Wikipedia talk:No original research/Archive 26

Archive 20Archive 24Archive 25Archive 26Archive 27Archive 28Archive 30

An important distinction

Isn't one of the points that is trying to be asserted with type sourcing is that the analysis we are seeking is authenticated analysis rather than uninformed opinion? I think this is the point of trying to recognise when an opinion is stated in a primary source (too close to the event, unchallenged, not peer reviewed) as opposed to secondary sources (supposedly validated).

However, it seems to me that, if that is the case, source typing by itself fails to resolve the issue, as we need to test whether the specific information being used has been validated. Again we come back to the issue of this being more about reliable sources. In a secondary source, the quote may be repeated in the secondary source and unless it is used in a secondary source to support an analysis, it may still be unusable, and only is usable if taken in context with the surrounding analysis.

So again, is source typing the solution, or do we look to define the issue of reliable sources more carefully. After all, source typing is really an issue of reliable sources, rather than originality.

Am I going down a rabbit hole here? Spenny 15:54, 30 September 2007 (UTC)

You are going down a rabbit-hole. The issue is not whether the opinion is well-informed or ill-informed - this goes back to the "verifiability, not truth" dictum; it is not for us to pass judgement on the veracity or plausibility of the view. What is important is (1) that we can identify a view (thus the need to peg it to a source, the source helps identify whose view it is) and (2) that it be a notable view. We all hope that the better-informed views are the most notable. But if a view is notable it can be included in an article ven if you or I happen to think it is an ignorant view. This is simple compliance with NPOV. And in our edits of an article we cannot word things so as to promote or arrgue for those views we personally consider best-informec - this is simple compliance with NOR. Slrubenstein | Talk 16:13, 30 September 2007 (UTC)
More specifically, this is not directly related to NOR--OR is editors publishing previously unpublished facts, opinion, or analysis. The issue of whether the previously published facts, opinion, or analysis are right or wrong are not only outside the scope of NOR, they are outside the scope of WP editors, as Slrubenstein points out above. We are only concerned with editors going beyond the sources, which would include taking them out of context. Editors should not be judging what published material from reliable sources are more or less correct (outside of not repeating obvious errors). To use source typing in the way you suggest would actually be promoting OR, not restricting it. Dhaluza 16:41, 30 September 2007 (UTC)

OK. So following that line, why are we worried about primary sources vs. secondary sources? Surely this is a test not of verifiability, but of making a judgement as to the quality of the information? The assumption is that primary sources do not contain analyses, but in fact they may well do - "I saw it, gov. He crashed because he was going too fast." It is not OR to repeat that analysis, but it is unwise for reasons of policy other than OR. I would have said using source typing in this way was one of the things that justified it being in policy - how to weigh uninformed opinion over verified analysis. Spenny 17:08, 30 September 2007 (UTC)

Actually Spenny, the example you give is not necessarily an example of a primary source. The source is the document from which you take the quote. If this quote is taken from a court record, then it is a primary source, but if it is taken from a newspaper article reporting the court case, then it is a secondary source. The source is not "what is said", the source is the "document where you got it". Alun 06:22, 2 October 2007 (UTC)
Alun, I most certainly agree with that. The point is that regardless of source typing, the statement cannot be weighed as useful in its own right. The source type of the statement is not the test. We would want it to be part of a wider discussion and for example: if the statement came from granny at the scene or a police officer (used to judging speeds and therefore an opinion of some weight) but more importantly the whole context to go with it, was it the judge quoting the officer in his reasoning to support his conclusion ("I find the officer's statement consistent with all the available evidence and I find his reasoning compelling" vs. "The Sun Says ban demon wheelchair drivers"). If the citation as a whole supported the premise and the context is a reliable source, then a writing style that made the point through the use of presenting the opinion unchallenged may be satisfactory, but if the opinion was unsupported (a la Princess Diana conspiracy theorists) then it is useless without all the surrounding information. I think this is exactly the sort of OR people are concerned about, taking that statement, unchallenged, and making the obvious extension from it, explicitly or implicitly when we know the statement is untested. The test is in the context. Source typing is one element of judging the context, but surely more a test of a reliable source than originality. Spenny 07:39, 2 October 2007 (UTC)
Dhulaza has already made clear his opinion about PSTS as being unnecessary. However, the need to distinguish between sources close to the topic under discussion frequently enough arises that it needs to remain. That is to say, the difference between on the one hand reporting things that anyone could understand by reading the sources, and on the other hand more oblique renderings of analysis and interpretation, is sufficiently important that PSTS serves a necessary function in this policy. We run into it fairly commonly in topics that are highly theoretical, multidemensional or otherwise highly complex in its own right, or otherwise complicated by a range of opinions on the topic. ... Kenosis 18:21, 30 September 2007 (UTC)
I think PSTS is unnecessary to NOR, and I still have not seen a cogent argument for why it is necessary to discern original research. Close sources do need to be treated differently, but how is that relevant to this policy? That bit is a significant change that was only added recently, and without significant discussion. I think it represents scope creep that only makes this policy harder to understand and apply, which is ultimately counterproductive. We should stick to the basic intent of the NOR policy, and not try to stretch it to cover other things, however important they may be. Dhaluza 19:19, 30 September 2007 (UTC)

This discussion has gone around in this circle several times now. Neither "stick to the sources" nor "PSTS" adequately describes WP:NOR by itself. Both were seen to be needed from early in WP's development, and both are still needed. ... Kenosis 18:21, 30 September 2007 (UTC)

Kenosis, your statement that "the need to distinguish between sources close to the topic under discussion frequently enough arises that it needs to remain". I actually agree with that proposition, but not because of OR, rather, because of RS and NPOV:
Sources closer to the topic under discussion are (1) more likely to be reliable, and (2) less likely to be biased (with some exceptions). For example, a journalist recording an interview with Abraham Lincoln is more likely to record that interview accurately, and accurately express Lincoln's intent, than the editor of a pro-slavery newspaper commenting on the interview in an editorial. Also, the original journalist is probably less likely to be biased, unless the journalist has some personal connection to Lincoln that would cause him to misrepresent the interview.
As to NOR, closeness or farness from the original source nas nothing to do with whether a cited fact in a Wikipedia article is original research. Therefore, while the primary/secondary distinction might be valuable (if confusing) in the RS and NPOV context, it really has no place in NOR. Going beyond a source is bad, whether that source is primary or secondary. We don't implicitly make special allowances for sources just because they can be construed as "secondary". COGDEN 23:01, 1 October 2007 (UTC)
I am worried about losing the distinction between primary and secondary sources because one of the most common ways people violate NOR is to use a primary source as if it were a secondary source i.e. promoting an actual argument. In some cases they are explicitly using a primary source to argue against a secondary source, in some cases i think they really think that the assertion of fact is also asserting a specific interpretation of that fact when in fact it is the editor's interpretation. I do not know how else I can say what i just said without using "primary sources" and "secondary sources" which is why I think the policy needs to use these terms. Slrubenstein | Talk 18:30, 30 September 2007 (UTC)
I think you spot on here Slrubenstein. And like you I think explaining what a primary source is and what a secondary sources is, is fundamental to WP:NOR--Cailil talk 18:46, 30 September 2007 (UTC)
But isn't that simply going beyond the source, that is using it to say something it does not? How is classifying the source as primary and the use of it as secondary help make this clearer? We have seen that leads to a distracting side discussions along the lines of: the source is primary (no it's not) and your use creates a secondary source (no it doesn't). Dhaluza 19:19, 30 September 2007 (UTC)
I just explained it. Why should I repeat myself? Slrubenstein | Talk 19:24, 30 September 2007 (UTC)
But isn't that the same for secondary sources as well? Secondary sources also contain statements of fact (that probably came from a primary source) that can be misused in this way, and interpretations from secondary sources can be misinterpreted in a similar way, such as by extending the interpretation beyond what it actually says. SamBC(talk) 21:00, 30 September 2007 (UTC)
Slrubenstein, I think this is a useful thread to explore. To be a dog with a bone, I would have thought you have expressed a basic concern that again is independent of sourcing, it is that basic principle of understanding the whole gamut of reliability, point of view and so on of the source. I guess one of my frustrations here is why all these bright people find, seemingly as a matter of principle, it impossible to express these concepts in other terms. Logic tells me the essence of this issues exist whether source typing is used or not.
If we take a step back and think about that, isn't half the problem here that source typing is about trying to express a lot of things at the same time, and there is something of a principled position that people want to use it so closely tied to OR.
To Cailil, this is circular really. In my view, the expressions of source typing on the policy page have not captured the essence of source typing and why it is useful. Over the last year there has not been a clarity of concept that passes the "Ah! So that's what it means" test, and various people have tried to adjust the definition which has been scoped as some test of OR. I think there is a consensus that there is something of merit in source typing, but trying to fit it to OR (only) is actually a poor fit. It does not test originality well. Interpreting sources inappropriately is not simply a feature of primary sources, but I can accept that at a high level, it is a tool for taking an overview of the "safety" of the source. Spenny 19:37, 30 September 2007 (UTC)
Of course, "safety" is likely to involve NPOV issues, as well as NOR ones, and non-PSTS concerns (e.g. party, partisanship, etc.) Jacob Haller 22:07, 30 September 2007 (UTC)
Slrubenstein, you said, "I do not know how else I can say what i just said without using 'primary sources' and 'secondary sources', but what about the following (in your own words):
"[I]n some cases i think they really think that the assertion of fact is also asserting a specific interpretation of that fact when in fact it is the editor's interpretation."
Which is well said. Why don't we just say that? I think everybody would agree with you if you put it that way. You sell your ability to express yourself short. COGDEN 23:11, 1 October 2007 (UTC)

Simultaneous with, or somewhat after, the otherwise reasonable move from PSS to PSTS, PSTS lost language clearly defining "primary sources" as non-interpretive sources, and editors like myself started understanding "primary sources" as "the sorts of things we call 'primary sources' elsewhere." I was never formally taught PSS, and picked it up from professors who may not have been formally taught PSS either... The idea that you can't draw interpretive conclusions from non-interpretive sources does make sense. But there are very few completely non-interpretive sources, and there are many interpretive sources which may be considered primary sources. Jacob Haller 22:07, 30 September 2007 (UTC)

So in addition to difficulty in creating precise definitions of primary and secondary sources, there is difficulty in classifying sources as either primary or secondary even with a particular definition. So how is this source classification useful in a policy? Policies are supposed to be definitive with few exceptions. PSTS is a puzzle wrapped in an enigma. It may be useful in a guideline like RS that is subject to interpretation, but not in a policy that needs to not only be easily understandable, it must be straightforward to apply to a wide variety of subjects. Dhaluza 23:17, 30 September 2007 (UTC)
..we came in? Isn't this where... - Pink Floyd, The Wall. Spenny 07:54, 1 October 2007 (UTC)

Well Spenny, the idea of distinguishing between primary sources and secondary sources seems to have worked well in academia for a very long time. I don't think there is any sort of problem with the distinction. Your claim that the distinction between primary sources, secondary sources and tertiary sources is not easily drawn for many people editing Wikipedia may well be correct, I don't know. But just because it cannot easily be explained does not therefore mean that we should dispense with it. Here's a list of web definitions for "primary sources", "secondary sources" and "tertiary sources" none are very different from the Wikipedia definition as far as I can tell. Alun 06:17, 2 October 2007 (UTC)

I am warming to the idea myself, but not in this context and with some clear understanding of what it can and cannot be used for. I think that there has been an attempt to transplant the concept without the supporting culture and understanding.

It is not about OR, it is about source typing being one element of the judgement call of whether a source is reliable for the purposes it is being used for. As has been said elsewhere, that is not the way the principle has been used by those who "know" policy. The simple challenge of "Primary source" is not in itself anywhere near sufficient to judge whether information is used appropriately or not. It's a warning sign, nothing more. Spenny 07:45, 2 October 2007 (UTC)

Jumping in here, and running the risk of being redundant, I see that failing to distinguish properly between primary and secondary sources and using the former rather than the latter when the latter is available, raises concerns of WP:NOR, WP:NPOV and WP:UNDUE, especially when it comes to scientific arguments and theories. One can fall into that trap by sourcing to some sub-group of researchers, and if they haven't validated (through a secondary or better, a tertiary source) that the opinions of this group of researchers are actually representative of the field, thay can very well present a minority opinion as mainstream. I've seen that happen often to be able to say it's one of the more common traps of researching and reporting on a scientific subject. Also, SLRubenstein's pont about using a single primary source to try to rebutt a secondary source is something that I've seen all too often: divergent views should be presented if they are notable, but minority views (and in this case a single primary source pitted against a secondary source such as a subject review certainly qualifies as a minority) shouldn't be presented as having equal validity with majority (or mainstream) views.--Ramdrake 13:35, 2 October 2007 (UTC)
I agree with respect to WP:NPOV and WP:UNDUE, but not WP:NOR. If even one researcher has published the idea in a reliable source prior to inclusion in Wikipedia, it is not original research. Citing a sub-group of researchers may very likely be biased, and it often is, but it's not OR. COGDEN 18:10, 2 October 2007 (UTC)

Proposal : Recognized expertise trumps other claims

Proposed addition to policy:

Recognized expertise trumps other claims

Recognized expertise trumps other claims. Examples:

  1. That the Earth is billions of years old is consensus belief/opinion/fact among recognized experts so other claims have no place in Wikipedia except on articles about those claims and the people/organizations/books that make them.
  2. Published claims by experts on claims they are recognized as experts on are citable by Wikipedia regardless of the otherwise unreliable nature of the publication (eg blog, Usenet) so long as there is no serious question as to expertise (eg bias) or authorship (eg site is a wiki).
  3. Expert authored peer reviewed literature trumps journalist authored literature.
  4. That a source authored by a recognized expert is or is not a primary source does not count against it.
  5. Wikipedia should only contain claims published elsewhere regardless of the expertise of the Wikipedia editor.
I agree. WAS 4.250 04:32, 2 October 2007 (UTC)
"Official government conclusions trump journalist authored literature." - This standard would reinforce systemic bias in favor of government denial of corruption, police brutality, frame-ups, etc. Consider the Dreyfuss Affair, the Watergate Burglary, the Downing Street Memos, etc. Jacob Haller 04:23, 2 October 2007 (UTC)
OK. Removed. WAS 4.250 04:36, 2 October 2007 (UTC)
Oppose. This proposal appears to be inconsistent with WP:NPOV, which not only states that Wikipedia doesn't exclude non-scientific points of view, it actually uses creationism as a specific example of how Wikipedia must take non-scientific points of view into account. The policy cites Larrys Big Reply, which gives the rationale for doing this. Larry explains:
We should not impose our values on other thinking people. You are all liberal-minded people, I trust--not liberal politically, necessarily, but liberal in the sense that you want to free minds. I enjoin you to think carefully about the best way to achieve this. By failing to take stands on controversial issues, we aren't demonstrating weakness--in fact, we are demonstrating the strength of our faith in the minds of our fellow human beings. We should let them arrive at their own conclusions. We should trust them to use their own minds--just as you want to be trusted. More benighted souls than our enlightened selves will appreciate our stance and be more apt to listen when we hand down the truth.
This is core policy, fundamental to Wikipedia. If you want to change our core WP:NPOV policy, go to Wikipedia talk:NPOV and argue for change there. Don't try to change a long-standing core Wikipedia policy by making an argument in forum for a policy that's supposed to be about something completely different. I respectfully suggest that all questions of which of multiple sources are preferred should be matters of guidelines, not policy, and discussion belongs in e.g. WP:RS. Citing to sources by people one disagrees with or regards as non-experts may be good or bad, but I don't believe it's original research, which has a narrower and more basic meaning. More guidelines may be helpful, but I recommend keeping them guidelines and not making them policy. In any event, this proposal directly contradicts WP:NPOV and I cannot support it for that reason as well. Best, --Shirahadasha 06:02, 2 October 2007 (UTC)
It is exactly that sort of argument that NPOV + V required a crackpot physics theory referred to in a reliable source to be on article pages not about that theory or its supporters that caused the NOR policy to be created in the first place creating NPOV + V + NOR as our article content policy trifecta. And since you don't see that in the current NOR policy wording, that simply proves the above wording does in fact need to be added to the NOR policy page. WAS 4.250 06:41, 2 October 2007 (UTC)
If it's verifiable in a reliable source, in context and accurately cited, there's no original research involved. However, if it's a "crackpot" theory, it is probably also an extreme minority that shouldn't be included. This seems like an issue very well covered in Wikipedia policy. Additionally, when NOR was originally marked official policy, it specifically detailed circumstances not appropriate for exclusion.[1] The "extreme minority" issue was later moved to NPOV, as it does fit more appropriately under the aegis of that principle. The inclusion and balance of viewpoints verifiable in reliable sources is a question of NPOV, and reliability to some degree. The use of those claims outside of their context, or rather to advance a position not supported by the sources, is original research. This proposal affects WP:UNDUE directly, while outside of the final point it does not touch on original research at all. Vassyana 17:08, 2 October 2007 (UTC)

WAS, you are being unfair to Shirahadasha, and this is the third time at least in the past week or so that you insist on creating false dischotomies as ways of polarizing the discussion in unconstructive ways. NPOV requires that we represent all notable views. NO Wikipedia policy ever allows one kind of view to "trump" another view. I agree that in an article on physics the views of physicists are most relevant - but if newspapers make it clear that there is a notable view not shared by physicists, the article has to include it. And it needs to identify the different views i.e. "Virtually all physicists agree that ... but among the general public two other views are widely held (use newspapers to support this)" in any event, this discussion does not even belong in NOR. You are pissed off at some other editors so you want to disrupt this discussion. Sorry, the issue you raise has to do with - I said this already - NPOV and RS, take the discussion there. But Wikipedia representas all notable views. The bulk of the theory on evolution explains how evolutionary scientists see evolution BUT the article itself acknowledges that non-evolutionary scientists have other views, and there views are summarized and linked to other articles. All notable views get expressed. Slrubenstein | Talk 09:07, 2 October 2007 (UTC)

