enough I'd probably at least consider editing it in. (Back then they made some errors, back then they got some things right. If I quoted something that arose because they got something wrong I would not do so to counter current understanding - but what they got wrong could be notable. Paul Seminon wrote a whole book about what they got intentionally wrong and the subject matter of that book is notable: it's little-known early history of the United States.) No doubt some early (or "primary") material could be misused but that's an editor relies on early (or "primary") is not evidence that the editor is trying to introduce OR. Those who are not in any way trying to introduce OR when they use primary sources neither need nor want some other editor breathing down their neck about their choice of sources.

I've looked at contemporary (to the event) accounts of some murders of native Americans by white men in the early history of the county where I was born. I can't recall which look "primary" and which look "secondary" but for all of them it's pretty clear that it was established (way back then) that the white men did commit brutal, cold-blooded murder. (I've edited none of this into Wikipedia - but I could, if it were pertinent to an article.) Perhaps all the sources are secondary, perhaps not. They aren't contemporary (meaning now) "established" secondary sources but I see no reason to mistrust them.

The best guess I can make is that this insistence on broadening the policy so that it deprecates some types of source all arises from particular cases and that the intent is to bend the overall policy to make it a weapon that can be used in those particular cases. In any case, "Verifiability" shows no preference for "established" that I can see. If you want "Verifiability" to say that go there and edit it and see if it flies. Things are already hard enough in NOR without it also being used to alter "Verifiability" in some strange, indirect fashion. --Minasbeede 02:33, 11 September 2007 (UTC)

The NPOV policy says this about facts; "By "fact" we mean "a piece of information about which there is no serious dispute."" For the modifier for "fact" I guess you would not object to changing "established" to "undisputed?" Slrubenstein | Talk 09:54, 11 September 2007 (UTC)

Taking out "established" would leave a fully accurate statement. Put "undisputed" in and that's an invitation to, well, dispute.

I think an encyclopedia ultimately is a "best effort." It's not perfect, it's not likely to ever be perfect. The words you want to put in may result from laudable goals but their effect is very likely to be continued and continuous discomfort, discomfort that will show up as edit wars on WP:NOR or as recurring complaints in WT:NOR. Putting the words in have close to zero practical effect on the material that gets included. The underlying notion seems to be that there's some subset of all sources that really has properly weighed every aspect of every notable thing and that Wikipedia will achieve perfection by echoing just those sources. That's unrealistic. Nor is it the foundation Wikipedia policy. The EB, after WW1, did not try to present an unbiased report of the causes of that war: the EB editors recognized that they couldn't, there was no such thing and explicitly stated it couldn't be done. Instead they included a section on causes from the German point of view and a section from the British point of view. They didn't aim at absolute perfection, they aimed at NPOV. Wikipedia aims at NPOV, and Wikipedia recognizes that for many issues there is no single "established" point of view. NPOV covers what you would do with "established" adequately, and in a way that works. I'd say leave Wikipedia the way it is and cease trying to find a formula for making it perfect (in the "established" sense, for instance.) It can't be done. --Minasbeede 10:18, 11 September 2007 (UTC)

This (page) is NOR. The above seems to be concerned with a mix of V and NPOV issues. It surely grates for someone to introduce something meant to contradict "established" thought (and while that's in itself a complex issue let's let it lie) but that it grates and that it is fully proper within the spirit, policies, and guidelines of Wikipedia to act to remove the material doesn't really makes it an NOR issue unless it's a novel interpretation. "Novel interpretation? Sorry, we don't do that here. It goes." It would seem novelty is already covered quite well by NOR. If reinforcing language is needed care must be taken that the reinforcing language doesn't go beyond reinforcement into creating new restrictions. The threshold isn't one that applies to the material, it applies to the (in the example, improper) use of the material. The csame material from a primary source may be proper and welcomed used one way, improper and destined for (rightful) speedy deletion in another. It's the same material (in the hypothetical example.) It is the way it is used that makes it "bad" and it is the way it is used (misused) that makes it subject to deletion. That's enough, NOR covers improper use and does so adequately. NOR isn't about what kind of material is used to provide verifiability, it's about how the material (which we can assume is verifiable) is used. If the material represents original research it should go, by NOR. If the material is not "notable" then that's an issue outside NOR - but the material should go. Cover the notability outside NOR, wherever it best fits. --Minasbeede 19:14, 9 September 2007 (UTC)

The current intro discusses the relationship between NOR and NPOV and V (just as the intro to NPOV discusses V and NOR, and the intro to V discusses NPOV and NOR). The above is only meant to replace the introduction as I stated before. I think it is reasonable to take a conservative approach and continue to place this policy in the context of the other two major content policies. Slrubenstein | Talk 09:54, 11 September 2007 (UTC)

Disapprove:

Points 2 and 4 misdescribe NPOV. Jacob Haller 17:58, 9 September 2007 (UTC)

You are wrong. I quote NPOV: "representing fairly and without bias all significant views (that have been published by reliable sources). This is non-negotiable and expected on all articles, and of all article editors." "NPOV says that the article should fairly represent all significant viewpoints that have been published by a reliable source, and should do so in proportion to the prominence of each." "A common type of dispute is when an editor asserts that a fact is both verifiable and cited, and should therefore be included. In these types of disputes, it is important to note that verifiability lives alongside neutrality, it does not override it. A matter that is both verifiable and supported by reliable sources might nonetheless be proposed to make a point or cited selectively; painted by words more favorably or negatively than is appropriate; made to look more important or more dubious than a neutral view would present; marginalized or given undue standing; described in slanted terms which favor or weaken it; or subject to other factors suggestive of bias." Jacob Haller, I doubt your vote should even count since you reject the very existence of anNOR policy and, it seems, NPOV too. Slrubenstein | Talk 18:15, 9 September 2007 (UTC)

That is a very good summary of why we need these aspects well described. Lately I hear a lot of comments about "this is properly sourced" as an argument for inclusion of a claim, when it actually violates other policies or their spirit. ≈ jossi ≈ (talk) 18:18, 9 September 2007 (UTC)

Slrubenstein, you said "Jacob Haller, I doubt your vote should even count since you reject the very existence of anNOR policy and, it seems, NPOV too." That is a personal attack, and it is false.

I am concerned with improving the quality of references on Wikipedia. I have made several suggestions for improving the policy. I simply oppose privileging secondary sources, because I believe that some primary sources are better than some secondary sources, and in my experience, secondary and tertiary sources are more likely to lead to improper synthesis.
I am also concerned with reaching neutrality. There is an ongoing debate about the purpose and nature of WP:NPOV (see the talk page). I follow WP:ASF as the guiding principle.
I have noted that some secondary sources (e.g. the Catholic Encyclopedia) add bias (e.g. labeling people heretics), and have used both primary and secondary sources to reduce this bias. (e.g. Photinus, using primary sources, sum of my edits). We may disagree about the best approach, but this was source-based bias-reduction.
The proposal states that "Our neutral point of view policy (NPOV) forbids editors from inserting their own views into articles, and demands that Wikipedia balance the relative prominence of differing viewpoints based on their prominence in the relevant field." which seems to remove the critical context of "representing fairly and without bias all significant views" (emphasis mine) and "NPOV requires views to be represented without bias. All editors and all sources have biases."
- I suggest adding ASF ("Assert facts, including facts about opinions—but do not assert the opinions themselves.") into point two, either in addition to or instead of the current phrasing.
- Adding ASF would resolve my concerns. Except for the description of NPOV, I agree with the proposal.

Do people agree that ASF would be a better description of NPOV in the context? Jacob Haller 19:36, 9 September 2007 (UTC)

No. This is just another one of your transparent attempts to make an end run around NOR. Give it up. Slrubenstein | Talk 19:46, 9 September 2007 (UTC)

Why don't you want the standard "simple formulation" of NPOV? How does it become an "end run" around NOR? Jacob Haller 19:56, 9 September 2007 (UTC)

Okay, I do not believe I am misrepresenting NPOV. However, you do single out one important point of NPOV and in the spirit of compromise I have, as you asked, added it to point two of my proposal. I modified the wording a bit, because it has been taken out of context and thus needs a little explaining, but i think what i did honors the spirit of your request. Slrubenstein | Talk 00:59, 11 September 2007 (UTC)

Okay, I like the new version - especially the reference to undisputed facts. I would recommend:

Undisputed facts - include.
Disputed facts - include if relevant; association is helpful if there are substantial disputes.
Any opinions - include if relevant, but associate with specific authors, schools or traditions.

Strictly speaking, NPOV concerns the last one.

I might suggest:

"Our neutral point of view policy (NPOV) encourages editors to add undisputed facts, including descriptions of various people's opinions. It forbids editors from inserting their own views into articles, and demands that Wikipedia balance the relative prominence of differing viewpoints based on their prominence in the relevant field."

If you prefer. Jacob Haller 17:38, 11 September 2007 (UTC)

The point about directly related is in bold in the original, so I've boldly bolded it here to retain the emphasis on this essential point. The "undisputed facts" seems confusing when it's divorced from the more detailed explanation, after all the age of the earth is hotly disputed, the point is that for our purposes it's a fact that various verifiable sources give statements about the age of the earth. To me it would be much clearer to have "encourages editors to add verifiable published facts..." ... dave souza, talk 09:39, 12 September 2007 (UTC)

Well, in this case the age of the earth is not a fact. It is a view scientists have based on other facts (data) and you are saying scientists have different views (i.e. interpretations of the data). Slrubenstein | Talk 10:32, 12 September 2007 (UTC)

Okay, I have made changes based on some comments, and others have made some edits. For the sake of moving on, can I replace the current introduction with this? As anywhere in Wikipedia, nothing is permanent, I really only ask if people agree this is better than the current intro. if so, I will make the change and we can continue discussing other issues. Slrubenstein | Talk 14:10, 12 September 2007 (UTC)

Fine by me, as you say that particular word can fine-tuned in future after due discussion. .. dave souza, talk 14:22, 12 September 2007 (UTC)

Looking Better

I'd pare back the paragraphs on Verifiability and on NPOV to the minimum. Referring to the other two policies is about all that is needed here and it is in those other two policies that the explication of each need be done. I think lurking in all this is the principle that V doesn't trump NOR and NPOV (etc.), which is true. The policies stand together. If the language in NOR can be made smoother by whittling some of it away that's all to the good. I am not advocating whittling away V or NPOV, only simplification. As these are core policies it would be almost essential that if language about any other policy appears here it would have to exactly mirror the policy as explained in its own section: anything else leads to confusion. So if it appears and is done right it's just an echo and that ought not to be necessary: you don't succeed as an editor by following only one policy or just two of them: it's all three that have to be satisfied. None of the policies has to carry any weight for the others: they carry their own weight. Expressing the need to adhere to all three is surely something that ought be expressed in all three, but that's about it.. Even that's a bit redundant but it's probably appropriate and it shouldn't add a large number of words.

I have an issue with "in any way" but I'd leave it in. "In any way" is almost a stand-in for the synthesis section (and the extreme interpretation it represents) and that section could (and surely will) be discussed once the current issue is resolved. That discussion can determine the fate of "in any way."

There is visible forward motion. Thank you, Slrubenstein. --Minasbeede 15:44, 11 September 2007 (UTC)

I think I understand the concern over the NPOV point. It should lead with balance. Adding facts is about NOR or V. A reorder might work well to resolve.

There is ambiguity over the way it says you cannot add your own point of view. So I want to write that the world is round - that is my view point. Again the point is to ensure balance, which a group of editors might achieve. The danger is that you may not be able to evolve an article with this writing.

Final comment is that I would phrase and editors should try to familiarize themselves with all three simply as content must comply with all three. That avoids the lack of prescriptiveness which is actually a requirement, it is weasely at the moment. Spenny 17:33, 11 September 2007 (UTC)

Please. Why complicate the description of NPOV and why push NOR into NPOV in the NOR policy page? NPOV doesn't speak to original research, NOR does. This is NOR. Do it here and do it as NOR (and not claim that NPOV does it. NPOV shouldn't address NOR, that's not it's focus.) If NPOV says any of that (in the current draft) clean up NPOV. It would still be true that all editing ought to comply with all three policies. As all three policies start with "This page documents an official policy on the English Wikipedia. It has wide acceptance among editors and is considered a standard that all users should follow" there is already language that asserts the policies as being required.

Just say: "Our neutral point of view policy (NPOV) demands that Wikipedia balance the relative prominence of differing viewpoints based on their prominence in the relevant field." Any further expansion of what NPOV requires belongs in NPOV. NOR excludes OR. NPOV can rely on that and doesn't have to further exclude OR: that's not what NPOV is about.

I can understand that it is known or believed that there have been instances of editors grabbing something from a primary source and using that to push a novel point of view. That fails on the basis of the novelty. Such edits violate the NOR policy and the policy doesn't need to be encumbered with a lot of added language aimed specifically at such offenses: it's wrong because it's novel. --Minasbeede 20:34, 11 September 2007 (UTC)

I'd agree with your suggestion. I was trying to nudge in that direction, as the worst thing about the NPOV point was the first thing it did was not explain NPOV. It could be very light-weight. The more you explain, the more the explanation becomes the problem rather than the policy. That difficulty is best addressed in one place. Spenny 21:04, 11 September 2007 (UTC)

I do not. What I wrote is a pretty straightforward summary of NPOV - it includes more than minasbeede wants, because there is more to NPOV than Minasbeede suggests. And if you do not know by now, you should: NPOV is non-negotiable and dictated by the Foundation. If you think there is some overlap between NOR, V and NPOV, you are right and there is a good reason: NPOV is the core, the rootstock, and NOR and V grew out of it. I think it is essential to explain NPOV in a way that people see how NOR has its roots in NPOV, and what is what I did - while being guided by what NOR actually says.

I think people are - with good intentions I am certain - quibbling right now. Vassyana has just appealed for mediation. I would like to propose adding my new introduction to the Policy today, so we can actually move forward. No one (including myself) is going to be 100% happy - we just need to get used to that here (I already made some changes incorporating other people's ideas - including ideas of people with whom I have vehemently disagreed; we all need to be willing to do that). Slrubenstein | Talk 07:47, 12 September 2007 (UTC)

NPOV says what NPOV says. Whether it's non-negotiable or not I doubt that the foundation insists that the particular wording you favor appear in the introduction section of NOR - and that is what is under discussion. I'm not appealing for things to be removed from Wikipedia policies, I'm appealing for things to be neat and orderly. You claim NPOV laps over into NOR and that you are putting that into the introduction. (I don't see what you claim NPOV says when I read NPOV, but I did read it pretty quickly.) I'm saying let NOR be about NOR and NPOV be about NPOV and not meddle in each others principles. They still interact and interlock, they still must be followed (with editors understanding that "must" is moderated by "Be Bold" and IAR.) --Minasbeede 14:36, 12 September 2007 (UTC)

The mediation is, I believe, about the primary/secondary source issue. I do not have a real problem over your reworded introduction: my notes here were just intended to a) record that the NPOV point could be better by addressing POV as its first concept, which I see you have addressed; and b) as policy is non-negotiable, I was trying to suggest a formulation that reflected that - "should" is weak, content "must" comply is strong yet allows for the newbie not feeling that they are being told what they must do before editing - it allows the wriggle room that a newbie needs (their actions are not incorrect per se, they are encouraged to dive in knowing that others will sort them out) without it being a matter for dispute that it is legitimate to remove their edits for non-compliance with policy. I'd suggest that is not a quibble, though it is a small point.

For the avoidance of doubt, I was not thinking of objecting, even in its current form. Spenny 08:34, 12 September 2007 (UTC)

I agree. No one (or at least very few) are going to be 100% happy. But I also think that what's been haggled back and forth lately on the Introduction is better than what is currently present in the policy. It may still not be perfect to all paeople, but it is a step in the right direction. I would prefer to move beyond this stage and begin to address and work out what I see as more substantial and contentious issues. Everyone involved should be willing to do a little "give and take", that's how concensus is finally achieved, though everyone usually isn't 100% happy about the concessions. wbfergus^Talk 11:05, 12 September 2007 (UTC)

The core issue isn't 100% happiness, the issue is having policies that are not a constant source of irritation (and the resulting "edit warring" and discussion on talk pages.) For example, if an offense against NOR is that the editor has introduced novel material then the novelty is the offense and there's no need to claim, within NOR, that in some indistinct manner NPOV forbids that. The offense is directly against NOR and that's sufficient. There has never been any need to add specific language about that specific offense to the NOR policy. My current issue is about style and method of presentation: it's about neither source-typing (the example cited is illustrative) nor about synthesis: make the introduction clean. Get the introduction solid, then proceed to source-typing and (eventually to) synthesis. If problematic material is left in the NOR policy page then it will continue to be a point of contention all the time it is there. That's not something to be facilely swept away by asserting that, well, nobody is going to be 100% happy. --Minasbeede 14:36, 12 September 2007 (UTC)

The issue is that people won't be 100% happy as long as problematic material is left in. The issue is that people will never under any circumstances be 100% happy because people have widely varying ideas of what is and is not problematic material. Go ahead and delete all the materal you know to be problematic. The result may be that you will be 100% happy, but most other people will be very unhappy. That is why we get things done through compromise.

All three major content policies say something like this: "NPOV, V, and NOR are Wikipedia's three principal content policies. Since NPOV, V, and NOR complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all three." My version of the intro mentions NPOV to the extent I think necessary to explain why this is so. It is also meant to help educate people as to why we have an NOR policy. You want a succint statement of what the policy is and what counts as an offence - whell it is there, in the "policy in a nutshell." Can't get more succinct than that. But hey, if you think the current intro is better than mine, well, I'll just drop it. I was trying to be helpful. Slrubenstein | Talk 14:49, 12 September 2007 (UTC)

This isn't all that complicated. I don't think NPOV prohibits editors from adding their own opinions, I think NOR (this policy) does that. NPOV requires level treatment of all opinions, analyses, etc. that comply with the other two policies, including this one, NOR. It's inaccurate and unnecessary to say NPOV forbids editors' own opinions: that's done by NOR. NPOV addresses the issue of editors giving unfair emphasis to a particular point of view. If that point of view is an instance of OR then NPOV never comes into play: that point of view shouldn't appear at all. There is no issue over fair treatment (according to NPOV) of material that shouldn't appear at all. The introduction to NOR should not say that NOR (or some aspect of it) is a subset of NPOV. It isn't. --Minasbeede 14:57, 12 September 2007 (UTC)

No it is not complicated at all. NOR existed for years before NOR, and during that time as now NPOV prohibited editors from adding their own opinions. This policy just took that one element from NPOV and expanded it. be that as it may, the policy (including my version0 does not say that NOR is a subset of NPOV. It says that it cannot be understood independent of NPOV or V. Slrubenstein | Talk 15:01, 12 September 2007 (UTC)

(Should the first "NOR" in the above paragraph be "NPOV"?) What you wrote is "...forbids editors from inserting their own views into articles." I think editors are thus forbidden; my point is that this is done by NOR, not by NPOV. IF NPOV historically said what you claim I can understand what you are saying much better but now that there's a specific NOR policy I think the historical reference needn't appear in the introduction. I'm not trying to change anything with respect to policy, I'm just trying to make the wording better and less confusing. You have been doing a very good job. --Minasbeede 15:28, 12 September 2007 (UTC)

I wasn't questioning your good faith, I do not think you are criticizing the policy. My understanding is this: NPOV has always forbidden editors from expressing their own views in articles. NOR was devised to address a particular way in which editors were seeming to obey NPOV and V yet were actually putting their views into articles (by combining sources in ways that moved beyonjd what individual sources were saying to forward an argument the eidtor wanted to make). Anyway, for now - should I put the proposed new intro into the policy page? I'll wait another 12 or 15 hours before I do anything. Slrubenstein | Talk 15:57, 12 September 2007 (UTC)

By the way, thanks, Slrubenstein | Talk 15:57, 12 September 2007 (UTC)

That NPOV has always forbidden editors from expressing their views in articles may be historically accurate but when NOR was created I think that prohibition passed over into NOR. I don't see the prohibition in the current NPOV (and I'm glad I don't, I think that when NOR was created it was entirely proper to move that entirely to NOR. Leaving it split between the two would be horribly cumbersome.) It's not deceptive to say NPOV says (although really, it "said") that in the Introduction but it is confusing: a person reading that might go to NPOV and search for it and not find it. The statement in the Introduction could be modified with the word "historically" or there could be a footnote or a parenthetic expression but that tends to make the Introduction (well, maybe the word "historically" doesn't do it) cumbersome. Leave it in if you must; be aware that is a source of confusion. Confusion in policies is not good (but I assert that in an entirely OR fashion.) I would not myself want to be responsible for causing confusion in a policy statement. I don't for a moment believe that's your intent but I do warn that it could be a result. --Minasbeede 16:46, 12 September 2007 (UTC)

Synt check requested

Can an article have a SYNT problem by implication? The article George Washington and religion contains what I find to be a subtle form of synthesis... using original documents in a way that very strongly "implies" that Washington held certain religious beliefs (specifically that he was a Deist). However, the article does not actually come out and say: "Thus, Washington was a deist". Is this a NOR/SYNT situation or not? I would appreciate an uninvolved opinion. Blueboar 15:30, 7 September 2007 (UTC)

If that's a matter of legitimate controversy (both words matter, I'd say) (and relevant to the article in question) then NPOV requires both sides of the issue to be presented, so a step backward can be taken and should be. It looks like this is a small part of something much larger. What's the best way you can find to resolve this - or can it be resolved? --Minasbeede 20:35, 7 September 2007 (UTC)

If this question refers to the Abercrombie quotation it would seem that it is verifiable that Abercrombie said it. I don't think anyone would claim that Abercrobmie is the definitive source on Washington and religion and it's likely that Abercrombie meant no compliment by the words (and could have meant that Washington's actions sure looked like the actions of a deist, which Abercrombie deplored - both deism and looking like a deist by one's actions. The quotation as given doesn't indicate that Abercrombie spat but that seems very possibly to be the sense of his words.) The article itself seems to indicate the question of Washington being a deist is still "up in the air" and that overall message of the article would seem to neutralize any false interpretation or promotion in importance of what Abercrombie said. I'd not automatically delete the passage. --Minasbeede 20:54, 7 September 2007 (UTC)

I was not focussed solely on the Abercrombie quote... but it is a good illustration. The issue that keeps popping up in my mind is that Abercrombie is a primary source... a witness to Washington's life. Without a reliable secondary source that takes Abercrombie's statement and uses it to say "based on this, I conclude that Washington was, or could have been a deist" doesn't the use of Abercrombie amount to a Synthesis? Please understand that the idea that he might have been a Deist is not really the issue for me... the issue for me is that the article implies what his religious views were based on primary documents, instead of citing a reliable scholarly sources that say he had this or that view. Isn't that OR? Blueboar 15:06, 12 September 2007 (UTC)

Scholarship, aware of Abercrombie, hasn't (apparently) decisively concluded that Washington was a deist (nor that he was not.) I think what Abercrombie said is relevant, I think it is worthy of inclusion in the article, I think that if some readers have the notion that Washington was a deist reinforced in their own minds by that inclusion that there's no problem. The Wikipedia editor can't himself make that conclusion and shouldn't nudge the reader toward it. As Abercrombie was an eyewitness for Washington's behavior I think it is completely appropriate to cite his observation. Wikipedia is not (I'd hope) supposed to block or impede thinking by its readers. It's a plus if readers think about the implication of Abercrombie's words. It's a greater plus if they decide to investigate more deeply, to find out why it is that scholarship doesn't take this one piece of evidence as definitive. In any case the reader is educated by learning that Washington routinely avoided communion (I was.) There is content there, and part of the content, which Wikipedians I suppose can't enunciate in that article, is that even in the 18th century men thought, made decisions, and acted on those decisions and on principle. They were not mere play-actors acting out 21st century scripts for what they were, did, and represented (as written by various "playwrights.") --Minasbeede 15:21, 12 September 2007 (UTC)

But that is part of my problem with the article, There is no discussion or any citations to show that scholarship is awair of Abercrombie, much less not drawn a conclusion. Isn't all that a form of OR - without a discussion of what modern scholars make of the material, doesn't if become a form of sythesis? I am not saying that the article should cut the Abercrombie quote... I am questioning whether using it - in the way it is being used - amounts to a subtle form of synthesis, and thus OR. Don't you need to cite a secondary source that has examined the issue?

Let's move this beyond the specifics of any particular article. Can you have an implied conclusion, based on what a primary sources says, that equates to original research? If an article stated "Thus, we can conclude that such and such is true" without citing to some reliable source that reached this conclusion, we would be all over it for being original research... my question is: Can an article be written in such a way as to lead the reader to form a specific conclusion, effectively saying "Thus..." without actually saying it? Blueboar 15:55, 12 September 2007 (UTC)

Using the Abercrombie quote I think that the reader might very easily be induced to think. Wouldn't it be a --Minasbeede 16:37, 12 September 2007 (UTC)violation of NPOV to suppress Abercrombie in order to favor the apparent attitude of scholarship? Presenting evidence that goes against the settled view ought to be possible within NPOV (and suppressing evidence that opposes the settled view seems to be a violation of NPOV.) I think this is an issue of fair importance and I'm glad you brought in the example. An editor might believe Washington (probably) was a deist and might use the Abercrombie quote to bolster the evidence for that belief and give it greater strength within Wikipedia. If the article (I'd have to look again and I haven't so I say "if") otherwise basically agrees with what scholarship has concluded then I think the Abercrombie quote is proper: the quote balances the main part of the article. The NPOV question would be whether the editor who includes the Abercrombie quote is obligated to search scholarly articles and find whether or where Abercrombie is discussed and dismissed. It wouldn't be wrong if the editor did that, doing that would be NPOV-consistent behavior. But if the article already mostly reflects scholarship need the editor go that extra mile?

We could also speculate on what the significance of the Abercrombie quote would be if it were newly discovered (say in a diary passed down through generations and just now revealed and published or by the editor happening to find it in an old and overlooked work.) This would be a situation in which (as this example is supposed to indicate) scholars have been totally unaware of the material. If the citation is valid, if the work is published, including the Abercrombie quote surely satisfies verifiability. The Abercrombie quote is evidence perhaps pertinent to an ongoing discussion (whether or not Washington was a deist.) Indicating Washington was a deist isn't novel within that discussion: that's what the discussion is about. The editor hasn't manufactured the evidence, hasn't twisted the significance of the evidence. In this hypothetical example the real question probably is whether Wikipedia should hold off on revealing the (verifiable) material until after the scholarly world has had a chance to discuss it and to publish analyses (on the scholarly timetable.) Is the NPOV prohibition on new ideas or on new (but published, verifiable) evidence? One could be hyper-literal and insist on suppressing new, valid evidence but I can't see very much harm (in the NPOV sense) with publishing the evidence, particularly in a manner that avoids expressing an opinion. If the evidence tends toward inducing an opinion within the mind of the reader that's not bad: evidence is not supposed to be neutral. If (again for the purposes of example) the next paragraph in the source of the quote about Washington says that Abercrombie was a raving maniac it would probably be contrary to NPOV to leave that out. --Minasbeede 16:37, 12 September 2007 (UTC)

I agree completely about the NPOV aspects of the issue... and again, I am not arguing that the quote should be removed. It's how the quote is used that bothers me, not that it is used. I would feel a lot happier if there were some modern scholar to cite who discusses the quote, but I can not find one, and no other editor has provided a citation. In fact, there isn't even a reliable secondary source that claims he was a Deist (what we do have are several citations to slightly POV sources that go to lengths to refute the idea, which tells me that someone somewhere says he is... but no one has been able to actually cite that someone.) That is why I am focussed on the OR question. To stay with the specific article for a moment, it used to be much worse... When I first discovered it, It contained a section entitled "Evidence that he was a Deist" (or something similar) that basically consisted of a bunch of quotes from primary sources as "proof" that he was a Deist. While there was no overt statement of conclusion, stating that he was a deist, the section title itself sort of formed a conclusion. Its existance made it more clearly a case of Synt and OR... especially when combined with the fact that there was little to no presentation of any other POV. It took a while, but most of that OR has been fixed ... other POVs have been introduced and the more blaitant OR violations have been dealt with (Although I feel that some of the fix entailed counter ballancing with other OR based on other primary sources with different POVs). In any case, We are now looking at the grey zones... places where the degree of OR involved is more a matter of interpretation of this policy and less obvious. Thus my raising the issue here.

Again, let's move away from the specifics of one article for a moment and discuss the general concept of implied OR and Synthesis. Leaving NPOV issues aside (if we can), to what degree can we use primary source material to plant a conclusion in the mind of the reader? If an article that relies heavily on primary material is phrased in a way that clearly leads the reader to reach certain conclusions, but avoids stating those conclusions outright, does it still violate OR? Can you have a synthesis without actually stating a conclusion? If the answer is 'no', then my concerns about a specific article are completely off the mark. If the answer is 'yes', then we can explore whether the the specific article fits that model. Blueboar 17:47, 12 September 2007 (UTC)

Wikipedia has answered the question: "Be Bold." What this really gets tangled up in is whether the material fosters NPOV (and if an article basically claims Washington was not a deist then "Evidence Washington was a deist" seems to be compliant with and in the spirit of NPOV) instead pushes a fringe attitude (oh, like "Washington was a robot" or "Washington was from a distant solar system.") For the case given the quote from Abercrombie alone is enough to convince me that it isn't fringe to think Washington might have been a deist. I think NPOV has been well served by that quotation: it gives reason to think. I again assert that it is not Wikipedia's function or goal to avoid material that engenders thought, even if that thought is contrary to "established" views. It is proper for Wikipedia to avoid being the tool of those with fringe beliefs. There's no default way to treat any particular case: thought is necessary every time. (Thought can sometimes fail or go astray. What else isn't new? That's life. There's no formula that's going to work here.) --Minasbeede 18:01, 12 September 2007 (UTC)

Minasbeede, I don't think that being Bold is an excuse to violate policy... The question is does it? But, please, think about this separate from the specific article ... and separate from NOPV. I understand what you are saying, and agree with you. But I want to explore the broader issue from strictly a NOR view point. For the moment, let's assume that the specific article does not have any OR issues. Let's make it a hypothetical... Does implied OR or implied Synt exist? Blueboar 18:30, 12 September 2007 (UTC)

I also don't think "Be Bold" can be cited as an excuse to violate policy. In the current example I don't think "Be Bold" is a violation of the policy. The line between admissible and non-admissible is unclear and I say "Be Bold." Yes, some may say you crossed the line and others may say you didn't. The "Be Bold" policy seems to protect you from severe criticism: the line is unclear and you chose "Be Bold" over "quaver in indecision" (which defaults to "do nothing.) "Be Bold" would seem to be a conscious encouragement to you to act rather than refrain from acting. Rest assured: someone will be watching to prevent you from doing wrong.

As to implied OR, if the subject matter of the article is X (and affirms that over not X) then any facts cited that favor "not X" are not OR (since they are facts, which, in the context of this discussion, are also sufficiently verifiable or they'd be rejected because they failed that test.) "Not X" is already an idea implicit in the article, citing the facts to support the idea is not the introduction of that idea, not OR. In an article about Washington being or not being a deist you don't introduce the concept of Washington's being a deist by citing a fact that lends support to the idea he was one: that is not a novel idea introduced by you. Seeing some facts support some POV is not the same as creating and asserting a new (novel) POV. That they counter the thrust of the article is at least nominally adherence to NPOV and in fact brings NPOV to an article that needed it. (Always with the caution that NPOV is not a tool for bringing in fringe positions.) The current rules of Wikipedia are such that even if you have a citation from a person who asserts "I am not X" you can't directly say that those are in error who claim the person was X. You can only show the facts: to introduce the statement that the facts contradict those other assertions requires that you quote someone else who says the statements contradict those other assumptions. More briefly: Implied OR? Haven't we had enough creep? --Minasbeede 19:27, 12 September 2007 (UTC)

I get what you are saying. In other words, in an article on "Washington and religion" (taken to mean the question of what his religious beliefs might have been) it is appropriate (and not OR) to mention what a person who knew him had to say on the matter, and that you can quote directly from the primary source and do not need to cite a secondary source that discusses it. I'm not sure I completely agree... but it does make me pause to think about it some more. Thanks for your patience. On the broader issue of implied OR... I did not mean to introduce any creep. I was actually simply looking for clarification of existing policy on what I see as a grey zone. I think you can fall into OR by wording things so as to lead the reader to a conclusion and without actually stating that conclusion... but I wanted to see if my intepretation of this policy matched what others thought. I gather that you do not think it is a problem. Blueboar 20:16, 12 September 2007 (UTC)

Another word for "original research" is "thinking." I don't think (uh-oh) that there needs to be an electric fence with fatal voltages around all thinking with respect to Wikipedia. The concept of NOR started out as being a way to reject quirky physics theories. "Bad" thought. It e x p a n d e d to also forbid all personal conclusions of an editor that were not backed by a source that reached the same conclusion. Well, OK. Now it's being twisted all out of shape: NOR is, by a strange focusing on source types, supposed to prevent inclusion of possibly suspect (in some minds) primary material (that can be properly sourced: it is verifiable): the prohibition on original thought now is to extend beyond the Wiki editor to the person who first had a thought or made an observation. Sometimes. The verifiability policy would accept the material but an effort is underway (has been for quite a while) to undermine verifiability in NOR and to alter the Wiki concept that it is verifiability, not truth, that is paramount. It's not enough to be verifiable, the source used has to meet additional criteria (as yet still not enunciated clearly enough to not be the source of long debate over meaning.) When do we say "enough" and remove the faulty section? That some primary source material is "possibly suspect" becomes the justification for well-meaning editors to search out all the ways that primary sources might be defective or might be used inappropriately and to then codify that. All the codification rests on the simple NOR concept and the simple NOR concept, alone, serves to reject material that needs to be rejected. Trying to exactly enumerate how (and how not) material is flawed is a hopeless task. That source typing remains in the policy while that hopeless (thus, never-ending) task goes on is a source of frustration, at the least.

I think (there's that naughty word again) that there is nothing in any way objectionable about doing simple math to show that a 1 degree (Celsius - or Kelvin) change in average global temperature is under a 0.4% change (on the Kelvin scale, which is the proper scale to use for a percentage calculation.) Yet if you look you'll see that is tarred as being "original research." It's simple math. This statement adds an illuminating fact to the debate (while a one degree change in average temperature is predicted to have major effects it's still a change of under 0.4%.) The underlying concept, which I'm apparently forbidden to say in Wikipedia, is that its not all that big of a change and it's not all that unreasonable to say that the added CO₂ in the atmosphere that comes from the burning of fossil fuels could easily cause change of less than 0.4%. The math can be verified, that the Kelvin temperature scale is the proper scale to use can be verified. The information enhances Wikipedia. It's forbidden, according to some. It's forbidden because of a too strict interpretation of a policy. That policy has been encumbered with two concepts that further widen the zone of prohibition: source typing and synthesis. Both of those are the subject of continued dispute. That continued dispute is evidence of something. Thanks for your patience. I'm pleased that something I've said has been the spur to thought. --Minasbeede 14:53, 13 September 2007 (UTC)

I thought you might be. :>) By the way, it has long been held that OR does not apply to talk pages... we are allowed to think for ourselves - and even (horrors!) speculate on talk pages. We just shouldn't carry our thinking into the article. :>) Blueboar 15:16, 13 September 2007 (UTC)

Recent POV-pushing

I reverted recent POV-poshing aimed at giving more slack to artcles based solely on primary sources. This encourages a deplored practice to create aricles by "quote farming" from google searches. The reverted significant changes in policy must be thorougly discussed in the policy talk page. `'Míkka 15:46, 7 September 2007 (UTC)

per another editor's request, I protected the page again in the form it was in before it was unprotected. Slrubenstein | Talk 15:53, 7 September 2007 (UTC)

I agree with Mikka. I would prefer having the article protected in the last version by Mikkalai, which was the last consensus version before Cogden started edit warring to get his non-consensus version in. The article should revert back to the last version by Mikkalai instead of reverting back to the non-consensus, contested version by Cogden. If a version is to be picked to protect on, this was the last consensus version before the edit war. Dreadstar † 16:11, 7 September 2007 (UTC)

It is not often that I agree with Mikka, but this is one case which I do. There is no need to dilute the spirit of NOR that would encourage the use of primary sources. ≈ jossi ≈ (talk) 16:16, 7 September 2007 (UTC)

Jossi hits it right on the mark, there is no need to change the wording to encourage the use of primary sources, a change which would indeed dilute that important aspect of NOR. Dreadstar † 16:20, 7 September 2007 (UTC)

I think this substantially mis-characterizes the essence of the debate. I have not seen any arguments encouraging the use of primary sources. The question is whether dividing sources into primary and secondary categories, then treating them differently in this policy is supported by a broad consensus of the Wikipedia community. Many editors have questioning the wisdom, necessity and/or relevance of the disputed text in the recent comments, and it is important to remember that policy must reflect consensus of what editors actually support, not what some editors think other editors should support. Dhaluza 00:31, 8 September 2007 (UTC)

I restored it to the version that was last protected in part because it seemed like the only disinterested way I could act, given that I have argued for my own views here. If an editor who has not been involved want to change the protection, I won't object - but check with the page protection policy. We are uspposed to protect it at whatever version it is in when the conflict calls for protection. I think one could reasonably argue that with policies, reversion should be to the pre-conflict version. Anyway, it could be a touchy subject. i did what I thought was prudent and won't object if another admin has a better idea. Slrubenstein | Talk 16:33, 7 September 2007 (UTC)

Strict rules of editing policy pages supersede formalities of protection ("metapolicy supersede policy"), so I regretfully have to restore my reversal of non-discussed changes to previous consensus version. Policies are highly visible, influential pages and we cannot afford "The Wrong Version" here. `'Míkka 17:10, 7 September 2007 (UTC)

Míkka, I would take a look at the archives. This issue has been very heavily discussed, and it's clear that there is no consensus as to any particular decreed rules distinguishing sources based on whether they are deemed "primary" or "secondary". The strict rules of editing policy are that any policy rules must represent and describe widespread consensus. Given the controversy, the status quo is 'no policy. If we can agree on positive language relating to the merits and disadvantages of particular types of sources, we can put that in, but in the meantime, we risk damage to the Wikipedia institution by maintaining a non-consensus and controversial policy. COGDEN 18:05, 7 September 2007 (UTC)

I don't want to look into archives. It is your job to present your arguments now, when you intend to edit (including exact references to relevant past discussions, rather than simply saysing RTFM. And you have to endure the discussion of your changes, not of what was spoken in the past. `'Míkka 19:05, 7 September 2007 (UTC)

COGDEn is bein disingenuous. The distinction between primary and secondary sources has been a stable component of this policy for most of its existence. Slrubenstein | Talk 18:26, 7 September 2007 (UTC)

Actually, the "primary source" language was around for only about 25% of the policy's existence. But that's not the point. The point is, it does not describe Wikipedia practice and consent today. See Wikipedia:Policies and guidelines. Rather, it prescribes what one faction of Wikipedia editors want Wikipedia to be someday. We need consensus now, not a year from now. COGDEN 18:54, 7 September 2007 (UTC)

Moreover, people can't agree on what it means. Jacob Haller 18:56, 7 September 2007 (UTC)

Cogden, If you persist in unilateral modification of the policy, then, rather than protecting the page, you will be blocked from editing until you understand the basic rule of editing the policy pages: seek for consensus before significant edit. What is more, you have discuss each change separately, to avoid confusion and conflation of items. `'Míkka 18:58, 7 September 2007 (UTC)

This is the rule for Wikipedia content pages. Policy pages have no such inertia. Please read Wikipedia:Policies and guidelines, which is an actionable policy that I am enforcing here. If non-consensus and controversial items in policies and guidelines could not be removed without consensus, then non-consensus policies would be permanent in every case, since by definition, there will always be a faction opposing the change. COGDEN 19:02, 7 September 2007 (UTC)

I am afraid you are badly mistaken: policy pages have the largest inertia, otherwise there would be a complete chaos. `'Míkka 19:06, 7 September 2007 (UTC)

"faction opposing the change": you are mistaken again: wikipedia is not democracy. Policies are shaped by arguments, not by factions. `'Míkka 19:09, 7 September 2007 (UTC)

Indeed. COGDEN, if you want to change the policy make a formal proposal. In the meantime, just because you don't like the policy doesn't mean it should go. You do not have a veto over all of wikipedia. And since policies are the framework that guide the composition of articles, of course they have the greatest inertia. Slrubenstein | Talk 19:11, 7 September 2007 (UTC)

None of us can veto true Wikipedia policy and consensus. But I do have the right to enforce Wikipedia policy with respect to how controversial policy pages are created and maintained, which requires that policy pages reflect current widespread Wikipedia practice and consensus. COGDEN 19:17, 7 September 2007 (UTC)

In this case, the disputed sections don't seem to guide anything, they just feed edit wars. Jacob Haller 19:15, 7 September 2007 (UTC)

Agreed. And by definition, any policy page section that is that controversial is not a true Wikipedia policy. Policy pages are supposed to encode broad areas of agreement and practice among Wikipedia editors. When something is so controversial as this, it's a sign it doesn't belong in the policy page. COGDEN 19:19, 7 September 2007 (UTC)

controversial policy pages? This is official policy. Current widespread Wikipedia practice? Is exactly what this policy reflects. ≈ jossi ≈ (talk) 19:21, 7 September 2007 (UTC)

Not true. Look around. See how primary and secondary sources are used. Jacob Haller 19:26, 7 September 2007 (UTC)

To Jossi: Obviously not, if you read the archive pages and the discussion above. You should also review the history of WP:ATT, which was, supposedly, an "official policy", but got demoted after it was reviewed by a wider body of Wikipedia editors and it became clear that there actually was not consensus for the principles described in it. Had WP:ATT required a consensus to demote it, it would still be an official policy today. Rather, the standard for promotion was to show that there was no consensus for the controversial statements, just as there is none here. COGDEN 19:27, 7 September 2007 (UTC)

Claiming that there is no consensus because you do not agree is at best disingenuous. ≈ jossi ≈ (talk) 19:23, 7 September 2007 (UTC)

That sword cuts both ways... Dhaluza 18:57, 9 September 2007 (UTC)

I was so confused when I first came here. I pretty much am still confused, but at least I have a bit of background now to make a slightly more informed judgement call though, aand I still think the disputed section "Sources" should be on it's own page, so these "discussions" can happen there instead of here. I also think that the "Synthesis" section will eventually have the same sort of problem (maybe not though) and should also have it's own page. A policy should be fairly stable and have the overwhelming majority support, not be a place of constant edit wars and disagreement. I still fail to understand why a policy on "No Original Research" has definitions of what a "primary source" or "secondary source" is, or when to use one over the other, etc. It has nothing to do with "No Original Research". It's a completely different animal, though it may be relevant to some examples of how original research can occur (not neccessarily does occur). "I" think that as long as the "Sources" section stays, this policy will be a place for edit wars. The original inclusion was probably well-intentioned at the time, but it looks like all the arguments back and forth semm to center around this one small (on this page) section. If you go to the linked pages on the various sources, not only do you see better (more complete) definitions and examples, but you will also notice some discrepenicies between this policy and those pages. wbfergus 19:31, 7 September 2007 (UTC)

I concur with your assessment. I created the WP:PSTS shortcut to this section back in June, because when I needed to refer to it, I couldn't find it where I expected to find it in WP:RS. Since then, I have wondered why these definitions were here rather than there. It really seems out of place. Dhaluza 00:44, 8 September 2007 (UTC)

protection

I do not kinow why the page was unprotected, but it was protected for a reason and i protected it again in the version it had originally been protected in. Mikka, do not screw with this. we have page protection rules for a reason and you have just proven the importance of our policy - bias the protection and you just unleash a new edit war. Now it is protected, nothing in the page has changed since it was firsdt protected by another editor, let's hash out conflicts here and reach consensus before unprotecting and changing the policy page. Slrubenstein | Talk 19:29, 7 September 2007 (UTC)

While the page is protected, does anybody object to keeping the disputedtag template introduced to the section by User:ChazBeckett? This would at least temporarily alert readers that this section is the subject of dispute. COGDEN 19:35, 7 September 2007 (UTC)

The very fact it is protected makes this very clear - the disputed tag is onloy needed for articles that are not protected, as people might think such articles are entirely uncontroversial. Slrubenstein | Talk 20:07, 7 September 2007 (UTC)

It's clear that there is a dispute, though finding out which particular section is disputed would take some wading through the talk page. COGDEN 23:52, 10 September 2007 (UTC)

There are also rules about who should and who shouldn't protect, and those engaged in the conflict are in the "shouldn't" class, are they not? --Minasbeede 19:37, 7 September 2007 (UTC)

If you think I have a ocnflict of interest you are free to accuse me of that, but please note I protected the version that was protected before I became active here (i.e. I just extended the protection initially made by someone else). Would you have done it any other way? Slrubenstein | Talk 20:07, 7 September 2007 (UTC)

Is't it up to the individual administrator to carefully determine whether or not that administrator has a stake in material for which there is an edit war? --Minasbeede 20:19, 7 September 2007 (UTC)

Well, I made a careful determination, acted and explained myself on the talk page. Slrubenstein | Talk 20:22, 7 September 2007 (UTC)

I see: "During edit wars, admins should not protect pages when they are involved as a party to the dispute, except in the case of simple vandalism or libel issues against living people." You are or are not a party to the dispute. --Minasbeede 20:27, 7 September 2007 (UTC)

Plus I protected it at the version that was first protected at the time of the original edit war. In other words, I protected a version I had no stake in. All i was really doing was extending a protection another admin had made, because the issues that led to protection had not been resolved. But it is evident that the version I protected did not reflect my own edits or preferences. Slrubenstein | Talk 20:30, 7 September 2007 (UTC)

P.S. I'm a relative newbie. What do I see (rhetorical)? --Minasbeede 19:39, 7 September 2007 (UTC)

As an admin myself, I see nothing improper with Slrubenstein's protection. Slrubenstein initially protected the article, before becoming involved in the discussion, and the re-protection just returns things back to that point when it was first protected. There was an obvious need for protection again, and since the first protection was very proper, returning it to that state is still proper, even though Slrubenstein has now taken sides in the discussion. COGDEN 23:52, 10 September 2007 (UTC)

Just wanted to note that I only placed that tag to replace one that I felt wasn't as appropriate. [1] Chaz ^Beckett 19:38, 7 September 2007 (UTC)

Note: "Protection is not an endorsement of the current version (protection log)." Slrubenstein | Talk 20:11, 7 September 2007 (UTC)

This sniping over the technicalities of the page protection are not productive. The real issue is that the policy has to be repeatedly re-protected due to continued edit warring, which is disgraceful. Besides the disputed language, the discussions and related scrutiny have identified other facets of the policy that also need to be addressed, and it is impossible to make progress on these issues as well. Dhaluza 04:24, 8 September 2007 (UTC)

Can we at least keep the page stable enough that editors can mark all disputed sections? - for the sake of those reading disputed sections as well as to improve the debate. Jacob Haller 04:34, 8 September 2007 (UTC)

As usual, the admin has protected the wrong version. <Sigh> -- But|seriously|folks 04:57, 8 September 2007 (UTC)

As usual, an editor is too lazy to read. Please take the monumental effort to shift your eyes up seven lines. A protected page is never the right version, and never the wrong version. It is the protection of a page in the middle of an edit war. It is by definition under contention. Anyone who wants to see what it looked like before protection can with the flick of a button. Awww, I guess that is too much work for some people. Slrubenstein | Talk 10:25, 8 September 2007 (UTC)

Just wanted to point out that Butseriouslyfolks was making a joke not a criticism - "the wrong version" is a humor piece.--Kubigula (talk) 20:45, 9 September 2007 (UTC)

Thanks, I realize that now ... Slrubenstein | Talk 20:46, 9 September 2007 (UTC)

Back to the question of what policy should be

I was just alerted to this debate via a notice at Wikipedia:Village pump (policy). As one who has not participated in the policy debate, I can say that somebody has made a royal mess of things here.

I'll bypass the questions of who has consensus, what version is protected, etc. People seem to be playing king of the hill regarding which version of policy language is consensus and which version is the new proposal, and I am in no position to decide who pushed who of that hill.

Returning to first principles, I believe it is a founding principle of Wikipedia that we are built on secondary sources. It is one of the five pillars. Jimmy Wales speaks on it at length. It is his reason for setting this up, and our reason for being here.

It is also indisputable that there are some occasions where primary or tertiary sources are allowable, and than a complete ban on them would be inappropriate. Also, there is room to disagree on the definition of what is primary versus tertiary, and whether various ways of gathering or surveying primary sources constitutes synthesis or original research.

My opinion is that those occasions should be very limited (you could use the word "rare" or whatever other descriptor), and not used when there is a suitable secondary source. They should be assigned to narrow categories for which criteria can be applied, and any individual instance of breaking the rule should be subject to a common sense test of whether it is really original research.

The page is protected for ten more days, which should be long enough to break a log jam. For those ten days it does not matter what the old consensus was or what language is approved. To focus the discussion, what should those "rare" occasions be? What occasions, if any, are uncontroversial (forget about old consensus - what has consensus here and now)? Which are the ones that need some debate? Where are the disputed boundaries of what is and is not a secondary source, or how to enforce and administer the rule?

If you think that's not worth restating for a know-nothing like me, fine. It's just hard for an outsider to know what's going on, and I see that some of the people in the thick of the debate may have lost sight of the broader question of what the policy ought to say, not just who followed correct procedure. Wikidemo 23:44, 7 September 2007 (UTC)

PS: one final thought. With respect to this edit, the version on the right is a little wordy. I've found it's best on policy pages to avoid extended definitions, expositions, expositions, rationales, examples etc., and to put them if necessary in their own separate section or associated guideline so you can see clearly what rule is being made. I'm don't think the comment about "source-based research" adds anything. The statement that the "rare exceptions" are "descriptive" cases that any educated person should agree on is on the right track but seems too broad. We don't want to require our educated reader to visit the primary source to actually observe and describe what is seen there; merely to confirm it says what the article claims it says. Any more work than that and it's not really a good source. Wikidemo 23:56, 7 September 2007 (UTC)

While I agree with your assessment that people have lost sight of what this policy should be about, and are apparently trying instead to defend territory, I was not aware that using secondary sources is a founding principle. A quick search of WP:FIVE did not find the word "secondary" in the current text. Perhaps you can cite some sources to help us verify that your assertion is not OR. Dhaluza 00:14, 8 September 2007 (UTC)

Ha ha, I get your drift. Maybe I'm wrong. I just thought that secondary sources was part of the bedrock formulation b/c compiling information from original sources is by definition OR. If that's not true then that subject's up for discussion too. Incidentally, it is permissible to use OR and TS when arguing policy, just not for writing articles.Wikidemo 00:24, 8 September 2007 (UTC)

It's perfectly possible to aggregate primary source information without introducing any "original thought", which is what the founding principles involved concern, IIRC. SamBC(talk) 00:42, 8 September 2007 (UTC)

Thanks for your help, Wikidemo. While I think you were being honest and sincerely trying to help I think Dhaluza has it right: "secondary" was not part of the founding principles - and that was not a case of overlooking something that should not have been overlooked. It's also not true that relying on primary sources is "original research." It's "literature research" in almost all cases and that literature research is the foundation of Wikipedia. That means no special language or justification is needed to cover the use of primary sources: all the existing language applies to them just as it applies to secondary sources - or, more simple, just "sources." Furthermore, the sole issue that has been identified as the reason for the policy is the occurrence of some instances of a type of use of primary sources that amounts to (at best) original research. It's not the sources that are at fault, it's the use made of them in a subset of editing behavior. Attacking the sources shifts attention away from the use (which seems already to be forbidden, and rightfully so) and also clutters up the policy. Having the policy mandate reliance on secondary sources also results in objections to the wording of the policy: such objections keep recurring. To make matters worse, "primary" means different things in different fields so that the source characterization isn't really that well focused. The existing (i.e., pre-edit-war) version of the policy didn't make a clean distinction since it only used the word "secondary," which likewise has different meanings in different fields. There seem to have been muddled (in my opinion) attempts to promote "primary" scientific sources to (Wikipedia-specific and as a result rather confusing) "secondary" status but that's so convoluted that I suspect few understand it. The promotion of "secondary" sources to preferred status seems to be an affront to the "be bold" principle and an invitation to ignore all rules. The policy would work just fine of all the source-discriminating language disappeared forever.

Nor do I think that that many editors come to WP:NOR for source-choice guidance. With the current wording any that do are likely to leave confused. I thought that the secondary-sources-are-preferred policy demotes Darwin's "On the Origin of Species" since it's a primary source. (Please, no discussion of how it might sometimes be primary and sometimes be secondary. It's already been said enough times.) That apparently is a misconception - but the wording used induces that misconception. That's surely a major flaw in policy wording. --Minasbeede 02:41, 8 September 2007 (UTC)

Just to help cut through some of the "wordiness" mentioned above, it seems that the key dispute revolves around these two sentences. IMHO, this isn't about defending territory at all, but a dispute on how restrictive the wording should be on the use of primary sources. I agree we can make the policy more concise, but not at the expense of eliminating or diluting restrictions on the use of primary sources. Dreadstar † 00:37, 8 September 2007 (UTC)

I don't think there is significant objection to the restriction on using primary sources only for descriptive claims without interpretation because that is central to NOR. It is those two sentences that do not appear to represnt a broad consensus opinion. By defending territory, I was referring to the comments supporting continued inclusion of these sentences based on inclusion over a period of time, and the assertion that consensus must be demonstrated to remove them, rather than to include them. Dhaluza 03:38, 8 September 2007 (UTC)

I think the primary/secondary source issue is an unhelpful contrivance as much as the word research is an unhelpful word to use when we do not really mean research. The principle is more "no original thought". Some observations require no analysis, but do require to be validated, say the height of the Empire State Building may get corrupted over time due to Chinese whispers. The best source to prove such a fact is a primary source: any secondary source which refers to the original source introduces the possibility of corruption - especially as we drift into the world of the press and so on. Similarly the principle of synthesis is a contrivance as it is a special case of the basic principle.

I would rather we had a much simpler statement of principle which avoids these terms, as I know darn well that the principle of primary/secondary sources is being used to disrupt Wikipedia - rubbishy secondary sources are held above good quality sources using this principle (and debating whether the source is a primary or secondary source is used to obfusticate). Any source can be examined to understand whether the fact, or analysis or point of view is contained within that source. A single source is not sufficient to verify whether it is a valid viewpoint, with the exception of sources such as peer reviewed journals. The principle of dividing sources breaks down once you go outside peer reviewed journals or other sources where it is a clear body of opinion (government inquests for example).

I would say the focus on secondary sources is damaging as it potentially holds dubious press articles in higher regard than the documents on which the press reporting was based, with no mechanism to validate the analysis. Which is better, the text of a Government Inquiry, which is a review of the evidence presented, or a press report on that Inquiry? I would hold the inquest is best, even if we are reporting the findings of the Inquiry - using it as a primary source. Only if we assert the findings are flawed, or accepted, do we move into OR territory. Spenny 09:22, 8 September 2007 (UTC)

And who is to say a secondary source is "rubbishy?" You? Your examination of primary sources reveals the secondary source to be rubbish? We should publish your own research rather than provide an account of what professional researchers have done? You have just proven that we need to make the distinction between primary and seconsary sources. In every criticism of the distinction that I have read, the argument is always, ultimatey, that "I" (the editor of Wikipedia) "know better than the secondary sources." That is original research, and it violates NPOV. It just proves that the people who are most frustrated by this policy are those who want to violate it. Well - good!! Slrubenstein | Talk 10:22, 8 September 2007 (UTC)

Yes, as an editor, I evaluate the sources. The primary/secondary source concept only works if you have a tightly defined definition of a secondary source which is of high quality and peer reviewed. As soon as you expand that concept into, for example, journalism, you have to introduce the concept of evaluating the quality of source, getting multiple citations to prove that these lower grade sources are reasonably reliable. The concept does not extend from the academic world to the general world of Wikipedia. See example below. Spenny 16:01, 8 September 2007 (UTC)

Spenny implies some secondary sources are "rubbishy" and that these particular rubbishy sources are used by some other editor to overcome non-rubbishy primary sources. The characterization of "rubbishy" is surely Spenny's own but unless there's some way to prove that what he implies cannot ever be true we owe him the benefit of the doubt: what he alleges has happened may have happened. (Perhaps Spenny can provide an example.) The policy wording in question implies that all primary sources (which, once again, I will point out is still a vague concept within the policy) are inferior to all secondary sources. The wording in the policy implies that primary sources are "rubbishy" - all of them. The sin committed in that policy section is orders of magnitude greater than the sin alleged to have been committed by Spenny - and it's the same sin, with "primary" and "secondary" interchanged. The editor who it is alleged argues "I know better than the secondary sources" has at least looked at the primary and secondary sources and made a judgment. He has not alleged, a priori, that one particular source is, by being primary or secondary, better than another. Not so the policy.

Nor is every instance of reliance on a primary source an instance of using that source to counter the content of a secondary source, yet all such reliance on a primary source (in a manner that fully satisfies the verifiability policy, the policy that mandates sourcing) is put under a cloud by the policy. --Minasbeede 14:51, 8 September 2007 (UTC)

Example: The BSE crisis in the UK had a Government Inquiry. It spent some time investigating the causes and specifically eliminated some speculative ones. It came to a definitive conclusion, based on its review of various scientific data. It was published. It summarised its conclusions. It is a highly regarded piece of work. First question, is that a primary source or secondary source? I would say that it's conclusions are peer reviewed and even if I quoted from it directly, to show what the report said, there should be no issue that I am not advancing an original viewpoint. Others would say that this should be discounted as it is a primary source on the subject of BSE disease. We then take a Reuters report which says that scientists blame Factory Farming for BSE.

what you say about the inquiry isn't accurate. It blamed factory farming for the epidemic. You have to read the entire report to find that, of course (and it's long), but journalists were briefed, and those journalists reported that factory farming was being blamed, which led Germany's chancellor to call for an end to it. We can't use our own interpretations of primary sources (and the inquiry is a primary source for our purposes); we have to go with the interpretations of reliable secondary sources. Yes, the primary way the disease was spread was feeding cattle to cattle, but it was the entire system allowing that practice (and feeding supplements to calves instead of milk, spraying animals with pesticides, and on and on) that was identified as the facilitator. Bottom line: we can't insert our own opinions. We must stick to what reliable secondary sources say.