Slrubenstein, you are being unfair to me, and this is simply a repeat of prior insistences of false accusations and personal attacks on me as ways of polarizing the discussion in unconstructive ways. Good riddance. You asked me not to give up on this conversation. For what? So you could insult me some more? Your disagreeing with me does not prove me wrong. It only proves we disagree. Anyway, as I said to you before, actual practise at Wikipedia is not that bad; so improving the wording is not that critical. It certainly isn't worth any more of my time. Good luck to everyone here that is working to improve Wikipedia; and shame on those who can't imagine that those who disagree with them might possibly be right. WAS 4.250 09:42, 2 October 2007 (UTC)
Slrubenstein, as an Aunt Sally, it is an interesting proposal. I think it highlights the difficulty of trying to codify the principles. No we cannot let government off the hook, but be cannot let the crackpot gain undue weight. We cannot evaluate the balance of sources without understanding the nature of the sources and what biases they contain. There has been a view that we somehow must take the sources at their word, but actually we must balance and evaluate them - we must do our own research on the sources and allow the consensus process to establish whether the interpretation presented is reasonable.
As people try to codify NOR, the codes are challenged each time by the context. As we cannot define all contexts, it is at least brave to try and define the codes. Spenny 09:52, 2 October 2007 (UTC)
I think this proposal actually reveals the real problem with OR and synthesis, but not in a positive way. Those who actually create content on varied subjects have probably found conflicting information in reliable sources when researching a new subject, and have had to struggle with what to do with this information. Trying to resolve the conflicts by yourself is OR. Whether you ignore certain references, deny them, or somehow rationalize them, that is engaging in original thought, not simple reporting (aside from correcting obvious errors). This proposed rule would just try to make this sort of thing acceptable if it is done in a certain way. Sometimes we just have to report that the jury is still out, and even though we may think the weight of the evidence says they should rule in a particular way, we have to present the conflicting arguments and let the reader decide the case on their own.
I agree with Shirahadasha that we need to trust the reader, and not try to act like parents treating them like children. Once we cross the line into trying to decide what is right and what is wrong for the reader (i.e. what trumps what), we are deep into the realm of original research. It's not about what type of source we are using, it is about what we are trying to do with them. If we are not simply trying to collect all available information, assign appropriate weight, and present it in a coherent fashion, it's OR. It's ironic that sometimes people use the literal text of the OR policy to justify doing OR ostensibly to eliminate it. Dhaluza 10:58, 2 October 2007 (UTC)
There is a need to evaluate sources and sift out the rubbish, especially in the context of the web. There are those who believe that Captain Pugwash really did have crewmen called Seaman Stains and Master Bates and Roger the Cabin Boy. To take you at your word, you are suggesting that to apply judgement to the editing process is OR. That only works in the imaginary world of reliable sources. Yes, we have to trust the reader, but at some point we have to trust the editors too. We cannot make editing a no brain, innocent activity. Judgements have to be made, we just want to direct those judgements. Spenny 11:07, 2 October 2007 (UTC)
The point is, what is being represented is not "the truth" but a point of view of the truth, or multiple and different views of the truth. Our standard for inclusion is relevance and notability, but all relevant and notable points of view should be included and properly idnetified and contextualized - not judged as being more or less true, but classified as representing point of view x, POV y, or POV z. Slrubenstein | Talk 11:46, 2 October 2007 (UTC)
I am not asking for that. What I am saying is that to do the reasonable process as you suggest, requires a judgement call. The Captain Pugwash example has calmed down now, and the fact it came out of a spoof is fairly well exposed, but there was a time when there was strong credence for the blatant misrepresentation which was very easily refuted by going to the primary source - the TV show themselves. To go down the path of being innocently non-judgemental is to expose yourself to manipulation. How do you assess what is a proper point of view of X amount of weight? With hindsight we can report the Captain Pugwash spoof as that, but at one point there was a belief system that this was indeed true with secondary sources (albeit of dubious reliability) to back it up. [2] Note the reference to the Guardian ascribing it as fact. Spenny 12:03, 2 October 2007 (UTC)
I see the Captain Pugwash example as a red herring. A hoax is simply a different thing from a notable viewpoint one disagrees with. Same with non- or trivially notable viewpoits. The person who invented the Flying Spaghetti Monster may think the account is logically as good as the Bible's, but this subjective belief doesn't mean it has the same level of notability in terms of historical influence or current adherence. Notability is not the same thing as truth. Best, --Shirahadasha 14:09, 2 October 2007 (UTC)
I don't want to split herrings, but the point is that at one time it was not perceived as a hoax. From my being a self-proclaimed expert in BBC Childrens' programmes c. 1964-1981 I had the knowledge, and could have dug up sources, but others might have used the Guardian to assert this POV against a host of unreliable fan sites, and it is suggested that we might have been obliged to present this misinformation as a plausible POV. Someone who knows nothing about a subject and attempts to edit with the available source might not be able to detect the hoax - or deliberate distortion or POV push of the source. You need to know your topic and apply skill to write properly on the subject, and no amount of rule-writing is going to solve that, it seems to me. Spenny 15:10, 2 October 2007 (UTC)
In principle, I'd almost go along with this (proposal). However, in actual practice I can think of several examples where this would lead to all sorts of problems, jusy within the realm of military history. I don't know about other disciplines or subject areas, but at least in the military history there are numerous "experts" who clearly have a POV to push, and do so quite often through their books (which aren't peer-reviewed), and their articles in various publications with a pronounced left- or right-wing bent. This proposal would somehow make these non-Wikipedia POV-pushers, who are acclaimed as "experts" by those who like their POV, to carry greater weight than those who don't have a POV to push, but just happened to published something along that subject objectively, more in line with what Wikipedia is "supposed" to be about. Even though he's now pretty much disgraced, look at how someone with a very clear POV like Ward Churchill could have had his work super-imposed over many other sources of information, since many proclaimed him as an expert in American abuses of American Indians, until he finally stuck his neck out to far and it was finally found that he not only extensively plagarized others, but he also "made up events and facts" to support the claims he made in his books.
So, while there does probably need to something somewhere, to help editors gauge the appropriate amount of weight to be given different 'sources of information' (probably NPOV), I don't think that it's a 'good fit' for the NOR policy and really would open the way for much abuse, more than there currently is, which is still quite a lot. wbfergus Talk 14:11, 2 October 2007 (UTC)

I'm not sure what this modification is intended to change. To my knowledge we still do not have an a priori ban on wikis, blogs, and other such things as reliable sources - merely a note that one has to be careful with them. What, exactly, is this intended to add? Phil Sandifer 16:31, 2 October 2007 (UTC)

It looks like a good snippet of an essay to me. Not that I agree with all of it or think that it's the best way to explain how we handle verifiability, sourcing, and NPOV, but if it helps some people put things in place that's great. I can't really see how this would formally fit into any guideline or policy page, though. Wikidemo 16:47, 2 October 2007 (UTC)

Oppose. While I don't disagree with it, it has nothing to do with OR (other than point #5). This would be a more appropriate addition to a policy like WP:UNDUE or WP:RS. OR is about the inclusion of new research by the editor. If research has been published prior to inclusion in Wikipedia, even in one source, it isn't original. COGDEN 18:41, 2 October 2007 (UTC)

Oppose. Experts should be allowed to blog speculatively without everyone assuming every word they utter is fact. What we need is a mechanism to distinguish their serious, rigorous work - i.e. peer review. Hesperian 23:09, 2 October 2007 (UTC)

NOR, discussion pages, and 'Explanation'

Firstly please excuse my temerity at adding another strand here, but allow me to add to the above just this: not in my experience it doesn't, but it should, as and when verifiable. In that regard I still hold to my suggested solution to all these article problems, visible at discussion for Vatican Bank.Now, I haven't reappraised the WP guidelines on the above, which seemed eminently clear when I read them, but I wonder - as I am now again allowed to wonder- at what point does NOR displace 'Explanation' ( the WP/jimbo written concept thereof) on discussion pages? I imagine that both guidelines referred to the mainspace articles, not to discussion pages. It is simple to understand NOR within mainspace, and that it is undesirable, but having been 'busted' and punished for it, and since I never entered it into mainspace, I have to think that I was busted for 'explanation' and NOR on discussion pages. Any research I presented there was in any case verifiably attested and linked, and was provided in support of discussion necessities rather that with a view to altering mainspace articles to include it per se. It was presented by way of explaing why mainspace articles could not be properly edited, rather than for inclusion in any way, and generally in rebuttal of another editor's unsubstantiated claims. To keep it simple, can someone tell me if NOR also applies to discussion pages, and how, and how such application would not destroy the project if denying it hinders even discussion 'explanation', let alone article 'explanation'? ~~ —Preceding unsigned comment added by EffK (talkcontribs) 11:22, 2 October 2007 (UTC)

It was probably not this policy that was intended. Instead, it was probably WP:NOT (specifically WP:FORUM and WP:SOAP). While discussion of the topic is essential on talk pages, the further we move away from discussing the article in relation to its sources in the context of policy, the closer we come to violating that policy. If the discussions revolve around improving the article and its content, within the bounds of policy, there should be no problem. If the discussion is intended, or perceived, as an attempt to justify original research, prove an unpublished point, advocate for a particular point of view, et al, then it will likely be considered unacceptable (and potentially disruptive). One should remain on the former end of the spectrum. Vassyana 17:19, 2 October 2007 (UTC)

The evolution article

Slrubenstein said:

You are going down a rabbit-hole. The issue is not whether the opinion is well-informed or ill-informed - this goes back to the "verifiability, not truth" dictum; it is not for us to pass judgement on the veracity or plausibility of the view. What is important is (1) that we can identify a view (thus the need to peg it to a source, the source helps identify whose view it is) and (2) that it be a notable view. We all hope that the better-informed views are the most notable. But if a view is notable it can be included in an article ven if you or I happen to think it is an ignorant view. This is simple compliance with NPOV. And in our edits of an article we cannot word things so as to promote or arrgue for those views we personally consider best-informec - this is simple compliance with NOR. Slrubenstein | Talk 16:13, 30 September 2007 (UTC)
I believe that everyone in this discussion approves of our wonderful article evolution in terms of accurately representing the scientific fact and theory of evolution. But there are uncountable reliable secondary sources (eg newspapers) that make assertions about the scientific fact and theory of evolution quoting young Earth creationists. These notable views about the scientific fact and theory of evolution (properly) are pretty much ommitted from the article. I believe that the two sides in the debate on this policy page both think their way of wording the article best explains the actual mental process that goes into identifying why the evolution article is a better encyclopedia article the way it is rather than stuffed with notable views about the scientific fact and theory of evolution from young Earth creationists as provided in reliable secondary sources. I don't think NOR is that mental process. I think "newspapers are an unreliable source for science" is far closer. I think very very reliable primary sources on the scientific fact and theory of evolution trumping so-called reliable sources like newspapers is far closer to the actual editor mental processes that tell all of us that the evolution article is more encyclopedic the way it is rather than stuffed with notable views about the scientific fact and theory of evolution from young Earth creationists as provided in those so called reliable secondary sources. WAS 4.250 06:58, 1 October 2007 (UTC)
It is worth saying that the reason I am here having this argument is that someone involved in writing this policy appeared to be using journalistic and POV sources to push a point of view by discrediting contradictory sources as invalid solely due to those being claimed as primary sources. Plenty of secondary sources have a POV to push.
Those who are asking for clarification in this policy are being rubbished as POV warriors, including by notable admins using condescending and abusive comments, yet one of our major concerns is the undermining of a sound principle of policy by having a wording that, as far as we can tell, has the potential to prefer journalism to academic or other notable sources. Spenny 08:12, 1 October 2007 (UTC)
Spenny, Could you give us an example of an academic source that is being challenged on the grounds of being "primary"? That might help people see what you are talking about. Blueboar 11:27, 1 October 2007 (UTC)
Look at the first sentence of Factory farming and read its footnotes. Then see Talk:Factory farming/Archive 3#Random defs of Factory Farming from decent sources and read as much of that debate as you wish. WAS 4.250 11:55, 1 October 2007 (UTC)

I do not see how this becomes a journalism vs. scientific article issue. Both newspaper articles and peer-reviewed journal articles are secondary sources. I have seen many editors claim that newspaper articles are more notable than scientific/academic publications, I guess because they reach a wider audience. This is an important issue but belongs on another policy page discussion. We need to be clear that "notability" is not just between points of view (there are more journalists than anthropologists), the concept is important within a view (i.e. is this a notable view among journalists; is this a notable view among agronomists; is this a notable view among farmers). NPOV insists on notable views but what makes a view notable is not the number of people in absolute terms because a crucial point of view in a debate - say, the view of molecular geneticists in debates over evolution - may be a view shared by all molecular geneticists ... but there just may not be nearly as many molecular geneticists as, say, journalists. But this is something that needs to be addressed on the NPOV and RS pages, I think. It is not relevant to NOR. If someone claims it is, tell them they are wrong - but don't bring their misconcepts to this page. (unless of course it is - I have certainly seen many cases of people drawing on articles published in peer-reviewed academic journals who were yes violating NOR because they were taking data or even elements of arguments out of context and using them to forward their (the editors') own arguments .. is this what is happening at Factory Farming? I thought it was more a notability issue.) Slrubenstein | Talk 12:19, 1 October 2007 (UTC)

It is impossible to say with certainty exactly what is going on in the heads of the disputants at factory farming because each side appears to say things that the other side claims makes no sense to them. But it appears to me that what is happening is that the people who desire its first sentence to claim what it claims (roughly that factory farming=intensive farming=industrial farming=concentrated animal farming instead of some being a subset of another) believe that making claims based on original interpretation of secondary sources is not only ok but trumps contrary statements in primary sources where any expert testimony is considered a primary source. As for "just tell them ...". We have tried for half a year now. Ever been up against a group of admins that back each other no matter what? WAS 4.250 12:31, 1 October 2007 (UTC)
I joined the debate later, but I had a specific example in mind. [3] discussed it on this page previously. My point was that the Inquiry was not just a sound bite source, but a long and detailed discussion of the issues. They specifically dismissed certain points and defined the areas of uncertainty and was itself not original research but a review of other research together with expert testimony. Based on my own views, and the wording of policy here, even as it stands now, it is not a primary source, and yet it was dismissed as such. Note the quote from the editor concerned.
A lot of the argument was about NPOV arguments and synthesis, the correct interpretation of the sources being open to some debate. However, the one point that should not have been the case was the Government Inquiry being an impeccable source being set aside under the guise of being a primary source and hearsay press articles being used to make strong POV assertions about the issue. Spenny 12:37, 1 October 2007 (UTC)

Well, no policy can or should anticipate all situations to which it will be applied - some things need to be debated on talk pages. I am more concerned with how above, WAS seems to love false dichotomies. "newspapers are an unreliable source for science" he writes, and suggests as if it the only alternative, primary sources by scientists. No. The former (newspapers) is indeed an unreliable source on science. But the latter (primary sources) violates NOR. Does this mean we are screwed? Not at all. We use reliable secondary sources by scientists or about science. There is no problem with this. —Preceding unsigned comment added by Slrubenstein (talkcontribs) 12:46, 1 October 2007 (UTC)

I absolutely agree. The problem to me was that there was an arbitrary interpretation of what WAS felt was a sound scientific source into it being a primary source, which is a problem of dubious editing rather than a false dichotomy.
The frustration of that page was a lot to do with one group of editors taking a viewpoint that an ethical debate should dominate the article, because in their interpretation of the subject, supported by a lot of dubious sources, that was the most common viewpoint. Unsurprising that a controversial topic, the controversy should dominate coverage of the topic, but that is not a good test of NPOV.
I wanted a factual exposition of the topic, followed by reporting the controversy, but that was deemed OR, NPOV and all the bad things that you can do in an article. And, as WAS said, there were a lot of bananas in ears, because arguably, in the context of factory farming as a pejorative term, the ethical debate could have been a sound approach for this subset of farming. That argument was undermined by other incompatible assertions.
Trying to set aside the frustrations of that debate, we can see then that having a rule which allows someone to rubbish sound sources simply by application of the rule rather than a healthy, intelligent assessment applying analytical editing skills to assess the worthiness of all available information has to be suspect. Spenny 13:28, 1 October 2007 (UTC)

Slrubenstein, defining a primary source as any source close to the object/claim means some people claim such sources are primary cuz the expert is close to the research/claim/object. SlimVirgin told me I did not understand OR after I told her that just because it comes from an expert does not make it a primary source. WAS 4.250 13:00, 1 October 2007 (UTC)

Anmd then there is a right and a wrong way to use it. I am all for clairfying the policy. But the issue is simply not the absurd choice between using only newspaper articles, or only primary sources. Presenting the dilemma in these polarizing terms is unconstructive, as I have already explained on your talk page several days ago. Slrubenstein | Talk 16:38, 1 October 2007 (UTC)
I think you are dismissing his argument by setting up the extreme case and knocking it down as a straw-man. Take the situation where you have a newspaper article up against a peer-reviewed research paper. The newspaper is secondary, while the paper is primary. If secondary trumps primary, what happens in this case (assume there are no secondary review articles available)? Dhaluza 02:56, 2 October 2007 (UTC)

(<---) (edit conflict) Spenny's BSE example concerning a factory farming edit (war) can also be looked at as using a source that says factory farming is concentrated animal farming and another source that says industrial farming caused BSE and another source that used both factory farming and industrial farming in the same sentence (we say cuz one is a subset of the other, they say no that proves they are the same thing) to justify what I claim is the original research claim that factory farming (in an article that covers primarily concentrated animal farming) caused BSE. WAS 4.250 12:54, 1 October 2007 (UTC)

In practice, several textbooks and other summary sources, such as Futuyma's book Evolution, were used for presenting a number of the basics in the article on evolution. Futuyma's book arguably is a tertiary source that draws on combinations of primary and secondary sources (including prior integrative work by Futuyma). Many sources that plainly are properly termed secondary sources in WP use of the word are also used in the article. On the whole, despite the many primary sources in the form of credible, reliable, published experiments and studies, I'd say what the editors have done at that article is a reasonable mix. To me, the main motivation for avoiding too many primary sources in a broad topic like "evolution" is that it's an extremely large body of primary sources that goes into a concept of this kind, one which requires many redundant primary sources to support the conclusions associated with evolution. AFAICS, no primary sources are used to draw analysis or make additional conclusions directly from the primary sources used in that aritlce. If there are any, it would be useful to cite PSTS and request that the editors find secondary or tertiary sources that report the particular analysis or conclusion of that primary source. At present, to me there appears to be adequate use of secondary and tertiary sources to bind the article together into a coherent presentation without original research. Thus, far from being an exception to PSTS, the article on evolution is a very reasonable example of applying PSTS in practice.... Kenosis 14:56, 1 October 2007 (UTC)
Half the trouble at Factory farming is the lack of a good core source to hang the article on - sourced based research based on Google does not make a good methodology. In part, I think this is why there is a disconnect between those here instigating for change against those who don't see the problem. Those who see problems see a world of unsatisfactory sources combined with a variety of editing approaches and are looking to the guidance to send that in the right direction. If everyone used good sources and good editing, we wouldn't be having this discussion. Spenny 15:15, 1 October 2007 (UTC)
What??? The best sources aren't the first ones Google returns??? Heaven forbid! Okay, sarcasm off. At times it does appear though that many editors feel that way. Whatever they find first has to be the best. wbfergus Talk 15:29, 1 October 2007 (UTC)
Holding that thought, I think a couple of years ago, Wikipedia could cope with that environment. SOmeone who knew enough about the subject would marshal an argument and hold it together on a page. Consensus would be used to batter the worst excesses out of the topic. Moving to the citation model, there are a raft of subjects which are quite notable but may not have a key text. Perhaps we haven't got the right approach for coping. By trying to use a citation model in the initial building can lead to POV by source manipulation which seems to be to be harder to deal with than a good old argument. Spenny 17:35, 1 October 2007 (UTC)
Yes, the approach apparently does need to have several prongs involved. WP:NPOV takes one slant. WP:V takes another. We forget sometimes that WP:V says "verifiability", not "already verified in a citation immediately upon placement in an article". So, originally at least, the expectation was that in response to a questionable statement or set of statements in an article, someone else could come along and say "OK, prove it". As the wiki quickly grew to the present stage, drive-by "prove it" templates (unreferenced, "weasel", and "citation-needed" or "fact", etc.) became more common. Enter WP:NOR, where the statement "I gave you a citation" often is not adequate to fulfill WP's mission. Which brings us quickly to the kinds of debate we've seen on this talk page. And consensus process runs throughout. Since there's no one way that all articles develop across the wiki, sometimes we get to calling an article or set of statements within an article "original research" without even passing through WP:V on the way. Sometimes the discussion at the article goes "OK, here's a source", followed by a response like "That source only verifies fact A but not conclusion B", and/or any of a wide range of variations on this. All this must be mediated by consensus to arrive at compliance with the three core content policies. And the approach does seem to work on the whole, especially taking into consideration that no one's getting paid a dime to do this. ... Kenosis 19:11, 1 October 2007 (UTC)
The main issue is that articles ought to comply with all core content policies, as said in the lead of this page: 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. ≈ jossi ≈ (talk) 20:58, 1 October 2007 (UTC)
I very much agree, and a lot of this discussion has not been contradicting that: it has been about saying that the presentation as it stands has undermined that, especially when versions of the policy have been bent into saying 'ban primary sources'. I'm coming to a view that says that providing people with a tool to set against all policy through the use of source typing could be helpful. I just think it is a distortion to try and fit source typing to a solution to OR. Spenny 22:12, 1 October 2007 (UTC)

In the evolution and factory farming articles, I have not seen any true argument relating to OR. Nothing that anybody has argued in either article is original. It's all been said before. There may be issues of WP:V and WP:NPOV, but not OR. When arguments start cropping up that cite the OR policy page for disputes such as these, you know there is a problem with how the OR page describes the policy.