References: The Inquiry [2], a DEFRA summary of the Inquiry (a reliable secondary source)[3], the secondary source used to cite the analysis:[4]. Now, given those 3 sources, which would you go with? At least three editors (two admins) have edit warred to maintain the insistence that factory farming should state that BSE is caused by factory farming based on that single last citation. Spenny 16:01, 8 September 2007 (UTC)

Here's an article (List of speeches) that appears to be based entirely on secondary sources. There are no references, though there are eight external links, two of which appear to be "Paid" web sites (actually, different "pay-to-view" articles on the same website). Much of the content is linked to a tertiary source (another Wikipedia article). Some "research" or "decision making" had to have been done, as I followed the link for Top 100 Speeches of the 20th Century, and it lists "Adopting the Declaration of Human Rights" by Anna Eleanor Roosevelt as number 100, but it is not in the article. What that site lists as #5 is also missing from the article. Therefore, without using any primary sources, and relying soley on secondary sources, the author(s) has taken the decision making process upon themselves to determine which speeches, as they put it, "This is a list of speeches, that have gained notability. They are listed in chronological order". The author(s) have arbitrarily decided for themselves where to draw the line in what is "notable", which surely must be OR.

So, the type of source used for an article is, by itself, insufficient to gauge whether or not the article does, does not, or may violate the NOR policy. Furthermore, if the "source type" issue is insufficient to guage this, why does this issue belong in the NOR policy, other than as a brief explanation that improper usage that may lead to OR? Wikipedia has an article on policy (in general, not specific to Wikipedia policies), that has a section called Policy content. The inclusion of "Sources", in it's current form, seems to violate the last principle stated in that section (Some policies may contain additional sections, including), "Definitions, providing clear and unambiguous definitions for terms and concepts found in the policy document".

I would therefore contend that this policy does not, and cannot, define the 'source issue' (my term), in a "clear and unambiguous" way, without making the policy more confusing than it currently is, and leading to many more edit wars in the future. In Wikipedia, with all the various users and editors there are of it, I would further contend that Policies need to be short, focused, and clear in their intent, to avoid, or at the very least minimize, confusing the users/editors. Convoluted, long, and/or contentious examples and definitions should be referenced as applicable guidelines, but not included within the scope of the policy itself. This is a policy on "No original research", not a policy on "Source type definition and usage", and this subject doesn't belong here since it's inclusion does nothing to identify clearly and succinctly how the types of sources inhibit or prohibit OR. wbfergus 11:56, 8 September 2007 (UTC)

Lists are a special case, probably because they don't actually present material of a type that needs to be sourced. Unless there is some commentary about an item (which in many cases should be sourced), a list just points to things and makes no factual claims about its items, except implicitly that they're members of the category being listed. As such, a list drawn from primary or tertiary sources is usually fine. Even the minor claim, though, that an item is a significant member of a class, often needs secondary sources. Take a look at (and PLEASE don't nominate for deletion - it keeps the newbies from messing up more important articles) list of Internet phenomena. It's a constant battle to keep spam off that list, and everybody wants to add a list item for something they just found on Youtube. The only way people keep the list orderly at all is to demand that every entry have a reliable secondary source. A primary source won't work - people are constantly claiming that the video itself, or its Youtube view count, establishes notability, but that would be a mess. Nor is the link to a Wikipedia article a good source to say something belongs on the list. The people watching the article would have to do an assessment of each remote article to see if it's really notable, and proper, and belongs on the list, i.e. further analysis that is inconvenient and inconclusive. The best way for this particular list is to require a reliable secondary source that states that a particular meme is an Internet phenomenon. But again, even though lists sometimes need secondary sources they are not normal articles. Wikidemo 15:34, 8 September 2007 (UTC)

Yes, lists are a special case. but this is a manually compiled list, not an automatically generated one from what different articles have as a category tag. In the of this particular list, I would tend to agree with you if it wasn't for the fact that roughly half of them are linked to Wikipedia articles and the other half aren't. So, somebody took it upon themselves to create a list of the greastest speeches, or speeches that have gained natability, and then made their own decision what to include in the list, ignoring some of the speeches listed in their external links section. It seems disingenious to cite a web site with the "100 greatest Speeches", and then ignore some of those speeches that are listed (because they didn't agree with them maybe?). The author(s) made a decision, based upon their own POV, as to what should be included. So, this example shows a violation of both NPOV and NOR without using any primary sources at all. I merely threw out this example to show that the "Sources" section really doesn't need to be included in this policy, as OR and other issues can just as easily appear in an article with or without primary sources. The inclusion of primary source material does not always constitute "original research", just as basing an article solely on secondary sources does not mean that the article does not contain "original research". I am also not advocating for the article to be deleted either, based upon a violation of NOR, NPOV or any combination. I'm fairly sure that this article is merely a work in progress, as are most articles on Wikipedia, but this shows how easily misuse or misinterpretation of existing policies can be applied, especially if the policies being touted are themseleves vague and/or under contention. wbfergus 16:47, 8 September 2007 (UTC)

I'm not familar with the specific UK BSE report you cite, but I am famialar with US NTSB reports on accidents, which are both primary and secondary source. Their format presents several sections detailing the results of their factual investigation, closing with a single section giving their determination of probable cause. The former is primary source because they are reporting on observations and evidence, without drawing conclusions; the latter is secondary source because it is an opinion drawn from analysis of the evidence presented. An article on the accident would present the material with similar weight--it would draw heavily from the primary source material, which is something the disputed language would discourage.Dhaluza 16:44, 8 September 2007 (UTC)

...which I think goes on to make my point - it would be inappropriate to discount a solid source like an investigation into an aircrash, where the investigations are generally held to be of the highest quality, and it would hardly be appropriate to discount such sources purely on the basis that they were in some way tainted by being primary sources. If I were an expert writing an article, rather then a Wikipedia editor, would I discount such information? I might not start with it, but if something said in a secondary source did not ring true, I would go to the source, review it myself and decide, for myself, whether the press report or whatever did reasonably extrapolate from the original source. If it did not ring true, I would cite the original source. It is an assumption of good faith to assume a press report is a fair reflection which time and again we know is not reliable. NOR is about avoiding original thought - this spurious primary/secondary nonsense gives us nothing useful to assess this. It is a nice sounding theory which does not stand up to scrutiny outside the domain it was defined for. Any supporter of it is comparing apples and pears to make the argument stand. It is not even verifiability over truth, it is falsehood over verifiability - I can provide a source that shows the conclusion of a press report is flawed and as it is spuriously defined as a primary source, rather a peer reviewed conclusion, it is discounted. Remember, I am not talking here about advancing a novel theory - I am saying I want to give published fact and there is a policy that is being used to stop this - especially where a secondary source itself synthesises.

...which is a good point, is it a secondary source if a report provides its own novel conclusion? You need primary sources, to prove that the secondary source is a secondary source! Spenny 17:21, 8 September 2007 (UTC)

That's all kind of metaphysical. The sort of critical analysis of sources described, whereby secondary sources are verified for truthiness against primary sources, is not appropriate for Wikipedia because it is original analysis. A reader has no way of knowing whether they should trust it, or the editor who did it, and different people would come to different conclusions. If anyone wants to go into that much detail they can write an essay somewhere and see if anyone will publish it. That very issue, that some people's work is published and some is not, is one of the gatekeeping functions of secondary sources. I am not so sanguine that we get to truth faster through primary sources than secondary sources anyway. Primary sources are often biased in a systematic way. Do you want to source information about the Iraq occupation on the US government's field reports, or on the body of secondary journalism? In practice newspapers, treatises, books, etc., have a level of editorial control and comprehensive coverage that primary material, even academic scholarship, does not. Government inquiries are notoriously biased and rigged. There are secondary sources that are more reliable than others, just like some academic theories are better than others. The function of a reliable sources policy is to sort that out. True, that means Wikipedia's accuracy can be no better than the sum total of all of the journalistic and documentary writing out there, as filtered through our verification process. But whose knowledge is any better than that? If we let people advocate for claims based on primary research we have less controls against false information, not more. Moreover, the function of Wikipedia is not to state the truth....when you start claiming truth you open yourself up to all kinds of trouble. We simply summarize the accepted wisdom on various subjects, and ideally source/link all of the claims to their sources so an interested reader can chase that back as far as they want.Wikidemo 17:43, 8 September 2007 (UTC)

I think in there is the flaw: the issue is not taking the sum total of all the writing out there, but actually we have to first filter out the rubbish, then sum it. Policy has the concept of reliable sources: we have at this point introduced editorial judgement. What tools are available to assess whether a source is reliable? The same tools we use to assess if a Wikipedia article is reliable, check the sources and see if it stacks up. In the end, editorial judgement is required; we are just trying to circumscribe the allowable types of judgement. Wikipedia doesn't work without a brain.Spenny 00:21, 9 September 2007 (UTC)

(ec) That seems to be using "clever" semantics to argue that the classic historiographic definitions of primary/secondary sources are self contradictory, and you're not the first person to argue that. It seem pretty easy to understand that primary sources are "root" sources, and secondary sources are those which aggregate and interpret primary sources. That said, I disagree with the preference for secondary sources as a blanket thing - however, secondary sources are necessary to include any commentary or analysis that isn't OR. Given the concise definition of primary/secondary I give above, that seems to be a pretty uncontroversial and obviously true statement. Whether it needs to be made in the policy is more questionable. I lean towards "no", instead it should be in the guideline that RS should turn into. SamBC(talk) 17:47, 8 September 2007 (UTC)

I'm not really trying to be clever, just pointing out that there is a simple observation: we can find statements and essentially copy them into Wikipedia - primary are essentially factual, secondary are essentially analysis if you like the oversimplification. That we can find statements elsewhere satisfies NOR, it may not satisfy RS. It is the Wikipedian philosophy that introduces the problems not the "classical" interpretation. It comes back to: is the idea, fact, analysis, synthesis whatever published for the first time here? If yes, forget it, regardless of source, if no, make sure the source contains the claim. That is easy to deal with. What are the key tests? Is it sourced? Is the source respected? Is the source contradicted? If the source is contradicted, is it a viewpoint which also requires documentation, or is it an error that requires further analysis of sources to justify a conclusion based on understanding of the topic? Spenny 00:21, 9 September 2007 (UTC)

It is not necessary for the editor to decide which source is more reliable, that can be left for the reader. I think the prescriptive restriction on primary sources is based in a fundamental mistrust of editors and readers to be reasonable. While we all have had to deal with unreasonable people, it's important to remember that most people are in fact reasonable, they just don't leave as strong of an impression. For example in a WP article, I quoted and cited a secondary source saying that a company received a million customer complaints on their product. Another editor added a primary source reference from the company saying that they had not yet shipped a million copies of the product. So we have these two apparently contradictory facts in juxtaposition, and that's just fine. Who knows what the truth is? Maybe they got complaints from non-customers, maybe customers made multiple complaints, maybe the secondary source made a mistake. But using a primary source to contradict a secondary source in this manner is completely appropriate. It is a purely factual statement without synthesizing a conclusion--that is up to the reader, and if they draw the wrong conclusion, that's their "problem" alone. Dhaluza 19:30, 9 September 2007 (UTC)

Arbitrary section break

I think what you are mainly proving is that this policy cannot stand alone. Which is true. NPOV and V remain as independent policies that must also be followed in editing articles. A final point: NPOV will never be achieved by just one editor or one source - if an article is lopsided in its use of secondary sources and thus is not NPOV, the solution is for other editors to do more research using other sources and add to it. There is no getting around that - no policy can address this issue, it is the very idea of a wikipedia: articles are always works in progress, never finished, and anyone can, should even has to edit to make it work. Slrubenstein | Talk 12:08, 8 September 2007 (UTC)

Thank you. Since no single policy can address the issue/example I supplied above, and as you say (paraphrased from here), it seems beyond the scope of this policy to attempt to cover such a diverse subject as "Sources". There are simply to many convolutions and interpretations to begin to address the issue here (this policy), especially with all of the newer users/editors on Wikipedia. This policy would suffice on it's own addressing "No original research" by linking to a "Types of Sources" article, which in turn would point to Wikipedia:Verifiability and Wikipedia:Reliable sources as major guidelines, with [[Primary source, Secondary source and Tertiary source appropriately linked in for further definitions and examples. This is a classic case of how no single policy can stand by itself, but is supported by other policies and guidelines. It seems that the inclusion of the "Sources" section is an attempt, however well intentioned, to make the policy more self-supporting. wbfergus 12:43, 8 September 2007 (UTC)

It makes a lot of sense to move the sourcing part of the policy to WP:RS because that's a more logical home for it. Of course that requires WP:RS accepting the addition. If they do not it will have to be a new guideline; the material shouldn't simply be dropped because nobody will have it. Also a good idea to minimize the definition-making both in policy and guideline pages. Wherever we put it there is a strong need to avoid primary and tertiary sources in articles. If you say Haydn wrote 94 string quartets and 55 of them are in a major key, or Martin Luther King made 287 speeches in his life, or the Bible mentions sin 2,000 times, is it really a valid source to link to the primary material? I wouldn't think so. HOw is that different than saying "I measured it and you can go out and measure it too" to support a claim that a hill is 2,000 feet tall, a city 5 miles wide, or a pond contains 1,500 organisms per ml. All of this stuff isn't verifiable without considerable effort, and someone else will come up with a different result. In the haydn example, there are some things that may or may not be a string quartet depending on how you think about it. If a book about him makes that claim, though, at least you can trace the fact to someone who presumably knows what they are doing. You have found someone who has done the research and stands behind it. For tertiary sources, I don't know which ones we are considering but Wikipedia itself and its forks are certainly a poor source for anything, not because it's got low standards but because of the feedback loop that propagates errors and makes them virtually untraceable if articles can cite each other as sources. Wikidemo 15:34, 8 September 2007 (UTC)

One more comment. Secondary sources are crucial for establishing notability. Primary sources just sit there. They don't make assertions of their own importance, generally (e.g. if you didn't know what the Bible was, linking to it doesn't tell you in any straightforward way why it is so important). And allowing tertiary sources to prove notability would let things bootstrap their own way into the encyclopedia without actually being notable. Whether you consider this a matter of WP:OR or not is another matter. Wikidemo 15:37, 8 September 2007 (UTC)

There are different reasons for using primary sources. In your examples, they are all either unsourced ("I measured it") or original analysis (number of things written, which requires more thought - when is a speech a formal speech to pick one). What is the best source for an article on the 2006 Formula 1 Championship results? - the primary source of the FIA published results. There surely cannot be a debate to say "Alonso won the 2006 Formula 1 Championship" and ref that to the FIA website is a reasonable approach, and for that fact actually the best citation we can find. There is no original thought, no POV pushing, and it is a source that given the FIA define the results is of the utmost reliability. To have a policy that excludes a reliable and verifiable source of this form and prefers some secondary reporting, and potentially holds some slovenly reporter's transcription of any old website over the original unadulterated source, would be bizarre. I quite understand the sentiment of the approach, but I don't think it really works when you try and turn it into an absolute rule. I'm quite happy as a commentary to discuss the principle as a guide, but at the moment there is clearly enough ambiguity in its interpretation for it to be unhelpful to be a hard and fast rule. Spenny 17:37, 8 September 2007 (UTC)

Yes, that was very perceptive of you, Wikidemo. The need for secondary sources is definitively related to notability, and the misuse of primary sources often leads to OR, being that the reason we ask for secondary sources in WP ≈ jossi ≈ (talk) 15:42, 8 September 2007 (UTC)

But, as I pointed out above, the misuse of sources can also happen just as easily without any primary sources being used. wbfergus 16:55, 8 September 2007 (UTC)

Just for the record... the section on types of sources was duplicated at WP:RS at one point. It was removed because people kept editing it in a way that contradicted what was said here.

Which raises an important point... WP:V and WP:RS are closely related to each other and to WP:NOR. What is said on one page affects what is said on the other two. Both V and RS are currently in flux. It is vital that those of you who care about this page get involved in the changes being made at the other two, to avoid the potential for conflict (don't assume that those editing on the other two pages share the same intent as you do). I think that a joint effort is needed to sync all three pages to ensure that they all agree with eachother. Blueboar 16:01, 8 September 2007 (UTC)

Good point. I didn't know that WP:RS had previously made a decision to remove a "Source" section. If those two policies or guidelines are so closely related to this, and the same type of section appears (or appeared) on multiple locations, it seems much confusion could be eliminated by having that ("Sources") as a common guideline for all to link to. The applicable policies and guidelines could link to a guideline page on "Types of sources and their use", which in turn would link to the three different pages that describe each source with definitions and examples. wbfergus 16:55, 8 September 2007 (UTC)

The only sense in which secondary sources were ever a significant part of the Wikipedia movement is in the fact that a Wikipedia article itself has to be a secondary source. It can't be a primary source about any topic, but it certainly can cite primary sources. All of Jimbo's statements about original research are about making sure that new ideas don't make it on Wikipedia unless first published in another source. The first source of publication is a primary source. If some crackpot physicist wants to write an article about a new physics, he has to publish first in a reputable journal. If he can do so, people are free to cite that journal article. When they do so, that citation is a primary source for the new idea.

The only other factor in which secondary sources has come to play a role in Wikipedia is the issue of notability. According to an apparent consensus, the existence of secondary sources is good evidence of notability. Once notability is established, however, one may cite primary sources as well. The notability criterion was never intended as a restriction on what sources you can cite concerning a notable topic. COGDEN 00:21, 11 September 2007 (UTC)

The core content policies are predicated on using reliable third-party publications with a reputation for fact-checking and accuracy. That is exactly what primary sources are not. We cannot use primary sources lacking mention in third-party sources without violating a number of principles and explicit policies, undue weight being a prime example. You cannot address this issue in a vacuum. You must take the body of content policy and guidelines into account. Vassyana 01:15, 11 September 2007 (UTC)

But isn't this the feedback loop that ties the policy in knots? We often reduce the discussion of OR down to an original concept/idea/whatever. If we consider the old chestnut of school results table, we then may see an analysis of this in say the Times, who draw some observations. Those observations are a secondary analysis of the results, but the concepts the analysis puts forward are found for the first time.

I think in the real world we are comfortable with the analysis itself being the secondary view, and as long as the place of publication is viewed as having been likely to properly check its facts, then we have achieved sufficient for this information to be reported in Wikipedia, albeit being concious of the need to avoid undue weight for what might be a minority view (at this point). The perspective argument gets dangerous and I think overcooks the policy because it becomes overly pedantic for this very reason of not understanding that it is the job of a secondary source to provide original analysis. So School X came top. The Times says X came top because it spent 10 times the amount per pupil on resources than school Y. We do not need anyone to review the Times review for us to quote, yet clearly the Times is the primary source for that idea. This is only a problem if someone claims the Times is unreliable (it may have a political agenda for drawing its conclusions).

Going back to my BSE example, although the Inquest was a review of many sources, (it did not do its own research, it evaluated evidence and drew conclusions), it was the first time these conclusions were drawn decisively (based on others opinions) and therefore it was suggested that this was a primary source and should be dismissed. If we cast aspersions on the quality of the Inquiry, then that itself is novel and needs a source, but to introduce analysis from the Inquiry is not original research. Spenny 09:39, 11 September 2007 (UTC)

Whoah, Vassyana, are you seriously saying that no primary source is ever (or is from/in) a reliable third-party publication with a reputation for fact-checking and accuracy? Government publications? Reports from the UK's ONS (office for national statistics)? Data from published research? BBC transcripts (from radio or TV interviews)? SamBC(talk) 09:59, 11 September 2007 (UTC)

Yes, by literal definition primary (first-party) sources are not from a third-party. Also, note I did not say that all citations must call upon reliable third-party references, simply that primary sources used should be noted in them. The UK's ONS and the US Census Bureau are both reported to be highly reliable sources by third-party publications. Raw data from published research should only be used in context, drawing upon evaluations of that data. If an interview is notable, it will be commented upon by a third-party source, but there should be little problem with a direct and relevant quote from said interview. Vassyana 10:29, 11 September 2007 (UTC)

This may be true on a trivial analysis, but fails on closer analysis. To take an example, there are sources (first-hand accounts) regarding the first Quaker shops, generally noting how weird it was that they had fixed prices. These accounts are primary (they are the reports of the people who went there) but they are third party (they are about the shop and the person was not directly connected to the shop). Regarding ONS and Census Bureau, are you saying that if a third-party reliable source considers them reliable, then they are permitted forms of primary source? SamBC(talk) 11:35, 11 September 2007 (UTC)

I think we might be talking past each other here. I'm indicating first/third party in relation to the sources/claims themselves, not towards the event/subject itself. Make sense? See below regarding the rest. Vassyana 11:51, 12 September 2007 (UTC)

Given that RS basically says all reliable sources are third-party, that interpretation of the first/third party distinction creates significant difficulties. The usual interpretation I've seen is that third party means independent of the subject. Otherwise, you're essentially redefining first/third party essentially equivalently to primary/secondary, which would seem intuitively incorect, seeing as both are regularly used in complimentary fashion. SamBC(talk) 16:55, 12 September 2007 (UTC)

Quite a reasonable and coherent interpretation. I could agree with that. Vassyana 01:21, 13 September 2007 (UTC)

This gets the the heart of our policy: we editors do not forward our own arguments or views, no matter how many primary sources we marshall. This is not the place for original research. Slrubenstein | Talk 10:52, 11 September 2007 (UTC)

I don't think there is any suggestion that there is anything but consensus for the principle of no original research (the interpretation and practice being another matter!). I think that there is also consensus that editors should not put forward their views, although to be able to edit effectively they need to use their understanding of the topic.

What is not a given, and where there is not a consensus, is the view that a primary source is a de facto indicator of putting forward points of view, or that in some way they are irredeemably useless and tarnished and can never provide evidence of a proposition that is, for example, nearly obvious.

Put another way: This gets the the heart of our policy: we editors do not forward our own arguments or views. This is not the place for original research. needs no qualification and is as meaningful as saying This gets the the heart of our policy: we editors do not forward our own arguments or views, no matter how many rubber ducks we play with in the bath. This is not the place for original research. Spenny 12:41, 11 September 2007 (UTC)

Sigh... of course it is a sequitor, it follows from what came before it. First, Sambc wrote, "Whoah, Vassyana, are you seriously saying that no primary source is ever (or is from/in) a reliable third-party publication with a reputation for fact-checking and accuracy?" Then, Vassyana wrote, "Raw data from published research should only be used in context, drawing upon evaluations of that data. If an interview is notable, it will be commented upon by a third-party source, but there should be little problem with a direct and relevant quote from said interview." Then I meade my comment, which in context I though was clearly the explanation for why Sambc is wrong, and Vassyana is right. Slrubenstein | Talk 13:00, 11 September 2007 (UTC)

Apologies for cutting into the flow of conversation with my reply; I didn't realise, Slrubenstein, that you comment was a direct reply to Vassyana, as it wasn't indented, and I didn't want my response to seem a reply to you instead of them.

My point remains that first-hand doesn't mean first-party. An eye-witness to an event is not necessarily a participant of the event. Thus primary does not imply not-third-party. SamBC(talk) 13:27, 11 September 2007 (UTC)

That's okay, Sambc, it happens. I don't want to get bogged down in semantics (what does third party mean etc) - the point is, it is not our job to be fact-checkers for published sources. We can present facts even from primary sources, as long as we do so simply to report the facts - but not to argue against a published analysis/explanation/interpretation etc of those or similar facts. Slrubenstein | Talk 13:45, 11 September 2007 (UTC)

Absolutely - on those points, we appear to agree completely. Vassyana, do you agree with this as well? Am I misunderstanding when I read what you've said as meaning that there should absolutely be a moritorium on the use of primary sources, except in "rare" (but not specified) situations? SamBC(talk) 17:19, 11 September 2007 (UTC)

Yes and no. (Helpful answer, eh? :D) I believe primary sources should not be used independently of third-party sources, except in extremely rare circumstances. Independently, as I use it, breaks down in a two different ways. First, if third party sources taken as a whole vouch for the accuracy and reliability of a primary source, then it is a valid source for complementary information using direct citation. Second, if a reliable third-party source uses primary sources, using them in same context as the reliable source is perfectly valid, since it relies on the third-party source. Direct citation of quotes, figures, etc from those primary sources is probably preferable, for the reference of the reader, but the use should clearly rely on third-party source. This cuts to the heart of no original research, as well as the other core content policies, in my opinion. The independent use of primary sources relies purely on the opinions of the contributor, which is plainly original research to me. That original research also leads to NPOV problems, as without third-party reliable references we cannot make a determination of appropriate weight. Vassyana 11:47, 12 September 2007 (UTC)

But weight is given to opinions/points of view, given NPOV, and a great number of primary sources do not give an opinion or point of view. As long as they're used without any analysis (which we plainly agree would be OR) then this is fine in that sense. As to the opinion of the editor, I don't believe it does require that - it requires the editor's judgement, which we use all the time, being as how we're not mindless automatons (honest!). SamBC(talk) 16:59, 12 September 2007 (UTC)

An overemphasis on certain facts is certainly a violation of NPOV. I do however think this is far less problematic, or just less of a concern, when dealing with boilerplate base stats from a census or citing a complementary fact. I may have to disagree with you on the opinion/judgment position. In the absence of third-party sources commenting on the primary source, it really is nothing more than the opinion of the editor that the source is reliable and accurate. It's not a judgment because there are no supporting sources/facts to base a judgment upon. (It's not that I don't think we shouldn't employ some judgment and discretion using a little common sense. In fact, sometimes I think it should be a requirement to edit. *chuckle*) I hope that helps clarify where I'm coming from on this count. I really don't think we're too far off from each other here. Vassyana 01:21, 13 September 2007 (UTC)

Quick check

Okay, this is the third protection (because previous ones expired, not because of a concensus) of the page in as many weeks, with no progress. This page is currently at around 173,000 bytes (characters), and most of it (I'd estimate 2/3) deal primarily with the "Sources" section. If other policies or guidelines had a similar section, I don't know, but it seems those would have had just as contentious a problem as we have here (with no end in sight). Just for this version of the talk page, we have over over 300 edits in just over a week.

So, rather than skirt around the issue yet one more time, with discussions about "Should this policy distinguish between primary and secondary sources?" or whatever, how about just one simple question:

Should this policy attempt to define the types of sources and their potential uses, or should this subject be more adequatley covered in a separate "article" (probably a guideline)? wbfergus 17:43, 8 September 2007 (UTC)

As long as a warning as to the use and potential mis-use of primary sources is included here, I have no problem defining the terms "Primary source" and "Secondary source" elsewhere. Blueboar 17:54, 8 September 2007 (UTC)

yes (it should be moved) but... only after we make sure the matter is adequately addressed on a guideline page; and further, this page should endorse that secondary sources are normally required except for certain situations described in the guideline. Keep in mind that this talk page is frequented by thoughtful, experienced people who are unlikely to write a poorly sourced article but that the policy is also written for the mass of less cautious Wikipedians, and the default rule should be that unless you know better and have a good reason, stick to the secondary sources. Otherwise a weakening of sourcing requirements could open a floodgate to lots of bad material here. Wikidemo 17:57, 8 September 2007 (UTC)

Wikidemo, that is another splendid, insightful comment. I fully support what you've outlined regarding use of secondary sources. Dreadstar † 18:34, 8 September 2007 (UTC)

I agree. I would support a statement to the effect of: "Wikipedia encourages new people to become contributors, but new editors should work from mainstream secondary sources for their initial contributions to avoid confrontations over the nuances of this policy that could spoil their first impressions." Dhaluza 19:43, 8 September 2007 (UTC)

I don't see any smileys, but I take Dhaluza's comment to be tongue-in-cheek. The two problems I have with Wikidemo's suggestion are: (1) it does not describe Wikipedia policy, since nearly all our featured articles rely mainly on primary sources, not secondary sources, and (2) nobody seems to agree on what primary and secondary sources are. COGDEN 00:47, 11 September 2007 (UTC)

I think this article should be self-contained, and, if anything is necessary to it, that thing should be in the article. I must also protest that some other editors make side-swipes against primary sources in a section reserved for "one simple question." Jacob Haller 18:09, 8 September 2007 (UTC)

Protest all you want, but that's part of my answer and it is quite specific as to what to do and why. No policy is self-contained; they all relate to each other. Wikidemo 18:13, 8 September 2007 (UTC)

We could have a brief inline definition of the terms used but the detailed definitions belong in a page about sources, not this policy about the end result of how they are used. Dhaluza 19:43, 8 September 2007 (UTC)

Okay, I think that most everybody can agree that primary source material is more easily abused in regards to introducing OR. I also think that most everybody can agree that even secondary source material can fairly easily be manipulated into OR, though it takes more thought on how to do it. Accomplishing this with primary source material is much easier, and can be easily, if inadvertently, done by newer editors. Neither source types are immune from OR abuse, it just takes a bit more effort with secondary sources. wbfergus 18:17, 8 September 2007 (UTC)

Also, I would tend to not go with the self-contained approach, as I see that as what's caused this mess. The policy itself does state "Since the policies complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all three.", it seems to me as if the self-contained approach has already been addressed as not viable. wbfergus 18:28, 8 September 2007 (UTC)

Well, your statements about "most everyone" don't reflect what I'm seeing here. I figure that about as many of the people here disagree with your claim that primary sources are easier to abuse as agree. However, this goes far beyond "one simple question." Jacob Haller 18:40, 8 September 2007 (UTC)

Well, most doesn't mean all. Besides, some people just like to argue. :^) wbfergus 18:43, 8 September 2007 (UTC)

Whether "most everyone" agrees that primary sources are more easily abused (a conclusion I find impossible for anyone to make, but let's assume it anyway) the question is one of what a policy page should say, not one of the relative reliability of source types, determined a priori. I don't know if "more easily" means that primary sources are 500% more easily abused or 10% more easily abused (and I doubt anyone wants to see that issue pursued.) It may be an interesting observation, it may be a true observation. I don't see any legitimacy in using such a loose probabilistic assertion as the foundation for the secondary-source-favoring wording in WP:NOR. The policy is that OR should not appear in the material added by edits. Look at the material added, determine if it is OR. If it is, remove it. If not, leave it alone. It is not necessary to have a discussion of the hypothetical relative merits of different types of source appear in the policy. The OR can arrive in any number of ways.