That said, I do agree with Spenny and others that some helpful guidelines relating to using primary and secondary sources might be helpful in the RS and NPOV realms. They just really have nothing to do with OR, and shoe-horning the primary-secondary distinction into OR is counterproductive. It's bad to go beyond any source with new material, be the source primary, secondary, or otherwise. No exceptions, no distinctions, no weakening of language. If you cite a fact, your source has to contain that fact. If you cite a conclusion or opinion, your source has to have that conclusion or opinion. It's a simple rule that even the simplest editors should be able to follow. COGDEN 23:35, 1 October 2007 (UTC)

Yes, you've made your opinion clear. I already agreed there's an argument to be made for PSTS having more of a shared relationship between WP:NOR and WP:V. This is why I had no objection to the transclusion suggestion made earlier. However, it is definitely not the case that, as you've said, it has "nothing to do with OR", because it has a great deal to do with OR, and also something to do with WP:V and even WP:NPOV. When the issue revolves around which sources constitute valid verification of a statement or set of statements in an article, it's still at least as squarely within WP:NOR as it is within WP:V. And where the issue is conflicting opinions of sources (often also involving differing POVs among editors), it's still at least as squarely within WP:NOR as it is within WP:NPOV. Under no circumstances should PSTS be reduced to an adjunct status or completely split out of this page, because it's integral to WP:NOR, as it has been for three years now (originally as "primary/secondary"). The underlying crux is that WP is neither a publisher of primary-source material (raw facts or new opinions), nor of original synthesis of any kind. Both of these issues are explicitly part of WP:NOR, as they have been for a long time now. And, what you refer to as "shoehorning" is not counterproductive, but rather is quite integral to the policy. ... Kenosis 00:21, 2 October 2007 (UTC)
And you have made your opinion clear too. But you have not made a cogent argument that rises above it's really important because it's been here for a really long time. The fact is that originally it said "Wikipedia is a secondary source," and the current definition of primary source was significantly changed recently without significant discussion. If NOR means Wikipedia does not publish either primary or secondary source material as you assert, then why do we need to make the distinction? You have just made the case for it being an unnecessary distraction. Dhaluza 01:07, 2 October 2007 (UTC)
OK, tell you what. If you present a brief, cogent argument for why a tripartite government (executive, legislative and judiciary) is superior to something simpler, like, say, "Vinnie here is the Boss", or "Vinnie and the Committee here are the boss", I'll consider another significant effort in response to the way you've framed your query. But, you must also explain cogently why it is necessary for the legislative branch alone to be divided into two branches, one of them proportional to the population and the other one proportional to the number of states, and also explain why the judiciary's authority is respected despite that it has no inherent enforcement power. (Hint: it's not as simple as "because that's the way it was decided".) ... Kenosis 01:57, 2 October 2007 (UTC)
If you were trying to make a point, it was completely obscured. My point was much simpler. If you say C can be divided into A and B, but you can't do A and you can't do B, why not simply say you can't do C? Dhaluza 02:43, 2 October 2007 (UTC)
I think Kenosis is saying that many people, including himself, believe that experience shows that our current content policies, including NOR, are working pretty good and they don't want any major changes that might possibly mess that up (which is what killed ATT also). I those those with a problem with the word "primary" in this policy have had experiences that shows that the use of that word in this policy made articles worse and so is something that should be fixed. Maybe the proposal in the next section can be something that fixes this policy for us without breaking it for you. WAS 4.250 04:22, 2 October 2007 (UTC)
Actually, Kenosis, if your major concern is as you state it above, I would agree. I have no problem with the following principle:
  • Wikipedia is neither a publisher of primary-source material (raw facts or new opinions), nor of original synthesis of any kind.
If what we are saying is that Wikipedia is only a secondary source, never a primary source, I agree with that. I don't think it's necessary, but I have no problem stating the policy that way. If we can come to consensus along those lines, then I'm on board. COGDEN 18:33, 2 October 2007 (UTC)
Sure. If the participants here can come up with a way of expressing this so that inexperienced users (and some of the more experienced ones with "blinders on", so to speak) aren't more likely to be mislead into confusing original sythesis with original prose and editorial decisionmaking about how to present topics to the reader, that's certainly fine by me. The rest of the particpants would need to speak for themselves. But I imagine something like that can probably be worked into the page in such a way that it helps users' understanding without changing any significant aspect of the existing policy. ... Kenosis 20:24, 2 October 2007 (UTC)
Do you think that the proposal below by Wbfergus does this? Maybe this is a good starting point, and if we want to add more material, we can do it a piece at a time, just so we can get something that's a consensus up on the page. COGDEN 21:33, 2 October 2007 (UTC)
In my opinion, the way you said it is just fine: "Wikipedia is neither a publisher of primary-source material (raw facts or new opinions), nor of original synthesis of any kind." To which one might add something like "Original research, or original sythesis, should not be confused with original prose and editorial decisionmaking about how to present topics to the reader in the editors' own words," Two sentences. I think surely we can find a place for something like that. ... Kenosis 21:47, 2 October 2007 (UTC)
I especially like that second sentence. I think there is something useful there which is not currently in the policy page and it is an issue that should be discussed (if only because good editing and synthesis are potentially so close together, especially when you add in NPOV and fairness of tone). Spenny 23:39, 2 October 2007 (UTC)
What has to be recognized with respect to the evolution article and with respect to the public debate is that the evolution article is about evolution-within-science. Science has a set of disciplines and practices. Opponents of evolution very often come to the debate from a philosophical point of view (where philosophy and religion can overlap.) That a particular creationist attitude is "notable" from the philosophical point of view does not make it "notable" within science. Science doesn't have to deal with such attitudes, claims, theories, objections: they aren't within science, do not adhere to the principles of science. Similarly, the evolution article need not mention such non-scientific material. Perhaps Wikipedia cold or should have an article about the conflict between philosophy and science but there is no NPOV requirement that every scientific article need make not of philosophical objections to the concepts of that article.
In general, the anti-evolution attitudes, concepts, or arguments arose outside science and have not been subjected to scientific examination in any of the appropriate journals. quite often (maybe always) this is largely because the attitudes, concepts, or arguments cannot stand up to scientific examination. The sources for such material are not notable scientific sources. Within a scientific article all these fail verifiability.
Philosophers may not like the above, may want to insist that their intellectual tradition is the prime intellectual tradition and rules all others. That is unsubstantiated, particularly in the case of science (and in particular of philosophy being superior to science.) The evidence is on the other side: philosophy gave us phlogiston, philosophy gave us earth, air, fire and water as the principles. Knowledge was held back for millennial because of failed and incorrect philosophical notions. Philosophy's track record is a bad one when it comes to science.
Evolution is a concept within science. Scientific articles need not treat as notable anti-evolution ideas from outside science. The evolution article is a scientific article. --Minasbeede 02:01, 4 October 2007 (UTC)
Minasbeede, science depends on the formation and replacement of provisional hypotheses. That's what the four-element theory, phlostigon, etc. were: provisional hypotheses of different eras. Other kinds of philosophy gave us logic, math, Occam's razor, Galton's problem (the distinction between correlation and causation), etc. Jacob Haller 02:23, 4 October 2007 (UTC)
that is the first (probably only) time I heard them called "provisional hypotheses." As late as the time of the 1st Encyclopedia Britannica the four principles were cited - in the article on chemistry. They were not cited as "hypotheses," they were cited as fact (although surely as "revealed fact.") The author of the EB article did show some doubt, for which he deserves great credit.
The operative phrase for well over 1000 years was "Aristotle dit." That didn't mean "Aristotle has formulated this provisional hypothesis." In practice it meant that inquiry was stagnant for centuries. Philosophy's track record in science is poor. --Minasbeede 02:31, 4 October 2007 (UTC)

Hopefully, Final version of this proposed addition

Taking into account the three evolutions of this proposed additon, and the various comments following each one, does this finally take into account each person's interpretation?

Wikipedia is not a primary source. Original research that creates primary sources is not allowed. Original research consisting of novel synthesis, or new interpretations, of reliable sources is not permitted. Conducting background research that consists of collecting and organizing information from existing reliable and verifiable sources is, of course, strongly encouraged. All articles on Wikipedia should be based on information collected from reliable published sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. However, in presenting such information in a Wikipedia article, great care should be taken not to take the information contained in the sources out of context or present it in a manner that violates the provisions of this policy.

If this is the correct and final interpreation of the various comments above, do we have consensus to add this to the article? wbfergus Talk 14:32, 2 October 2007 (UTC)

Agree

Oppose

  • Please explain how this statement would weaken the policy. This is not really a "poll" or a "vote". Unless you have reasonable explanation behind what you are saying, it doesn't count much for purposes of determining whether the above language is consensus language. COGDEN 19:04, 2 October 2007 (UTC)
  • Per above. Much, much simpler version: "While Wikipedia does not allow original research, it does encourage original prose." That's all that needs saying. Marskell 17:06, 2 October 2007 (UTC)
    • I like that formulation much better, it's very clear and sets out the balance between WP:COPYVIO and WP:NOR well. I would like to see this part about what to do put back as well: "Collecting and organizing information from existing reliable and verifiable sources is strongly encouraged – this is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia." Dhaluza 23:25, 2 October 2007 (UTC)
  • Is your concern that the proposal is too wordy? If it's just a style issue, I don't think that's much of an issue. We can always implement the language and tweak the style later. What we want to know here is if there is any fundamental disagreement with how the policy is described above. If not, I don't think stylistic issues are impediment to obtaining consensus. COGDEN 19:04, 2 October 2007 (UTC)
  • I'm not so sure we can edit it later based on the edit and protection history here. Some misguided editors are liable to say that this exact text is the consensus, and you need to show a new consensus to change it. In addition to wordiness, it's too complicated. It needs to be something someone can get a basic understanding from immediately on the first pass. Most readers will not take the time to parse this out. Dhaluza 05:43, 3 October 2007 (UTC)
  • It's usually much easier to get consensus on stylistic and readability changes that don't alter the fundamental meaning. I'd just like to get something in place. COGDEN 18:30, 3 October 2007 (UTC)
  • Fhat the wuck? "Wikipedia is not a primary source"? Are you daft in proposing this as an explication of policy? What has that to do with anything? Wikipedia cannot be used to support Wikipedia. Period. Besides, I can find loopholes all over the place in that detrital piffle. &#0149;Jim62sch&#0149; 22:29, 2 October 2007 (UTC)
Why do you consider yourself above the house rules on civility? Spenny 23:15, 2 October 2007 (UTC)
  • I would like to see the lost points restored, but this is just too wordy. For example, see how two wordy sentences can be combined in my comment above. I also think this should be up front, before the PSTS section, and therefore should not rely on its definitions. Dhaluza 23:29, 2 October 2007 (UTC)
    • This isn't computer code or a legal document - we don't have to define terms before we can use them. Links are also very useful. SamBC(talk) 23:50, 2 October 2007 (UTC)
      • No, it's a policy that needs to be understandable to someone who may not be spending as much time pouring over the text as we do. Making the content clear and easily understandable should be a primary goal, not a secondary one. Dhaluza 00:04, 3 October 2007 (UTC)
        • Personally, I don't care whether or not it's too wordy. That's just a quibble in my view. Some people will always disagree on the best stylistic or literary expression—we shouldn't let disputes on style or readability bar progress towards getting some sort of consensus on the fundamental idea. COGDEN 18:18, 3 October 2007 (UTC)
  • Ah, I see the trolls are back again. wbfergus Talk 16:45, 3 October 2007 (UTC)

Comments and questions

  • I may be missing something, but what does this add to the policy that isn't already in place? Phil Sandifer 16:48, 2 October 2007 (UTC)
  • It's a handy summary... I'm not certain, but I believe it's largely replacing some stuff that was lost a while back, and we couldn't find discussion of the removal. I might be mixing it up with something else.
    • As the above poster said, this somehow was lost from the policy during a series of edits a while ago. Something similar to this was present in the policy for well over a year before it suddenly vanished, with no mention on the discussion pages. We can only assume that it was an inadvertent deletion during the series of edits, and nobody ever caught it. This one little paragraph clarifies some of the 'sore points' various people have had with the policy, where some people claimed that the current policy "almost" prohibited using primary source data. This is all covered several 'discussion' sections above for further detail. wbfergus Talk 19:01, 2 October 2007 (UTC)

Finally, if this meets consensus, where should it be placed? At the beginning of the "Sources" section , before the "Reliable sources" sub-section, or someplace else? wbfergus Talk 14:32, 2 October 2007 (UTC)

As I've noted in my "agree" above, I think it would fit nicely as an intro to the sources section. SamBC(talk) 17:28, 2 October 2007 (UTC)

Short version

Here is a shortened version that will hopefully address some of the concerns above. Thoughts? Vassyana 16:17, 3 October 2007 (UTC)

Wikipedia is not a source for original or novel claims and facts. While Wikipedia does not allow original research, it does encourage original prose (or paraphrasing of published reliable sources). However, great care should be taken to present the information in context and in a manner consistent with this policy.

Erm, well it doesn't really speak to sources (not that I am going native, you understand!), it is more of a nutshell for the policy. Spenny 17:48, 3 October 2007 (UTC)
I'm fine with this version too. I don't care how tersely or baroquely we express it. I'm on board with anything that says roughly the above, if it will make progress toward consensus. COGDEN 18:23, 3 October 2007 (UTC)
Nice, neat and clean. I would also like to see the "...this is source based research..." statement too. Dhaluza 00:27, 4 October 2007 (UTC)

{{Editprotected}}Could some administrator include a (interwiki) link to Swedish Wikipedia article "Wikipedia:Ingen egen forskning". Add [[sv:Wikipedia:Ingen egen forskning]]. Best regards, Pierreback 20:28, 3 October 2007 (UTC)

 Y Done--cjllw ʘ TALK 23:56, 3 October 2007 (UTC)

A separate problem with the synthesis section

This would never have occurred to me, but now that I've seen it happen, I understand how it could. The example that's used for explaining what constitutes inappropriate synthesis happens to be about plagiarism, which is a related concept to the subject of this page, original research. This has apparently caused someone to come to the conclusion that plagiarism is synthesis -- see Wikipedia:Articles for deletion/Shadowrun timeline. Like I said, this isn't something that ever would have occurred to me as a potential problem, but since it's happened at least once, maybe we should find a different example that's very clearly unrelated to the actual topic of the policy. Pinball22 13:59, 3 October 2007 (UTC)

I've been resisting, as we are all a little bamboozled, but over the past 6 months there have been a few attempts to clarify this section. It is not plain English and there is a significant viewpoint that the example is more confusing than the issue that it is trying to clarify.
Unfortunately, there appeared to be some strong ownership suggesting that the example was the pinnacle of perfection, so a number of efforts to address the issue fell on stony ground. The archive shows something of the last set of discussions. Sections 15 & 16 cover the last attempt at an example, and some thoughts on wording. Spenny 16:35, 4 October 2007 (UTC)

Identifying the ground floor

We've had quite a large amount of discussion about WP:PSTS. A lot of it has gone in circles and much of it is muddled and buried in endless debates. We need to focus on what we're centrally trying to achieve in such a section, or appropriate replacement. We can focus on how to address those points after we establish what we're trying to say. My previous comments help clarify my position and concern in this.[4][5][6] Let's focus on determining the ground rules of what we're trying to address here. Please address the following questions, without addressing questions of terminology, as briefly and succinctly as possible. I've split them into two groups, one primarily addressing sources themselves and the other addressing issues that have repeatedly come up in relation to the discussion. Thanks! Vassyana 17:01, 3 October 2007 (UTC)

Sources questions

  1. Are references that are mainly organized collections of "raw data" a particular concern in relation to NOR?
  2. Are sources close to the subject especially a concern for NOR?
  3. Are historical references related to the subject a notable concern for NOR?
  4. Are obsolete sources a serious concern in relation to NOR?
  5. Are sources that make novel or original claims particularly a concern for NOR?
  6. Are there other types of sources that may be deeply problematic in relation to NOR?
  7. What is the best way to handle sources of concern?

Discussion

  • Given that there are over two million articles of widely ranging approach and type of topic on the wiki, the questions are essentially unanswerable in anything less than a tome. "Yes" to #1 through #6, with the exception of #4, to which the answer is "not normally". For question #7, use the existing WP policies of WP:NPOV, WP:V, WP:NOR, and the guideline WP:RS, in keeping with WP:Consensus and other applicable policies. ... Kenosis 17:33, 3 October 2007 (UTC)
  • Short answer: yes to #1, no to #s 2-6. Long answer:
    1. Collections of raw data can be a particular problem in relation to NOR. But it's not a widespread problem: citing raw data is pretty rare, and improper citing even rarer. Much more common are improper synthesis of multiple sources, and citing previous research and conclusions and taking them a step further.
    2. Closeness of the source to the subject is a possible NPOV concern (the source may contradict more distant sources), but has nothing to do with OR.
    3. Historical references related to the subject have nothing to do with OR.
    4. Obsolete sources have nothing to do with OR. If it's obsolete research, it's old research, not original research. There might be NPOV concerns, however (e.g., failure to express the modern POV, or giving the obsolete view undue weight).
    5. Sources making novel or original claims are OR themselves, but citing them is not.
    6. I don't know of any deeply problematic categories of sources. OR abuse happens with all types of sources.
    7. Possible Solution: Make it clear that you can commit OR by using any source improperly, that no category of (reliable) source is inherently bad, it's all in how you use the source, and give examples of how OR can arise in the use of various types of sources (like raw data, quotations, scientific conclusions, newspaper editorial pages, reference works, etc.).
COGDEN 19:07, 3 October 2007 (UTC)
My answers:
  1. Merging data requires special care. It's not inherently OR but is likely to lead to OR.
  2. NPOV yes, NOR no.
  3. NPOV sometimes, NOR no.
  4. It depends on the field and its development. Old sources are not necessarily bad ones, but obsolete sources, or obsolete claims in still-useful sources, should not receive undue weight (I think we can speak of NOR undue weight as well as NPOV undue weight).
  5. Not as such.
  6. I've noted the "special definition" problem. Numbers claims in ancient military history. Dictionary definitions.
Jacob Haller 21:44, 3 October 2007 (UTC)
  • Rather than obsessing on PSTS, I think we should stick to basics, and instead work on what NOR is about and what it's not. Don't assume everyone already knows, because the discussion on this page and throughout the project show it's far from clear. I think if we work on that first, PSTS will be easier to sort. Dhaluza 00:35, 4 October 2007 (UTC)

Ancillary questions

  1. How does verifiability and reliability relate, or not, to a prohibition on original research?
  2. Is it helpful for this policy to note the preference for independent reliable published sources (aka "reliable secondary sources", "third-party published sources with a reputation for fact-checking and accuracy")?
  3. How does NPOV relate, or not, to a prohibition on original research?
  4. Is using information in context important to preventing NOR?