Furthermore, the policy that seems to be the main and ruling policy here is that of verifiability. If a primary source not otherwise shown to be unreliable is given as the source of an edit then that seems to fully meet the most relevant policy requirement: it is verifiable (and Wikipedia explicitly denies it is concerned with truth,, so considerations of "truth" are not relevant to the discussion. Wikipedia has made its bed (truth is secondary), let it lie in it.) If the usage made is an OR usage then it is the improper usage that justifies the removal and it is irrelevant for the purposes of removal how the material was justified or sourced. I fully understand and do not deny that primary source material can be misused in a manner that creates OR. It is the misuse that is the offense, not the source. Misuse of any type of source does not blacken that type of source and it is inappropriate for any policy to assert or even hint that it does. Any such statement has to have so many qualifications attached that it ends up looking unwieldy. If a particular type of abuse relies 99% of the time on primary sources that's an interesting statistic but it's not reason to say use of primary sources thereby falls under a cloud. The statistic ignores the times when the usage of that type of source is fully appropriate, and the statistical approach to classifying sources is in any case weird. Asserting that material from secondary sources is always (or should as a first approximation be assumed always) more suitable for use is neither valid no useful nor likely to have the slightest effect on editing behavior. It probably will be the case that most material in Wikipedia is from secondary sources, just by the nature of things. But that's a statistic, not a policy. --Minasbeede 19:41, 8 September 2007 (UTC)

Some people think that primary sources are more easily "abused", but there's clearly no consensus about that. Even if there were, that's no basis for a blanket deprecation of primary sources. Quite to the contrary, there are good arguments that secondary sources are much easier to abuse, simply because they are more biased and likely to lead to WP:NPOV violations. There's a lot of give-and-take, and a lot of grey area on the subject of which sources are "better" than others, and we aren't anywhere near reaching a consensus on when secondary sources are better than primary sources, and vice versa. We might eventually come up with a guideline to that effect, but that has to be a separate article. This is a policy page, which really isn't meant for that kind of fine-grained balancing. Policy pages say, "this is the rule". Guideline pages say, "these are the factors to consider". COGDEN 00:48, 11 September 2007 (UTC)

Remove the source typing from the policy. It neither works nor is even clear enough to be understood. The potential for harm is greater than the potential for good. --Minasbeede 22:04, 8 September 2007 (UTC)

I agree with you on the 'self-contained approach' wbfergus. Due to the nature of these sourcing and content policies, there has to be some overlapping areas between them, we just need a better means of making sure that the overlaps match, and are not confusing or contradictory to each other. Dreadstar † 18:44, 8 September 2007 (UTC)

There is material about sources in these policies and guidelines:

I do not think we need more than that. If there are any discrepancies, please point them out. ≈ jossi ≈ (talk) 18:54, 8 September 2007 (UTC)

That's plenty sufficient for me! And I'm not aware of any discrepancies; I was just speaking in a general manner regarding some of the comments on this talk page about confusion and discrepancies. Perhaps the current system of checks and balances is fine; I'm open to anything that makes things better, simpler and clearer. Dreadstar † 19:06, 8 September 2007 (UTC)

≈ jossi ≈, Thanks for those links. I have all four open in different windows now so it's easy to swith between them all. To me, a realative newbie to all this discussion about NOR and sources, those links help highlight what I've been talking about. While they may all say pretty much the same thing, they all say it completely differently. They are most definately not a mere cut and paste across all four articles, giving the quick visual impression they are all completely different. It's not until you start reading them in their entirety that the similarities become a bit more obvious, though to those new Wikipedians who may be referred to them, that's a difficult assumption to make without spending an inordinate amout of time going back and forth comparing this sentence to see how it's described on the other links, etc. They all have good information, though they are all "wordy" in their explanation of that information. Just by a quick perusal of those pages, I think a merging of the text and style of the first two is the most "self-evident" of what the information is trying to convey.

Imagine how much less confusion there would be (especially for newbies), if each of the pages addressed the 'sourcing' issue strictly in relation as to how sourcing applied to that policy/guideline, and then had a link to one 'master' page (maybe a guideline), detailing in just one way what the different types of sources are, etc. It would give quite a bit more consistency to Wikipedia, and eliminate what appears to be both a confusing and therefore contentious 'subject area', as that 'subject area' would exist and be explained fully in just one place. Further details and examples could be linked to each of the different types of sources, like how primary, secondary and tertiary currently are. Your input, and those of other contributors to this policy, are greatly appreciated, as you have been involved in this for a long time and can apply your historical knowledge of this and other subjects to the discussions. But I also beg of you (all of you), please try to look at the subject from the eyes of a newbie. Imagine their confusion trying to figure all of this out, and what can be done to improve this policy (and maybe the other realted policies and guidelines), by keeping things simple, clear and consistent. Currently there is to much overlap trying to explain the same things different ways in different places. Thanks. wbfergus 12:06, 9 September 2007 (UTC)

I think that the policy should discuss sources in relation to original research. It should do so in harmony with existing policy and guidelines. Generally, there should be a distinction between modern reliable third-party sources and everything else. I think the primary/secondary distinction is problematic and widely misunderstood/abused. "Modern reliable third-party sources" would cover all the secondary and tertiary sources that should be used, as well as permit widely demanded "exceptions" for certain primary sources, such as scientific journal articles. Vassyana 19:18, 8 September 2007 (UTC)

I agree that the primary/secondary distinction is problematic, but I don't think "modern reliable third-party sources" is much better. Why always modern? In many cases, it's the ancient or original source you want, not the modern one. When you quote the Declaration of Independence, you should cite the original source, not some paraphrase by the American Enterprise Institute. When you have to quote the Bible, you'll get far fewer edit wars if you quote some unbiased translation, or even the original Greek, than if you cite a paraphrase by the Jehovah's Witnesses. Also, why always third-party? If you want to describe what Tom Sawyer looked like, why not cite The Adventures of Tom Sawyer? If you want to describe Darwin's theory of Evolution, why not cite Origin of Species? COGDEN 00:59, 11 September 2007 (UTC)

It occurs to me that situations where a large number of policies/guidelines have equivalent material, especially where it becomes important for those to be in harmony, it makes sense to have this (be it the sources matter here, or anything else) defined in just one place - then it can never become contradictory. It can be linked or transcluded as necessary. With reference to this specific case, the definitions of sources should be in once place that all policies/guidelines that need to can refer or transclude, while this policy should then discuss the application of this (if any) to the policy on original research. Well, that's what makes sense to me. Everything being self-contained is a recipe for contradiction between them. SamBC(talk) 19:44, 8 September 2007 (UTC)

It's the same with article content--when we have redundant material in different articles we split the content out and merge it into a new article (that probably gets nominated for deletion, but that's another problem....). Dhaluza 19:47, 8 September 2007 (UTC)

I am not sure that your proposal to consolidate is a good one. Each policy have specific wording related to sources in the context of the wording of the policy. ≈ jossi ≈ (talk) 20:02, 8 September 2007 (UTC)

Agreed. It's necessary to have some duplicated material that provides additional focus for the specific intent of the relevant associated policies. For instance, elements of NOR source evaluations can have a slightly different focus than WP:V source evaluations. While generalities about the sources can be spun out into a general sources guideline, I think each source/content related policy has to have its own sources section that specifically applies to that policy's particular concept. Dreadstar † 20:12, 8 September 2007 (UTC)

I agree, and I think that's in my comment - they should address the implications seperately, but not define anything seperately. Seperate definitions for seperate purposes within wikipedia policy and guidelines will readily lead to confusion during discussions, especially where people are trying to apply multiple of the policies/guidelines. SamBC(talk) 20:15, 8 September 2007 (UTC)

Is anyone preparing draft proposals for this? I'd like to participate. I know there are several drafts being worked on now, not sure if there are any over-arching Source drafts..? Dreadstar † 20:12, 8 September 2007 (UTC)

That was the original purpose behind the creation of Wikipedia:Classification of sources, although someone retagged it from "proposal" to "essay". SamBC(talk) 20:16, 8 September 2007 (UTC)

Cool! Thanks! Dreadstar † 20:24, 8 September 2007 (UTC)

I can tell you from my experience in writing technical specifications that it is important to never define a term twice, because you do not get to use the sum of the two definitions, you only get the least common denominator. It should also be self-evident that if we are defining terms in different places, there is a real risk that the definitions will be different, and that will only lead to confusion. I still think the source type definitions are out of place here. Dhaluza 18:51, 9 September 2007 (UTC)

I would demote the idea of types of sources to a background essay on reliability of sources. I would also demote the synthesis section in the same way. There is a clarity in No Original Research that is obscured by these different attempts to give examples of undesirable forms of OR. Spenny 00:36, 9 September 2007 (UTC)

I very much disagree on demoting the Synthesis section. Synthesising what sources say (whether primary or secondary) to form a conclusion is a major form of OR, and we need to make it clear that this is not acceptable. The whole point of this policy is that we don't put our own thinking into articles, but simply report the thinking of others. Any systhesis that is included in an article has to be citable to a reliable third-party source. We can re-word and clarify the section if needed, but it definitely should be included.Blueboar 12:19, 9 September 2007 (UTC)

I agree Slrubenstein | Talk 12:50, 9 September 2007 (UTC)

I view the sythesis section as something 'related' to "No original research", but a different topic that needs better 'coverage'. The policy can have a statement or two about synthesis, but it shouldn't define what synthesis is. How is that part of "No original research' except as yet another example of ways OR can be identified or creep into an article? I've seen (though not on Wikipedia - yet) policies that state the policy also includes the applicable guidelines, etc. Why can't this? It's not that difficult, and helps to "compartmentalize" the different subjects so that they can appropriately be defined and explained, while still part of the overall 'subject'. The main 'subject' is practically universally short and simple, with many other 'subjects' appended as needed or appropriate, not re-worded and included. wbfergus 13:23, 9 September 2007 (UTC)

This is a good objective. The question is, is it time to spin off the bulk of the content to a separate page? Our general practice is this: we work on main pages (whether articles or policies) until specific sections are good enough to spin off. As long as some editors feel that the no synthesis component of the policy can be improved - whether just in the style of the writing, or in the clarity of the conceptualization - it makes sense (and is common practice) to continue working on it as part of the main page (two reasons: it is the easiest way to ensure that, prior to spinning off, it is fully consistent with the main page - you know how easily it is for separate pages to become inconsistent; the more stable the version prior to spin-off the less danger of this); second, most people have a natural limit to the length of their watchlist and more people will pay attention to and comment on changes at NOR than at spin-off pages. So while I agree with you in principle, i think right now it is a premature move. Slrubenstein | Talk 13:43, 9 September 2007 (UTC)

I will agree with you on this. I have only seen the synthesis subject a couple of times in the last three (or so) weeks, so it seems fairly uncontentious at this point. I merely broached the subject of moving it a while ago as part of looking at the policy in current form, and identified it as another 'subject' that really didn't need to be here, though the subject itself is related to the NOR policy. It doesn't seem to be causing any conflicts yet, though who knows as more "newbies" come to Wikipedia. And for the record (since at times my posts may appear to be otherwise due to bad wording on my part), I am not opposed to the NOR policy, I think it is a good policy. But, I also think the current policy (specifically the sources section) needs cleaned up to alleviate confusion and contention. This seems to be the single biggest issue people have, both new and old. wbfergus 14:02, 9 September 2007 (UTC)

Well, I in corporatede COGDEN's suggestion into my proposal for the intro. If people agree it is an improvement, I would ask their permission to replace the current intro with my revised proposal - and thus move step-by-step. I would not unprotect the page (yet) - but neither would I add my revised proposed new intro unless there were general agreement that it would be a positive move. Slrubenstein | Talk 16:09, 9 September 2007 (UTC)

It sounds good. To help clarify exactly what it looks like now, how about adding a new section on here with it, and asking for votes of approval or disapproval, with something like a one week wait, so everybody has a chance to chime in? wbfergus^Talk 16:46, 9 September 2007 (UTC)

The last thing this talk page, already too long and in need of archiving, needs is another section. it is there, people are commenting. I agree with you it is worth waiting a week. As a more neutral party would you keep an eye on it and let me or another admin know if/when you think it is appropriate to make the change to the policy page? Thanks, Slrubenstein | Talk 17:03, 9 September 2007 (UTC)

Archive away..please...;) Dreadstar † 17:13, 9 September 2007 (UTC)

Math calculations

Are taking two sourced numbers and performing a math calculation original research? See Talk:Jatropha‎. For instance calculating any data on a per capita basis by dividing the value by the published population. --Richard Arthur Norton (1958- ) 22:01, 12 September 2007 (UTC)

The case in question is definitely OR. In general, taking numbers from a bunch of different sources and combining them in this way leads to the question of whether the basis for the numbers fits. Mangoe 22:15, 12 September 2007 (UTC)

So then saying a 1 degree change in average global temperature is less than a 0.4% change in average global temperature is also OR? That takes no number from any source, uses the fact that for temperature in radiative processes the proper scale is the Kelvin scale, and uses the definition of percent. --Minasbeede 22:23, 12 September 2007 (UTC)

I think that natural, obvious arithmetic is not OR, but that would be OR because it's a pretty novel comparison for someone without specialist knowledge (the Kelvin scale is pretty specialist). SamBC(talk) 22:37, 12 September 2007 (UTC)

Calculation that are so trivial that they cannot be accepted for publication in a peer reviewed journal are not OR. We cannot consider calculations that are too difficult to follow for lay people but are trivial to experts as OR, because then such calculations could never be included in wikipedia. An example is this calculation of an integral. Count Iblis 17:04, 13 September 2007 (UTC)

I think in that case, the fact that it can reasonably be looked up fairly trivially and isn't central to the article, it's not objectionable. Common sense rules. Of course, that section is also written like a textbook, but that's beside the point. SamBC(talk) 17:10, 13 September 2007 (UTC)

Agreed. I didn't ask my question for guidance, I asked it to illustrate the problem with the current wording - and was rewarded: I got my illustration. I had considered referencing the article on Pauli matrices, which I suspect is a very good article and also above the head of the guy who doesn't know about Kelvin temperatures - and maybe my head, too. That some readers may not understand or even know where to begin does not turn simple math into "original research." Yet with the current policy some lust to identify and remove anything that can remotely be characterized as "original research." The policy no longer serves Wikipedia, Wikipedia (all of its editors) must serve the policy. Do we have consensus on the notion that policy trumps consensus, once the policy has been written (snarky rhetorical question)? --Minasbeede 18:03, 13 September 2007 (UTC)

In my opinion, any calculation that could reasonably be challenged is probably OR, unless it's copied directly from a reliable source. Something that's truly trivial such as 2+2=4 will never be challenged. The per capita example given above could reasonably be challenged and shouldn't be included. Chaz ^Beckett 17:19, 13 September 2007 (UTC)

I am inclined to agree with Count Iblis on this topic. Math formulas that are considered trivial to an expert in the said field of research are not OR. The decision of what is trivial must be made by consensus of contributors who are knowledgeable in the discussed field. This is similar to the consensus that is needed to verify the validity of sources or the use of such sources in any Wikipedia article. There can be no line drawn at what is "truly" or "absolutely" trivial as even a case such a 2+2=4 can be disputed on the grounds that the author does not assert a base 10 system. In a trinary system 2+2 != 4, rather 2+2 = 10. Thus in the absence of such assertions, one must have familiarity in the field to understand the assumptions being made, however trivial or complicated they may appear. Gsonnenf 02:25, 18 September 2007 (UTC)

Refocusing on the source discussion

Reading over the discussion, the main issues people seem to have with primary sources revolve around 1) raw data, 2) historical/obsolete sources and 3) sources lacking analysis/interpretation and/or lacking editorial review of those claims. I am not necessarily proposing these are distinctions we should use in the policy. But rather, I'm inviting discussion of these points to see what distinctions we might arrive at to find a consensus for the policy.

Raw data. Using raw data to support a claim outside of reliable third-party publications seems to be an obvious violation. Using raw data to "correct" or contradict a published source also seems to be an obvious violation. There seems to be strong agreement on this particular point.
Historical. Using historical sources without third-party references to place them in context and provide analysis of the claims contained within such references seems to be a violation of this policy. Without third-party sources to provide such information and context, particularly about the reliability and historical importance of such sources, the use of such sources seems to be a de facto violation of no original research.
Inadequate sources.
1. Lacking analysis and/or interpretation. In the absence of analysis and/or interpretation, the source lacks context and information about its importance. In some cases, such context and importance is self-apparent, or plain on its face, such as boilerplate census data. However, this seems to be the exception, rather than the rule.
2. Lacking editorial review. Reliable sourcing is predicated on modern models of editorial and/or peer oversight to ensure reliability and accuracy. Sources lacking such oversight are by default non-reliable, unless endorsed by the general body of reliable sources as an accurate and trustworthy source. However, such endorsements are rare and often heavily qualified. For example, Caesar's Gallic Wars is widely regarded as a pinnacle of ancient reporting, but is also acknowledged to contain unfounded hearsay and as a masterpiece of propaganda. Without modern reliable references, we cannot determine which claims are factual, inaccurate, hearsay or propagandized without engaging in original research and potentially introducing gross errors.

Just some thoughts. What are yours? :) Vassyana 01:48, 13 September 2007 (UTC)

Raw data

Agreed - the use of raw data to draw conclusions is counter to WP:NOR (but would be ok for filling tables or referencing raw data like automic number or weight. 02:01, 13 September 2007 (UTC)
I can't see much reason to use or forbid raw data
- Using raw data without interpretation is rarely useful but may be useful under certain circumstances.
- Using raw data for interpretation is OR. When we make logical arguments, we take certain hidden premises for granted. Another editor may suspect our "logical deduction" of equivocation, of involving false assumptions, etc. We have to have specific sources backing our claims, they don't need to have specific sources backing removal. Jacob Haller 02:20, 13 September 2007 (UTC)
Agree. ≈ jossi ≈ (talk) 02:22, 13 September 2007 (UTC)
I have reservations with this. In part it is the terminology used to describe the issue which makes statements in a scientific context obvious and agreeable. My point is that in other contexts the issue is not so clear cut, (the stretching of the terms raw data and reliable source causing the contention). So taking sport for an example, results of races or matches and the publication of these by the official bodies might be interpreted by some as being raw data and thus inappropriate. However, they are statements of undisputed fact (except in rare circumstances) and in the specific context of being used as a statement of fact, the source is not only very appropriate, but in all rational arguments would be held above other sources that would contradict them. It would be perverse to hold a third party source that contradicted those facts as a better source through blind application of policy and any rational editor when presented with a contradiction where a third party source gave no indication of there being a dispute of interpretation would hold that source as suspect. In reality, such sources are likely to be journalistic and the editorial oversight likely to be lightweight and so there should be no issue in allowing that the source may not meet the grounds of reliability - and therefore it should not be an issue of using raw data to discredit another source, but using a sound source to replace unreliable information from a dubious source. Also, it is likely that there will be other contradictory sources, so the "raw data" would be a determining factor in assessing the reliability of the source. In other words, we must allow editorial judgement in the final analysis. Having said that, any interpretation of results, other than the most obvious of statements such as the simple arithmetic totalling of races won, must be original research and is discounted. In summary, I am not really talking about true raw data here, and I am not really talking about reliable sources: I am wary of falsely extrapolating sound policy to other areas. Spenny 09:45, 13 September 2007 (UTC)
- Hm, I quite see your point about how this could become problematic and I understand your closing statement. Any suggestions on how to disentangle the differing concepts? This is something to ponder. Thank you for your insightful comments. Definitely food for thought. Vassyana 10:17, 13 September 2007 (UTC)
I would make the very basic, bland statement that editors may not make analysis of or draw conclusions from raw data unless it's analysis or conclusions that come from a reliable source. Raw data does have plenty of uses. It's worth using a wording that indicates that this is exactly the same as drawing new conlusions or making new analysis using such from reliable sources as a priori premises - we do not analyse, we report. As such, I agree in principle but the wording of anything to this affect that goes into policy must be considered very carefully. SamBC(talk) 12:57, 13 September 2007 (UTC)

Isn't that simply saying NOR in a particular context? NOR is a policy. Is there a need to enumerate all the ways (or any ways) that OR can be done, but shouldn't (in Wikipedia) or is it overall far better to just say NOR? --Minasbeede 14:10, 13 September 2007 (UTC)

Disagree. Vassyana said: "Using raw data to "correct" or contradict a published source also seems to be an obvious violation. There seems to be strong agreement on this particular point." Perhaps I would need a definition of what you mean by "raw data". If the "raw data" has been published, I think it's important to be able to use it to correct or contradict published sources which are in error about it. For example, if the raw data is a piece of text and has been published, and a source says of it "It puts the words 'probability' and 'message' in the same sentence where it says 'probability is the message'", then a Wikipedia article should be able to use the published raw data to say "However, the raw data as published says "probability. Is the message ..." (if any of this is important enough to include in the Wikipedia article at all. Where there is significant controversy over a relatively minor point, often the best thing to do is to just leave it out of the article.) --Coppertwig 17:15, 13 September 2007 (UTC)

Historical

Please define historical source usage above - The United States Constitution is a historical primary source for content of the document, there is nothing more accurate then the original. Are you suggesting that a resource older then a specific age suddenly becomes unreliable? Jeepday (talk) 02:09, 13 September 2007 (UTC)
Agree. The keyword here is historical sources without third-party references, which is not the case in your example above. ≈ jossi ≈ (talk) 02:23, 13 September 2007 (UTC)
I suggest using, not avoiding, ancient historical sources. However, editors should use modern sources if they combine historical sources or to interpret ambiguous passages.
- I recommend that editors take numbers with a grain (or 100,000 grains) of salt. But this doesn't belong in NOR.
- I recommend (from experience) that editors use modern sources for timelines where possible. But this doesn't belong in NOR.
- Where ancient sources conflict, editors may use recent scholarly sources to pick one ancient source over another, or they may summarize both sources and their claims.
- Where recent scholarly sources conflict, editors should summarize these sources and their claims. Jacob Haller 02:38, 13 September 2007 (UTC)
Aside from the definition issue, which we may be able to fudge, this (Vassyana's point) seems reasonable. SamBC(talk) 13:00, 13 September 2007 (UTC)

Inadequate sources

Agree. Context is everything. ≈ jossi ≈ (talk) 02:24, 13 September 2007 (UTC)
Two additional concerns involve:
- Sources which are reliable for one topic, and refer to another topic in passing
- Sources which focus on one topic, but use special definitions for key terms... Jacob Haller 02:55, 13 September 2007 (UTC)
It seems to me that one of the issues is the concept of editorial review.
- In an academic journal, there is a professional need to be seen to properly assess the worthiness of the paper. Outside the academic world, there is not the same editing process: a newspaper will have a number of editors who are qualified, but that editing process is not designed to ensure that what is written is neutral, or even correct in the worst cases. On any given subject a newspaper will have an agenda.
- In reality, the world is not made up of perfect sources, but a collection of imperfect sources, whether through age, political bias, agenda and so on. We must be careful that policy is pragmatic enough to account for this. One of the underlying issues on many a dispute is that a good editor can recognise a poor source, and can justify this but the policy does not give guidance on quality of sufficient robustness outside the context of the academic world to allow the edit in question to be challenged without this leading to contentious disputes.
- We need to explore these concepts for articles which are encyclopaedic but outside the academic sphere. For example, when we look at an entry on a notable town, there are a lot of sources which are based on a tourist view of a place, and play up those aspects - a majority view and the only ones the editors have ready access to. It may be very difficult to find other sources without doing some research... Spenny 10:16, 13 September 2007 (UTC)

Reasonable points, but what is the relevance?

Talking about various types of sources and their relative merits and demerits is interesting, but this does not address the subject of the policy - that editors should not add their own data and interpretations to articles. What types of sources publish useful and accurate data and interpretations, or what the editor's own ideas are based, on is not really all that relevant to this simple principle.

I think this discussion of various types of sources does not belong in the policy at all. Tim Vickers 03:32, 13 September 2007 (UTC)

I'm a bit confused, since my initial post in this section clearly describes the relevance of the various points to this policy. Are you disagreeing that those points are valid? Are you indicating they are extraneous to the discussion? Are you saying these issues should be addressed elsewhere? Please be more clear. Vassyana 04:47, 13 September 2007 (UTC)

First, I think TimVickers (correct me if I'm wrong), is saying that eventually, hopefully pretty soon, "Types of Sources" and their definitions, with examples, should be moved out of the various policy pages for several reasons, one being consistency of definitions. It is fine for this policy to address "original research" principles as they relate to reliable sources vs. questionable sources, and how drawing any conclusions from the source (and then writing that into the article) if it hasn't already been published, is an example of OR. But it is beyond the definition of this policy to begin encroachment into a different subject area, mainly the types of sources, their definitions and the whole "good" vs. "bad" debate. Keep this policy focused soley on what OR is and why there is a NOR policy. A few examples of how various sources can be used to infiltrate OR into an article may be appropriate however, and can be gleaned from previous discussions here. wbfergus^Talk 13:50, 13 September 2007 (UTC)

Thank you for explaining, or at least providing your understanding of his comments. It's helpful to have a better idea of what's being expressed. :) Vassyana 15:58, 13 September 2007 (UTC)

To make wbfergus' point in a simple way: "The sources may be poor, they may even be wrong, but it doesn't make it original research if they are used, just a bad article with poor sources." Spenny 16:23, 13 September 2007 (UTC)

Rant on special definitions

e.g. Ludwig von Mises wrote a work arguing against centralized economic planning. He called centralized economic planning (regardless of whether it involved worker-ownership) socialism and worker-ownership (without central planning) syndicalism (and hastily dismissed the latter). Syndicalism proper, mutualism (economic theory), etc. and most other forms of libertarian socialism get grouped into syndicalism and excluded from socialism. Historically, all these have been considered socialism. This gets messy when Mises makes categorical statements condemning "socialism." It may not be clear whether these arguments apply to socialism as a whole or to socialism exclusive of syndicalism (etc.). Commentaries may inherit these definitions and face the same problem. Jacob Haller 02:54, 13 September 2007 (UTC)

I understand your point. One of the roadblocks on factory farming was the inability to come to consensus on the term as to one party it was quite clear that in different contexts it meant different things, to another party it was interpreted more strictly as just one of those meanings, and then this strict definition was used to interpret other sources. Good analytical editing is required to avoid an issue here. Spenny 08:31, 13 September 2007 (UTC)

Another proposal

Per discussion, I revised the introduction, taking into account coments and edits by other people active here.

I would not like to propose to change the following section, "What is excluded." I have three problems with this. first, it opens up not with a statement about what is excluded, but with the motivation behind the policy, so the title of the section and part of the contents are inconsistent. Second, the list of what is excluded seems ad hoc. Third, following sections explain in greater detail what is excluded and why.

Here is what I propose: since we are still debating primary versus secondary sources and what is actually included and excluded, I suggest that instead of calling this section "What is excluded" let's follow the lead of the first sentence and make it a section about what motivated the policy i.e. its origins. Such a section can discuss things that have traditionally been excluded, but in the context of the history of the proposal. I think this explanation of the history would be educational.

ALL I am proposing right now is to change the second section. I am not proposing anything about the third or following sections (on sources and synthesis).

here is what I propose:

Origins of the policy

The core policy of Wikipedia, NPOV is meant to provide a framework whereby editors with diverse, often conflicting, even opposing points of view can collaborate on the creation of an encyclopedia. It does so through the principle that while it is often hard for people to agree as to what is the truth, it is much easier for people to agree as to what they and others believe to be the truth. Therefore, Wikipedia does not use "truth" as a criteria for inclusion. Instead, it aims to account for different, notable views of the truth. First codified in February 2002, the objective of the NPOV policy is to produce an unbiased encyclopedia.

In the year that followed a good deal of conflict on article talk pages involved accusations that editors were violating NPOV, and it became clear that this policy, which provided a philosophical foundation for Wikipedia, needed to be supplemented. Wikipedians developed the concept of "verifiability" as a way of ensuring the accuracy of articles by encouraging editors to cite sources; this concept was established as a policy in August 2003. Verifiability was also promoted as a way to ensure that notable views would be represented, under the assumption that the most notable views were easiest to document with sources. Notability is especially imortant because while NPOV encourages editors to add alternate and multipe points of view to an article, it does not claim that all views are equal. Although NPOV does not claim that some views are more truthful than others, it does acknowledge that some views are held by more people than others. Accurately representing a view therefore also means explaiing who holds the view and whether it is a majority or minority view.

Soon it became evident that editors who rejected a majority view would often marshall sources to argue that a minority view was superior to a majority view - or would even add sources in order to promote the editor's own view. Therefore, the NOR policy was established in 2003 to address problematic uses of sources. The original motivation for NOR was to prevent editors from introducing fringe views in science, especially physics - or from excluding verifiable views that, in the judgement of editors, were wrong .^[1] It soon became clear that the policy should apply to any editor trying to introduce his or her own views into an article (and thus a way to distinguish Wikipedia from Everything 2). In its earliest form the policy singled out edits that:

introduce a theory of method of solution;
introduce original ideas;
define terms; or
introduce neologisms

for exclusion, and established that

ideas have been accepted for publication in a peer reviewed journal; or
ideas have become newsworthy: they have been repeatedly and independently reported in newspapers or news stories (such as the cold fusion story).

as criteria for inclusion.

As before my attitude has been conservative, to try to preserve as much as the previous content as possible. I have attempted to make the layout more consistent and clearer. Slrubenstein | Talk 10:53, 13 September 2007 (UTC)

I appreciate the changes, and agree 100%. This improves the page considerably. But doesn't it send the wrong signal when the administrator who imposed edit protection and chided another for removing it is now serving as gatekeeper for edits to the page? Even in the hands of a benign neutral administrator that becomes de-facto arbitration over page contents. One thing edit protection does is to calm edit wards by forcing people to come to their senses when they realize that if they won't play fair they don't get to play at all. If edit protection is just a way to keep everyone but administrators from editing because nobody else can be trusted, that seems exclusionary.Wikidemo 12:09, 13 September 2007 (UTC)

Very good changes! Vassyana 14:33, 13 September 2007 (UTC)

More than a day ago people were expressing support for the change, and I asked anyone if they objected. No one registered objection. If someone else wants to act as "gatekeeper" they can but I haven't made any changes without ensuring that people active in discussion on this page approved. That is what is important, I think. Slrubenstein | Talk 12:18, 13 September 2007 (UTC)

I think the page is still protected because someone has gone to mediation on primary versus secondary sources - as I have made clear, none of my proposals have addressed that issue, the main source of contention, Slrubenstein | Talk 12:19, 13 September 2007 (UTC)

Mediation should not require the page to protected. So long as edit warring won't result, I don't see the need for continued protection. Vassyana 14:33, 13 September 2007 (UTC)

Thanks for the explanation. Mediation? Where? I hadn't noticed, and don't see a link. Can policy decisions even be made by mediation? Wikidemo 12:42, 13 September 2007 (UTC)

Policy decisions cannot be formed by mediation. However, mediation may help conflicting editors reach a consensus that can then be reviewed by the broader community. Vassyana 14:33, 13 September 2007 (UTC)

I've said I have a problem with certain wording and I've said go ahead and make the change, with that wording if you want it. I don't think that wording will persist so putting it in is just the seed of further controversy but go ahead, put it in. Change the Introduction and we can move on.

I'm also conservative. I want the policy returned to how it was before the contention-causing source typing and synthesis sections were added. It was never necessary for the changes to occur, it was necessary (under one definition of "conservative") for those changes not to occur. There's no need for a "my conservatism is better than your conservatism" spat. The point is that labeling a position as "conservative" doesn't say much with any practical utility.--Minasbeede 17:48, 13 September 2007 (UTC)

Like Wikidemo, I also agree 100%. Let me also add that I appreciate the work you have put into this and your patience through this long and tedious process. We may not always agree on different points, but you are not being bull-headed or anything either. I feel that your slow approach of addressing things one at a time is more beneficial to the entire process than attempting to wrap everything up all at once. This way the different salient points are being addressed individually instead of collectively. I also think giving the history helps to not only put things in perspective, but also adds another (needed) dimension to the explanation and policy. wbfergus^Talk 12:36, 13 September 2007 (UTC)

Here is where the sources issue is under mediation. I hope that the talk on this page can be refactored in order to work with the mediation process. in the meantime - I look forward to commnents on my proposal, above, to replace the "what is excluded" section with the above "Origins of the Policy" section - and I repeat: I wrote this with an intention of not addressing the current debate over sources, but simply an attempt to rationalize and more clearly explain what is currently in section 2. PS thanks wbfergus, and Wikidemo Slrubenstein | Talk 13:34, 13 September 2007 (UTC)

Last I looked 12 of 18 had agreed, 6 hadn't, and 2 of the 12 had withdrawn their agreement. It's not going to happen: the rules require unanimous agreement for mediation to occur. --Minasbeede 17:49, 13 September 2007 (UTC)

Firstly, thanks for implementing the change to the lede in a fully correct and well discussed manner, with some last minute improvements which meet some concerns I was thinking about. Secondly, the proposed "Origins of the Policy" section seems to me to be a bad move as it looks like background and introduces a large lump of text which by implication is no longer current – anyone reading through the policy is likely to skip over it to get to the current requirements. The "What is excluded?" section is more concise, is clearer and the heading matches the NO original research article title – proposals for a better title or other improvements could be considered, of course. .. dave souza, talk 18:01, 13 September 2007 (UTC)

Thanks. Of course people can skip over any section they want to. But the point is not that the content is no longer current, the point is that the origins of the policy help explain why it exists and why people think it is important. Anyone who wants the "rules" in a nutshell can ... well, look at the "policy in a nutshell." Precisely because we provide it in a nutshell, I think there should be space in the rest of the page to provide more of an explanation. For some people, history is a very valuable part of an explanation. For those who don't care ... as you say, they can skip it. Giving readers this choice does not to me seem like a bad idea. Slrubenstein | Talk 14:40, 14 September 2007 (UTC)

Bravo, Slrubenstein. Good edits, and excellent historical perspective on the evolution of our content policies. ≈ jossi ≈ (talk) 03:57, 14 September 2007 (UTC)

Bravo, indeed! Very well written and an excellent summarization of the history and purpose of the content policies. Wonderful! Dreadstar † 19:57, 15 September 2007 (UTC)

Why we DO need some sort of a statement on primary sources usage

Leaving asside the debates on current wording and definitions for a moment, I want to focus for a moment on why I think having a section that discusses primary and secondary sources (in some form) is important and useful. As we currently define the term, primary sources are things that do not contain analysis, interpretations and conclusions ... things like raw data, eye witness reporting, original documents etc. I think we are all agreed there are situations where quoting such a source is both useful and appropriate. However, because primary source materials do not contain any analysis, interpretation or conclusions, they have to be used with great caution. It takes a very skilled editor to use a primary source material without interpreting, analying or making a conclusion from it. Since whole point of this policy is that editors should not put their own thinking and conclusions into articles, but should instead report on and cite the thinking of others (preferably the best experts in the field) we need to caution editors about this potential problem. We need a statement that explains the pitfalls of using primary source material in an article. It is a caution and explanation, not an out right ban. Blueboar 13:01, 13 September 2007 (UTC)

Hear, hear. This puts the situation as I see it very clearly and concisely. It must simply be very clear that it's a caution and explanation, not something saying "primary sources are bad". SamBC(talk) 13:26, 13 September 2007 (UTC)

I think that is a clear perspective. I am more than content with the conclusion, but whilst it is a suitable generalisation, I would stick with my point of the day which is that sources are rarely perfectly formed, classifiable things, so I am not really comfortable with trying to use classification as a mechanism to assess sources. I would make an analogy with the UK 4x4 debate: should these vehicles be banned? The shorthand discussion knows exactly what is meant, but it misses the point, the issue being excessive emissions, and there are genuine applications for the vehicles which is appropriate use and there are vehicles that are not 4x4 which have the same problem. Once we label rather than consider the underlying issues we confuse and misdirect. Spenny 13:54, 13 September 2007 (UTC)

WHOA WHOA WHOA WHOA Let's pause and think

I need to learn to read carefully. Above it says "Since whole point of this policy is that editors should not put their own thinking and conclusions into articles..." (Italics added.) No, that is not what the policy says or is about. The policy, if it intended to reject thinking, would have identified thinking as something that is excluded. It doesn't. It identifies novel interpretations as being excluded. There is not a one-to-one correspondence between "thinking" and "presenting novel interpretations." This is a very noxious form of policy creep and is the reason I'm in this discussion in the first place. The difference between "thinking" and "presenting novel interpretations" exists and no policy wording should extend the prohibition to cover all thinking without a very clear consensus (which I hope and believe could never be attained.) Thinking is not out of place in an encyclopedia. I understand and agree with the principle that says Wikipedia is not a place for the introduction of novel ideas. I understand that thinking is needed to create a novel interpretation. It is still hugely untrue that "thinking" and "presenting novel interpretations" are the same thing. An editor has to think to compose a sentence. An editor has to think when deciding on the order in which to present material. An editor has to think when deciding which material to use, which to ignore, which to give prominence, which to downplay. An editor might think of a way to present an existing idea in a new way. That's not identical with "present a novel interpretation."

Be alert to improper novel interpretations made by editors. Alert means alert. It is not a sign of alertness to automatically blacken and reject all signs of thought. The policy was not presented originally as a rejection of thought by editors, it has not become that (although things have gotten added on that approach rejection of all editorial thought.) That's a bad direction in which to be headed. --Minasbeede 04:13, 15 September 2007 (UTC)

Minasbeede, there is no point in a discussion about rewording a policy if we limit ourselves to its current wording. The very reason we are having this discussion at all is because some people think the current wording is unclear or misguided. If you mean to limit NOR to rejecting "novel interpretations" and to allow for editors to forward their own views, I think you are misguided, and if you think this is what the policy says, then I think we need to reword it to make it clearer. Slrubenstein | Talk 12:46, 15 September 2007 (UTC)

If I do that, sure. But I don't, and I find it insulting and improper for you to put words that misrepresent what I am saying into my mouth. It is very clear, from my repetitions, that I am objecting to the equating of "thinking" with "presenting novel interpretations" and it also very clear that the original formulation of NOR was meant to codify that Wikipedia was not the venue for "novel interpretations." I have repeatedly asserted that I agree with that principle. Let me make it more simple: it is not and should not be Wikipedia policy that "no edit made to Wikipedia should ever reveal that the editor engaged in thought." That would be an absurd policy. What I objected to asserts otherwise: editors are not to think, the material above equates all "thinking" with "original research." I deny that. Shall we see what the consensus is on "thinking" in Wikipedia?

I think the current source typing wording is unclear, always has been. I think that is misguided, always has been. I think the wording does not belong in a policy and it is well past time for those who want the wording to start giving active recognition to the opinions expressed here that also say the wording does not belong in the policy. I approve of the effort to clean up the wording, I made an early proposal that said that the first order of business should be to change the terminology. That is now proceeding, haltingly, and I am glad to see it. I think ultimately the source (or "material" typing language should be ejected form policy. I think the process ultimately will reveal that such rejection is by far the best action. --Minasbeede 14:03, 15 September 2007 (UTC)

No one is suggesting that thinking is prohibited, and to suggest that is simply absurd. You're really starting to border on disruptive here. What is being said is that editors should not insert their own thoughts and conclusions into articles, and I have a hard time believing that you do not know that. To be utterly blunt, if you honestly couldn't see that, you shouldn't be taking part in this conversation, as it would belie a deep lack of comprehension. Vassyana 14:16, 15 September 2007 (UTC)

The word "thinking" was used. I quoted accurately, did I not? The quotation directly suggested that thinking by editors was prohibited. We might perhaps be able to have a long discussion about the difference between "insert their own thoughts" and "thinking" but unless you insist there is absolutely no difference I assert that my point was valid. Rather than attack me (twice) the better course might be to agree that I have a point. I appreciate your assertion that thinking is allowed and, I guess, your implicit agreement that it should be allowed. I did make my point forcefully and it was my intent to do so. I see that sentence as a sign of creep and I wanted to noticeably object to the creep. I should be happy: it was noticed. --Minasbeede 14:31, 15 September 2007 (UTC)

The word "thinking" has several meanings. Here are a couple: v.intr. 1. To exercise the power of reason, as by conceiving ideas, drawing inferences, and using judgment. 2. To weigh or consider an idea: They are thinking about moving. I think in context it is clear to any reasonable person that the meaning of "thinking" in this instance is one's thoughts on the subject of the article. I think it is just as obvious to any reasonable person that in context it does not mean one should not think at all. For you to suggest so seems unnecessarily argumentative, especially with the melodramatic WHOA WHOA WHOA. Vassyana is right: it is disruptive. Let me be blunt: after accurately quoting text, your entire reaction is based on various mis-quotations. The quote does not say that "thinking by editors is prohibited." It just does not. You are making that up. The quote does not say "no edit made to Wikipedia should ever reveal that the editor engaged in thought." It simply does not. You are making that up. The quote does not say "editors are not to think" - it just does not. Again, you are making things up. Am I misrepresenting you? no, i am not. You are misrepresenting the very passage you quote, and that insults all of us. Slrubenstein | Talk 17:42, 15 September 2007 (UTC)

I was indicating that the quotation represented policy creep. There has been policy creep. I have already acknowledged to blueboar that I did not interpret his remarks as meaning he opposed all thought. The point is that if there's a continual equation of "own thinking" with "forbidden original research" that goes too far and it will end up being that all thought is forbidden because that is "original research," and that could happen because editors were insufficiently attentive to what words in or about policy really mean. Simple logic, a form of thinking recognized and valued for millenia, is currently forbidden by the "synth" section of the current NOR policy. That to me is policy creep and is harmful to Wikipedia. I think that policy creep was done by a minority and should be undone. I very much oppose further policy creep and I am clearly fully entitled to do so and be so. I did screw up: I should have said in my comments what I finally did say to blueboar, which is visible to all on his talk page.

If what I have said has served to alert other editors to the danger of policy creep I am more pleased than sorry. My regret is the failure to adequately indicate that I was not aiming my remarks at blueboar, who is doing a superlative job, even though he and I appear to disagree. --Minasbeede 19:47, 15 September 2007 (UTC)

okay. Slrubenstein | Talk 19:53, 15 September 2007 (UTC)

To get rid of bad thinking we reject all thinking. To get rid of bad reliance on primary sources (or reliance on bad primary sources) we encumber all use of primary sources. Where does it lead? During the period cited in the history above there surely were instances of faulty "original research" appearing in Wikipedia. It's a wiki, what else do you expect? I suggest that during that same period there were far more instances of "original research" that created sound Wiki material and improved the quality of Wikipedia, accepted by all. Those edits weren't forbidden by any policy, they were not grating, they were not particularly noticed because they reinforced the overall quality of the encyclopedia. Because of some number bad wiki edits all such contributions are now to be rejected - and the criteria for rejection are slowly expanding. Objections arise to source typing, have been arising since 2004. The objections get beaten down, only to arise again. This appears to be a classic example of what is described in the Consensus article.

Tim Vickers was asked to come to this page and comment. His comment indicates that he sees no reason for source typing. That opinion, from a practicing scientist, deserves far more attention than is getting. After all these years there is still insistence on having source typing be a part of policy (with the force of requirement, not just guidance) and yet it's still not well enough defined to be understood by all. I agree (as should be obvious) with Tim Vickers. --Minasbeede 14:12, 13 September 2007 (UTC)

Yes, and I was the one who asked him. And he and I often disagree - it doesn't mean I don't value his views. But his views carry no moe weight than any other editor. I provide people with my reasons for my views (as does Tim); I never ask that my views be given more weight than others because of my credentials. Slrubenstein | Talk 14:44, 14 September 2007 (UTC)

I think the vague description of the type with a warning, clearly written as anything but prohibition, serves a purpose. It might be better, however, to simply advise caution when using sources that "provide no analysis or commentary" or suchlike, to avoid the use of the primary/secondary terms whatsoever. SamBC(talk) 14:08, 13 September 2007 (UTC)

Yes. The combination of the words "policy" and "vague" is a formula for disaster. But that makes me think: is the combination of "vague" and "guideline" all that much better? This comment would become inoperative (hooray!) if "vague" were gone. Can that happen? --Minasbeede 15:22, 13 September 2007 (UTC)

Well, how about getting rid of the definitions of primary sources, but say, for example, "Sources that provide simple information without commentary or analysis, such as many primary sources, are particularly difficult to use without inserting your own commentary or analysis. Editors should thus be careful when taking information from such sources." This could easily be followed by a description of the situations in which it tends to be safe, if people want that. SamBC(talk) 16:51, 13 September 2007 (UTC)

You know, I am beginning to think that the problem is more with the word "sources" and less with the word "primary"... what the policy intends (as I see it) is a caution about certain types of primary material (usually found in primary sources) ... That we should not analyse or draw conclusions from such material ourselves, but should cite secondary sources that have done so. Does the distinction of the word "material" clarify the issue? Blueboar 17:18, 13 September 2007 (UTC)

Looks useful to me, as it's common for a source to include both "primary" data about the writer or the subject, and at the same time "secondary" analysis or opinion about another source. .. dave souza, talk 17:40, 13 September 2007 (UTC)

The policy is "no original research." What is special about some material that requires it to be singled out so that it can be said "don't do original research using such material"? As Tim Vickers says, this is all very interesting, but does it need to be in the policy? --Minasbeede 17:53, 13 September 2007 (UTC)

The common problem is people assuming that the raw data is the "research", and it's ok to put together published "research" in a way that reaches a synthesis or analysis that's not published – which is why "secondary" sources providing such assessment of the "primary" data are generally, though not always, more useful. ... dave souza, talk 18:05, 13 September 2007 (UTC)

It's original research, it's not allowed. How "people" did it does not seem to me to make it necessary for the policy to single out that particular type of OR and say it's forbidden. I see what you're saying and understand the concern but a look at the history of the article shows that however noble the intent the actual wording has been a source of dissatisfaction for essentially the whole time the wording has been there. And in reality all this says is "and original research of this type is included in what is forbidden." Plus see the synth discussion above: including material that speaks to an already-introduced issue is not the introduction of a novel view: the view is already there. the Abercrombie quote sheds light on the issue, even though it isn't and can't be regarded as definitive. It's policy creep to forbid that usage of primary (or any type) material. Literature research is allowed, literature research is demanded. Using NOR automatically to nullify valid literature research is an injury to Wikipedia, not a benefit. --Minasbeede 18:14, 13 September 2007 (UTC)

Which is why it can be useful to set out the "primary" data in the context of "secondary" view, as that example seems to be doing. In my view these are preferences and aims rather than absolutes. As it happens, today I noted discussion on an article about a book where it was questioned why we didn't summarise the book, then show critical analysis separately. The response pointed to the need to utilise the (readily available) secondary analysis rather than making our own assessment of the book. There were further NPOV issues, but the basic point was useful. .. dave souza, talk 18:31, 13 September 2007 (UTC)

A lot of editors don't understand that using primary material as a source for analytical or conclusionary statements is a form of OR... we have to spell it out for them. Blueboar 18:48, 13 September 2007 (UTC)

Do we have to spell it out in a policy?

Plus, at one level, I don't even care if it is in a policy - but if you spell it out it has to be clear enough that most understand it (else it isn't actually "spelled out" at all.) Currently the evidence suggests it isn't spelled out at all well. That means it fails for what it supposedly is intended to do and also muddies the policy. Muddying the policy could lead to edit wars and long, long, long discussions. And, glory be, that is exactly what we have. --Minasbeede 18:58, 13 September 2007 (UTC)

Blueboar, it's not that we don't understand, its that we don't agree. When you are the author of a new analysis or conclusion, and somebody cites you, they are citing the primary source: you. You are the primary source of your own original ideas, even if those ideas build upon the ideas of others. As to the raw ideas of others minus your novel insight, you are a secondary source. This is how the term "primary source" is generally used by librarians and historiographers. COGDEN 19:02, 13 September 2007 (UTC)

I support having some sort of statement concerning classification of sources somewhere: I just don't think the place is here and the time is now, because we don't have a consensus, and these factors concern not just WP:OR, but WP:NPOV, as well as WP:Notability. I think there are a lot of good suggestions that could quickly be crafted into a guideline, however. COGDEN 19:02, 13 September 2007 (UTC)

What we do have is a policy, which to many editors is useful and important. It's fine if you feel you know it all well enough to do without clarification and guidance, but this policy has to be read and understood by newcomers. in my opinion the distinction between "primary" and "secondary" as used is worthwhile. .. dave souza, talk 19:12, 13 September 2007 (UTC)

Cogden, I think you have a fundamental misunderstanding of what WP:OR is saying. WP:OR is about what Wikipedia editors do... analysing material and drawing conclusions from it - and doing so in Wikipedia articles. It isn't about what the authors of the sources do. Citing others is what we want Wikipedia editors to do. It does not matter if that source makes an "original" anaysis or draws "original" conclusions... since the analysis and conclusions are external to Wikipedia, we can cite to the source that contains them (assuming it is reliable). This is why I draw the distinction between primary 'material' and primary 'sources'. It is using primary material that leads to OR.

Minasbeede - in my opinion, yes, we have to have it in policy. It is important enough to include it. How we phrase it is something I am willing to discuss, but I do think we need it. Blueboar 19:33, 13 September 2007 (UTC)

It sounds like we agree. Citing others is exactly what we want Wikipedia editors to do. And whether the "others" they cite are primary sources or secondary sources of information, in neither case is it original research. As long as the Wikipedia article itself is not a primary source, OR is satisfied, almost by definition. If the purpose in making a distinction between primary and secondary sources is limited to that simple principle, I would be in favor of addressing it here. Where I have a problem is when this policy pages starts going beyond the concept of OR and addressing the relative pros and cons of various types of sources outside of Wikipedia. COGDEN 19:50, 13 September 2007 (UTC)

Quite often, a primary source includes both the raw data and some context and interpretation, and fairly often, a secondary source includes some raw data as well. I don't see why it's a primary/secondary issue. Jacob Haller 19:47, 13 September 2007 (UTC)

Focus on the intent of the policy... When writing a Wikipedia article, using raw data, original documents, eyewitness accounts, and other such materials (what for the sake of argument I am calling primary materials) to support an analytical interpretive or conclusionary statement - THAT is OR. Citing someone else's analysis, interpretation of data, or conclusions is not. It's that simple. Blueboar 19:59, 13 September 2007 (UTC)

Blueboar says "Its that simple." I think the original NOR policy was "that simple." That simplicity is a meritorious goal. In a policy "goal" means more than simply "something that considered abstractly would be nice." --Minasbeede 20:46, 13 September 2007 (UTC)

Exactly. I see the primary/secondary language and its interpretive problems as a confusing distraction from the intent of the policy. It says one thing (primary/secondary) when it means another thing (we need to source interpretations as well as data). Jacob Haller 21:01, 13 September 2007 (UTC)

"Data". Now there's a nice simple and rather ambiguous word. In the context of a book review, one book (the primary source) is "data" and another is "interpretations" (the secondary source). Except the "interpretations" can readily be data, selected, analysed, interpreted or synthesised to apply to the first book. Most people seem to find the primary/secondary distinction a useful way of establishing the relationship of the source material to the subject. Why are you having so much difficulty with it? If you've spelt it out earlier, do please summarise it or provide diffs. .. dave souza, talk 22:27, 13 September 2007 (UTC)

Ammianus Marcellinus' histories are considered primary sources in history. They include both claims and interpretations of these claims. According to NOR they are partly primary and partly secondary sources. According to Wikipedia:classification of sources we need to know what sources Ammianus may have had, and sift through to tell which parts are primary and which parts are secondary, because his new ideas are primary and his copied ones are secondary. And, according to that essay, when a secondary source comments on a primary source, it becomes a primary source. Jacob Haller 23:18, 13 September 2007 (UTC)

It's only a primary source of the new commentary. It's really a simple concept: the primary source of an idea is the source in which the idea originated. It's true, though, that determining who is the primary source of something is sometimes difficult. That's why historiography is such a high-falutin academic field. If it were easy, you wouldn't need a Ph.D. The difficulty in application, in my view, is a very good reason not to base this policy on this distinction. Whether a source is the primary or a secondary source of an idea really doesn't matter, anyway. In either case, it's not OR. COGDEN 02:19, 14 September 2007 (UTC)

Regarding that essay, This document is extremely rough and not of any use for anything. Don't link here except from talk pages. User:Slrubenstein/NOR looks much better to me. Always, it's a judgement call: The Origin of Species is clearly a primary source for Darwin's ideas, but I've found it of use as a secondary source providing historical information about Baden Powell (mathematician). ...... dave souza, talk 23:53, 13 September 2007 (UTC)

We keep coming back to disagreement over the word "source"... If we get rid of that word, and substitute something like "material" (or some other word) does that change anyone's feelings about the section?

Cogden, you say you disagree with my assertion that using primary material (defined at the moment as including "archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs") to support analitical, interpretive, or conclusionary statements is a form of OR (correct me if I misunderstand your statement)... can you give me an example of a situation where you think using such material to support such statements is appropriate and not OR? Blueboar 13:02, 14 September 2007 (UTC)

I'm a clean-language-in-policies-advocate. Of course Cogden can't give an example: you specify (by the words "to support analytical, interpretive, or conclusionary statement" that the usage is OR. As a clean language advocate I see no need for a long list of examples of primary sources. It's not the sources that create the OR, it's the usage. That's adequately covered without any list. Creating language in which some part of the list is used to feed an example of OR doesn't make the list (or the entire gamut of source typing) necessary in WP:NOR. Editors are not to make their own "analytical, interpretive, or conclusionary statements." That's NOR. No list is needed, no source typing is needed. The source typing creates confusion, not understanding (and confusion will always attach to using the words "primary" and "secondary" to mean something in Wikipedia different from how the words are used and what they mean elsewhere. Not totally different, but different enough to cause confusion, as can readily be seen by looking at the history of this talk page. There is no option available to change the readers who are confused. The only viable option for eliminating the confusion is to change - or, better yet by far - remove the language.) --Minasbeede 13:29, 14 September 2007 (UTC)

I understand what you are saying... and in the abstract ideal I agree... but unfortunately we are not dealing with an abstract ideal, but with real life... a lot of editors don't seem to understand that using such material in such a way is OR. We need to spell it out for them. I think it causes more confusion and argument when it isn't spelled out. Blueboar 13:42, 14 September 2007 (UTC)

Well, OK, it looks like you do agree that using the word "source" is part of the problem. I'll hazard a guess that the editors who don't understand have the attitude "I cited the sources of what I used: it's verifiable" and plain don't see (refuse to see) that if they are using the sources to draw a conclusion (etc.) they are doing OR. Maybe they say "I went back to an original source that is verifiable and it proves..." That's still OR. The OR is indicated by the words "proves", "shows", "indicates", "establishes", etc. The type of source used isn't material, it's the usage. It isn't that what they are doing is invalid that's the issue, it's that Wikipedia isn't the place to do it, valid or not. It might be that you or I would agree, in some few cases, that what they are doing would improve Wikipedia, were it allowed. It's still not what Wikipedia does. Given the huge breadth of human knowledge excluding (which should really only mean delaying until the improvement in knowledge appears elsewhere) new concepts is a small price to pay. Maybe someday Wikipedia will embrace new concepts a little more than it does now. We can wait.

I'm very willing to see altered language that does what it is claimed is needed and avoids the confusion the current language causes. I don't think it's likely to be productive (or successful) to require me (and others) to first agree the language needs to be there before any attempt at improving the language is initiated. I'm not going to pre-agree - but if you propose alternate language I might (very likely will) stay silent while the advocates of source typing work on it. It's probably the silence that confers the major benefit. --Minasbeede 14:33, 14 September 2007 (UTC)

I think a warning to focus on what the source says instead of what it "implies" or "proves" but doesn't say, would cover most of this. Jacob Haller 15:51, 14 September 2007 (UTC)

I think that many of us feel that we absolutely must end up with a policy that doesn't create a situation where anyone can be told "OR, because you're using a primary source", rather than "OR, because you're drawing novel/unsourced conclusions". SamBC(talk) 13:53, 14 September 2007 (UTC)

Well said. --Minasbeede 14:14, 14 September 2007 (UTC)

That I can agree with completely. Just to be pedantic... I will note that the policy (as currently written) never say using primary sources is OR... in fact it says exactly the opposite: "Primary sources that have been published by a reliable source may be used in Wikipedia, but only with care, because it is easy to misuse them." (bolding mine). In other words... it cautions the editor about misusing primary sources. That said, I do see how someone might focus on just the words "primary sources" and not actually read the policy... which could lead to misundertanding... and so I do agree that some re-wording is probably called for. With that in mind...