Discussion

  • To questions #1 and #3: They all dovetail with one another. To #2, absolutely yes, because many topics have a broad mass of primary sources that can too easily be cherrypicked to arrive at a presentation that is inconsistent with the summary analyses of the relevant secondary and/or tertiary sources. (See also the answer to questions #1 and #3). To question #4: Yes, using information in context is important to all three core content policies. ... Kenosis 17:44, 3 October 2007 (UTC)
  • My comments:
  1. The concept of OR is closely intertwined with verifiability. Anything that is original research is, by definition, unverifiable. (The converse is not true.) Also, many if not most concerns that people frame as OR issues are actually V issues. For example, an editor might interpret a published paleontology paper in a way that supports some crackpot creationism theory. If that crackpot theory has been published before, however, it's not OR. Unless the editor cites a prior source of the crackpot theory, the problem is V, not OR. A V problem arises when the editor fails to cite sources. OR arises when there are no sources to cite.
  2. I think that too much focus on reliability here is likely to confuse the issues. We should note that sources have to be reliable, but focus here on "going beyond" the reliable sources.
  3. NPOV is a separate concept that can be entirely divorced from OR. Unlike WP:V, something can be non-neutral without being OR, and can be OR without being non-neutral. I think it's more helpful to entirely separate the concepts, rather than muddle them together. However, NPOV issues can and do arise with respect to citing sources. Most source-related problems have to do with V or NPOV. True OR is relatively rare, compared to V and NPOV problems.
  4. Using information out of context is mainly an NPOV issue (because it fails to reflect the POV of the author). However, it's possible to support a novel conclusion by citing something out of context. Then it becomes a double NPOV-OR problem.
COGDEN 19:33, 3 October 2007 (UTC)
  • Mine:
  1. OR straddles V and RS. OR covers unreferenced claims, referenced claims with unnreliable sources, and referenced claims with reliable sources - especially if some of the sources have been misused.
  2. Yes. However, such sources may not exist, or two or more groups may have rival sets of secondary sources, regarding their own as reliable and their opponents' as misrepresentation.
  3. I'm not sure that it does.
  4. Usually.
Jacob Haller 21:28, 3 October 2007 (UTC)
To me:

I am a 'liberal, strictly convinct that only dialectic facing will be cleared issues about the stuff published by Wiki and no policy can make orders about reliability only descerning about 'sources'. It's the capability to elaborate and mind onesty that make the difference.

Now to the issues: let me allow to show what i think with an example: i have many old sources, like old articles, old magazines, old enciclopedies (do not forget, wiki.en took off thanks an 1911 encyclopedia, so before spit over sources, be careful). Now, if i grab moden sources like magazines, internet sites, and so on i will seen bi-dimentional descriptions and opinions about this and that. Example, the stuff you can read about a modern aircraft is not seldom totally different to that you can find on old sources: today its' difficult to resume all the materials, the issues, the developement datas and so on.

But with an old article you can find a lot of them. Example, A-10 Thunderbolt, before Desert Storm were considered almost a failure of aircraft (too slow, without night capabilities etc.), after they were better evalued and today are one of the 'oustanding' aircrafts.

AMX, on the other hand at mid '80s were seen as 'the aircraft that will conquered the world'(eng Da Silva, 1985, said). After 20 years of marketing there is still no export success and that aircraft was often at the center of fiercely polemics.

So if you take a modern article (to not to talk about websites) i'll find just the last developements and few or nothing about 'the old, already well known history'. In fact, it's worthing to remember 'old history' to learn how that aircraft was seen, and why today it is consiedered well or bad (who remember also the fierce debate about the cost of M1 Abrams, M2 Bradley, B-1B and so on?).

Obviousely it is needed to understand that you cannot base articles only on 20 years old article, but integrate with modern stuff. The results are tridimentional, better, quadri-dimensional. It's the classical reason that worths to study history, to know how and why things are like we can see today.--Stefanomencarelli 15:40, 4 October 2007 (UTC)

Archiving

I took the liberty of archiving the slightly older discussions to archive 25, the talk page was 325kb pre-archiving, it's now 177kb. Please restore any still relevant discussions as needed. Dreadstar 17:10, 5 October 2007 (UTC)

Request edit

Could two spelling mistakes be corrected:

Secondary and tertiary sources should be used in a way that does not give rise to new analysies, syntheses or original conclusions that are not already present in the sources. In short, the policy of "No original research" requres that wikipedia users '''stick to the sources.'''

to:

Secondary and tertiary sources should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources. In short, the policy of "No original research" requires that wikipedia users '''stick to the sources.'''

Only small things really but it doesn't look very professional on such an important project page. --Lo2u (TC) 20:00, 6 October 2007 (UTC)

Agreed. ... Kenosis 20:50, 6 October 2007 (UTC)
Done. Dreadstar 21:53, 6 October 2007 (UTC)
Thanks a lot --Lo2u (TC) 21:59, 6 October 2007 (UTC)

SYN

I think Wikipedia would benefit a lot from attenuating WP:SYN. —Preceding unsigned comment added by 70.248.146.109 (talk) 05:07, 7 October 2007 (UTC)

Focus and awareness

We need to decide what the focus of PSTS, or any replacement, will be set upon. Are we concerned about "raw facts" or "purely factual" sources? Are we concerned about sources close to the subject? Are we concerned about both? There seem to be two parallel trains of thought being expressed in regards to this.

I feel that PSTS, or any replacement since it will be built upon its legacy, needs to take into account the differing accounts of the P/S distinctions and their purpose as a framework for original research. We are trying to create unified definitions for a policy that discourages original research. This apparent dissonance to the original purpose of the terms/concepts and the variety of definitions needs to be consciously taken into account and addressed in the section.

We need to additionally bear in mind the context of the other policies, which put a strong preference on independent references that have solid editorial oversight. Ancient histories, diaries and similar sources are often considered "primary sources" and lack the editorial oversight that is expect of reliable sources. On the other hand, some typically "primary" sources (such as the U.S. census or EU economic statistics) are generally considered very reliable, but also "very raw".

Thoughts? Vassyana 23:42, 29 September 2007 (UTC)

Vassyana I agree with your position and assessments. Perhaps it would be helpful if we were all using the same definition for Primary and secondary sources. From where I stand and where I understand policy to be is that use of PS beyond description has to be original research because primary sources are "materials on a topic upon which subsequent interpretations or studies are based." this definition is held in The Scott, Foresman Handbook for Writers (1996, p. 547), it's a fact that I've lived with as an academic for years since it's the mainstream scholarly standard[7][8][9]. IMHO PSTS should explain how to use primary sources while avoiding OR--Cailil talk 00:43, 30 September 2007 (UTC)
I think we need to primarily focus on explaining what NOR is and is not, and all other considerations here are secondary. Yes we need to integrate this with other policies within tne existing framework, but we should not be overly concerned with what happens to things that may be useful, but are not useful in the context of NOR. Dhaluza 01:12, 30 September 2007 (UTC)
Improvements in explaining what NOR is and is not will be welcome, but we also need to be fully concerned with the overall effect of any changes on all Wikipedia policies taken together, and with the need to ensure continuity in editors' understanding of the terminology of policies. Improved explanations can be added without deleting useful and well tested aspects of the policy. .. dave souza, talk 08:52, 30 September 2007 (UTC)
Useful in what context? If it is useful in the NOR context, then it belongs here. If it useful in another context, then it belongs somewhere else, because it detracts from this page's purpose. If the contexts are related in some direct way, then we can have cross links to improve understanding. Yes the polices must work together, but if someone added useful content to this page in error, its usefulness is not a valid reason to perpetuate the error. Dhaluza 09:15, 30 September 2007 (UTC)
Do you want the "well tested aspects of the policy" evident in the first sentence of Factory farming to be made manifest at evolution? Leave this policy as it is and there is then precedent for changing evolution to say what newspapers say about evolution. Is that what you want? WAS 4.250 09:29, 30 September 2007 (UTC)
Not just that, a big problem I see is that the non-consensus policy would actually favor an evangelical creationist's published interpretation of a scientific article on a topic such as the second law of thermodynamics, above the peer-reviewed article itself. What kind of crazy policy would give a blanket precedence to what some crackpot says about new published research, above the raw research itself? COGDEN 22:33, 1 October 2007 (UTC)
I agree, Vassyana. That's a helpful way to look at it. We really need to focus on the fundamentals of the OR policy. Let's jettison all this secondary baggage that has grown up around the particular PSTS expression of that policy, and get back to the basics of what the OR policy is about, and not try to revolutionize the world here. COGDEN 22:33, 1 October 2007 (UTC)
Seems to me that the issue of whether primary or secondary sources is used must be orthogonal to the issue of whether a given article is "original research" or not ... so I'm one more person agreeing with Vassyana. These distinctions make sense for historical research, but not for harder sciences (and technology) where "primary" sources are strongly preferred and "secondary" sources are at best imperfect filters which can introduce errors. If the source is verifiable and reliable (and, for secondary sources, reputable), why should the distinction between primary and secondary matter at all? It's not like biases are are only introduced by primary sources, and never by secondary sources; or that only primary sources can be misused/twisted. Especially nowadays, where "traditional media" have largely abandoned the processes which earned their reputations, but are still often (and wrongly!!) treated as "reliable" sources, it's very important to be able to use primary sources to separate the chaff. --69.226.208.120 06:20, 8 October 2007 (UTC)

Direct quotes and special definitions

I'd like guidance on the use or misuse of direct quotes.

Currently, Criticisms of socialism directly quotes Mises' Socialism, where Mises makes a categorical statement about "all socialist ideas." I've disputed this section and discussed certain problems on the talk page. Basically, Mises uses his own definition of socialism for his critique and declares that most socialists aren't socialists, so it's not always clear what he's denouncing when he's denouncing socialism. Jacob Haller 22:12, 6 October 2007 (UTC)

I feel a factory farming discussion coming on: misrepresenting or misreading the meaning of sources by making assumptions about what particular words mean to advance a position.
It is another example of where, if you don't know the detail, source based research, especially cherry picking sources rather than using them as a whole, can lead you astray. No doubt those you discuss this with don't understand the problem and could proclaim your "specialist knowledge" is being used to interpret the source making it OR to dispute the source.
I guess we can see that Mises is a secondary source on socialism, or are we saying it is primary source as the subject is primary source for criticisms? Just wondering whether policy works for the example (assuming you are correct). Spenny 08:51, 7 October 2007 (UTC)
Mises, like Marx, is a notable far from center biased extremest expert whose writings push a political agenda. Quoting them is inappropriate if context does not make clear their non-consensus standing among experts and the fact that their writings on economics pursue a political agenda (Mises is right-wing and Marx is left-wing). WAS 4.250 15:00, 7 October 2007 (UTC)
That makes sense. That is "specialist knowledge". How does policy square that circle? Spenny 16:35, 7 October 2007 (UTC)
What's typical and is far from useful is wiki-lawyering and revert wars. What works is addition of data and sources so that by reading the added data and sources one becomes knowledgeable enough to make one's own mind up. Solutions that add content, in the long run, help build wikipedia. Solutions that involve deletion are sometimes necessary but are not a reliable strategy for building an encyclopedia and are best used as a last resort. Trouble is that deletion and locking a page down are so much easier than adding sourced content. Especially if one feels responsible for (owns) a hundred or so articles. Such a burden! WAS 4.250 17:09, 7 October 2007 (UTC)

Jewish Bolshevism

{{Original research}}

  • The above is just antisemitism pure and simple. Nevertheless, some editors have taken it upon themselves to try to explain why Jews were, or were not, Bolsheviks.
    • I find the Template, {{:Original research}} insufficient because it does not Flag that kind of improper editorial policy or behavior. I certainly am aware that no reputable scholar has found a need to explain why Jews (some or many) were, or were not, Bolsheviks. To do so tains the article in question not only with a POV violation, but also suggests to our readers that Jews have a need to explain their "misbehavior."
      • Accordingly, we need to have this template say that Wikipedians are not to bring in explanations which scholars themselves have not done. --Ludvikus 10:22, 7 October 2007 (UTC)

I'm quite honestly confused. If someone is inserting original research, or there are reasons to believe they are, the tag is perfectly appropriate. It doesn't change meaning based on the article, nor do I understand why you seem to present it in that light. Additionally, the article seems very clear in the that it is discussing "an antisemitic political epithet" that is also "the antisemitic conspiracy theory which blames the Jews for Bolshevism ...". I'm not at all sure what your concerns are. Vassyana 19:03, 7 October 2007 (UTC)

I have cleaned up some dubious material that was unsourced. The article now seems to be well sourced and neutral, presenting facts and attribution opinions as it should. ≈ jossi ≈ (talk) 19:07, 7 October 2007 (UTC)
The article should be merged with General Jewish Labor Union, to provide the necessary context for NPOV. Proposals for merge placed on these articles. ≈ jossi ≈ (talk) 19:11, 7 October 2007 (UTC)

Here's the problem:

  • The Jewish Bolshevism is the title of a pamphlet about which we already have an article.
  • 20th century Antisemitism blames the Jews for everything wrong with the world one could think of, including Bolshevism.
  • There is no coherent theory, or even a conspiracy theory; there's only the claim that Jews and Bolsheviks are one.
  • No reputable scholar has written any significant tract analyzing the relationship between Jews and Bolsheviks.
  • Nevertheless, Wikipedians are spending an a lot of "(electronic) ink" writing about such thing as the Jewish Bund in order to explain why Jews were, or were not, Bolsheviks. That the original research. It is not for Wikipedians to do such accounting. Furthermore, the subtext is that Jews have something to be ashamed about, something which requires explanations by Wikipedians: How many, and Why, Jews were, or Were not, Bolsheviks.
So fix it. ≈ jossi ≈ (talk) 01:43, 8 October 2007 (UTC)

A slight rewording of the proposal

Taking into account the comments above... I have slightly reworded the proposed text:

Per WP:NOT#OR, Wikipedia is not a primary source. Original research that creates primary sources is not allowed. Conducting background research that consists of collecting and organizing information from existing primary and/or secondary sources is, of course, strongly encouraged. All articles on Wikipedia should be based on information collected from reliable published sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. However, in presenting such information in a Wikipedia article, great care should be taken not to take the information contained in the sources out of context or present it in a manner that violates the provisions of this policy.

Thoughts? Blueboar 14:40, 30 September 2007 (UTC)

Again I have problems with "Wikipedia is not a primary source. Original research that creates primary sources is not allowed." I still prefer something along the lines of "Original research consisting of novel synthesis, or new interpretations, of reliable sources is not permitted." Can't support this in its present form but this is my only major objection, I would change a few words but those are minor concerns. On the whole I prefer this version to the other two above--Cailil talk 14:50, 30 September 2007 (UTC)
The first sentence is straight from WP:NOT#OR, so I don't think that should change. I wouldn't object to the other change though. wbfergus Talk 14:58, 30 September 2007 (UTC)
Same objection as I made above. Vassyana 16:07, 30 September 2007 (UTC)
Three points. (1) in response to Calil: I see nothing wrong with keeping this sentence "Wikipedia is not a primary source. Original research that creates primary sources is not allowed." I also agree with you that something like this sentence - "Original research consisting of novel synthesis, or new interpretations, of reliable sources is not permitted" is also needed, in other words, it doesn't have to be one or the other, it can (and should) be both. (2) I agree with Vassyana that the more accurate statement would be to say that we encourage research based on reliable sources. In this particular context, it is meaningless to say "primary and secondary sources" because (and in this I think I echo a point WAS has made) in this context it just doesn't matter whether something is a primary or secondary source. (3) Vassyana's prefered statement is really just a recapitulation in a highly abbreviated form of V and RS. I have no objection to including such a recapitulation in this context because NOR is related to V and RS. But Vassyana's prefered phrasing, which is prescriptive, does not get at the heart of NOR which is proscriptive. The reason the distinction between primary and secondary sources is important is because using the standard for appropriate use is different for each of them. To complete the NOR policy we need to explain what uses of primary sources are not allowed, and what uses of secondary sources is not allowed. Slrubenstein | Talk 16:30, 30 September 2007 (UTC)
I'm not sure whether to respond to this suggestion or the one above, but I would say that there's no need to quote WP:NOT#OR, or to specify that wikipedia isn't a primary source in this passage, but on the other hand it doesn't necessarily do any harm. I see no problem with having it in this passage, but not defining "primary source" until a little later, especially if we use a link within the page, with something like (see below), "below" being linked. "Wikipedia does not bear witness to facts" is quite ugly language, and subject to considerable misunderstandings - wikipedia makes statements of facts, that could be taken as "bearing witness". There may be another suitable phrase distinct from either of these, like "wikipedia does (should?) not contain statements of fact not taken from a reliable, published source". "Original research consisting of novel synthesis…" is good, and should be used as well. NOR must stay proscriptive, as it doesn't actually make sense as a prescriptive policy. Prescriptive statements and examples of what is allowed, however, do tend to aid understanding. I disagree with SLR, though, that it is necessary to seperately list what you're not allowed to do with each of primary sources and secondary sources, as the additional problems with primary sources are more to do with cautions than extra proscriptions to my mind; that is, you may never go beyond what sources say, at all. In the case of primary sources, this usually means that interpretive statements must be given prose attributions ("source Y claims that…") and suchlike, which is more of a clarification/caution than an extra pro/prescription. SamBC(talk) 16:39, 30 September 2007 (UTC)
After reading the above, I guess that I will pull my objection to the "Per WP:NOT#OR" pert, though I still really like the rest of the first sentence "Wikipedia is not a primary source". If this policy goes on to define primary, etc. sources, then at least a sentence similar to this should be present, it succintly sums up exactly what the whole policy is all about. wbfergus Talk 17:07, 30 September 2007 (UTC)
Re Wikipedia does not bear witness to facts ... doesn't that phrasing rub anyone else the wrong way? The only people I've ever heard talking about "bearing witness" like that are Evangelical christians (note that capitalization, reflecting the priorities I've observed). I'm glad to know that Wikipedia is not Conservapedia, but framing policies in those terms lends support to a worldview that's destructive of most things for which WP stands. Let's quickly remove that "bear witness" phrase. --69.226.208.120 16:29, 8 October 2007 (UTC)
  • I like it, and like it even better than the proposal in the previous section, though I don't oppose either. I'm surprised we are getting close to consensus on this so easy. Is this a good model for future discussions about the policy? Maybe we can just start with a clean slate and the above paragraph, and then just insert elaborations to the policy little-by-little, as we establish consensus. What we'll end up with will be a something with consensus, and maybe the article can come out of protection. COGDEN 22:44, 1 October 2007 (UTC)