Let's see where we agree as a base line for acheiving this... since this thread is getting over long... I will start a new one below. Blueboar 14:41, 14 September 2007 (UTC)

Good idea (the coming below thread.) The policy also doesn't say that secondary sources can be used carelessly. In fact, I'd venture to say that every caution that is appropriate for the use of primary sources is appropriate for the use of secondary sources. This is talk, I could draw a conclusion, I'll leave that for the reader. --Minasbeede 14:47, 14 September 2007 (UTC)

Defining terms by interpreting quotations

There's a dispute at the article on the term "Anchor baby" concerning whether "common sense" interpretations of quotations that mention the term can be used to define the term. Some sources comment on the term and how it's used, and there's no question that those sources are allowed. It's the sources that use the term without commenting on it that are in question. My belief is that doing so is the equivalent of interpreting a primary source. Some other editors believe that we'll never find a source saying that the term is used in a certain way and that NPOV requires we include that usage, therefore it is aceptable to use our interpretations of the quotations. Do regulars here have any thoughts on this type of issue? ·:· Will Beback ·:· 21:08, 13 September 2007 (UTC)

Glancing at it as a non regular, the problem seems to be that editors are finding examples of the term being used, then categorising each usage as with or without "pejorative" or "negative characterization", or as purely descriptive. Looks very subjective to me, hence original research unless a reliable source uses such terms to describe the usage. At the least some caution would be needed, saying "what appears to be a pejorative characterization" for example, but that's still interpretation and in my opinion should be avoided. It might be appropriate to find notable uses and give them in sufficient context to allow the reader to make up their own mind, but secondary comment on what the usage means would be much better. What do others think? .. dave souza, talk 22:56, 13 September 2007 (UTC)

To clarify, it's an assymetrical situation. We have reliable sources that say, "Anchor baby is a pejorative term", and no one is proposing interpreting quotations to buttress that definition. What we don't have are any sources that say "Anchor baby is not alwys a pejorative term". Editors would like to add the latter assertion based on their interpretation of quotations as using the term neutrally. ·:· Will Beback ·:· 23:21, 13 September 2007 (UTC)

We have sources which use "anchor baby" in an apparently neutral, news-reporting context, but these sources don't come out and say they consider the term to be neutral. Some of us say such sources validly illustrate the expression being used neutrally and are appropriate for inclusion (indeed, that they must be included for the sake of NPOV). Others of us insist, in effect, that unless a source explicitly, literally says something like "'anchor baby' is not a pejorative term", or "some people use 'anchor baby' in a neutral context", it is utterly useless as evidence either way, since we have no right to infer what the source says about pejorativeness and would be violating WP:NOR if we were to attempt to do so. The "it's always pejorative" camp then says that since no sources (at least, none they recognize) can be found to support any other view, we have a duty to say that "anchor baby" is pejorative without qualification. The "it's sometimes neutral" camp complains that accepting only neutral sources that explicitly call themselves neutral is a standard that no real-world source can realistically be expected to meet, and that failing or refusing to acknowledge that some people are genuinely using the expression in a neutral manner is a violation of WP:NPOV. We've been stuck on this OR-vs.-NPOV argument for several weeks now, despite a CfR, with each side insisting its understanding of the Wikipedia core principles is correct, and without any signs of anyone being willing to budge. Richwales 23:42, 13 September 2007 (UTC)

Just looking at sources [4] and [6], they're presented as "without characterization" but both refer to the babies as a burdensome expense to the taxpayer, so that's questionable. The most you might say would be that negative implications of the term are not always clearly evident, probably it would be better to say that it's commonly used as a pejorative term without claiming that it's used in a neutral way or that it's always pejorative. Very much a matter of interpretation, so it seems to me that a source explicitly saying that it's not pejorative is needed. Just my opinion ... dave souza, talk 23:47, 13 September 2007 (UTC)

If there are no sources that refer to the term in non-pejorative way, we cannot say in the article "these sources did not use the term pejoratively" simply because we do not know that, and as such it would be OR if we say that. What we can say, though, is that these sources used the term to describe children born to immigrants. That's it. ≈ jossi ≈ (talk) 23:51, 13 September 2007 (UTC)

Agree, my suggestion is it's better to say that the term is used to describe children born to immigrants [link to sources] and that x and y describe the term as pejorative. If z is found saying it's a positive term, that can be added later. .. dave souza, talk 00:08, 14 September 2007 (UTC)

I think this is a classic case where Wikipedia struggles with rebuttal. Initially I assumed this referred to children who were offspring of sailors in foreign ports. If that were so then I could see a trail that would lead me to viewing the term as offensive slang. However, if it is simply a term derived from the sense of anchoring the family in the country, then it is a harder call. I can both see that it could be offensive (with the implication of unwanted child and so on) though in that sense I could see that although it might be insensitive, it is potentially not pejorative. So as an editor, having Googled around, I am sympathetic and I can see that the term is widely enough used that it is reasonable to conclude that there should not be an issue with qualifying the statement.

This is a case where the OR principle struggles as with any potentially offensive interpretation it may be that there is a motivation to document the problem, where there is no motivation to assert the acceptability: perhaps a strongly nationalistic Welshman might find the term British offensive, but who will go around documenting directly that calling a Welshman British is not offensive, though they will document that it is a legitimate attribution. You simply will not find the direct information, especially if the view is not held to be worthy of rebuttal due to its obviously extremist position.

My view is that in this case, there are good grounds for recognising that the policy is being used (with good faith) to assert a point of view. There is a conflict between NOR and NPOV, there has been a good faith attempt to justify the dispute and a sound editorial judgement would allow for a phrasing that recognised the contentiousness of the assertion that the term is pejorative and it should therefore be qualified. Spenny 09:18, 14 September 2007 (UTC)

I agree with much or all of the above. You can't say "the term is sometimes used neutrally". That would violate OR. What you can do is (try to) use NPOV wording such as "x and y describe the term as 'pejorative'". You do not have a duty to state "The term is pejorative". That is disputed, unverifiable and (arguably) apparently likely false. What is verifiable is that someone said it was pejorative, not that it (actually, always and everywhere) is pejorative. --Coppertwig 17:36, 14 September 2007 (UTC)

I was unware of this discussion. I am one of the principle parties involved in the dispute, along with two other editors above, at "Anchor Baby" and the one that agrees with the above assessments. I've tried to argue this case but to no avail. Some editors are completely insistant on declaring the term is absolutely pejorative even when presented with sources that might indicate otherwise. My main concern was and is NPOV for the article. Your input along with Spenny's would be helpful in clarifying the issue and I invite you into our discussion if you wish to join - if only to clarify the situation with WP:NOR as you see it. --Northmeister 00:23, 16 September 2007 (UTC)

Is this original research?

I created the article Blackburn, Aberdeenshire. It's a sizable commuter village for Aberdeen but if you read the article you'll see that I said that it has an industrial estate. I could not find a citation referring to this specific fact, but it is completely true. There are online sites for the companies in the industrial estate, but nothing actually mentions that Blackburn industrial estate exists. I may have violated NOR here, what do you think?--h i s ^{s p a c e} _{r e s e a r c h} 13:52, 14 September 2007 (UTC)

By strict interpretation of policy, yes, this is original research. Whether a strict interpretation is appropriate in this case, I'm not sure. SamBC(talk) 13:57, 14 September 2007 (UTC)

This seems like a good situation to apply WP:IAR... no harm arises from stating this fact, even if it is technically OR to do so. Blueboar 14:16, 14 September 2007 (UTC)

Isn't this really a verifiability issue? Seeing that there is an industrial estate is not the same thing as concluding there is an industrial estate. There's the "to a man with a hammer everything looks like a nail" phenomenon but that this page is about the "hammer" NOR doesn't make everything an NOR question. He saw it, he didn't cite a published source that says what he saw. It has flawed verifiability (again, by strict interpretation of the policy.) Probably somewhere there is a government document that indicates that the area in question is an industrial estate. I'd suggest that HisSpaceResearch put finding that on his list of priorities, of things to do. Very low on his list. --Minasbeede 14:25, 14 September 2007 (UTC)

Well, I just did a Google search for +aberdeen +"industrial estate" and got 353,000 hits. It shouldn't be to difficult to find a source that says it's there. wbfergus^Talk 14:41, 14 September 2007 (UTC)

It also is fairly apparent on Google Maps. As is a big piece of lint. --Minasbeede 14:50, 14 September 2007 (UTC)

What do we agree on?

We seem to be making a bit of progress ... So it might be helpful to see where we agree as a basis upon which to discuss things further.

To start: Can we agree that the intent of the Sources section is to caution users about misusing certain materials in a way that violates NOR? And can we agree that the policy should include such a caution? Blueboar 14:57, 14 September 2007 (UTC)

That's the stated intent. The issue is whether the policy need have that sort of caution. Up to now the wording has been a source of frequent discomfort (as shown by edits, edit wars, and discussion) and a source of confusion (in fact, as seen here. Perhaps the tendency to be confused lies in those who are confused but even with 100% consensus that such people ought to change they won't. The only option available for removal of the confusion is to change or remove the wording.) If the wording were moved to a guideline that would be a step forward - but the confusion-causing parts still would need repair. Whatever is done the confusion must end. If the wording is to remain (as some insist) then the wording has to be changed to eliminate the confusion. It would seem the ones who want the wording bear the greatest responsibility to improve the wording. Please do it. Please do not ask me to accept a pig in a poke: show me the best effort at good wording. What currently exists isn't acceptable. --Minasbeede 15:15, 14 September 2007 (UTC)

We need it for this reason: the sources can come from verifiable sources, and thus comply with V. And the view they represent (let's say, Mussolini's view of Fascism) can be identified, and thus comply with NPOV. But different sources can be used selectively to forward an editor's viewpoint that does not exist in any scholarly literature e.g. books or articles by histortians or political scientists. NOR fills what many people involved in edit wars have seen as loopholes in NPOV and V. the loophole has to do not with whether the sources are verifiable or not, but how they are used to provide what kinds of content to the article. Slrubenstein | Talk 15:22, 14 September 2007 (UTC)

Yeah, and it does that without enumeration of examples. New idea not sourced? OR. That's simple, that's adequate. NOR isn't (and needn't be) concerned with how OR is done. It's NOR that fills the "loopholes in NPOV and V." It does that without having to identify particular source types used to produce OR. It doesn't matter, with respect to OR, whether the viewpoint exists in scholarly literature or doesn't. If it's OR it isn't proper for Wikipedia. If a view is pertinent and does appear in other sources then the advocate of that view can introduce it if he does so in compliance with V and NPOV. If it doesn't, he can't, and it is not a matter of concern for NOR how the editor formulated the view nor how the editor backs the view. The editor can't introduce it: it's novel. It is the novelty that makes it improper, not how the editor backs the novelty. It is asserted and understood that all the policies have to be followed. That the evidence used to support a novel idea is verifiable doesn't suffice: the idea itself has to be attributable to a source to appear. That seems very clear. Adding source-typing language doesn't make it clearer. It is also obvious that source-typing is also being used to favor the notion that secondary sources are superior to primary sources, with both "secondary" and "primary" given Wikipedia-specific (in the details) definitions. --Minasbeede 17:49, 14 September 2007 (UTC)

I would agree with Blueboar's description of what the Sources section is about. I'd even go a little further to say that the purposes of the section are (1) to prohibit new, unpublished interpretations of (both primary and secondary?) sources, and (2) to prohibit the use of sources (again, both primary and secondary?) to imply original, unpublished ideas by the editor. Can we all at least agree on these two principles? COGDEN 15:54, 14 September 2007 (UTC)

That sounds like a step in the right direction, COgden, Slrubenstein | Talk 15:57, 14 September 2007 (UTC)

So far, so good. Jacob Haller 15:58, 14 September 2007 (UTC)

'Tain't necessary. NOR forbids OR. That's enough for a policy. --Minasbeede 17:59, 14 September 2007 (UTC)

I agree in principle so far as well. So far, this hasn't broached the subject of whether this policy should have yet again another definition of what the types of sources are, nor do I think it should. Keeping the policy (and the source section) simple and straightforward with how sources relate to OR and how they can be abused is sufficient. Maybe something along the lines of "to avoid the appearance of original research, cite (with reliable sources) any conclusions, interpretations, or theories presented"? wbfergus^Talk 16:27, 14 September 2007 (UTC)

OK... If I understand the comments, it seems that we have agreement that some sort of cautionary statement is useful and perhaps even needed. Now we move to areas where I expect disagreement will begin to crop up... but let's see where it happens. next two questions: 1) Can we agree that, in practice, it is common for editors who violate Cogden's two principals to do so by misusing what I call primary material (taken from any type of "source"). 2) Can we agree that it is more common for them to do so using material taken from what the policy now calls a "primary source"? Blueboar 16:43, 14 September 2007 (UTC)

(1) No. We have no way of knowing that. Satisfactory evidence of that has not been presented. In the body of Wikipedia that statement would be OR.

(2) No. Same reason.

This is begging the question. (For most of my life I didn't use that phrase and it still doesn't feel natural to me, but that's what it is.) You're saying assume what it is that we need to prove.

It hasn't been proven, it likely never will: the issue always gets ducked. Even if proven it does not constitute a reason to encumber a perfectly good NOR policy (without source-typing language) with source-typing language. Nobody is going to get away with something because the language isn't there. The language is superfluous, and it should be a major goal to keep policies clean and simple. Note that had the policy been kept clean and simple there'd not be all this discussion and the NOR page would not end up locked for 24 days (and counting.) --Minasbeede 17:56, 14 September 2007 (UTC)

Okay, maybe this should be on another page where there's room for all the discussion that's sure to follow this. I'd say that the source type doesn't matter, OR (and also POV), can be intentionally injected regardless of source used. It may happen more often through "good faith" edits unintentionally by editors using primary sources, but that's also another issue not for here. wbfergus^Talk 16:58, 14 September 2007 (UTC)

No disagreement from me on those points at this time. Good consensus-building work, Blueboar and others. :-) --Coppertwig 17:04, 14 September 2007 (UTC)

Me! Me! Me! (waves hand in the air). First off, I don't see how we can get consensus when using the primary/secondary terminology, because regardless of intent, it is ambiguous - I am not clear on what people are saying or reading. Do we mean the first appearance of a concept, or do we mean a general description of a source that is essentially analytical or data or whatever? Second, I simply do not get this fixation with being able to ascribe a particular behaviour to a type of source. My reason for saying this is that I do not think that the main offenders would recognise there actions as fitting this mode of conduct, (can your average Britney fan spell secondary?) and even if they did I don't think it is useful to explain it in this way. Recognise that many editors are not academics in general, so will not have that academic viewpoint (however much there may be a wish within some elements of the Wikipedia movement that it should really only be within the domain of the academic world). Policy pages shouldn't be for the mental stimulation of the cognoscenti but to guide Those Who Do Not Know Better.

I would agree with COGDEN's statement without the superfluous text in parentheses, but I don't think that adds anything that was in dispute anyway. Spenny 17:31, 14 September 2007 (UTC)

I would agree with #1 (basing original research on primary sources is common), however, I would disagree with #2 (it's more common than basing it on secondary sources). While #2 might technically be true, that's only because the use of primary sources is more common in general, for all purposes. When secondary sources are used, it's not clear to me that the probability of original research is either greater or less than the corresponding probability for primary sources. Even if it is, that doesn't assist us in drafting a policy, because there are other factors and policies to consider, such as the fact (it's clear to me at least) that the use of secondary sources is more likely to violate NPOV than the use of primary sources. COGDEN 16:26, 18 September 2007 (UTC)

We might, perhaps, be able to agree that what the kind of editor who is the source of all this might be doing is something that could be perfectly valid were the editor to do it in a scholarly journal. Wikipedia is not to be the place where new ideas first appear: that's what NOR is about. Wikipedia, through the NOR policy, excludes all such material without paying attention to whether it would or would not be valid in a scholarly journal. It might also be that the editor is doing something that no scholarly journal would ever publish. Excluding whatever it is the editor is doing from Wikipedia is unrelated to the quality of the material. If it's novel Wikipedia doesn't want it, Wikipedia is prejudiced against it solely because it is novel. It is not Wikipedia's business to analyze the quality of the material, to conclude whether or not it is a proper or improper use of primary (or any other) material. If the result is a novel idea Wikipedia rejects it, good or bad. --Minasbeede 18:11, 14 September 2007 (UTC)

We might, perhaps, be able to agree that what the kind of editor who is the source of all this might be doing is something that could be perfectly valid were the editor to do it in a scholarly journal. Uh? ≈ jossi ≈ (talk) 18:52, 14 September 2007 (UTC)

Translation: "Taking primary material and analysing it, interpreting it and drawing conclusions from it is something that is approved of if you are writing in a scholarly journal, but not here." Blueboar 19:19, 14 September 2007 (UTC)

Yes. That's what I meant. Thanks. --Minasbeede 19:47, 14 September 2007 (UTC)

What does "novel" mean?

Maybe this is the heart of the source-typing issue (if not, clue me in.) It would reasonable to me to accept that if a Wikipedia article says (X is Y) or says (X is not Z) that NPOV would make proper any verifiable evidence that indicates (X is not Y) or that (X is Z.) It's reasonable to me because once you categorize X in some way there is an implicit opposite category. The idea of the potential for the existence of that implicit opposite category is not novel. If it's asserted that John Doe is hairy then either all the evidence indicates that or it doesn't. The assertion creates the issue of whether or not John Doe is hairy (and if the assertion exists in Wikipedia we can assume that the topic is notable, else it shouldn't be there.)

But if a long line of secondary sources all say (X is Y) or (X is not Z) then there would be those who assert that the opposite view, if unsourced as a view, is OR. If Wikipedia claims that John Doe is hairy is it OR to produce evidence that shows John Doe to be essentially hairless even if there is no source that explicitly states "John Doe is essentially hairless"? Is it a "novel" idea that John Doe could be hairless when the statement has been made that he is hairy? That he is hairy is a POV. An editor finds verifiable evidence that counters that POV. He wants to put it into the article to counter the apparently erroneous assertion, as is proper under NPOV.

I understand that John Doe could have been hairy once and essentially hairless another time, I understand that "hairy" might mean, to some, "has some quantity of hair, no matter how small," I understand that there could even be confusion over which John Doe the statement was made about. All of those could be considered in deciding what the article should ultimately say. The question is whether saying John Doe is hairy creates the implicit concept "John Doe is not hairy" and since that concept exists it is not, for Wikipedia, a novel idea. Evidence for "not hairy" as applied to John Doe is fully legitimate within the Wikipedia policies of V, NPOV, if Wikipedia claims John Doe is hairy - even if there are dozens of secondary sources that make that claim. --Minasbeede 18:50, 14 September 2007 (UTC)

I do not know what is that you are arguing for here, Minasbeede. If your argument is that we should allow the presentation of evidence in a primary source to counteract a claim made in a secondary source, I do not think that it would be acceptable as it goes against the grain of NOR. ≈ jossi ≈ (talk) 18:56, 14 September 2007 (UTC)

I don't see how the primary part goes against NOR. I'd sum up my own understanding of NOR as: "report what the sources say, not what you (the editor) think they prove."

Minasbeede, I don't think any rules-set can cover everything. I'd start with disputed tags on the original assertion, and counterevidence on the talk page, and ask for sources contradicting the original assertion. You can also cite the counter-sources for what they do say. Jacob Haller 19:18, 14 September 2007 (UTC)

"It's reasonable to me because once you categorize X in some way there is an implicit opposite category. The idea of the potential for the existence of that implicit opposite category is not novel."... I very much disagree with this. You need to show that a source actually has placed X in some other category. To categorize X yourself is OR. Blueboar 19:54, 14 September 2007 (UTC)

I understand. Is this the heart of the source-typing issue? That's why I put it on an already over-long talk page: I think it may be. Someone has material that satisfies V and that contributes to NPOV by presenting evidence that counters an existing POV in an article. The evidence visibly counters what the article claims; the editor inserted it with a sincere belief that it contributed to NPOV (and perhaps the editor's POV corresponds to the POV the implication of the material that was inserted: it's not forbidden for an editor to include same, only forbidden to misrepresent any POV.)

Is this the heart of the source-typing issue? --Minasbeede 20:21, 14 September 2007 (UTC)

If this is the heart of the source typing issue... then there is fundamental misunderstanding of how the core policies work going on. V, NPOV, and NOR are all core policies. One does not "over rule" the other. If a statement or citation has a flaw with any one of them, the statement or citation may not be used. In this case... you may not put something that violates NOR into an article, just to satisfy NPOV. Blueboar 20:42, 14 September 2007 (UTC)

Well, you are expressing my question in a different way. Even if we all agree that NOR is violated in the example the question is, still, is this the heart of the source-typing issue? It appears to me that the edits that are supposed to be the historical reason for source typing being in the NOR policy is that someone took a verifiable primary source item and inserted it into an article that asserted something that was contradicted by the primary source item. I assume the editor who did this thought he was acting in full compliance with NPOV and V and since the editor was introducing evidence that related to a topic already in discussion it did not appear to be OR to the editor. (Again, the question isn't whether this is OR, the question is whether this example serves to illustrate the situation.) I fully understand that you may not put something that violates NOR into an article just to satisfy NPOV. I asked it this is the heart of the issue. That is, have I come up with a hypothetical situation that fairly well represents what this is all about? I know what many would say. I don't ask that, don't want to argue that. Is this an example that illustrates what is at issue? I could argue it much better than I have but arguing this isn't the point. Have I identified the cause (a cause) of the concern that led to the introduction of source typing or have I not? That is, have I, to use your words, illustrated the nature of the "fundamental misunderstanding of how the core policies work"? That's what I'm trying to: illustrate the nature of the problem. --Minasbeede 21:12, 14 September 2007 (UTC)

Sorry I took your question as rhetorical to mean that this was the heart of the debate for you. My mistake. I think this is a yet another reason why we need to have have the sources section (or something like it)... but it is hardly the only reason and probably not the key one. To me the example is fairly obviously OR even without the language we are debating. So no, I would not say this is the heart of the debate. At least I hope not. Blueboar 22:34, 14 September 2007 (UTC)

It seems like some of what has been said about the need for source typing arises from cases very similar to the example. I was trying to construct an example consistent with those statements.

I fear I have not been paying close enough attention so I can't figure out what the actual key reason is nor even what other reasons might be for wanting to have a sources section in the policy.

OK, I'm looking at Slrubenstein's Mussolini example. I can readily understand that (just like in physics) there could be a kook theory about Mussolini's view of fascism. I readily accept that. I find it harder to understand how someone can put forward an interpretation of a primary source (any source) without using a word like "therefore," "thus," etc. or a phrase like "this shows," "we can see," etc. (with that word or those words being a tip-off that the editor was inserting OR.) I can even see that maybe Mussolini had an "off" day and said something like "fascism is like a big plate of pasta" (which is a nonsense statement as far as I can determine.) I can see that an editor could quote the Mussolini "plate of pasta" statement to counter what historians generally agree was Mussolini's view of fascism - which would be nonsense, not in any way supported by NPOV. But I can also see that an historian could have an off day and that historian could say "Mussolini saw fascism as being not different from a plate of pasta." Still nonsense, although maybe in the context the historian might just have been using the "plate of pasta" in order to convey an idea: a weird metaphor. I can't see how such a statement is more likely from a primary source than from a secondary source (whatever those two words mean.) What am I not seeing? --Minasbeede 02:50, 15 September 2007 (UTC)

Novel basically means that completely mechanical rearrangement of information is permissible. Strictly speaking, unless you copy a source word for word, you need to do some minimum of understanding, comprehension and research. If I can dig up sources that Johnny Nobody was born in Cochrane, Ontario in 1971, I can say he was born in Canada even if I cannot find a source that says "Johnny Nobody was born in Canada". Similarly, if my source says "Alpha Centauri is 1.3 parsecs from Earth", it is not original research to say "Alpha Centauri is 4 lightyears from Earth", which I might want to do for style reasons or something. Or that completely trivial, unavoidable synthesis is acceptable. If I can source that Johnny Nobody is a physicist, I can reasonably put him into Category:Physicists from Canada, even if I can only source those two facts independantly. Don't create any new knowledge, don't draw any new conclusions and its not original research. Wily D 20:36, 14 September 2007 (UTC)

An illustration of the proper vs improper use of a primary source...

Suppose I were to write an article on "Assault weapons" (there is one, but for my illustration assume that I am writing it fresh)... In that article I want to discuss the different laws about assault weapons around the world. Now suppose I write:

The Second Amendment to the United States Constitution states: "A well regulated militia being necessary to the security of a free State, the right of the People to keep and bear arms shall not be infringed."<cite to US Constitution> Noted legal scholar, I.M.Pompus has stated that this guarentees the right of all US citizens to own Assault Weapons.<cite legal journal where I.M.Pompus saying so>

This is a proper use of a citation to a primary source document - the US Constitution. No OR is involved because I am using it simply to reference itself. The interpretive statement (that the Constitution guarantees the right to own assault weapons) is attributed to a different source - one external to Wikipedia.

But if I were to write:

The right of US Citizens to own assault weapons is guaranteed by the Second Amendment to the Constitution which states: "A well regulated militia being necessary to the security of a free State, the right of the People to keep and bear arms shall not be infringed."<cite to US Constitution>

This would be an improper use of the citation to the primary document. I am interpreting the Constitution for myself, and concluding that it applies to assault weapons. It may or it may not... but, as a Wikipedia editor, I can not make that determination in my article. It is OR to do so.

Unfortunately, I see far too much of the second type of useage of citation. This is the kind of thing that the Sources secton is trying to prevent. I hope everyone would agree that we want to discourage such usages. However, because the primary source (the constitution) can be used appropriately (as per the first usage), we are getting caught up in semantics and debates over the terms like "Primary" and "source". Blueboar 20:34, 14 September 2007 (UTC)

"A government inquiry researching the issue for the first time found that many people thought that it was a constitutional right to own an assault weapon." "Sorry you can't cite that, its a primary source for that research." (An illustration of an improper use of primary source policy) This is the sort of thing such a policy encourages. Spenny 21:46, 14 September 2007 (UTC)

I would say that this depends on whether the inquiry actually includes the conclusionary statement or not. If the inquiry does not actually say that many people thought this about assault weapons, but just included raw data that the wikipedia editor writing the article interpreted as saying that... then the the person saying "sorry" is correct. On the other hand, if inquiry did contain that conclusion, then the person saying "sorry" misunderstands what this policy actually says (even in the current version) ... At no point does the NOR policy say you can never use a primary source... in fact it says explicitly that you CAN use them. However, it says you should be careful when using a primary source and that you should not use it for analytical, interpretive, or conclusionary statements. There is no such beast as the "Primary source policy"... what there is is a section of the OR policy that cautions the use of primary sources. Your example does not contain any OR. Simply a statement of fact (assuming the inquiry actually concluded that many people found this to be true) backed with a reliable source. Now... it you went on to make further statements about the inquiry, interpreted what it said or drew any conclusion from it, we would be back to OR.Blueboar 22:24, 14 September 2007 (UTC)

If the primary source contains "analytical, interpretive, or conclusionary statements" then it's OK to use it for those statements. I think it would be better to go back to your idea of talking about primary and secondary "material" rather than "sources", or as I put it earlier, "sources of facts" and "sources of interpretations".

(inserted comment: I thought this is what I was saying. :>)Blueboar 23:34, 14 September 2007 (UTC))

In the good example by Blueboar above, the primary and secondary sources are being used in exactly the same way: that is, the Wikipedia article is stating pretty much exactly what is stated in the source. In the bad example, the Wikipedia article states something different from the primary source. One could also (but should not, of course) do that with the secondary source, like this: "Commentator I.M. Pompus supports the right of individuals to own nuclear weapons when he states, 'the U.S. Constitution guarantees the right of all citizens to own assault weapons.'" As Minasbeede points out, it's wrong to go further than the source, regardless of the type of source. --Coppertwig 22:41, 14 September 2007 (UTC)

This example is not about original research. This is about WP:V and WP:NPOV. That the second amendment can be interpreted to guarantee the right to own assault rifles is not a novel idea. It's not original research, and presumably that point could be cited to some source such as the NRA. The use of the Second Amendment is not the problem here. The problem is failing to attribute the interpretation of the Second Amendment to a verifiable source (such as the NRA), and the fact that the interpretation is presented non-neutrally. COGDEN 16:32, 18 September 2007 (UTC)

In Spenny's example, if the government survey asked the question "Does the Constitution guarantee the right to own assault weapons?" or anything like that ("...rifles, handguns, and even assault weapons?") and the answers showed that 55% of those surveyed said "yes" it's hard to accept that an editor can't represent that 55% as "many people." Even if only the raw data were presented and (for simplicity with the numbers) 55,000 said "yes," 40,000 said "no," and 5000 said "don't know" it would still seem that an editor could say "many people" had the belief. I recognize that survey methodology can be attacked but the editor didn't design the survey, he's just reporting what it said, and "many people" does not seem to misrepresent the results of the survey. The editor has engaged in no "original research" by any reasonable interpretation of the term. If the editor had said "an overwhelming majority" that's a misrepresentation. If the editor had said that "many people, well-versed in the principles of the Constitution and vigorously defending their rights under it" the editor would clearly be going beyond the content of the source and adding personal opinion. There's also nothing about the example that requires the editor to have found those numbers in the government survey itself. The editor might have found the numbers in a secondary source. (If the editor extracted survey results from a secondary source there is the possibility that the secondary source engaged in "original research" in its presentation. That seems to be why historians prefer going to primary sources.)

If the question asked about "personal firearms" it would probably misrepresent the survey to say the people believed the Constitution granted a right to own assault weapons. Assault weapons may be personal firearms but the question was more general. It's a misrepresentation to claim that the opinion of the many about "personal firearms" was meant by the respondents to include a particular class of "personal firearms."

I agree with Spenny's point. If the survey asked about "assault weapons" and can be seen to have produced answers consistent with "many people" then I think it is fully proper to include that material even though the source doesn't have the precise words "many people thought that it was a constitutional right to own an assault weapon." It is not original research to report, using common mathematical techniques or different words with a compatible meaning, what the report actually said. I think Spenny's point is that some over-literal individuals would attempt to mischaracterize a simple, valid, straightforward restating of the reports content as forbidden "original research." I share his concern, and I think that there has been scant attention to this point of view on the part of those who want to retain source (material) typing. I see no wall to prevent such over-reaction, I may even sense a desire to institutionalize such over-reaction. This is a legitimate concern and it largely is being ignored, which is improper in a quest for consensus. --Minasbeede 16:00, 15 September 2007 (UTC)

My point was intended to be simple. Let us say that my cited quote was not a misrepresentation of this government investigation. We look at the source and find it is actually a) a summary of other research pieces, or b) government commissioned "market research". (My point was to say it was the first time that this conclusion had been drawn, and it was correctly paraphrased). In either case, it is an analysis which draws a conclusion.

There seems to be a movement afoot which has got confused and wishes that any source used has to be first filtered through a further level of analysis, even when the source is solid. The problem gets worse when we allow journalistic interpretations to override reporting of solid, original material. Clearly, we can then get into derived sources who use such research to draw their own biased conclusions, which we can then attempt to cite to force our own point of view. e.g. we might find that a reliable but conservative press cite the research as saying that people are in favour of assault weapons, a fairly typical statistical misrepresentation, and a fairly typical reaction of the press to government published material, cherry picking for the interesting, controversial bits. We see that this secondary exposition quotes the survey and we go to the survey and see that it is a misrepresentation. 1) The original source is a valid analysis, drawing conclusions, 2) the derived source is an invalid journalistic misrepresentation but in say the Wall Street Journal which for our purposes we will assume is treated as a reliable publication (within the bounds of journalistic competence). Which is the source that should be discounted? Which supports NPOV? Does classification of sources help us make a valid judgment on the matter? Are we, in our editing, allowed to think, check and do research based on these sources to validate them or are we expected to blindly work with what we are given? Why do we think that a further level of review adds quality rather than being an opportunity to distort? Again, don't think academic world, think about the wider sources for the wider encyclopaedic content.Spenny 19:38, 15 September 2007 (UTC)

You ask: "Does classification of sources help us make a valid judgment on the matter?" That's a very good question.

Elsewhere I complain that rather than requiring a layer of filtering the source typing appears to disfavor raw data and to accept analysis and conclusion with hardly a look (by comparison.) That's for analysis and conclusion as it appears in original scientific publications: the analysis and conclusions are promoted, in the Wikipedia definitions, to privileged "secondary material" status. The result is the exact reverse of what I was taught, which was that the data probably are reliable but that the conclusions need be regarded very skeptically until they've had a chance to be given a thorough checking out.

I'm quite willing to lay silent for a while and let the proponents of source typing work all this out, if they will. I suspect that some of the proponents don't really have a full grasp of the nature of scientific publication, and I attach no onus to that - it's just an observation and explanation. --Minasbeede 21:45, 15 September 2007 (UTC)

Blueboar, you had said, "However, it says you should be careful when using a primary source and that you should not use it for analytical, interpretive, or conclusionary statements." That's what I was referring to. Less importantly, this does use the word "source" rather than "material". More importantly, if people continue to pop up from time to time to make definitive statements on this talk page that peer-reviewed scientific articles are "primary sources", and if the policy does not explicitly state that that is not what it means, and if the policy says what you just said in that quote that it says, then this would mean we couldn't report the main conclusions of peer-reviewed articles -- which would mean huge amounts of information would be deleted from Wikipedia articles. Which is not what we want. --Coppertwig 21:22, 15 September 2007 (UTC)

It's also a bit weird. The caution is against using raw data: conclusions can be passed through with hardly a thought (by comparison: no caution.) While scientific papers do contain conclusions and are refereed that's not by any stretch a strong assurance that the conclusions are valid: they may appear very early in the course of a research project: "publish or perish." Future research (and publications) may confirm the original conclusions or modify or overthrow them. But the caution is issued against using, as I understand "primary material," the raw data. Conclusions are let by with scarcely a look. I think that W every editor needs to use care in everything included in an article and that such a caution is at least as needed for the conclusions in primary articles (that in Wikipedia are promoted to "secondary material" status.) This looks a lot like an attempt to find a way to back-justify source typing: make the definitions fit the intent, whatever it is. It does not appear that the source-typing language grew out of the nature of the material - unless the policy was originally conceived within the narrow and particular meanings of "primary" and "secondary" in history (although I still really do not know what the limits are there insofar as Wikipedia policy is concerned.) --Minasbeede 21:36, 15 September 2007 (UTC)

What if we just swaped words?

Since people like my phrase "primary material"... I thought I would see what the current section would look like if all we did was substitute "material" for "source" (with a minimum of tweeking of other bits to make it make sense). This isn't really a proposal... but take a read and tell me if this is more along the lines of what we want to say in this policy: Blueboar 23:28, 14 September 2007 (UTC)

Primary, secondary, and tertiary materials

Research that consists of collecting and organizing material from existing sources within the provisions of this and other content policies is encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia. However, care should be taken not to "go beyond" the sources or use the material in them in novel ways. Such material may be divided into three categories:

Primary materials are documents and data that do not contain analytical, interpretive or conclusionary statements. An eyewitness account of a traffic accident is primary material. United Nations Security Council resolutions are primary materials. Primary materials that have been published within a reliable source may be used in Wikipedia, but great care must be taken when doing so, because it is easy to misuse such material. Any interpretation of primary material requires citation to a source that has done so. Examples of primary materials include archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.

Secondary materials are those that draw on primary materals to make generalizations or interpretive, analytical, concusions or synthetic claims. A journalist's story about a traffic accident or about a Security Council resolution is a secondary source, assuming the journalist was not personally involved in either. An historian's interpretation of the decline of the Roman Empire, or analysis of the historical Jesus, are secondary materials. Wikipedia articles should rely on reliable, published secondary materials for any statements of analysis, interpretation or conclusion.

Tertiary materials are publications such as encyclopedia articles that sum up other secondary materials, and sometimes primary materials. (Wikipedia itself would fall into this category.) Some tertiary materials are more reliable than others - some articles may be more reliable than others. For example, articles signed by experts in Encyclopaedia Britannica and encyclopedias of similar quality can be regarded as reliable secondary materials instead of tertiary ones. Unsigned articles may be less reliable, but they may be used so long as the encyclopedia is a high quality one.

A Wikipedia article or section of an article that relies on primary materials should (1) only make descriptive claims, the accuracy of which is easily verifiable by any reasonable, educated person without specialist knowledge, and (2) make no analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on primary materials should be careful to comply with both conditions.

This is a good step forward. I'd hope to see more work along these lines. Of course, as noted below, a major primary component of scientific articles consists of theories advanced or conclusions reached. It is recognized within science that these may be the most tenuous parts of a scientific publication: some theories and conclusions survive further research, some don't. The original publication of a theory is the primary source (it can happen that in time the "primary source" cited is a restatement or expansion of the original idea) for it and one of the principle motivators of scientists is to be the author (or one of the authors) of a paper that does advance knowledge, does present a novel idea: true "original research." You can't say make a blanket statement that primary material in scientific journals consists only of data. It's untrue, it amounts to trying to twist reality to fit the desired wording in the NOR policy. That is objectionable, both because it grossly misrepresents the nature of scientific publication and because doing such a misrepresentation masks the deep flaws in source- (or material-) typing. "Primary" can very definitively include analysis and conclusions. Such do not appear only in "secondary" sources. --Minasbeede 16:17, 15 September 2007 (UTC)

Which is why I got rid of the word "sources"... For scientific topics, the raw data compiled by a scientist is the primary material... the scientist's analysis of it and the conclusions he/she reached (no matter whether they are "original ideas" or not) is secondary material - stuff that is based on the primary material. What we want the policy to make clear is that an editor can not quote the data and then insert his own interpretation of that data... editors needs to cite to a reliable source that interpreted it in the manner described. This might be the same source that contained the raw data... or it might be another source that used the same data to make a different conclusion. The point is that any conclusion that is stated in a Wikipedia article was made by someone else rather than the wikipedia editor. Blueboar 17:29, 15 September 2007 (UTC)

Which I appreciate and for which I thank you. I advocated a change in terminology (to get rid of the confusion, at least) over two weeks ago. You've done that. I still think the entire material-typing section is unneeded but I am content for there to be continued steps toward improving it. I disagree with something in your paragraph above but it would quite likely be far more productive for you to continue in efforts to improve the section and not have to bother at this time with that disagreement. (I consider the issue to be deferred, not abandoned.) --Minasbeede 19:57, 15 September 2007 (UTC)

Would you be willing to identify what you disagree with (if you are worried that it will distract us, you can do so "privately" at my talk page... I have no problem with deferring discussions, but it might not be something I feel strongly about... and if not, I may be able to address it before we have to discuss it. Blueboar 20:04, 15 September 2007 (UTC)

If we're going to continue working with the distinctions, I disagree with making a distinction between secondary and tertiary. The line is rarely clear and some tertiary sources (such as textbooks) are among the most reliable sources available. Vassyana 21:09, 15 September 2007 (UTC)

I don't think swapping "sources" with "materials" helps. Also, there are very, very few published "materials" that do not contain conclusions or analysis. Why would it be published (i.e., be a "reliable source") if the author or publisher didn't say something new and original about it? Thus, the new category "primary materials" would effectively just be anything that isn't a reliable source. Besides, since we have a policy called WP:Reliable sources, we should probably use the same word, "sources", to avoid confusion.

Plus, when it's defined this way, the policy has nothing to do with original research. It may or may not be a good policy, but it doesn't address original research, which is the whole point of having it here. COGDEN 16:55, 18 September 2007 (UTC)

The famous(?) "OMG WTF, this isn't what I learned" argument

I'd like to see some sources for this division of sources.

When I was a first-years in university, when studying "use of libraries and sources", I learned:

Primary source: Article in peer reviewed journal (Is that what's called a secondary source here?)
Secondary source: "Review article" in peer reviewed journal, doesn't provide new information, merely summarizes a number of primary sources. (is that a secondary or tertiary source here?)
Tertiary source: A textbook or other large document, putting together information from many Articles and Review Articles. (what'd this be called on wikipedia, a *quaternary* source?)

I'm somewhat puzzled that scientific sources aren't even *mentioned* in the above discussion of sources.

Maybe I'm just a Science-biased crazy-type person, but like, can someone explain to me?

And how can we pretend to be a reliable encyclopedia on scientific subjects if we don't know how to deal with scientific sources?

Or am I just stupid, and has this been handled someplace way back in the archives? (can someone point me to the relevant archive?)

--Kim Bruning 01:56, 15 September 2007 (UTC)

Kim, Yup, you are a nerdy Science-based crazy-type person... :>) I have a feeling that the policy section was mostly written by humanities majors... the definitions fit fine with what we real scholars learned in university. :>)... seriously, in the humanities, both peer reviewed articles and "review articles" would be considered secondary - primary is reserved to the original documents and data used in writing the secondary stuff. In any case, I think we are trying to get rid of the word "sources" to avoid exactly this kind of confusion. Does primary "material" etc. work better for you? Blueboar 02:16, 15 September 2007 (UTC)

Another vote for "nerdy science-based crazy-type person". As a mathematician, I get to just make stuff up and publish it in peer-reviewed journals, so it counts as a primary source. I don't think that historians, for example, have that luxury. --Ramsey2006 09:09, 15 September 2007 (UTC)

Not good historians at least. Blueboar 13:32, 15 September 2007 (UTC)

An article in a peer reviewed journal is a primary source if it's publishing original data (as opposed to ideas) - as in most "hard" sciences. This is a primary source in Wikipedia, too, because these kinds of articles are essentially "written or recorded notes of laboratory and field experiments or observations". An article in a peer reviewed journal of one of the "soft" sciences like history is generally a secondary source. I agree the wording Wikipedia is using isn't entirely clear, but I believe it does follow the conventions in both "hard" and "soft"-type science disciplines. Lyrl^Talk _C 15:42, 15 September 2007 (UTC)

The point of a scientific article is quite often that the data presented lend support to a particular idea that is being presented for the first time in that article: the researcher isn't engaged merely in a quest to find new raw data, he wants the data to enable a better understanding of some facet of nature. The article is also a primary source for the idea, not simply for the data. If in time that idea becomes established and if it is important then that article will be cited as the original place of publication of the idea. Some scientific articles do simply report previously unknown results: X-ray crystallographic articles often being a good example of that kind of article. What is reported is simply the locations of the atoms in a crystal. (Of course a crystallographer can collect a lot of bond length data from such articles and formulate a general theory about chemical bonding, particularly if that crystallographer's name is Linus Pauling.) --Minasbeede 15:52, 15 September 2007 (UTC)

I think there is some confusion in the use of the word "Article".... to make sure we are all on the same page... in the NOR policy, this word refers to Wikipedia articles... not articles that appear in Scientific, or other academic journals. Presenting original ideas in journal articles is common... in fact it is almost mandatory. What this policy is saying is that Wikipedia articles should not contain the original ideas of Wikipedia editors. Talking about, and citing, the original ideas of those writing in academic journals is fine... the idea did not originate with Wikipedia. Blueboar 17:40, 15 September 2007 (UTC)

The point is that primary sources do and can contain analysis and conclusions. It's deceptive to claim that there's a neat line between "primary" and "secondary" by which "primary" means only data and "secondary" means/includes all analysis and conclusions. That's not so, not in science. The motive behind scientific publication is to get original ideas accepted. The primary sources (material) include analysis and conclusions. Those are the authors analysis and conclusions and not the Wikipedia editor's analysis and conclusions. That means that both primary sources and secondary sources do contain analysis and conclusions. You can't diminish scientific publication (journal articles) to being merely the publication of data and observations. The interpretation is a vital part of the publication. In short, it's just plain false to claim "primary" excludes "analysis." Any conclusion based on that claim is based on a false premise. --Minasbeede 20:09, 15 September 2007 (UTC)

Well... in the humanities, primary sources rarely include interpretive statements... but you certainly have a point in reguards to other study areas. And from the comments, I don't think anyone has disagreed with you. Again, the problem is with the word "source" and less with the words "primary" and "secondary". Blueboar 20:51, 15 September 2007 (UTC)

Lyrl said "An article in a peer reviewed journal is a primary source if it's publishing original data (as opposed to ideas) - as in most "hard" sciences." A number of other people have also asserted on this talk page that scientific articles in peer-reviewed journals are primary sources. However, peer-reviewed scientific articles practically always (in my experience) include discussion, interpretation and conclusions, and usually a summary of previous studies in the same topic, as well as data. More importantly, peer-reviewed scientific articles are an important and often-used source for Wikipedia articles. Therefore, the policy needs to either explicitly make clear that when it says "primary source" it does not include the typical peer-reviewed scientific article, or else avoid mandating that primary sources are to be used in a manner which is "rare" or which does not involve "relying" on them, etc. --Coppertwig 21:13, 15 September 2007 (UTC)

I think the second suggestion is the only way we can go here, because defining "primary source" in a non-standard, Wikipedia-only way that excludes peer-reviewed journals would only lead to confusion. We should use the terms the way they are used in academia, or not use them at all. COGDEN 17:04, 18 September 2007 (UTC)

Scientific papers are mostly standardized the world over, and they always contains analysis. I did a quick google for "structure of a scientific article", and the first random hit I checked was already pretty decent: [5]

Since these documents are made to exacting standards, I tend to trust them quite a lot more as a source than practically any other form of document. Especially once they are peer reviewed. --Kim Bruning 21:33, 15 September 2007 (UTC)

I think you've got the definintions exactly right, and the definition is the same in the humanities, as well, except in addition to peer-reviewed journal articles, primary sources include other sources where original ideas and data are found, such as biographies, diaries, interviews, newspaper editorials, etc. COGDEN 17:04, 18 September 2007 (UTC)

In the study of history, historical sources = primary sources and original ideas published in journals =/= primary sources. I'm sure those who have studied humanities in university have a slight bit more of a clue than you do on that count. Vassyana 19:39, 18 September 2007 (UTC)

I do not think there is any question that scientific articles published in peer-reviewed journals are - according to the current definition in the policy - secondary sources. It is also true that they may present data i.e. segments of primary source material. This is not unusual: a history article may quote from a journal or archive and in that sense contain primary source material. A sociological or anthropological article may contain within it transcription of an interview, which can serve as primary source material. The distinction between primary and secondry (or data an analysis) however remains important for both Wikipedians and academics. It is important for academics because it is always possible for scholar B to publish an article drawing on the data in another article by scholar A, but to make an orignal analysis. This is acceptable in academic publishing and indeed a good scholar enables it by providing segments of raw data in his or her book or article. We on the other hand have to discourage drawing on that material to make original arguments. But whether one is encouraging or discouraging a new analysis, the distinction between the data and the analysis remains. The problem here is that some people seem to think that these are mutually exclusive categories of texts. They are not. They need not be. indeed, it may not be possible to come up with definitions that are both meaningful and mutually exclusive. The distinctions lie in how the texts are used, not what the texts "are." Slrubenstein | Talk 19:55, 18 September 2007 (UTC)

OK, how about these questions?

A. Is the typing of sources that is advocated

(1) intended to be for the education of editors before they edit

or

(2) intended to be used by other editors to challenge material inserted by other editors alert other editors to a problem and correct material that is OR?

B. Why must this be part of a policy rather than part of a guideline?

--Minasbeede 02:30, 15 September 2007 (UTC)

Speaking solely from my own personal experience, I first ran across Wikipedia several years ago when a Google Alert showed there was an article here on my old Army unit. I read the article and saw that it needed some mojor work, and started editing it. I becam involved in correcting it and several other articles as well, and then finding additional sources to back up what I'd written. Fast forwarding about two years, one of the things I'd written was criticized by another editor as OR. Not knowing what that meant, I came here. So, regarding "A", in my case it was definately "(2)". Regarding "B", I fail to see why anything more than a brief explanation of how sources (actually their misuse) pertain to this policy should be included. wbfergus^Talk 11:36, 15 September 2007 (UTC)

Thank you. --Minasbeede 14:09, 15 September 2007 (UTC)

As for A: Both... ideally (1) but too often (2) (because too many editors don't read our policies before they start writing)... I would quibble with your language... instead of "challenge" I would say "alert other editors to a problem and correct material that is OR". As to B: Because using such material incorrectly is a far too common form of OR - it makes perfect sense to discuss the problems of using such material in the NOR policy. As I have stated before, many editors don't realise that using primary materials incorrectly is a form of OR... we have to spell it out for them in the Policy. We don't have to be as long winded as we are in doing so... but it should be part of policy. Blueboar 13:30, 15 September 2007 (UTC)

Thank you. Wording changed per your suggestion. --Minasbeede 14:09, 15 September 2007 (UTC)

Math - where is the line between paraphrasing and synthesizing/interpreting?

I see it's been recently discussed by others (#Math calculations) which math calculations are allowed and which are not. It would be nice if a section on this topic was added to the article.

Some of the numbers I work with in birth control articles may need to be altered to fit the context of the article - a 2% failure rate is the same as a 98% effectiveness rate, for example. I don't think anyone would consider this type of calculation as OR; personally, I think it's more like paraphrasing.

Some of the contraception articles source the prevalence of the method in the United States to numbers from the National Survey of Family Growth. Table 60 (on p.111) shows that 27% of women age 15-44 using contraception are using female sterilization, 18% condoms, etc. The less popular methods were grouped into "other" in Table 60. Their prevalence rates can be calculated by taking the number from the total population from Table 58, and dividing by the number of the population using contraception (from Table 59, this is 61.9%). Is this original research? I would tend toward saying no, since it's just using the same calculation the paper authors used for table 60, and is using the same source data the paper authors used (another table in the same publication on the same survey).

But this use was challenged, and there doesn't seem to be an entirely relevant section on this policy page. The person who challenged cited the synthesis section, but the source is in relation to the topic at hand and the calculation is not being used to advance a position. The topic of mathematical calculations is obviously relevant to this policy, but I'm not sure the current sections provide sufficient guidance as to which calculations are allowed and which are not.

I feel a little silly not being able to interpret the policy page and having to ask on the talk page to see if edits I'd like to make follow policy (or not). In addition to answering my question above, I would appreciate opinions on adding a new policy section along the lines of this:

Trivial and obvious mathematical calculations are not original research, and may be used if they improve the encyclopedic value of an article. However, calculations that use data from multiple sources, that require specialist knowledge to understand, or that are used to support original ideas or particular positions must be directly sourced and not calculated by a Wikipedia editor.

Lyrl^Talk _C 16:43, 15 September 2007 (UTC)

You can say that and it's not all bad but for actual mathematics articles there may be steps in proofs that are the editor's own. I don't know, am probably unskilled enough that I'll never know. I'd not dream of flagging such an article as possibly containing unsourced statements, or steps, or whatever. Specialist knowledge is definitely required no matter what the source for the steps of a proof, etc. but I think such articles shine brightly in Wikipedia and are a justifiable source of pride in Wikipidia. There is a line between reasonable and justifiable computation or mathematical manipulation and improper computation or manipulation but I think it's very hard to draw that line precisely. What you say is very reasonable and might be good enough for the policy. Along with IAR (ignore all rules) there may be an implied "don't slavishly follow rules." "Be bold" (which is part of the Wikipedia system.) The focus should be on the product and on whether it is or isn't appropriate and not on the rules. The rules exist to help make the product be appropriate and to ease the process of making it be so. What you suggest looks like it could work, provide it's not taken as an iron law. --Minasbeede 17:00, 15 September 2007 (UTC)

I would say that re-stating numbers through relatively simple calculation is fine... 2% water is the same as 98% not water... however, to draw a conclusion from that calculation (saying 98% not water means 98% some other specific element) without a reliable source to back it up does equate to OR. Blueboar 17:48, 15 September 2007 (UTC)

I would also like to thank you for posting your specific example here. If you, as another relative "newbie", are confused by, or don't understand the policy, then it does help show that the policy needs better wording and/or clarification. Many of us have been debating this for around three weeks or so now without reaching any definitive concensus, though we are making some progress. wbfergus^Talk 17:58, 15 September 2007 (UTC)

statements concluding: "No conclusion can be made"

Here is an interesting issue. I am currently working up some material on the origins of the fraternal group know as the Freemasons (doing so off wiki at the moment - I'm still in the source gathering stage - so I don't have an article to link to for you.) The issue is this... there are a bunch of competing theories about where the Freemasons came from... some scholars say the fraterinty grew out of the Medieval stone mason's guilds, others that it developed from the medieval Knights Templar, some even place it's origins as far back as ancient Egypt or the Celtic Druids. In my article, I am attempting to discuss all of these theories seriously, making note of who says what, the popularity and reputation each theory, and what the pros and cons of each theory are. For each one, I have reliable sources that demonstrate its flaws and problems. But now comes the potentially OR part... I would like to finish up my article by saying something along the lines of "All of these theories have gaps, and none have been definitively proven to be true. For the moment the origins of Freemasonry remain "unknown"." I would be happier if I had a source that actually said this... but most of the reliable sources focus on either "proving" one theory or another, or "disproving" one theory or another... I don't think anyone has ever actually said that none of them are definitive. It is a logical conclusion when you examine all of the scholarship, but one that I would say is OR for me to make. However... the policy currently is silent on stating a non-conclusion (unless you call it a conclusion of "No Conclusion"). Would you agree that adding a conclusion of "no conclusion" is a form of OR, or do you feel that such logical comments are ok? (and should we add something more definitive about this to the policy, or is it already covered?) Blueboar 19:40, 15 September 2007 (UTC)

Bit late for the brain to work, but would it not be better to conclude that "none of the theories has gained widespread or general acceptance, leaving the question still open" or something on those lines? ... dave souza, talk 22:56, 15 September 2007 (UTC)

"No conclusion can be made" is incorrect as many people do in fact draw a conclusion. "Opinions differ" and the like is not OR as the article properly sources just that. See Existence of God for a good example. WAS 4.250 00:31, 16 September 2007 (UTC)

If you actually wish to make a statement about the state of consensus and agreement on an issue, and you can support that, go for it. If it's just throat clearing or wishy-washiness, no, that's not encyclopedic. The point in saying that opinions differ on the existence of God is to highlight the nature of differing beliefs, not to assert as an encyclopedia that God may or may not exist. Wikidemo 00:41, 16 September 2007 (UTC)

Thanks, this matches what I thought... but it is good to have that opinion affirmed. I will either have to find a source that says it, or not say it at all. Blueboar 11:57, 16 September 2007 (UTC)

Nothing wrong with saying something like what dave souza suggested. You just can't say that it isn't proven, since no doubt in some peoples' opinions it is proven. --Coppertwig 16:56, 16 September 2007 (UTC)

Moving the discussion forward

Are we agreed that what the policy should be talking about is material... and not entire sources. Such material might be the entire source (such as a court transcript or a historical document) or it might be part of a source (such as the data section of a scientific research paper) ... but what is OR is the misuse of the material? Blueboar 21:03, 15 September 2007 (UTC)

It's a step in the right direction. ~~Go for it.~~ Be bold. --Minasbeede 21:24, 15 September 2007 (UTC)

If others agree as well... can we please stop spending time arguing on and on about the use of "Primary Source" vs. "Secondary Source" and what they mean? And can we stop beating a dead horse over the point that some sources contain both primary and secondary materials. Let's agree to change the policy to talk about "material" instead. Now... Are we agreed that some types of materials are "primary" and some "secondary" (or "tertiary")? Blueboar 22:33, 15 September 2007 (UTC)

Suits me. .. dave souza, talk 22:51, 15 September 2007 (UTC)

It would be a positive step in the right direction. Vassyana 23:12, 15 September 2007 (UTC)

I think so as well. ≈ jossi ≈ (talk) 23:49, 15 September 2007 (UTC)

I'm not sure I fully agree. I do agree that sources may contain material that is either primary or secondary, or both. And I do agree that original research may occur when an editor makes conclusions based on material in a source, and those conclusions are not also found in that source or another source. However, the OR does not arise because of the material. The material itself is what it is, and if you cite it for what it is, you cannot commit original research. Where original research occurs is when an editor cites a source, or primary or secondary material within a source, and implies that the cited source supports their new, unpublished conclusions. COGDEN 17:18, 18 September 2007 (UTC)

How about just saying what we mean? Say "uninterpreted raw data" versus "interpretation of that raw data" instead of "primary" versus "secondary"? As in: Uninterpreted raw data is to be used sparingly and always in connection with properly sourced interpretation of that raw data. Any interpretation of the raw data that is not actually in any of the sources cited is called "original research" at Wikipedia and is not acceptable content for a Wikipedia article. WAS 4.250 00:43, 16 September 2007 (UTC)

I like it but suggest some changes/concerns:

"used carefully" instead of "used sparingly"
"always in connection..." poses certain problems
- it may be appropriate to cite the data without reference to interpretations, particularly in lists or infoboxes
- it may be that two sources provide the same basic data, one being more detailed, and another being more widely-available. I've read scholarly articles (e.g. Gilham and Marx) which comment on the opposing sides in the Battle of Seattle (N30) using less-complete sources while one of the editors has more-complete sources (e.g. police rosters from the city archives).
  - in this case, it seems perfectly appropriate to cite both sources, and then the interpretation of one source.
  - I'm not sure about citing details from one source and getting the interpretation from works using the other source. I suspect some limited use is NOR but it can easily cross into OR.
tightening "not actually in" to "not in" Jacob Haller 02:21, 16 September 2007 (UTC)
"raw data" might not be the best word for narrative sections of histories, but the same principles apply. Jacob Haller 02:23, 16 September 2007 (UTC)

Hmm. A birthdate is raw data. A quotation is raw data. All photos are raw data. So are you really sure any of this is what you mean to say? - Jmabel | Talk 02:43, 16 September 2007 (UTC)

Excellent points. How about:

Uninterpreted raw data is to be used carefully. Any interpretation of the raw data that is not in any of the sources cited is called "original research" at Wikipedia and is not acceptable content for a Wikipedia article. For example:
- Numbers are easy to misuse and should be presented in a context similar to the source, such as income and population data for a city given in an article on that city or a birthdate in an article about that person; rather than in an article or section comparing populations or birthdays to make some point not made in the sources (richest city; oldest Jewish poet)
- Quotations are easy to misuse and must not be used to make a point that in the original context was not explicitly being made.
- Images and their caption should not be used to make a point that is not made in the text of a proper source.
- Any use of old text or text written in a context other than for general current popular consumption can easily be misunderstood as to its significance or meaning and must be dealt with carefully by knowledgeable editors who have the prerequisites to evaluate. For example, math articles, science articles, and sources written in a language other than English require editors with specific prerequisites to properly evaluate the sources. Also words change in meaning over time and have different meanings in different contexts. "Jon is gay" means something different depending on context and era. "Suffer the children" in the King James Bible shows a famous change in the use of the word "suffer". The prerequisites must the common enough for editors to be able to check each other. Wikipedia can not allow raw data that requires requires the editor to be an expert in order to properly context that raw data, as we require all our content to be evaluatable by many (but not all) editors. WAS 4.250 03:11, 16 September 2007 (UTC)

User:WAS 4.250 "Wikipedia can not allow raw data that requires requires the editor to be an expert…" Since "context" isn't a verb, I may be misunderstanding, and also the phrase "allow raw data" is really confusing. Allow it in the article or as a source?