Question about WP:SYN

WP:SYN says that "A and B, therefore C" is unacceptable if not published by a reliable source. What if they leave the C up to the reader, so they synthesize information so a conclusion is obvious but don't state the conclusion? As an example (real cases would probably not involve something so uncontroversial and mundane), in an article about Weatherman Jim, someone creates a subsection "Criticisms of Weatherman Jim" and says something along the lines of "On 9/30/07, Jim said that the temperature was 30 degrees Celsius. The temperature was actually 35 degrees Celsius" and have that cited to temperature data, NOT an RS detailing Jim's speech and the actual data. Would this be a violation of WP:SYN? They have left off the conclusion "Jim is an unreliable source for temperature data" but it is an implicit in the way it is stated. - Jarn 03:34, 1 October 2007 (UTC)

I'd be first concerned about it being an issue of undue weight. I wouldn't argue that such a case violates SYN. I'd also be concerned that it might be used to violate WP:BLP depending on the context. --Ronz 03:43, 1 October 2007 (UTC)
This was really just an example. It's more the thing of "A and B" organized in a way that C is not stated but is intuitively obvious to the most casual observer. That was just an example and, yes, that example specifically is probably void under some of those other things, but think of the ways that it could be expanded. In political articles, for example. "President XYZ of country ABC said this(1). The truth is (2)" with (1) being a cite to the president saying that and (2) being a cite to something else (such as historical data). They're leaving out the implicit assumption that "President XYZ of country ABC was either wrong or lying" but it is intuitively obvious in the way it is phrased. Pretend, for the example, that it does not violate any of the other rules. Just consider how it would affect WP:SYN. - Jarn 03:57, 1 October 2007 (UTC)
The SYN rule helps other editors distinguish one editor's implicit interpretation from the sources' explicit interpretations, and disentangle the sources if the interpretation is flawed. Putting two claims next to each other doesn't pose nearly as much of a problem as merging two claims and their references. But that's just my understanding of it. Jacob Haller 04:05, 1 October 2007 (UTC)
(e/c)Synthesis can be implied, but putting anything in the policy about it would not be a good idea unless there is a concrete way to determine that it is implied synthesis. Without assuming bad faith, we can't know if a user tried to create an implied synthesis or not (unless they said so). There is also the problem of defining implied; what might be implied synthesis to one person may not be to others. If one goes looking for an implied sythesis or has a preconceived notion that something is synthesis, one is bound to find it somewhere. Sometimes it is close to impossible to avoid it too. Take your second example (about a president) as an example: Could that be worded in a way that avoids implied synthesis while still making sense and not being overly wordy? Mr.Z-man 04:23, 1 October 2007 (UTC)
Yes, the example you give violates WP:SYN and would most definitely be OR. Badgerpatrol 04:44, 1 October 2007 (UTC)
Can you elaborate? I reread the section and it looks perfectly clear that such cases don't violate SYN. I'd advise caution, but the SYN policy doesn't require it. It looks to me like the first Smith-Jones example, the one without SYN. Jacob Haller 05:56, 1 October 2007 (UTC)
The problem with the SYN section is that it elaborates on an element of OR in a specific way. This is probably a good example of why people should not read the letter of the law. The principle is clear: don't let an article say what hasn't already been said. To me that includes reasonable inferences. It is examples like this where I think the SYN section is actually damaging to policy, that someone can read policy and say a bad edit fits with policy. Real editing would not go down this path. Spenny 07:51, 1 October 2007 (UTC)
Not sure why I need to elaborate...the obvious inference in the above case that would be drawn and is intended (obviously extending disbelief for this simplified, hyper-realised example) is that Weatherman Jim is not a very good weatherman- otherwise, why parse the data? Why quote the temperature at all? Here the editor is taking two primary sources- a TV broadcast and a temperature record - and juxtaposing them to proffer a new interpretation (= that Weatherman Jim is not a good weatherman) that is not drawn from any secondary source - e.g. forming a new synthesis from primary sources. In the Smith-Jones example, the editor is simply quoting the contrasting interpretations already given in two secondary sources- thus there is no new synthesis. Badgerpatrol 11:20, 1 October 2007 (UTC)
Badgerpatrol, I had a similar discussion (above... perhaps in the archives) on this issue concerning the George Washington and religion article... where primary sources are quoted extensively to imply that Washington was or was not a deist... without supplying any reliable source that draws the information together and acutally claims he was (or was not) a deist. I felt that this was SYN ... "A + B" being stated (but "= C" being implied). I was told that this wasn't SYN. Now I am confused. Do me a favor and check out that article ... I would like to know whether to leave the OR tag or remove it. Blueboar 11:44, 1 October 2007 (UTC)
I explained my understanding of the role of WP:SYN above. As long as everything's in the open and properly cited, and NPOV, it's fine with me. But if it's hidden away, it's OR, if not outright deception.

Here's my interpretation of SYN and its intent. Suppose source 1 says A, source 2 says B, and A+B probably imply C. If an editor states A and cites 1, and states B and cites 2, the editor may imply C but does not commit SYN. If an editor states C and cites 1+2 the editor commits SYN. Why do I draw the line there?

  • It is possible that the original editor made a mistake.
  • It is possible that another source will state ~C, in which case we may need to cover the dispute.
  • It is possible that another source will contradict one of the underlying claims (A or B).
  • Other stuff might happen.

If any of these things happen, and an editor has merged the claims, other editors will find them harder to untangle; but if an editor has just set the claims alongside each other, other editors should have no trouble. Of course, NPOV may apply. Jacob Haller 18:53, 1 October 2007 (UTC)

Placing two related but apparently contradictory facts in juxtaposition is not OR. It is all we can do when we find conflicting information without engaging in OR. In your example, the statement of the temperature would have to be important. If the person reported the temperature every day, and an editor just picked one particular spurious data point, that would be undue weight. This case is rather contrived, so let me repeat an example I gave before that has been lost in the archives. I added a statement that a company received X complaints from customers on their new product, and cited a newspaper story that backed that number up. Another editor modified this to say that the company reportedly received X complaints from customers, but they had not sold X copies of the product at the time, and cited a company report that backed that up. This was contradicting a secondary source using a primary source (and a close source as well), but I thought it was a perfectly appropriate way to address the apparent contradiction. Dhaluza 00:25, 2 October 2007 (UTC)
So it's a case where it appears they received more complaints than products sold? --Ronz 00:33, 2 October 2007 (UTC)
The source was not specific about the inclusion criteria. The company could have received complaints from customers who bought its other products as well, or the count could have included complaints from non-customers, or they may have counted multiple complaints from individual customers multiple times. So there is no way to tell if the source is right or wrong, but the clarification puts the statement in perspective. Dhaluza 01:16, 2 October 2007 (UTC)
Then I'd say putting two different statements about the number of customer complaints is SYN if they are presented in such a way where the reader would tend to think that the numbers are the same measure. --Ronz 01:58, 2 October 2007 (UTC)
I'm not sure what you mean by that. The point is if you have conflicting facts, trying to resolve the conflict is OR--simply juxtaposing the conflicting facts is not. Cherry picking one over the other (assuming both are reliable) is NPOV (and possibly OR too if you are trying to make a point). Dhaluza 02:48, 2 October 2007 (UTC)
If information that appears to be comparable (X number of complaints...Y number of complaints) is placed together without indication that it may or may not actually be comparable, then it's SYN. --Ronz 15:43, 2 October 2007 (UTC)
No, WP:SYN is for synthesis to advance a position. If the only point being made is that there is a difference of opinion, or some doubt about the specifics, that is not synthesis, it's maintaining NPOV. Dhaluza 23:57, 2 October 2007 (UTC)
The position that concerns me is that it invites comparison. Yes, it's more an NPOV issue than OR, but I've found that looking and dealing at such situations in light of SYN is appropriate and helpful. --Ronz 18:17, 3 October 2007 (UTC)
I disagree strongly that inviting comparison is OR. When we have conflicting facts and no overall consensus of accepted truth, inviting comparison is necessary. Ignoring the conflict would be OR. Dhaluza 12:51, 8 October 2007 (UTC)

Issues with this policy

I'm simply going to say that I've always had a problem with this policy. The issue with Wikipedia and it's No OR policy is that things must be varified through text or otherwise it constitues as OR, however some things can't be verified through text, because sometimes what you're referencing is merely the absence of text or otherwise is a television show or a film where there is no text at all other then the creators info on it. I don't see this as a big concern, it's just a pain to edit the ariticles sometimes because of this policy and the surprisingly vague restrictions that are portrayed in it when talking about certain things like fictional media. --VorangorTheDemon 13:17, 2 October 2007 (UTC)

This is a false complaint. That is not to say you're lodging it dishonestly, but rather that is simply does not wash. Let's examine your examples of film and television. Your implication that print sources are not available is simply ludicrous. Script and transcripts can be found for the movie or program in question. There are numerous mass-market publications that provide reviews, actor interviews, "behind the scenes" articles and so forth. Further, there are a number of industry periodicals and other publications that touch upon many of the details in greater depth. Additionally, there are a number of scholarly and mass-market books and articles that address the cultural impact of various movies and television shows, along with their artistic "meaning" and similar analysis. There are plenty of sources available. Vassyana 17:28, 2 October 2007 (UTC)
I think this is partly an issue with WP:V as well as WP:NOR. The WP:V policy does not require verification to "text": it requires verification to a reliable source. That source could conceivably be non-textual, so long as the point you are making in the article is non-controversially apparent and obvious from the source. For example, I don't think there would be a problem citing Beethoven's fifth symphony and saying the first note in that composition is "G". COGDEN 18:52, 2 October 2007 (UTC)
Actually that example would probably run afoul of the "without specialist knowledge" provision. Being able to read music, particularly to interpret historical compositions, could reasonably be argued to be specialized. Dhaluza 00:13, 3 October 2007 (UTC)
The "without specialist knowledge" was removed by Marskell on 30 September. And I also see he replaced it at your request, which is fine. I wonder if there isn't a clearer way of saying it than just "anyone—without specialist knowledge", e.g."For this reason, any reasonably literate reader — with a general education but without specialized knowledge about the topic — should be able to verify that the Wikipedia passage agrees with the primary source." Just another thought. ... Kenosis 02:29, 3 October 2007 (UTC)
I only asked that he revert his edit because the page was protected. I actually would prefer it said without specialist training, which is more specific. Specialist knowledge could cover almost anything you didn't learn in primary school. I once added info on from a CLLI code lookup to an article with a cite to the web site. Was that specialist knowledge? True, I knew to look for it because of my specialist knowledge, but someone could look up what it is and figure it out in under an hour. So I would not consider that it required specialist training to verify. Dhaluza 05:33, 3 October 2007 (UTC)
True. Maybe a better example would be citing an authenticated, published photograph of some historical figure for the proposition that they were missing their front teeth. Or citing a published nude photograph for the proposition that the subject was male. This type of thing would be pretty rare, because usually if something is (1) obvious and (2) worth discussing in a Wikipedia article, somebody will have said something about it in prose. But some interesting facts might be so obvious that nobody has yet expressed them in text. If something in a reliable, non-textual source is so obvious that nobody bothers to comment on it, I don't see how it can count as original research. It's certainly less like original research than translating a French phrase into English, which is usually considered okay.
Another example is using a map as a source. When writing an article on a geographic place, you need to relate it to the surrounding places and wikilink them. This type of information is commonly found on maps; in prose, not so much. Simply translating what the cartographer depicted on the map into prose is not OR. So looking at the map and writing that East Northport, New York is actually south of Northport is OK, but saying that it got it's name because it is East of the old Northport junction along the railroad line would be OR using only the map as a source, because the cartographer could not depict that. Dhaluza 05:33, 3 October 2007 (UTC)
Dhaluza, thanks for pointing this out, because it presents an interesting borderline case that could potentially challenge our expression of the policy. In some cases, such as determining what letter corresponds to a note on sheet music, verifying the citation requires specialized knowledge, but it is a completely mechanical process that would be 100% non-controversial and obvious to anyone with that specialized knowledge. That kind of citation is somewhat common, especially in photo captions, and maybe such citations are okay. Plus, things like translation of phrases by Wikipedia editors are usually considered okay under WP:V, but they require specialized knowledge. I'll have to think about that. COGDEN 01:55, 3 October 2007 (UTC)
I agree that OR requires cognition. Something you could simply program a computer to do, for example, should not be OR. That is why I think we should change this to specialist training, from specialist knowledge, because that is more on target. Dhaluza 05:33, 3 October 2007 (UTC)
I think "training" is better. What about this?:
An educated and informed person who reads the source should be able to verify from the text itself—without doing additional research, using specialized training, or looking at other sources—that the source directly supports the facts, conclusions, and implications of the Wikipedia passage.
COGDEN 17:45, 3 October 2007 (UTC)
Training has little to do with it. The issue is that any person who can read and understand general usage of English should be able to verify that the primary source agrees with what's in the article, without any specialized familiarity, understanding or knowledge of the issue(s) involved in that aspect of the topic. ... Kenosis 17:53, 3 October 2007 (UTC)
I'd like to make sure that we don't run afoul of what has already been written in other policy pages. For example, WP:V says it's okay, if necessary, to include editor-created translations of foreign phrases. A translation such as that requires specialized understanding of the foreign language. But you might argue that it doesn't require special training. Maybe we don't even need the idea of "specialized". Maybe all we really need to say is that the reader shouldn't have to do additional research to verify the source. COGDEN 18:09, 3 October 2007 (UTC)
I just recently ran into this issue. The editor posting the sources refused to provide translations, despite that he was fluent in both languages. I happen to agree it shouldn't be permitted by WP:V to verify across languages without a reliable intermediary translation available for those who understand English only. In this particular instance, several editors were able to weed out the OR and confirm the sources, but it required another editor fluent in both languages and an online translator to help sort it out. I wouldn't want to make the case at WT:V unless I had more experience with other particular events of this kind. ... Kenosis 18:31, 3 October 2007 (UTC)
I think specialized research is a step in the wrong direction. Any specialized knowledge you can acquire in a short amount of time should not be excluded. Doing additional research should be encouraged not discouraged. The distinction with specialized training is that it takes time and outside help to acquire, and so that should be where we draw the line. I would make an exception for multilingual training, though. Dhaluza 00:25, 4 October 2007 (UTC)

Proposed change to specialist knowledge

Based on the above discussion, I would like to propose changing the sentence:

For that reason, anyone—without specialist knowledge—who reads the primary source should be able to verify that the Wikipedia passage agrees with the primary source.

to:

For that reason, any reasonable educated person without specialized training who reads the primary source should be able to verify that the Wikipedia content is consistent with that source (understanding a foreign language is not considered specialized training in this case).

-- Dhaluza 13:03, 8 October 2007 (UTC)

I don't think it should be limited only to primary sources. Secondary sources must be verifiable as well. Otherwise, it looks good, but this is a bit untested. I'm always worried that there is some counter-example that shows this language is either restrictive or not restrictive enough. Provisionally, however, I'd support such a change, if it's applied to all sources, not just "primary" ones. COGDEN 21:48, 8 October 2007 (UTC)

Primary vs. secondary: facts vs. analysis

Let me try to put just a little structure on the types of sources dispute which I think should clear up most of the issues involved.

Most of my contributions have been on specific things; of late, for instance, I have been doing articles on lighthouses in the Chesapeake Bay. These last are very much standardized in their construction: a very brief opening sentence with a notability point if possible; an infobox with a picture (if available) and data about the light; and a short history. There is a standard set of sources I've used for the latter, and there are sometimes disagreements; since none of them are true primary source, however, it is simply a judgment call as to which is the most reliable in a given instance. However, for one datum I do have access to (essentially) primary sources: the location of the light. One of my history sources gives lat/long positions for extant lights; they are not, however, perfect, and in some cases I have found noticeable errors, including one descriptive error on the order of Krakatoa, East of Java. In general, where there is any existing structure I've been able to locate it using NOAA charts, Google Earth, and other aerial photography servers. One person might call this synthesis; another, fact-checking. I don't see how one can condemn it on anything but excessively dogmatic grounds.

I also monitor articles about alumni of my high school, because the fact of their attendance is often mentioned in their articles. I also police the school's website to keep its students from including each other prematurely. Thus have I have gotten involved in two disputes about the facts of these people's lives. In one case, there was a question as to whether the person was a "II" or a "Jr.". The erroneous source in this case was a remark (in my opinion flippant) from his son (also the subject of an article) which was easily citable-- but wrong. After much batting about, I finally resorted to looking in his high school yearbook (which is available on-line). In the second case, there was dispute as to whether a current TV celebrity was actually an alumna of the school. The erroneous source was a semi-interview piece in which the write identified the subject with a different school of the same name. This did not seem to be consistent to me, as the other school is located several hours from the subject's home town. So again, I went to the yearbooks, and I was able to find the subject and give a specific date for her graduation (as well as pinning down another perhaps not entirely accurate fact).

All this leads me to a simple three way division.

Quotations

As far as possible, quoted material must be attributed to the original source. Failing that, a definitive, authoritative source should be used. Either must be preferred over secondary citations.

The reason for this is simple: what is important in this situation is accuracy of reporting above all else. Therefore, the primary source has precedence as a reference, with authoritative testifiers used only as necessary. More secondary sources are to be discouraged as potentially unreliable (even the NYT citations-checkers have off days).

Facts

Statements of more or less uninterpreted facts should be cited from as authoritative a source as possible.

As a rule, the notion of "primary source" doesn't obtain here. Maps are secondary sources for the locations of objects, for example; a primary source would be one's own visitation of the spot with a GPS unit or a theodite. The primary source for someone's given name would be his or her birth certificate. It is unlikely that we would ever be confronted with a true primary source datum, but if there is no other source, and the datum deserves reporting, and its interpretation is is uncontroversial, then it could be used.

What is most important is the avoidance of secondary, unauthoritative sources for factual data. They of rights should be getting their information from authoritative sources; we should use those same sources, and not rely on potentially incorrect reportage.

Analysis

It is analysis that No Original Research is directed at. We do not want to present our own analyses of the the facts; rather we want to represent the consensus of the field.

At this point the notion of primary vs. secondary source gets complicated. Robert Massey's Nicholas and Alexandra is a secondary source when reporting about the facts of Nicholas II's reign as tsar, because he is repeating what primary or authorative sources say (and often choosing among them). The book is a primary source, though, for Massey's theory that the tsarevitch's hemophilia had a negative impact on the tsar's rule. In relating that theory, it is the preferred source, because description of that theory is a matter of fact. Acceptance of that theory, however, is analysis, and other secondary analyses and surveys must be appealed as sources for the theory's place within the state of the field.

In some cases what one would think are simple facts are caught up in analysis. For example, economic data are from time to time subject to controversy, because considerable interpretation is needed to produce a statistic. How great are Saudi Arabia's oil reserves? Well, this is not an easily derived number, and no doubt investigation would turn up some difference of opinion. In lieu of trying to evaluate the various values against each other, it is better to pick one authoritative, generally trusted source, unless there is clear evidence that there is an error or misrepresentation. However, that "clear evidence" obviously has to meet very high standards of reliability. A controversialist website is not a good enough authority on its own to overturn an official Saudi estimate of reserves; and if one turns to that website's sources (if it gives them), one is liable to present their analysis on one's own authority.

On the other hand, taking a "zero-tolerance" stance about trivial analysis is also original research. Most fields are subject to lingering disputes about minor items; raising the profile of such disputes by aggressively pursuing them here is in essence assertion of a peronal analysis that they are important.

Summary

This is of course my personal presentation, for which I haven't offered a single reference. However, I believe this represents standard practice in all fields and interprets our policies correctly-- that is, in a way that produces the most accurate encyclopedia. Mangoe 18:16, 8 October 2007 (UTC)

I'm afraid it does not accurately represent all fields in the least. As addressed (repeatedly) on this page, there is no such thing as a "standard practice" across fields in relation to the delineation between primary and secondary sources. The definition you use is a strict variation of the definitions used in library science. Using your example about the tsar theory, in its own field it is a secondary source. Original ideas within the work do not affect it's status as a secondary source. Vassyana 18:47, 8 October 2007 (UTC)
True, but non-original ideas in a work also do not affect its status as a primary source. Any work can be either primary or secondary; it really depends on how it is used. If you are citing the new idea, you are using the work as a primary (original) source of that idea. If you are citing secondary or quoted data, you are using the work as a secondary (derived) source of that data.
For example, a diary or interview (which you might think is a primary source, based on a lot of the simplistic definitions, right?) might very well quote information from a published newspaper article, which would make it a secondary source for the newspaper article and any commentary thereon found in the diary. Likewise, a book review in a magazine (clearly a secondary source, right?) might very well contain an interview the writer had with the book's author, making it a primary source for the author's words. This is why I think this distinction is way too confusing and complicated to be the foundation of the NOR policy. But as I said earlier, I'm not opposed to this policy formulation stating that Wikipedia cannot be a primary source, or that it must be a secondary source. That's non-controversial according to any definition of primary or secondary. COGDEN 22:53, 8 October 2007 (UTC)
It's probably better than the current almost-dicta of "never use primary sources". But I still fail to see why the WP:NOR discussion shouldn't be focussed on (a) use reliable sources (b) don't diverge from them. Research involves at least creating new knowledge, and especially doing so from unreliable sources. So diverging from either of those two simple -- and much more universal!! -- rules is one clear sign that the result will violate NOR. The NOR policy would be much more clear if it were just "what may be done with facts", rather than being tangled up with "what may be deemed a fact". --69.226.208.120 21:07, 8 October 2007 (UTC)
It's obvious to me the single biggest problem here, and why we can't get past this primary vs. secondary issue, is that way too many people have mistakenly vague notions about what sources do. The thing is that sources are always relative: that are sources of something. And in the context of Wikipedia's typical citation practices, they are sources of specific statements. A primary source, therefore, would be a text in which the statement originates; a secondary source would be one which repeats or recounts the statement from some other source. (There's no point here in distinguishing secondary from tertiary sources.)
In a lot of the cases I see above, these conditions are being identified sloppily or simply incorrectly. For example, the statement by the witness about the speeding driver is testified to by both the court and the newspaper as secondary sources, to be fastidious about it. We ascribe greater authority to the court because we are wont to believe that its reporters are unimpeachable witnesses to what is said before them, but their testimony is otherwise the same as that of the reporter to whom the witness spoke.
In the Benjamin Tucker case, the same sort of thing is occuring, but much more subtly. The problem has nothing to do whether Tucker can or cannot be cited as to whether or how much he should be considered a socialist. In the disputes back in the spring, part of the problem was that the quoted passage doesn't quite justify, on its own, some of what was being claimed for it diff. But the other side of the coin was that there appeared to be a contest going on (which may still be in progress) over how Tucker was related to socialism as a movement. The quoted passage is quite clear that Tucker considered as a kind of socialism the anarchism he espoused. That unimpeachable primary source has been essentially displaced by a secondary cite from a compatriot which superficially appears to contradict it. The missing piece (which is, I think, not entirely cited in the article) is that the notion of socialism has been narrowed with time. The controversial quotation actually makes this quite clear, even without reference to later explainers. But in the end, the issue wasn't really about sources; it was about a controversial attempt to link Tucker with other people who called themselves socialists whom Tucker quite clearly had contempt for-- and indeed, said contempt is the statement of the quote! Sourcing had nothing to do with the real problem; it was just something else to try to pillory the offenders with. As far as I can tell, the quote was largely removed because it was associated with a claim that someone disagreed with.
I'm a polymath without a commitment to a single field, so when I go back to your explanation of who different fields work diff, Vassyana, I don't find it compelling. In the only sense that matters here, peer-reviewed journal articles are primary sources only for the ideas that are first stated in them, and secondary sources for the ideas that they take from others. The field of inquiry has nothing to do with it. Peer review is a reliability guarantee, and it seems that we might agree that the nature of that guarantee varies considerably from field to field. The real OR problem with journal articles is that they are prone to require a lot of explanation to be comprehensible to those outside the field, so that some (maybe trivial) bit of analysis-- that is, research-- is required to turn them into encyclopedia-usable statements.
Massey's book, of course, received no such review. But that does not matter. It is the primary source for his contention. For his personal analysis, it is, by nature, the most reliable source. Mangoe 22:58, 8 October 2007 (UTC)
I'm not sure what you do not find "compelling" about it. It's pretty plain. Differing fields use differing definitions that are not compatible. Also, for being a polymath without commitment, you seem very committed to a specific definition. To be honest, I'm really at a loss regarding what you're trying to respond to or communicate in general, beyond advocacy for a specific definition of primary/secondary. Vassyana 00:15, 9 October 2007 (UTC)

Anonymous insertion at WP:RS

An annonymous edit has been inserting the following in WP:RS:

In some areas only secondary sources are viewed as reliable, since the primary sources are fragmentary, ambiguous, and/or contradictory. That is typical of historical topics; Wikipedia should hold no original research, which would be required to make use of such primary sources. However, in other areas primary sources may be preferred since no others can be authoritative, and other sources may be weak or scarce. That is normal in current technologies, which initially tend to not provide the kind of secondary sources which define historical narratives.

I've been following this discussion, and I've never noticed this passage being proposed here. Besides which, it isn't entirely true. Jacob Haller 21:05, 8 October 2007 (UTC)

I'd be that person-without-account. What parts are not entirely true, and why didn't you just fix those parts?
That specific passage was not proposed, however its content has certainly been discussed and numerous people have agreed to it. There's a fundamental flaw in the supposition that primary sources can't be reliable, or contrariwise that secondary ones are (more) reliable. Several counter-examples have been easily produced. That flaw needs to be fixed. Related, there's the fact that recent WP:NOR changes have reduced the effectiveness and validity of that guideline. One effective way out of that mess -- accepted by many people, especially those who edit technology rather than history -- delineates sourcing policies (which are demonstrably domain-specific) from NOR, and redresses that flawed bias against primary sources. The para you dislike would be part of resolving that ... by starting to clarify when a given primary or secondary source may not be reliable. --69.226.208.120 21:41, 8 October 2007 (UTC)
It is not typical of historical topics. Standard practice is to return to the earliest available sources, and use other sources, e.g. geographical references, logistical, demographic, or other objections to the primary sources, archaeological sources, etc., as appropriate. Jacob Haller 22:07, 8 October 2007 (UTC)
That standard practice is original research, which is highly encouraged in academic fields. However, it's exactly the opposite of what we should be doing here. Reiterating my perennial example, Caesar's Gallic War is considered a pinnacle of ancient reporting. It however is also considered a fine example of propaganda writing and is known to contain numerous errors of hearsay. We cannot use it to support facts independently of reliable secondary sources without violating the basic principles of no original research and neutral point of view. Vassyana 00:08, 9 October 2007 (UTC)
It's OR to do my analysis. It's not OR to cite another's analysis.
It's reasonable to say that Wikipedia practice is to use secondary sources, but not to say that the secondary sources are more reliable than the primary sources on which they depend. If we take the military historian Hans Delbrück, his army size estimates are well regarded (despite their age) and would be suitable for Wikipedia citation, but their quality comes not from their secondary nature, but from Delbrück's innovative use of new primary sources from his own time - road space requirements and the like - to check the traditional primary sources.
I'd suggest use of the ancient sources with minimal interpretation - with caveats about numbers, years, and other known sources of error. Jacob Haller 02:56, 9 October 2007 (UTC)
You've left me guessing about which of those are primary vs secondary; and in any case it seems like you're describing how to violate the NOR policy (by doing original historical research), not how to follow it. Other commentors on this page have said what I repeated: that for history articles, correct use of primary sources is tantamount to doing original research. (I'd be willing to believe that policy is as flawed for history as it is elsewhere, though. Maybe some folk like just like having a "bright line" policy, instead of something that could produce arguments?) Regardless, you haven't answered the question about why you didn't just fix what you found wrong, rather than start an edit war. Fixing things is standard practice... --69.226.208.120 00:13, 9 October 2007 (UTC)
I went here because I wanted to make sure that any edit has consensus. I have only made a few edits to policy pages - mostly marking disputes - and once boldly adding an entry to WP:WTA - but generally it's best to go to the talk page, work out disagreements on the talk page, and get clear consensus on the talk page before making the edit.
Making additions to the policy page is, in my understanding, the standard way of starting an edit war... I'm sure we've both had such experiences on the article page. Funny thing is, I can make some suggestion on the talk page, nobody objects for six weeks, I follow through on the article page, and all hell breaks loose. Jacob Haller 02:46, 9 October 2007 (UTC)

Can somebody weigh on a case of possible OR

Can somebody take a look-see at: [10] and weigh in as to whether the disputed passage being mentioned actually constitutes OR? If this is the wrong forum to ask, please just point me to the right place, and thank you.--Ramdrake 22:02, 8 October 2007 (UTC)

No consensus even among the proponents

I left and said I'd not be back. I see, though, that there still is not an agreement on source typing, that the proponents are unable to find wording that is satisfactory. If the proponents can't agree how on earth can it be part of an official policy? That the proponents don't agree shows that the concept is too vague (or flawed in some way) to be understood or expressed. Blueboar tried to redirect attention and focus but the discussion is back on what "primary" means, and that is still unclear. Whether I had made this comment or not the history of this talk page will be one that shows not only a lack of consensus on having source typing as part of the NOR policy but a lack of consensus on what "primary" really means - even among those who are for such typing. And even then all that the policy says is "be careful," but unless "be careless" is an implicit principle that's entirely superfluous - when is being careless ever proper when editing Wikipedia? ("Be careless" is not the same as "be bold.)

There's also a strong anti-wiki flavor to source typing: editors can't really be trusted to edit, they have to be told in meticulous (but unclear) language what to do. --Minasbeede 23:21, 3 October 2007 (UTC)

OK, so you still have strong feelings, but how would you suggest we address them? Dhaluza 00:40, 4 October 2007 (UTC)

That's up to you - as a non-participant/drive-by-poster I have scant authority to tell you what to do. Which I suspect is your attitude, which I think is the proper attitude, or close enough.

There is a standard Wikipedia concept to overcome such problems: consensus. The task before you is to reach consensus. Until you do I can flit in and point out that you haven't. What I do is hardly important: it's the lack of consensus on what you supposedly agree should be there that is the problem. Whether I point it out or not there isn't consensus, not even among those who favor some sort of source typing on what "source typing" means (within Wikipedia.) I doubt there's consensus on what "primary" means. The thing that lifts Wikipedia above Usenet ("above" seems to denote and is here meant to denote "better") is the consensus policy. If you don't have consensus what you have is hardly distinguishable from Usenet (by that I mean news group discussions within Usenet.) (That is not a compliment.) --Minasbeede 00:59, 4 October 2007 (UTC)


Yes, agree. The problem is that 'consensus' is often seen as 'consensus among who is already in agreement'.--Stefanomencarelli 15:42, 4 October 2007 (UTC)

But there isn't even that. --Minasbeede 12:27, 11 October 2007 (UTC)

Often is not always (what a mess..)--Stefanomencarelli 16:01, 4 October 2007 (UTC)

Use of (sic) and [] in a quote

Any opinions on adding "(sic)" to a quote? So far common sense seems to prevail, which is good; but is it original research to tell the reader what is wrong rather than presenting evidence and letting the reader decide? This relates to using "[]" to add or replace text in a quote. I think editorial discretion is called for in all good editing. NOR is not to be applied blindly making the edit-copy worse when no one is arguing that a representation is false. It is designed to be helpful in sorting out disputes. In other words, write it as accurately and truthfully as you can; but the minute there is a dispute, accurate sourcing takes precedence over "yeah but this is what is really true" What is "really" true takes precedence when no one is arguing that it is not true. Hence if everyone agree "not" was left out of a quote, then its ok to add "[not]" or "(sic)"; but the minute it is is an element of dispute, then it is not obviously true and we must fall back on verifiability and sticking to the sources and neutrality. Any opinions? WAS 4.250 14:18, 9 October 2007 (UTC)

If the quote doesn't say what it was meant to say, and you consider (sic) or [explanation] necessary, then it is not a good quote and you should not use it. Instead, summarise the source in your own words and reference precisely. This will also keep you clear of infringing WP:NOR through the creation (modification) of a quote that has not been used before. --SmokeyJoe 00:14, 10 October 2007 (UTC)
Using [sic] to point out that a spelling or grammatical error is transcribed literally from the original does not mean the original is not a good quote, and using [brackets] to change the parts of speech to make the quote fit in a sentence is not a problem either. Adding things to quotes, or otherwise "fixing" them is problematic, and it is probably best to leave it alone and let the reader decide. Rephrasing it in your own words can actually be a bigger problem, if the original meaning was ambiguous, and I would rather see the original and decide for myself. One way to do this is to put the literal quote in the citation so it appears in the references section. Dhaluza 00:43, 10 October 2007 (UTC)

Just wondering

Several of the distinctions drawn above are too fine for me to attempt to undertand them. I just wonder if this International Herald Tribune article is in any way related to the primary/secondary source discussion. At first glance it seems to play down the value of secondary sources and of consensus among such sources. Lima 12:10, 11 October 2007 (UTC)

I'm not sure either about its relevance here but it's a great article and certainly worth citing! - maybe we should add this to the NPOV tutorial WP:NPOV ? —Preceding unsigned comment added by Harald88 (talkcontribs) 10:52, 13 October 2007 (UTC)

In the beginning, when original research is defined, a link to the Original research article should be given. Banaticus 10:49, 13 October 2007 (UTC)

Why? Wikipedia makes clear to 13 year olds how to write an encyclopedia by redefining the term "original research" in terms of the redefined term "primary source" in order to convey the idea one should stick to the original meaning. Ironic? No? WAS 4.250 12:40, 13 October 2007 (UTC)

Option 2b

I propose to improve option 2, in accordance with my remarks above about option 2.

Perhaps something like this:

Source-usage proposal 2b: yes primary/secondary source distinction

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 them in novel ways.

One useful way to ensure that a Wikipedia passage does not "go beyond" the sources is to consider the difference between what are often called primary and secondary sources. A primary source is a document or person with original factual data. Primary sources are usually close to the topic at hand, and present information that is not interpreted by a another party. A secondary source is a document or person that obtains the relevant information from primary sources and provides an interpretation. Secondary sources may repeat or comment on primary sources, but do not provide first-hand information. Factual information should be backed up by reliable references, which in general implies citing either primary or secondary sources. Reliable secondary sources should be used for an interpretative synthesis. Wikipedia is itself a tertiary source. See WP:Source Tutorial for some examples.

Wikipedia may never be the original source of information. Wikipedia may repeat and quote sources, but may not present undocumented information or theories concerning those sources. To the extent Wikipedia comments on a source or analyzes it, such commentary or analysis must not be original, and must be verifiable. Conducting background research that consists of collecting and organizing information from existing sources is, of course, strongly encouraged. All articles on Wikipedia should be based on information collected from reliable published sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. However, in presenting such information in a Wikipedia article, great care should be taken not to take the information contained in the sources out of context or present it in a manner that violates the provisions of this policy.

Sources should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources. In short, the policy of "No Original Research" requires that Wikipedia users stick to the sources.

Harald88 13:01, 13 October 2007 (UTC)

Minor change above. Struckout a redundant sentence previously used in opening paragraph. wbfergus Talk 13:35, 13 October 2007 (UTC)
This version works for me... it places the distinction between primary and secondary sources in the context of how and when to use them appropriately. I have added the fact that purely factual information can come from either primary or secondary sources (in bold above to show the addition). Blueboar 14:32, 13 October 2007 (UTC)
This pretty good. I'd just point out that a source can be an object (i.e the painting, Guernica in the Picasso article). I would also find a better wording rather than "people" - this looks like it's inviting people to go out & do interviews but WP:NOR doesn't allow that - "authors" might be a better word. While in real life people can be sources, for WP's purposes only documents or objects can be (because people have to be interviewed & the interview has to be published by a reliable source and then it's teh interview that's the source). Otherwise this is fine--Cailil talk 18:40, 14 October 2007 (UTC)
So something like changing "A primary source is a document or person with original factual data." to something along the lines of "A primary source is a document, object, or person (author, eyewitness, researcher, etc.) with original factual data."? wbfergus Talk 19:20, 14 October 2007 (UTC)
Yes, as long as we make it clear that the views/accounts of persons must be verifiable (ie published in an WP:RS). Something along the lines of "Please note that all sources used in wikipedia articles must comply with the policies of WP:RS and WP:V." This could go in around about the final line that says "stick to the sources"--Cailil talk 20:05, 14 October 2007 (UTC)

<undent> Perhaps I'm not seeing it, but "A primary source is a document or person with original factual data" doesn't seem to cover a primary source which is the subject or is close to the subject of the article or section. A poem or novel need have no factual data, but in an article about the author or the work it's a primary source. Similarly, writings about a literary movement by a prominent leader of the movement can be used as facts about the author, but not as a secondary independent analysis of the movement. ... dave souza, talk 21:31, 14 October 2007 (UTC)

The primary/secondary distinction here does not and cannot cover all the sources which I have used, in particular (1) ancient histories (2) political works in articles about their authors or their philosophies (3) religious works in articles about their authors and their philosophies. (1) the advice above hinges on whether these sources are primary or secondary (2) if we make the common-sense assumption that the latter two groups of sources are primary, then the guidance in the source-typing section is completely backwards. These latter two groups of sources are most reliable for what their author believed and least reliable for matters of fact. I am convinced that source-typing must be removed from the policy, or it will lead to the opposite of its intent. Jacob Haller 23:55, 14 October 2007 (UTC)

This is not bad, but I'd suggest a couple of changes:

  • The statement "A primary source is a document or person with original factual data". I'd delete the "factual data" and replace it with "information". Something can be the primary source for a new theory, or a new interpretation, etc. For example, Charles Darwin is the primary source for his Theory of Evolution, which is a theory and a synthesis of data, not just data.
  • The statement "Reliable secondary sources should be used for an interpretive synthesis" should be deleted. Actually, primary sources can be uses for an interpretive synthesis. For example, I can use Darwin's Origin of Species (a primary source) to cite his interpretive synthesis of data from his tour of the Galapagos Islands.