I think it is entirely appropriate to have an article where expertise is required to make sense of the sources. Typically, we will have a number of people with such expertise, and their consensus should be good enough for us non-experts to accept that things have been sorted out. In the rare case where we have only one expert available, assuming we have no reason to doubt that person's claim to expertise, I'm usually inclined to assume good faith.

On the other hand, dropping obscure raw data into an article without context is certainly not generally useful.

As for "any interpretation" being illegitimate, I simply disagree. See, especially, my remarks in the next section, but let me give some concrete examples:

If I am reading a source about the Beatles, I am going to assume that "McCartney" means Paul, not (say) Linda, unless the context is clearly to the contrary. That is the kind of interpretation we all do continuously in reading any source.
Similarly, but less obviously, to take an area in which knowledge is not so universal: if I'm reading about beer and I run across the name Michael Jackson, I'm going to presume it is the recently deceased writer on beer and whiskey, not the performer. Again, this is necessary interpretation.
To refer obliquely to a case you gave above: if we know that A was born a decade before B and that they taught in the same university department at the same time, I think referring to A as B's "older colleague" is an entirely legitimate synthesis, and can save a ton of writing. As I discuss below, this was not the sort of thing that "NOR" was intended to discourage, and I see no reason to discourage it.

- Jmabel | Talk 04:29, 16 September 2007 (UTC)

Jmabel: would you agree that if some form of information is possibly subject to misuse that it isn't the form of information that is the problem but instead it is the misuse? That being the case should not the policy concentrate on the abuse and not on the form of information (if the policy must concentrate on anything at all)?

Or suppose an article already says X is Y. An editor finds a piece of data incompatible with that statement and inserts it (with a reference to it's source) into the article as a contribution to NPOV. The cant is that the editor is not to do that, the editor must find a source that indicates the data (which can be seen to be incompatible with the existing statement) is incompatible with the existing statement and cite that. The issue of y-ness already exists within the article. The editor who introduces the data isn't bringing in a novel concept (non-Y-ness), it's already implicit as a topic because of the existing "X is Y" statement. --Minasbeede 13:58, 16 September 2007 (UTC)

Minasbeede, here is where you and I very much disagree... I feel you can not assume a state of "non-Y-ness" without a source. I do not believe that it is implicit in the article (nor is it implicit in real life). The statement "water (X) is wet(Y)" does not at all imply that there is a state where "water is not wet". Unless you have a source that concludes that X is not Y, saying so in a wikipedia article is indeed your own original research. NPOV does not trump NOR, they have to word in tandem. Blueboar 14:23, 16 September 2007 (UTC)

The counter-example that comes to mind first runs the other way. An article says John Doe was not a communist. A person finds a list of members of a communist cell and John Doe's name is on it (and in the example the name isn't a common one: there's scant likelihood that it's another John Doe.) The editor says John Doe's name is on a list of members of a communist cell and gives the reference to where he found the list. I say that's not novel.

(To me this is clearly original research. If no one else has ever commented on this list, it certainly is novel to use it to state that John Doe was a communist. The conclusion originated with you and on Wikipedia, which is exactly what this policy is designed to prevent. In fact, the type of source is irrelevant, it has to do with you as a wikipedia editor making a conclusion.) Blueboar 18:37, 16 September 2007 (UTC)

This hypothetical would be a bit dicey. If the article just said, "The name John Doe is found on list X of Communists," then you might argue that this is not original research in itself, because it's an obvious statement based on a reliable source (assuming the list itself has been published), and doesn't make any new, unobvious conclusion beyond the source itself. The problem arises, however, in the implication that the John Doe on the list is the same John Doe as you're talking about, and that his presence on the list implies he was a Communist. If those implications are not obvious and inescapable to any reasonable reader of the list, then I'd say the editor has drawn, by implication, an original conclusion based on the list, and thus performed original research. COGDEN 18:29, 18 September 2007 (UTC)

Or here's a forward one. An article says that John Doe drove a 1984 Blue Screamer that was still in factory-mint condition and had a manual transmission. The editor finds a list that shows the production statistics for Blue Screamers from 1970 to 1989. For the years 1981 through 1986 the number of Blue Screamers shipped with manual transmission is zero. He puts that reference into the article and notes that it appears impossible for John Doe to have had a 1984 Blue Screamer with a factory manual transmission. Not having a manual transmission is not a novel concept with respect to the article. It seems absurd to require the editor to find a source that says that John Doe could not have had a 1984 Blue Screamer with a factory manual transmission because there were none. The editor has sourced the evidence. It is only because of the wording added to the NOR policy that the issue even arises.

There may be no statement anywhere that water is not wet but there have been courses in which water was studied as a "non-aqueous" solvent. So far I can't recall much but I do have a professor and a university associated with that in my mind. Nor is every statement in Wikipedia as secure as "water is wet." --Minasbeede 14:48, 16 September 2007 (UTC)

The Blue Screamer example seems like a textbook example of OR. Isn't it possible that a handful of 1984 Blue Screamers were specially made with a manual transmission for a few select customers. Perhaps these were kept "off-book", which means they weren't included in the production statistics the editors referenced. The point is the first article specifically states that John Doe does have such a vehicle. An editor should not be using a primary source such as production statistics to attempt to advance a position that "disproves" the first article. Chaz ^Beckett 15:05, 16 September 2007 (UTC)

In any actual discussion of this in the appropriate (automobile aficionado) circles wouldn't it be up to the person who hypothesized off-book Blue Screamers with manual transmissions to find a verifiable source? What's missing in what you say is WHY an editor can't use plain facts that contradict what's in a Wikipedia article. It's verifiable. It's official. All you're really saying is that the claim that got into Wikipedia first must be assumed to be true over any other evidence - if the other evidence is presented directly rather than as "car expert Richard Roe laughs at John Doe's claim because the production figures clearly show there were no such cars manufactured. John Doe may have had such a vehicle but it wasn't stock, it is not mint: that's an after-market modification." So if Richard Roe cites the figures that's OK for inclusion but if the Wikipedia editor cites them he's guilty of "original research" because he assumes the numbers say what the numbers say and mean what the numbers mean. Just because the John Doe claim arrived in Wikipedia first. Pray, what do we do it the John Doe article says he has such a vehicle and the Blue Screamer article says there was no such vehicle? --Minasbeede 18:00, 16 September 2007 (UTC)

You could use inter-article contradiction tags and/or note that no 1984 Blue Screamers were built with manual transmissions. (I don't know if this is true). I think you can use the word "however," but some other editors may object. Jacob Haller 19:49, 16 September 2007 (UTC)

If there were no Blue Screamer article I could refer to the source that has the production figures and put those figures on a new Blue Screamer page, right? I can use that data if creating an article and it's not "original research." It's verifiable. That's an utterly valid edit, right?

So I do that, there's now a Blue screamer article. As I understand the current rule some other editor cannot now use any claim that John Doe had a 1968 Blue Screamer, stock, mint, to show the incorrectness of the 1968 production figures that show no such car was built. The editor would have to form the conclusion that John Doe's possession of such a vehicle invalidates the 0 in the production table. He can't do that.

So it's whoever got to Wikipedia first, right? Either one of us could put the information we have (the production figures, the information about John Does vehicle) into an article, no problem. But neither of us can use that exact same information in the other article that is based on the other information. Either way each of us is using the same information. Either way once one item of information is used the other cannot be used in the same article. Is that a policy? --Minasbeede 20:10, 16 September 2007 (UTC)

ChazBeckett is completely right, which is why once again I have to wonder whether minasbeede doesn't just reject this policy. I am not sure why this is "moving the discussion forward" and no "back to basics" but let's be less hypothetical. A good number of academic disciplines do source-baes research. I won't quibbl over semantics like sources of materials, I am making a specific point. Historians and Comp. Lit people are what I am thinking of, but you can add some political scientists and economists. An historian has read the Gospels as well as other documents from the early perio of Roman occupied palestine (about 50 BCE to 135 CE) and publishes a book or article in a peer-reviewed journal making claims about Jesus. An editor says, "I have read the Gospels and the historian is wrong." Now, the historian will agree with the editor of Wikipedia that the Gospels say X. But the historian and Wikipedia editor disagree over how to interpret X and how much weight to give it. I say, to make this a superb encyclopedia we need to trust as authoritative (not "the truth," but having some value based on the training of the author and the research that went into writing the book or article, as well as the formal process such as peer-review through which the book or article was vetted) the views of published scholars. And no matter what the Gospels say, and how much of the Gospels the editor has memorized, this is not the place for the editor to argue that the historian is wrong. A literary scholar writes a book analyzing Shakespeare. An editor has read Hamlet, and has a different interpretation, and has arguments and evidence from not just t he version of the play widely published but early folios the editor read at the Folger library. I do not question that the editor read those folios. This is still not the place for the editor to add to an article "However, the x edition of Hamlet reads .... which contradicts this scholar's argument" or "However, if one compares Hamlet and Macbeth, another interpretation is compelling. No. If an editor wants to argue with a published researcher, they should wubmit an article to a peer-reviewed journal - not use Wikipdia to in effect publish his or her own views. Or, I have a graph indicating changes in GDP around the world over the past thirty years. I think this graph proves that democratic countries have healthier economies. An economist or political scientist may say, no, you don't understand the equations by which that chart was compiled, which are insufficient for your claims. And I say "No, the chart makes it very clear, look, I am just going to say democatic countries have x% growth or y% below the poverty line, and non-democratic countries ..." and a political scientist says that no, we can't interpret the chart to involve causality and here is why: .... These are the kinds of examples we actually encounter in editing Wikipedia. The point is that many fields of study require people to spend years learning how to analyze texts appropriately, and then diferent experts fight it out in trying to get their books and articles published. An encyclopedia should value the work that they do, and base itself on that work, when relevant. I've seen too many editors who think they understand what a text (source, material) means and conclude that the published secondary source (or whatever you want to call it) written by a scholar from a particular discipline is just mistaken - when in fact it is the editor who is mistaken, who does not understand the discipline or the source material. Aside from everything else, NOR (like NPOV) requires a certain humility on the part of its editors: we do not decide who is right or wrong, or what is the truth and what isn't. We defer to others - to verifiable sources. Slrubenstein | Talk 15:11, 16 September 2007 (UTC)

I agree with the above, and I'd like to point out that in the Gospels-interpretation hypo above, both the Gospels and the historian original peer-reviewed conclusions about the Gospels are primary sources. The reason an editor can't use the Gospels to directly contradict the peer-reviewed historian is simply because the editor is countering the scholar not with the Gospels, but with the editor's original or unverified interpretation of the Gospels. But if a peer-reviewed scholar directly disagreed with the Gospels, there would be no problem with an editor saying something like, "The Gospel of Luke says Jesus "gave up the ghost", but Professor Rosenbloom suggests he was just sleeping". This would be an equal conflict between two primary sources. COGDEN 18:29, 18 September 2007 (UTC)

Minasbeede rejects something, but it's not the policy. Minasbeede embraces the policy as worded before the source typing appeared. --Minasbeede 18:01, 16 September 2007 (UTC)

BS. You reject the policy. You are advocating original research. Make an honest proposal to the community then. Slrubenstein | Talk 18:17, 16 September 2007 (UTC)

Blueboar is doing fine. I regret the distraction from his efforts. --Minasbeede 18:23, 16 September 2007 (UTC)

You can't just reject four years of consensus like that. Slrubenstein | Talk 18:40, 16 September 2007 (UTC)

Blueboar is not rejecting any of it, as far as I know. He is working to improve the wording. At the end I'll probably say "I don't like it." I hope I can say "but it's much better." Then there will be a consensus, I'd hope. --Minasbeede 18:53, 16 September 2007 (UTC)

I find much of the discussion here very confusing ("non-Y-ness"?). Yes, I agree that the issue is misuse of sources. I also agree that, in general, "If an editor wants to argue with a published researcher, they should submit an article to a peer-reviewed journal; however, I also think that this cannot be followed blindly." Certainly this is the case for science, interpretation of history, etc.

However, I find it silly when (as occurred, for example) someone uses an outdated newspaper article to claim that there is a bicycle shop in Wedgwood, Seattle, Washington. I live there. It's gone. A dance studio is now at the same address, but it still requires some synthesis — or simply walking down the street — to establish that there is no longer a bicycle shop in the neighborhood. I want to make sure that we do not disparage this sort of common sense correction as unacceptable "original research". It is presumably impossible to find a "reliable source" that postitively asserts that the neighborhood lacks a bicycle shop. At some point, presuming good faith has to trump formalities.

Let me give a potentially more controversial, but similar, example. Edward Said's obituary for Ibrahim Abu-Lughod, which we cite in that article, misspells the name of Abu-Lughod's daughter "Deena" as "Dina". Deena Abu-Lughod is now (or at least when I last checked) an academic at the New School in NYC. Normally, Said would be an excellent source on his friend's daughter's name; however, he (or his editor) misspelled it. I happen to be quite sure of how to spell it, because we were in college together, and I'm quite certain she's his daughter because I had once been to their home in Chicago in May 1977. Given that I have no track record of falsification, how many hoops should I really need to jump through here?

In short, while I think this is a generally good policy, I'm very concerned about how it can be applied too mechanically and/or used by wikilawyers to make an awful lot of extra work for knowledgable contributors, wasting time that would be far better spent other ways… or simply driving knowledgable contributors away.

As I've written here, part of the problem is more or less epistemological. Some people seem to think that an encyclopedia article can be generated by following a set of rules, even by someone relatively ignorant of the topic at hand and of general principles of research. I think that is seldom the case.

We need standards that clearly exclude people from using Wikipedia to publish their novel scientific theories, their original interpretations of history, or even their factual claims in areas likely to be controversial. We do not need a standard that prevents people who appear to be operating in good faith from making simple statements of presumably uncontroversial fact, making presumably uncontroversial syntheses, or making minor factual corrections to inadequately researched or poorly edited/proofread sources. - Jmabel | Talk 18:41, 16 September 2007 (UTC)

I agree. If something's not controversial or likely to be challenged, according to WP:V, you don't need to cite any source at all on that point. Likewise, an obvious and inescapable conclusion is not original research, because it's not a new idea, but something inherent in previously-published material. The inescapable and obvious conclusions of prior research, in my view, are part of the prior research. Stating the obvious, straightforward implications of old research is not new research. COGDEN 18:36, 18 September 2007 (UTC)

Sorry. In the context of "X is Y" non-Y-ness means that the counter statement, "X is not Y" is implicit in "X is Y." In the context of this disucssion all I mean is that if someone says "X is Y" ("John Doe is hairy", with Y being "Hairy") it's not introduction of anything novel to cite a source (which I'll call "something") that shows John Doe is not hairy. Not being hairy is no-Y-ness here. The claim being made is that an editor can't do that, he can only say words like "Richard Roe says, based on (something) that John Doe is not hairy." Same source. The editor can't look at it and determine that it indicates John Doe isn't hairy, he has to quote Richard Roe, who looks at the same evidence and says the same thing.

It looks like we have areas of substantial agreement: your last two paragraphs, for example. --Minasbeede 19:01, 16 September 2007 (UTC)

Back to basics

Slrubenstein asked me for my views on the recent discussions here.

While I think the "primary material" vs. "primary sources" issue is valid, I think we need (1) to get back to basics and (2) to recognize that this cannot be reduced to formulae. The "NOR" policy was originally about keeping crackpot pseudo-science out of Wikipedia. It seems entirely appropriate to restrict opinions, syntheses, etc., to those attributable to a person or organization that would normally be considered an authority and to a general requirement that factual claims be cited from reliable sources (largely covered by other standards and guidelines, but an "OR" issue arises when we have to choose between conflicting sources).

In general, "OR" becomes a much more important issue in areas that are potentially controversial. The sophistic arguments I've often seen here questioning sources in matters where there is no legitimate controversy strike me as a juvenile waste of :time. While it is nice to have a citation for the statement that "avenues" in Midtown Manhattan run north-south, presumably there are enough people who can verify this from their own knowledge that the lack of a citation is no big deal.

On the other hand, there can be potentially controversial issues of fact, not only of interpretation. If someone wanted to claim that a particular Manhattan avenue carries the most traffic, that would absolutely call for citation from a reliable source. Many sources would be legitimate prima facie: a news article in a mainstream newspaper, a document from the city government, a book from a mainstream commercial or academic press on transport in New York, etc. If, however, someone else came up with a second source that would normally be considered comparably reliable and that disagrees, there is no "formula" to solve this: some original thought will be required. We would want to mention the clashing views and to cite both; it is still going to be a judgment call whether both go in the mainline equally, or one is consigned to a footnote, or what.

There are areas where primary documents are to be preferred; there are other areas where secondary documents are to be preferred. Consider, for example, writing about the Declaration of the Rights of Man and of the Citizen (a rather thin article at the moment). It would seem to me that:

When it comes to listing what particular rights are enumerated, the author should have (and cite) both the French original—the absolutely primary document—and a professionally-done translation (which slides slightly into secondary).
When it comes to discussing the role the Declaration played in the history of France, or its status today in France, or its role in the history of the rhetoric of human rights, this is inherently a matter of opinion, and mainstream scholarly opinion should be cited. Where mainstream scholars disagree, a range of mainstream scholarly opinion should be cited. Opinions that cannot be so cited don't belong in the article. (Note, though, that I say cannot be so cited, not are not so cited. Unless someone has a good faith doubt about poorly cited material, I say tag it and let it stay.) These sources are secondary relative to the Declaration itself, but primary for their authors' views.
When it come to how the Declaration was perceived at the time it was issued, probably the ideal would be a good overview derived from a scholarly source. Still, there is undoubtedly some primary source material that would be relevant and useful if well handled. It might well be worth quoting a political pamphlet or speech of the time, as long as there is some ability (presumably from secondary sources) to contextualize the politics of its author.

To come at this from a different angle, different sourcing standards apply in different areas, and even to different facts in the same area. Judgment will always be needed, and this judgment should not called "original research" in any disparaging sense. For example:

The Rolling Stones' own web site is a perfectly good source on the current membership of the band, but not on how their recent album was received by the critics.
An article in the Seattle Weekly is a perfectly good source on the party affiliation of a politician in that city, but not particularly citable on the Iraq War.
Historian Eric Hobsbawm is a fine source on world-historical facts, and quite citable for his opinions about history (identified as such, and typically with an overt mention that he is a Marxist), but is pretty useless on popular culture, and especially youth culture (e.g. in a passage in Age of Extremes, he seems entirely ignorant of the fact that Buddy Holly died in a plane crash, not because of some "life-style" excess).

I could go on, but I'll try to cut to the chase:

Writing an encyclopedia article requires making a certain number of judgments, evaluating one's sources, and having at least a decent working knowledge of the topics you are writing about.
While there are types of "original research" that we need to reject as inappropriate for Wikipedia, this should not become a ban on thought, a ban on bringing to bear one's knowledge of the subject at hand, or a ban on introduction of essentially uncontroversial material merely because the formal sourcing is weak.

- Jmabel | Talk 04:01, 16 September 2007 (UTC)

Jmabel: Is it fair to summarize what you say by extracting these two bits: "we need (1) to get back to basics and (2) to recognize that this cannot be reduced to formulae" and "While there are types of "original research" that we need to reject as inappropriate for Wikipedia, this should not become a ban on thought, a ban on bringing to bear one's knowledge of the subject at hand, or a ban on introduction of essentially uncontroversial material merely because the formal sourcing is weak."

May I also ask if you'd agree with "stop here, this is enough"? --Minasbeede 13:46, 16 September 2007 (UTC)

Jmabel, as near as I can tell we do not disagree. You disagree with your straw-men, as I also do. Surely when you stand up for an expert making an expert judgement saying you are inclined to AGF, you are not saying that a self proclaimed expert is allowed to assert his judgement against the consensus of the rest who dispute his expertise - if so then we do disagree. WAS 4.250 10:30, 16 September 2007 (UTC)

Responding directly to prior comments: "To context" as a verb means "to create context" referring to what you refer to when you say "dropping obscure raw data into an article without context is certainly not generally useful". "Allow raw data" in the Wikipedia article (as opposed to its sources) is what was meant. By "any interpretation" being illegitimate, I am referring to criteria for resolving a dispute. Undisputed interpretations are not being addressed. WAS 4.250 10:43, 16 September 2007 (UTC)

DIRECTLY RESPONDING to WAS 4.250 16 September 2007 (UTC), with further;

Material. Good. Good. Seems you ladies have advanced considerably since I last visited the locked discussion (please forgive me).

This seems all right: "Numbers are easy to misuse...", your second point;

But: What about Wikipedia holding mathematical deductions — results from already well known mathematical laws?

If I exemplify with "For example 1+3 = 4", you would perhaps say "that's ridiculous" if someone (a Wikipedia administrator) would demand that this exact form must have an already published source to be even SPELLED in Wikipedia — because it is easy for us to VERIFY AT PLACE. Right?

But (IF that would be the case) are you still that cocky too (I don’t mean to be harsh) in front of "simple examples" if the example is the DEDUCTIONS OF Planck radiation law and Stefan-Boltzmann law? These, EXPLICITLY NOT PRESENT IN ANY ALREADY PUBLISHED SOURCE, already exists in Wikipedia articles. Yes. but they are SPECIFICALLY not related to any other source than Wikipedia ITSELF. Yes again. What did you say about the ”1+3 = 4”?

Would you accept such »deriving editing with respect to already known basic mathematical laws» in Wikipedia?
Y/N?_

Because then, my friend, Wikipedia will soon develop into a huge NOVEL of NEW mathematics AND physics, never before seen in any other place on this planet Earth.

I personally would accept that, because it would ALSO OFFICIALLY (see further below) open much more of "encyclopedia" — and debate — to knowledge in our humanity than ever will be "published" by traditional channels:

Verifiability with consensus: use whatever you want to portrait the natural content of a topic (within the interest of human rights, the welfare of education and development of technology, science and philosophy in general within the limits of normal conduct and civil behavior, of course), but see to that it is verifiable (meaning already published sources on top, citations, references, u name it) and that the editors are in agreement (meaning a good editorial spirit); if they disagree, try to map the landscape there too (still in a good spirit). Easy. Make it RICH. Make it ATTRACTIVE. Make it THE BEST. Make it so, that every living soul wants to visit its story because of fascination, skill, solid references, reliable description and presentation. When the opinion changes, so does the encyclopedia.

Any ”crackpot pseudo-science” (term borrowed from Jmabel contribution above) or type ”WackoArticle” (term borrowed from earlier contributors on this talk page) will (in time) delete itself — due to lack of solidity and trustability in the presentation.

BECAUSE THAT IS ALREADY HOW (the discussed details of) MANY ARTICLES IN WIKIPEDIA ALREADY HAVE BEEN SET (Physics, Magnetic Field ...). If you haven’t noticed.

What say you? BMJ 11:11, 16 September 2007 (UTC)

I say this: BellMJ, in the month or two you have been here you have not contributed to any articles. I suggest you get some actual expeience researching and making contributions to articles that stand the test of time, and have more experience collaborating with editors working on aticles, before you try to comment on our core policies. Slrubenstein | Talk 11:27, 16 September 2007 (UTC)

Well i am there in the same time, i have contribued to many articles and still i would try to comment 'core policies' but with this terms:

"Be bold" (which is part of the Wikipedia system.) The focus should be on the product and on whether it is or isn't appropriate and not on the rules. The rules exist to help make the product be appropriate and to ease the process of making it be so. What you suggest looks like it could work, provide it's not taken as an iron law. --Minasbeede

Overall, apart this, if someone believes that 'policies' will override the competence, good will, good faith, proficience in specific fields and consensus+bold+IAR, then he wanted to build a burcocracy (with many arrows in the ands to who wants to demolish some unliked articles) and not an encyclopedy. By the beginnings of the time who writes an encyclopedia adds personal toughs, and surely now that there are 6 billions humans many of them linked with interet we will see a different, silly way to act, that scraps some foundamentals capabilities of human mind. As example you cannot seriously ask to not insert original syntesis/tough while you ask to editors to re-editing an existent source to avoid Copyviol or resume an entire book in a single article. It's simply impossible. Or Wiki buys already 'secondary sources' that, as example, resumed the 'War and peace' novel, or buy directly the right to post it. Asking to someone to resume this or whetever opera with own words and asking to do it without personal tough is simply impossible. Parithetic revision, comments, discussions are the only way to prune some 'excess'. If not, if there are, as usually are not so many, persons involved to talk about an article, then there is not policy that can be used instead, and NOR is IMO, the most absurd (as often used) to all. Let's understand this, it's simply a countersense ask to one to think enough to resume 100 pages of 10 different sources in an article and then accuse him to have made 'OR'! It's like to say that 'thinking' is a crime. Stefanomencarelli.

Re outlawing thinking: Well, why not? After all, policy is supposed to reflect practice. (Just kidding. :-) Seriously, I'm with Minasbeede against instruction creep. --Coppertwig 17:55, 16 September 2007 (UTC)

Yes, I have been fairly paraphrased above. But, no, these are not straw men. Bad-faith, juvenile wikilawyering over things like those I describe here constitute a large part of why I am much less active in Wikipedia than I used to be. - Jmabel | Talk 18:47, 16 September 2007 (UTC)

I totally agree with Jmabel's analysis above, which I think is a great summary of the way things work in Wikipedia. We don't want to over-generalize about different types of sources. Determining which source is the best one is not something amenable to simple rules like "primary sources should be rare", or "secondary sources should be primarily relied upon". While there might eventually be enough consensus for a guideline on these issues, there's certainly none now, and I think the best direction we can take right now is to focus on the fundamentals: what original research is, and what constitutes "going beyond" a source to inject into the article ones' own unpublished ideas. COGDEN 18:48, 18 September 2007 (UTC)

WP:PSTS

Where is the link to this Request for Comment? One should be provided here. Please feel free to remove my comment here after providing the wikilink. Thanks. ... Kenosis 19:22, 22 September 2007 (UTC)

As noted above in the tag, it will be automatically added to the list. The heavy lifting is done by RFC bot. Vassyana 19:34, 22 September 2007 (UTC)

Kenosis, this is the RfC. The bot will add a link here to the RfCs page, to rovide wider notification that broader consultation has been requested. The discussion still happens here. SamBC(talk) 23:28, 22 September 2007 (UTC)

Ahh yes, thanks. I hadn't known that policy RfCs had been set up to be conducted this way; unlike the way RfC's are generally set up on a separate page. ... Kenosis 03:03, 23 September 2007 (UTC)

I think all content-based (including content of policy) RfCs are done like this. User conduct and usernames are exceptions. I've been involved in this sort of RfC on normal articles as well. And if I don't fall asleep soon, I'm probably going to start seeing things... SamBC(talk) 03:19, 23 September 2007 (UTC)

Indeed. Sleep well -- argue later on, tomorrow, whenever ;-) ... Kenosis 04:57, 23 September 2007 (UTC)

^ Wikipedia's co-founder, Jimbo Wales, has described the origin of the original research policy as follows: "The phrase 'original research' originated primarily as a practical means to deal with physics cranks, of which of course there are a number on the Web. The basic concept is as follows: It can be quite difficult for us to make any valid judgment as to whether a particular thing is true or not. It is not appropriate for us to try to determine whether someone's novel theory of physics is valid; we are not really equipped to do that. But what we can do is check whether or not it actually has been published in reputable journals or by reputable publishers. So it is quite convenient to avoid judging the credibility of things by simply sticking to things that have been judged credible by people much better equipped to decide. The exact same principle will hold true for history." (Wales, Jimmy. "Original research", December 3, 2004)

[1] Wikipedia's co-founder, Jimbo Wales, has described the origin of the original research policy as follows: "The phrase 'original research' originated primarily as a practical means to deal with physics cranks, of which of course there are a number on the Web. The basic concept is as follows: It can be quite difficult for us to make any valid judgment as to whether a particular thing is true or not. It is not appropriate for us to try to determine whether someone's novel theory of physics is valid; we are not really equipped to do that. But what we can do is check whether or not it actually has been published in reputable journals or by reputable publishers. So it is quite convenient to avoid judging the credibility of things by simply sticking to things that have been judged credible by people much better equipped to decide. The exact same principle will hold true for history." (Wales, Jimmy. "Original research", December 3, 2004)

[1]