COGDEN 18:25, 15 October 2007 (UTC)

Sorry folks, but I am not getting any sense of what you are doing, or where are you trying to go with this. ≈ jossi ≈ (talk) 01:39, 17 October 2007 (UTC)

Discussion about particular attribution

I encourage Wikipedians who watch this page to comment about a new proposal at Wikipedia talk:Fringe theories#Appeal to particular attribution. Thanks ScienceApologist 17:29, 15 October 2007 (UTC)

Primary sources vs. secondary sources

Live on Radio 2, Noddy Holder reviewed his own article and pointed out that he had never made a record with Anna Ford. This is clearly reported in the press - Anna has three references, but they are clearly all based on the same source. So Noddy Holder, a primary source, denies it, and you can listen again in the UK for the next week to verify this statement. However, we can go to another source, a database of the chart positions and find indeed there is no entry for this record. So we can verify by another primary source that the newspaper report is inaccurate. The newspapers are secondary sources, but clearly not reliable. And Stuart Maconie and Noddy Holder have just congratulated me for my efforts. Presumably, my edits should be removed as not desirable as they were based on setting primary sources against secondary sources. Discuss. Spenny 20:38, 4 October 2007 (UTC)

I don't see what congratulations from the subject of a BLP has to do with it. The PSTS section in WP:NOR states:

Primary sources are documents or people very close to the situation being written about.. 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. For that reason, anyone—without specialist knowledge—who reads the primary source should be able to verify that the Wikipedia passage agrees with the primary source. Any interpretation of primary source material requires a secondary source. ...

An article or section of an article that relies on a primary source 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 sources should be careful to comply with both conditions.

This language is fairly plain. The editors of those two arcticles on recording artists you mention can readily reconcile any such contradictory sources by reporting the primary sources in keeping with the directions of PSTS and making note that newspapers have reported differently, noting the common source of those reports if at all possible. ... Kenosis 21:32, 4 October 2007 (UTC)
I don't think this language (which, by the way, is not the same PSTS language that existed from late 2006 to mid 2007) would allow what Spenny did to occur. First, Noddy's denial in the interview, a primary source, was an "explanatory claim", which supposedly cannot be made using a primary source. Furthermore, the wording says, "Any interpretation of primary source material requires a secondary source." Here, primary source material (the claim chart) was interpreted as omitting an entry for the record. But this interpretation was backed up by only a primary source (the interview). Therefore, what Spenny did would be prohibited. Which is not what we want. COGDEN 01:23, 5 October 2007 (UTC)
You're kidding, right? ... Kenosis 01:53, 5 October 2007 (UTC)
Dead serious. This example illustrates that the above language has two problems:
  1. It doesn't account for primary sources that make "explanatory claims".
  2. It requires that "any interpretation" of a primary source requires a secondary source, even when:
  • The "interpretation" is obvious, non-controversial, and minimal (e.g., that a record is not listed on a chart), or
  • An even better "interpretation" may be found in a primary source.
COGDEN 22:20, 5 October 2007 (UTC)
This is a good example of the use of primary sources to counter secondary sources, which was one of the "bad" practices people wanted to use PSTS to prevent, wasn't it? Jacob Haller 02:12, 5 October 2007 (UTC)
No, not when simple facts are involved. ... Kenosis 02:39, 5 October 2007 (UTC) ... If the words, "Any interpretation of primary source material requires a secondary source" are that much misunderstood, just delete them when the page is unprotected, and I'll support the edit as best I can, for my own part. Maybe better yet, rewrite that sentence so it reads something like "Any use of primary-source material involving significant analytic or evaluative expansions of the plain meaning of the primary source requires a secondary source" . Or something like that. ... Kenosis 03:07, 5 October 2007 (UTC)
I raised this as an example of a classic urban myth. For some reason some "fact" gets into the bloodstream of the press. It then circulates but is not contradicted, presumably because it is not apparently deemed important to do so. Wikipedia picks it up, arguably some trivia, but someone interpreted it as interesting and noteworthy. Citations are accurate, yet we find that it is disputed. I think the real issue is that, even though the claim is repeated, the press are simply not a enough of a reliable source for the source typing to work with. I don't think it is really an issue of source typing - but in these articles of pop culture, it is the world in which we live. It made an entertaining example as it played out on national radio and Wikipedia in front of our eyes. The claim appeared to be verified, but the sources were flawed, it was "not well verified, over truth". I would have thought that was the correct policy - and it shows that we should be extremely critical of any claim dependent only on the press. Spenny 07:09, 5 October 2007 (UTC)
Kenosis, your suggestion "Any use of primary-source material involving significant analytic or evaluative expansions of the plain meaning of the primary source requires a secondary source" still doesn't address the problem. In many cases, a primary source itself contains significant analytical or evaluative expansion. You wouldn't be able to use commentary and analysis by the author himself, if the author is "very close to the situation being written about". Noddy, for example, could not analyze his own song lyrics. COGDEN 22:20, 5 October 2007 (UTC)

<undent> It's certainly an interesting example of source based research pushing the boundaries of NOR. In a sense the section (which I've not checked out) is based on the secondary news reports, but in checking these reports against primary sources they're found to be unreliable and in error. It makes good sense to report what the various sources say without drawing conclusions, but as soon as we start drawing conclusions we're getting into the field of investigative journalism. Which is why it's preferable, though in my view not essential, to find a secondary source drawing that conclusion. Just out of interest I came across an interesting claim today and commented on it,[11] making it clear that we need a source other than the claimant before making too much of it. Whether the problem would be OR or RS is arguable, probably a combination of various policies. .. dave souza, talk 09:00, 5 October 2007 (UTC)

The case it strongly makes is that you cannot apply policy individually, and that in the end there is some editorial judgement to be made. I am not a fan of verifiability over truth in the sense that if you do not apply good judgement to verifiability you actually get neither. I've said before that my issue is in the wider arena of Wikipedia, we simply are not going to get learned journals and really reliable sources on this stuff, and we are a lot more dependent on the many eyes of Wikipedia to keep things straight. Spenny 10:54, 5 October 2007 (UTC)
If a conclusion is obvious and non-controversial, such as the fact that a record is not listed on a chart, it's actually better to cite the primary source than the secondary source. If a conclusion drawn from a source is obvious and apparent, a reader can verify it. If it is filtered through a third-party or fourth-party commentator, however, such commentators can make mistakes, and you can't verify whether there has been a mistake without going to the primary source. This example is an illustration of that. COGDEN 22:20, 5 October 2007 (UTC)

Color me confused ... about exactly what Wikipedia is trying to achieve here.

  • First, the "primary" source definition is so loose that it seems counterproductive in many contexts. Normally primary sources are treated as the originals, the sources of all facts. Applied to technology, a collaboratively developed specification (e.g. rev 1.2) is a primary source. But allowing people very close to that source, e.g. contributors, have equal weight introduces error: two people may interpret the same paragraph differently, even if both were in the room when it was drafted. (Differences might be just emphasis or context; or could be more significant. They might both "explain" the same thing differently.) Normally (outside of Wikipedia) those "people" would be viewed as secondary sources with respect to that specification.
  • Second, all "secondary" sources will almost by definition be sources of error, so articles ought to prefer primary sources over secondary ones. Using that specification example, only the primary source document could ever be authoritative about what it says. Yet various places (like {{primarysources}}) claim that using secondary (and tertiary) sources is the way to increase accuracy! Which is complete and utter nonsense in all areas where I work; which, I'll note, do not include history or social science. (More than one person has observed that policies can differ between domains of scholarship ... e.g. "hard" vs "soft" sciences, vs engineering.)
  • Third, the reasons to use secondary sources include: to reduce work, since primary sources are sometimes hard to use (complex, specialized/abstruse, fragmentary, inaccessible, wrong language, off-net, etc); to help establish WP:NOTABILITY; providing meta-context, such as peoples' reactions to those primary sources; introducing viewpoints, conclusions, synthesis, and context; and probably a few other things. But they can never include increasing accuracy, and must always be weighed against the common intent of introducing some sort of bias. (Two different trade rags might cover a consortium specification very differently, based on whether consortium members advertise heavily with them or not.)

Seems to me the distinctions should be straightforward, and not so radically different from what I learned in elementary school, high school, college, and the work world. At least in "hard" sciences and engineering, it's a Very Good Thing to use primary sources ... not to be discouraged (i.e. change that policy). If the author/editor of an article can't or won't do that, for whatever reason, then it's fair to report what "quality" secondary sources say, with enough context (citations etc) to highlight the fact that these are not facts from authoritative sources.

--69.226.208.120 01:13, 8 October 2007 (UTC)

"Primary sources" and "secondary sources" are commonly used, well defined terms in the humanities, especially historical scholarship. In sharp contrast, they are not used in the sciences (hard or soft), where the words "primary" and "secondary" have a huge range of uses, but almost never in combination with the word "source". Wikipedia is not a scientific enterprise, it belongs to the humanities, and the standard conventions of enclopedias should be used. Where some specific matter involves primary and secondary sources from a hard science, I suggest that you should do the right thing, use the appropriate source, and be aware that there is a cross-disciplinary terminology issue.

One issue that is not properly addressed in wikipedia policy, is that "reliable" is an adjective that applies to primary sources but not to secondary sources. A secondary source is only as reliable as the primary sources it contains. A better measure of quality of a secondary source is "reputability". Secondary sources make comment, analysis, criticism of something else tranformatative of primary source information. The interesting question is whether that comment/analysis/whatever is reputable. There is no sensible question of it being reliable. The comment/analysis is there, plain to see. --SmokeyJoe 01:33, 8 October 2007 (UTC)

Wikipedia is not a scientific enterprise ... not even when presenting scientific/technical topics? Broken! That's a formula for creating and disseminating falshoods/untruths, and FWIW not how I've seen encyclopaedias present technical topics. I really think that the guidelines should be explicit about the "terminology" issue, because it's just plain wrong for anyone to apply the methodology of history except to historical topics. (History of science is of course not science itself.) When the definition of A is A, no volume of commentors saying that A is instead B should be able to make an alternative story appear to be viable; this isn't Faux news redefining history. ("A is also B" is of course a different statement.)
That distinction between "reliable"/primary and "reputable"/secondary sources makes some sense, but again if one must present technical specifications in such terms, only a primary source could ever be "reliable" about itself. And I'm not sure there's consensus on your point; Wikipedia:Reliable_sources seems to talk about secondary sources, not primary ones, and it hardly touches on reputation. (But I did update that page to bring out point that primary and secondary sources have domain-specific reliability implications.) And if some editor were to try to de-rate an article, perhaps even argue for deleting it, on the grounds that it should instead rely on secondary sources in order to be accurate ... one shouldn't need to deep-dive into Wikipedia policy discussions in order to justify the reality that this is an area where secondary sources can never be authoritative. --69.226.208.120 05:01, 8 October 2007 (UTC)
Speaking of science subjects, I've just had occasion to ask an editor to find secondary sources as references for his proposals for the homeopathy article, all of which are based on primary sources (homeopaths, including himself) and support his claim that homeopathy is not pseudoscience. That certainly looks like classic original research to me, finding opinions and conclusions without a reliable third party source. ... dave souza, talk 07:25, 8 October 2007 (UTC)
Yeah, pseudoscience needs to get properly treated. In "soft" sciences, personal interviews can be essential, and are "primary" sources in the same senses as with historians and for the same reasons. They aren't the team-reviewed specifications of my example. They are not "reliable" in themselves, and turning a set of interviews into a "reliable" source is indeed original research which would to some extent need to bank on someone's reputation. (I'm sure there's a more apt distinction than hard vs soft science, less loaded, but it's late and I can't think of one just now.)
My specific gripe is with respect to technology, where an article about X that presents highlights of the foundational specs for X (primary sources, nothing else could possibly be more reliable) should never need secondary references to improve "accuracy" about describing X. (No opinions, conclusions, or viewpoints involved...) But that is, nonsensically, what the {primarysources}} tag says is the rationale for secondary sources! I could understand wanting secondary sources to help motivate the inference of "notability"; but in this context, not accuracy. --69.226.208.120 09:05, 8 October 2007 (UTC)
Are you suggesting that medical science is a "soft" science? It seems to me that those educated enough to be well versed in "hard science" usage of primary / secondary should be able to understand that there's the more general usage that WP draws on for PSTS. Regarding your example, the present policy explicitly allows primary sources for facts, secondary sources are required for the context including notability, or the editor drawing on the primary source is relying on OR to assert the notability of the subject and to draw conclusions from the primary material. It's not a black and white situation, but in my experience is a useful explanation for many editors including myself. .. dave souza, talk 09:23, 8 October 2007 (UTC)
There are fields of medicine which use "soft" methods. Homeopathic field trials sure seem like they they'd need be one -- if homeopathy is to be any kind of science at all. Re the "present policy", one of the problems many people have noted is that it strongly favors secondary sources over primary ones. I'm not sure it'd be excessive to call that an anti-science bias, but it's certainly contrary to the way most fields of technology operate. That bias is so extreme that it seems to be a real problem in its own right. Especially since so many people just blindly follow policies, without necessarily being versed in that more general usage. --69.226.208.120 15:54, 8 October 2007 (UTC)
Another example of this, to repeat a point above, is aircraft specifications from a manufacturer, which would be primary source under any of the various definitions and interpretations, but are considered the most reliable source within the field. Despite the general distrust for close sources, in aviation the opposite is true. Manufacturers are required to do rigorous testing to government and industry standards, and reliability is ensured both by professional ethics and strict liability for errors. For example, if an aircraft manufacturer overstated the performance for marketing purposes and this resulted in an accident, there would be severe financial penalties. We should never take the results of a casual flight test of one example of an aircraft published by an aviation magazine as more reliable than the results published by the manufacturer, but that is what PSTS would have us do. Dhaluza 13:39, 8 October 2007 (UTC)
I don't think there is a difference in how "primary" and "secondary" sources are treated in different fields. In the end, they are all historical. The purpose of sources is to convert an idea into a historical fact: to prove that some historical person had those ideas and contributed to the field. Therefore, a primary source is an original source of information or ideas, whatever their nature. In scientific fields, the primary sources are usually peer-reviewed journal articles, because that's where new ideas originate. Same with humanities and journalism fields, although new ideas and information can also arise in diaries, interviews, legal documents, etc., which are also primary sources.
In any field, scientific or historical, primary sources are more "accurate" (though not necessarily "reliable" in the sense of WP:RS) and less biased than secondary sources. Secondary sources are just primary sources filtered through the lens of a third party, a lens which often introduces inaccuracies and bias. However, secondary sources can themselves be used as primary sources—and usually are. The majority of sources used in Wikipedia are cited as primary sources, although many of them are also secondary sources. This is why the present formulation of the policy is so confusing, and why it causes so many problems. COGDEN 22:19, 8 October 2007 (UTC)
I strongly disagree with that when it comes to the field of history! Many primary sources (medieval annals, ancient texts, local histories, etc) have been proven to be highly inaccurate. And a primary source authors can be quite biased ... The annals of the Venerable Bede are a prime example. As a Christian Monk writing about (Pagan) Saxon England, he has a distinctly Christian bias in his accounts of things that happened. However, Bede would be considered a very reliable (as in WP:RS reliability) primary source for any article on Saxon England. Sure, some secondary historical sources can be inactuate and biased, but the vast majority are not. Blueboar 01:01, 17 October 2007 (UTC)
That's one example of an unreliable primary source, and there are many others. But on the other hand, things like diaries, tombstones, census records, immigration records, interviews, transcripts of speeches, church records, etc., are more accurate than secondary sources based on those records. In some fields, like religious history, where I work with a lot of articles, primary sources are markedly more accurate than secondary sources. The secondary sources try to create history in the image of the particular ideology or religious view of the secondary source author, and they cherry-pick primary sources to suit that view. Even when primary sources conflict, stating both primary sources together in an article is usually more accurate, and better reflects the uncertainty and inconsistencies of the historical record, than citing one particular secondary source that ignores one or the other and therefore manufactures certainty and consistency where none necessarily exists. COGDEN 18:48, 17 October 2007 (UTC)

I have been pondering on this issue for several days now. For the area that I do most of my contributions which are historical articles PSTS is clear, useful and a shield against charlatans both inside and outside the project. However on reading this talk page I asked myself the question what if there is an article on philosophy for which there are only published articles in journals, from the original paper through to the last published one. I chose philosophy as there need not be any data attached to the first publication of such a paper, (which can cloud the issue over primary sources,) and if there is any previous publication it tends to be yet another published article ("standing on the shoulders of giants" or "turtles all the way down")

In such cases the I can understand the problem people have with PSTS, but my answer is in that in such case all the reference papers for such an article can be treated as secondary sources, or secondary and tertiary sources (since once a Wikipedia article is written there is at least one tertiary source). I think that this policy document could be strengthened if a caveat is introduced into the PSTS section pointing out that in some disciplines such as philosophy and theoretical science, there may be a lack of what Wikipedia calls primary sources and in such cases editors should rely on what Wikipedia describe as secondary and respectable tertiary source when editing a Wikipeda article based on the papers of such an academic disciplines. --Philip Baird Shearer 11:48, 8 October 2007 (UTC)

I can't quite agree about that "lack of primary sources". That's pretty much definitional, since you can't have a secondary source without underlying primary ones. What I've observed is that it's secondary sources which are in short supply, since everyone uses primary sources directly. In my eyes the issue is that many people are wrongly focussed on primary vs secondary sources, as opposed to authoritative or even just reliable ones. Plus, any focus on academic disciplines and methodologies to the exclusion of professional ones (like technology practice) substantially undermines core goals of Wikpedia. --69.226.208.120 15:54, 8 October 2007 (UTC)
I think you have picked up on one of my assertions: this primary/secondary thing has undue weight. The real issue is reliable sources, and it grates to be told that a reliable source is discounted over some journalistic clap-trap. Somehow, the lack of reliability in the nature of that source is compensated for by the fact there is some editor doing his utmost to ensure only unbiased facts get into his paper...oink, oink, oink Spenny 17:01, 8 October 2007 (UTC)
Please that I did not say that that papers written for disciplines such as philosophy and theoretical science are secondary sources ,I said can be treated as secondary sources. The thing which I think you are missing Spenny is in the fields such as history there is a clear difference between primary and secondary sources, and as a lot of Wikipedia articles are concerned with subjects that have such a distinction, to deny that such distinctions exist is in my opinion not helpful. See the example I gave above over interpreting a sentence in international treaties. I'll give another example in article 23 of the IV Hague Convention (1907) it says "it is especially forbidden -... To kill or wound an enemy who, having laid down his arms, or having no longer means of defence, has surrendered at discretion;" it is probably clear what this means up to the word defence, but what does the qualification "has surrendered at discretion" mean? This can only be answered with the use of a secondary expert source. Another example of primary/secondary sources cropped up on the talk page of the River Teme.
The comment: "It would be helpful and informative to have more information about what is in the Journal of Railway and Canal Historical Society articles - assuming these are the best available articles at this time, it would improve Wikipedia to add more of their content. ... 172.143.168.154 08:59, 25 May 2007 (UTC)"
The answer: "The policy of Wikipedia is WP:NOR. That means that only published sources should be cited. Several of the references are to archives; other articles are too detailed to be set out in an encyclopaedia article. If you want to see the articles, I would suggest you find them in a library or order them via Inter-library loans. For example, this [Wikipedia] article formerly claimed that vessels went up to Bringewood with cargos of iron (or something like that). The account for Bringewood survive from 1733 to about 1778 and clearly distinguish freight (always on the river Severn) from carriage (always by land). This information is taken from the original accounts (now in Worcs Record Office), thus not an acceptable source for WP, and clearly contrdicts what was formerly claimed. Peterkingiron 21:23, 26 May 2007 (UTC)"
What I an trying to do above is to suggest that for some academic disciplines this dichotomy (of primary and secondary sources) is not appropriate and all sources in those disciplines should be treated as secondary sources. --Philip Baird Shearer 17:44, 8 October 2007 (UTC)
What are some examples of academic disciplines where this is the case? ... Kenosis 18:14, 8 October 2007 (UTC)
For example disciplines such as philosophy, pure mathematics and theoretical science --Philip Baird Shearer 18:45, 8 October 2007 (UTC)
Re all sources in those disciplines should be treated as secondary sources ... why not just come out and say the obvious: that what matters is the source's authority/reliability? Because treating them as secondary means more than "you won't get flamed for referencing them"; it also means they're sufficient to establish notability. But that's nonsensical, since would allow primary sources to declare themselves as notable. The problem is that this primary/secondary distinction doesn't consistently relate to reliability. So it's better not to use that distinction where that's the issue. --69.226.208.120 20:29, 8 October 2007 (UTC)
It has nothing to do with "you won't get flamed for referencing them" it has to do with the difference between how those sources designated primary and those designated secondary can be used within a Wikiedia article. --Philip Baird Shearer 12:25, 9 October 2007 (UTC)
--Philip Baird Shearer, I agree that the distinction has difficulty in scientific fields (I would argue that the difficulty is even more problematic in humanities fields, including history and journalism). But rather than creating an Wikipedia-specific exception to the meaning of "secondary sources", why don't we just use other terminology. Let's get back to basics. I really believe we can craft a PSTS replacement that (1) prevents editors from using sources to inject their unpublished ideas into articles, and (2) has a consensus, or at least more of a consensus than the PSTS formulation. COGDEN 22:31, 8 October 2007 (UTC)
How can you argue "that the difficulty is even more problematic in humanities fields, including history and journalism"?, because In my opinon in history and other areas like international realations, the distinction between primary and secondary sources is (usually) clear and is useful. See the example (article 23 of the IV Hague Convention (1907)) I gave above. I would like you to address how you would tackle such an issue without the primary/secondary divide. --Philip Baird Shearer 12:25, 9 October 2007 (UTC)
Spenny yes, absolutely: this primary/secondary detour ignores the (cough) primary issue. I don't think I've noticed anyone disagreeing that the issue is what's called "reliability". A thus-far unstated issue is that assessing reliability can unfortunately get political. Maybe some folk would rather have a "bright line" (primary/secondary) distinction than open up those sorts of problems any more than necessary? --69.226.208.120 00:36, 9 October 2007 (UTC)

Section Titles are Original Research?

Lately, a new line of logic has cropped up on WP:SPOILER. The guideline basically says that any section titled "Plot", "Plot summery", "Synopsis" or similar is assumed to contain spoilers and therefore needs no spoiler disclaimer. However, a small group of editors are now using the logic that because section titles are not "verifiable" through reliable source and are immune from said policy, they constitute original research and editors cannot makes assumptions on them (see WT:SPOILER#Removal of "spoilers implied by section title"). This has resulted in an edit war over that particular section of the WP:SPOILER guideline. --Farix (Talk) 15:50, 15 October 2007 (UTC)

Section titles can contain OR. Take for example, a section entitled: "Evidence the X is true", followed by a discussion based upon primary sources. In this case, the title could well constitute a synthetic conclusion. However, I think that in your case NOR is being misapplied. Blueboar 16:32, 15 October 2007 (UTC)
In the example of "Evidence the X is true", it appears that the whole section, and not the section title, would fall afoul of the NOR policy. --Farix (Talk) 18:28, 15 October 2007 (UTC)
A 'small group of editors'? There's just one.--Nydas(Talk) 18:39, 15 October 2007 (UTC)
This is querrelous rules-lawyering at its worst. Of course we can make assumptions about section headers. We can't write an encyclopedia without making certain assumptions about what people will expect to find in a given section. It is not original research to suggest that words have meaning, and will convey their meaning to their audience. Christ. Don't be stupid. Phil Sandifer 22:29, 15 October 2007 (UTC)
The only way an editor can prove they are not performing original research is to provide a citation. To suggest that unsourced headings are reliable completely ignores the policy on verifiability. The criteria for inclusion in Wikipedia is verifiability, not truth. Editors do not count as reliable sources, so headings that do not provide a citation that are written by editors also do not count as a reliable source. You can assume anything you want on Wikipedia, but it's wrong to recommend that readers should assume unsourced material is reliable. User:Phil Sandifer, please be civil. I could edit ==Section Titles are Original Research?== to say ==Phil Sandifer assumes unsourced material is true== and that would be true, but unsourced. The mere presence of equal signs before and after text is not an indication of accuracy. --Pixelface 03:34, 18 October 2007 (UTC)
Section headings can be OR, and they should not be OR, but they almost never are, and it's hard to make them so. If a heading is OR, then that probably means the whole section is OR, and should be deleted. Controversies like these are one reason why we need a more coherent statement of the policy. It is permissible to make obvious or non-controversial inferences from the sources. That isn't OR, and it's permissible in headings. COGDEN 23:20, 15 October 2007 (UTC)
No amount of clarifying policies will stop a rules lawyer hell-bent on getting his or her way, which is what this argument stems from. Phil Sandifer 02:06, 16 October 2007 (UTC)
If you think certain content on Wikipedia is exempt from the policy on verifiability, that's your right. But recommending to users that headings are inherently reliable is baseless. --Pixelface 03:34, 18 October 2007 (UTC)
Perhaps a specific example might help explain how it can be possible: Bans on ritual slaughter#Trends. I didn't see any sources for the proposition that the set of events and arguments described constitute a trend or trends. --Shirahadasha 03:46, 17 October 2007 (UTC)
True... and I have changed the section titling. It wasn't a case of OR, however, just a bad choice of words. Blueboar 12:25, 17 October 2007 (UTC)

Conversion and WP:SYNTH.

I have a question about conversion and WP:SYNTH. Suppose that source X says "John was a Jew when he married Tracy at age 25". Then source Y says "John converted to Christianity at age 37".

Using the above can we say "John converted from Judaism to Christianity at age 37"? Can we list John in List_of_notable_people_who_converted_to_Christianity#From_Judaism?Bless sins 20:06, 14 October 2007 (UTC)

No. He could have been agnostic or atheist or something else at age 36. WAS 4.250 20:25, 14 October 2007 (UTC)

Another example

Here is another example. Suppose source X says "John was a Jew at age 17". Then source Y says "John was a Christian at age 37". But no source says both things at once.

Can this be used to put forth the position that John converted from Judaism?

Can this be used to put forth the position that John converted to Christianity?

Bless sins 01:59, 15 October 2007 (UTC)

"Jew" and "Christian" can both be used to mean ethnicity rather than religion; so those quotes alone can not be used to source either of those claims. Further, one can claim allegiance to more than one religion for social, political or family reasons regardless of whether the religions are logically compatible or not. If one source talked about John no longer practicing one religion and another source talked about John thereafter taking up the practice of another religion then you can use both sources to claim that John converted from one religion to the other; although "convert" might be too strong a word if the reason for the change was other than a belief change (maybe just moved to a new community and then simply acted like the locals). WAS 4.250 18:41, 15 October 2007 (UTC)
Given the very large time gap, if that's the only data you have, it's hard to say what happened without speculation. For all you know, John may have tried paganism for a while. And with only two sources available, one of them might simply be wrong.
If I were writing the article, and these were the only sourced facts about John's religion, I might wonder about notability. If no source exists that explains how he went from being a Jew at 17 to a Christian at 37, perhaps it really isn't that important. Or perhaps it's important, but the facts just aren't known. If it's impossible to write about it without making huge assumptions, then the issue might not belong in an encyclopedia at all. Marc Shepherd 18:45, 15 October 2007 (UTC)

Another example

Wikipedia:When to cite says that "it is not necessary to cite a source in describing events or other details" when writing about the plot of a movie or book because "it should be obvious to potential readers that the subject of the article is the source of the information." What if, for example, a movie is adapted from a book and the book's story takes place during the 1950's and the movie takes place during the 1990's. Would it be ok to state in a "differences between the book and its film adaptation" section that the two don't take place during the same time period? Or would there need to be a source explictly mentioning that difference? Ospinad 17:35, 17 October 2007 (UTC)

In the first instance, use common sense: if it is obvious and uncontroversial then it should not be a problem simply to record the settings. Many book and film adaptations have this and I wouldn't rate it as "research" to make an obvious observation. What you should not do is draw inferences from that, e.g. don't say "The film was refreshed by the use of an anachronistic setting." just observe that they were set in different periods. Spenny 18:32, 17 October 2007 (UTC)
Thanks for your response. I only ask this because of a discussion I am having here about a section in the The Last Mimzy article about the differences between the film and the short story that it is based on. The person that I am debating with says that any comparison between the two would qualify as WP:SYN and I'm saying that it doesn't unless you are drawing conclusions or making assumptions that you aren't getting directly from the plots. Am I overreacting or is he not understanding the policy correctly? I wish I didn't have to be the only person arguing against him. Ospinad 14:22, 19 October 2007 (UTC)

Primary/secondary distinction

I think that in practice there can be no absolute (not relative) definition of a primary source, since the term "primary source" is used of accounts that are not those of eyewitnesses. Thucydides' history of the Peloponnesian War is considered a primary source for information on that war, though he was not an eyewitness of most of the events he describes and obviously drew on sources for what he wrote. His account is primary only in relation to accounts that are based on his.

In that case, Wikipedia rules on the use of sources should, I think, be exactly the same whether the sources are primary (in relation to one or more others) or secondary (in relation to one or more others).

The only distinction might be that the secondary character of a source should be noted whenever its primary source or sources can be consulted.

If these remarks are found irrelevant to the above discussion, please treat them as an aside. Perhaps indeed they have already been made in more appropriate form. Lima 04:44, 21 October 2007 (UTC)

I thought I had also put here a reference to an interesting International Herald Tribune article about the way secondary sources sometimes copy each other without checking the facts. I cannot find it now. Lima 09:30, 22 October 2007 (UTC)
I think you're exactly right. However, some editors here feel very strongly that we should use the terms "primary" and "secondary" in the policy description, and they want to say something different about each. So we're trying to work out some kind of compromise. COGDEN 18:53, 22 October 2007 (UTC)

Where's the consensus?

I don't mean to butt in to what is obviously a long and very detailed discussion, and I did read nearly the entire discussion, but... where is the consensus? The one thing I do like about Wikipedia is that a consensus among editors is like a steering column of a car, but in this entire discussion about sources what seems to be missing is any mention of consensus among sources. Just a thought. Metrax 06:32, 22 October 2007 (UTC)

That is exactly the problem... there isn't a consensus. There is no consenus among editors, and no consensus among sources (or to be more accurate, although there will be consensus among sources in a given field of study, there is no consensus between fields of study.) Blueboar 14:27, 22 October 2007 (UTC)

The recent discussion about drafts 1-6 has been productive, in that new ideas have been introduced, and there have been some efforts to compromise. However, I don't think we can adopt any of these choices at this point as a whole. But maybe we can move forward incrementally. We don't have to explain the whole policy in one big bite. We can start with a general introduction, then add more detail as it finds consensus. From the discussion above, I'm pretty sure that at least the following language represents a consensus:

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" what is expressed in the sources or use them in ways inconsistent with the intent of the source.

In short, the policy of "No original research" requires that Wikipedia users stick to the sources.

This language has been included in all the proposals, and I don't see anybody disputing this language. If anyone finds any fault with it, let me know; otherwise, I'll plan to put it into the policy and we can concentrate on the next steps for adding incrementally more detail. COGDEN 21:31, 19 October 2007 (UTC)

I like the language... and think it should be in the policy... but I would want to know how and where it will go before I fully agree. Are you thinking that it would replace something currently in the policy (in which case, what should it replace?), are you thinking that it would be added on to a paragraph currently in the policy (in which case, where do you think it should be added?), or are you thinking that it should go in its own little sub-section? The devil is often in the details :>). Blueboar 21:39, 19 October 2007 (UTC)
Please don't. Before you make any such changes, please explain where this would go and if it replaces or dilutes anything in the current formulation. ≈ jossi ≈ (talk) 23:10, 19 October 2007 (UTC)
This will go in place of the old "PSTS" section. Then we'll build from there. COGDEN 18:50, 22 October 2007 (UTC)
In view of the likelihood that certain people will insist on reverting anything, even consensus language, and replace it with non-consensus language, I just thought of another alternative: we can divide the PSTS section into two parts: a consensus part, and a non-consensus part. The consensus part would include the above language. The non-consensus part would material such as the what is currently on the page (with a tag noting that the language has not achieved consensus). Gradually, we could shrink the non-consensus section, and grow the consensus section. COGDEN 19:04, 22 October 2007 (UTC)
Your persistent enthusiasm for "shrinking" the present consensus version is noted, but any replacement for PSTS has to be clearly demonstrated to be fully equivalent in effect, and a translation kept in place for users returning to this policy who find it changed from what they're using properly and effectively. ... dave souza, talk 19:46, 22 October 2007 (UTC)
There's no "shrinking" involved here. I'm trying to expand the consensus language. Currently, there is no consensus language in the policy page concerning the use of sources. We need at least some. I don't quite understand why you think that the language in the current draft (as revised added in September, I believe, after an edit war) represents a Wikipedia consensus. You've been following this page close enough to know better than that. What I'm proposing is to start with some consensus language, then we can build on that. COGDEN 00:30, 23 October 2007 (UTC)

(Outdent) This language is not terribly distinct from the current version, if a bit more clear. The currently policy reads (here):

I would recommend replacing it with:

This is basically the same as COGDEN's language above, with an additional note about using information out of context and a shorter concluding statement. This is relatively simple change that seems to have a solid consensus. Thoughts? Vassyana 22:46, 23 October 2007 (UTC)

Looks good. I don't think anybody disputes anything in this paragraph. We can go in a number of directions based on this language, including toward a primary-secondary-type historiographic model (if we can reach consensus on something), or something simpler, or whatever. Everybody should be happy with this, even though it doesn't include all the bells and whistles yet. COGDEN 00:25, 24 October 2007 (UTC)

When can we unprotect?

Discussion has died down here a bit on this page. At some point, we need to unprotect the page and replace the presently-controversial PSTS section with some consensus language. Is everybody willing to suppress their natural inclination to revert while we start out with something clearly non-controversial that we can build on? If anybody is wholly unwilling to allow consensus language to emerge, I'd consider that to be an action against the best interests of Wikipedia, and if that's the case, perhaps the best way to remedy the situation would be to refer the matter to the arbitration committee. COGDEN 20:56, 23 October 2007 (UTC)

Unprotection is a good idea. No page should be on a long-term lock-down, particularly policies. However, being in a hurry to replace the language will only result in protection (once again). We need to be cautious and ensure that whatever replaces the section is consensus language. I support unprotecting this page if people will abide by civil and rational discourage and avoid editing controversial portions until clear consensus emerges. Trying to push through a change without establishing consensus is disruptive and this policy page has seen quite enough of that. Vassyana 22:33, 23 October 2007 (UTC)
One of the few virtues of protection is that it can bring these discussions onto the talk page - removing the problems of reversion and allowing us to propose successively-better versions. Jacob Haller 00:29, 24 October 2007 (UTC)
I agree, and think it's too early for unprotection. I don't sufficent consensus for the sort of change cogden wants to make; I certainly don't support it. FeloniousMonk 04:39, 24 October 2007 (UTC)
I think there is a problem with the way protection has been managed and there should be a greater enthusiasm to unprotect. I don't think that there is necessarily consensus for the change (that is not to say that an individual has the right of veto) but it is stopping a return to normal processes. I would also point out that the protection has little to do with the recent discussions. Last time around it was a spat over a tag and generally the protection has been supported as being related to not gaining a consensus on the PSTS debate, where there has not been an attempt to force a major change in place. The presumption of bad faith when every attempt has been made to follow processes - including magical unwritten ones that are supposed to be acquired by some unknown process - is unhelpful.
In this case, it could be quite reasonable for COGDEN to place his change in the main page with the expectation that it gets reverted or altered - though I am less than happy about the blind reversion and blocking that has been going on here under the guise of the Greater Good. The problem only comes with how people react. If they behave appropriately, there should not be a problem. Arguably this has been blocked for a month over a dispute tag, which has already been pointed out to be an ironic state of affairs.
Let's move on. After all, the worst that happens is a bit of an edit war - nobody dies. Spenny 10:59, 24 October 2007 (UTC)

My observation is that there is a much broader and fundamentally stable consensus for the basic principle of PSTS than would be apparent merely by reviewing the recent discussions on the talk page. I would speculate that unprotection would serve to reveal this, given some time. But given that the "metapolicy" matter of protection of policy pages is not well discussed and not settled at this stage of the wiki's development, I have no preference about the protection issue. My own position on the substantive issue is as I said before, which is that whatever difficulties there may be with PSTS in arguing whether certain sources may be precisely fit into one of these categories (primary, secondary or tertiary), the basic distinction is key in much of the appllcation of WP:NOR throughout the wiki, and many thousands have come to rely on it. I honestly don't see an adequate warrant to be changing it at this stage. ... Kenosis 12:04, 24 October 2007 (UTC)

Do you really believe that there is a consensus for the primar-secondary-tertiary model in the NOR page? Honestly, now. COGDEN 20:48, 24 October 2007 (UTC)
As I said weeks ago, yes, I do think there's a general wiki-wide consensus for it that hasn't changed as a result of the discussion among a dozen or so participants here (with a few additional passersby along the way). But ultimately the way to find out is to unprotect it and see what the longer term response is. Substantively, as I also said in earlier discussion, PSTS is important as a policy, because it directs us to avoid cherrypicking primary sources in those many instances where there are plenty of secondary and tertiary sources to draw upon. We see this issue come up repeatedly in scientific articles, philosophy articles, history articles and religious articles. If it were to be moved to WP:VER I'd have no objection, or even WP:NPOV (because it comes up in the context of negotiating NPOV in the many situations where the sources contradict each other in ways that cannot readily be addressed by stating "Source A says X, and Source B says Y")) -- but such an approach would require discussion on the other relevant talk page too, and also would require a sufficiently strong consensus to do so. Nor would I disagree with putting PSTS on a separate policy page that allocates the principle as needed among the three core content policies, though this too does not appear to me to have consensus at present. ... Kenosis 23:45, 24 October 2007 (UTC)
Okay. Maybe that's an opening for consensus. I at least, for one, have no objection to having a policy page somewhere that makes distinctions between primary, secondary, and (if someone really thinks it adds something meaningful), tertiary sources. I suspect that probably some kind of consensus on that could emerge. But the present expression doesn't have consensus, mainly because it's not precise enough. It currently has too many loopholes. Peer reviewed journal articles, for one, should never be discouraged. They should be encouraged, and I know there's no consensus for discouraging them.
Here's what I would suggest: Let's take the present language in the PTST section, move it to its own article (without changes), tag it as a policy page, replace it in the NOR article with consensus language like has been suggested, and go from there. (Possibly even have a link to it.) This would result in no net change, but might allow this page to come out of protection and remove the controversy. This might not solve all the issues, but at least it would be a step in that direction. COGDEN 00:39, 25 October 2007 (UTC)
That would be a step in the wrong direction as it will clearly dilute this fundamental policy. ≈ jossi ≈ (talk) 00:56, 25 October 2007 (UTC)
(chuckling, respectfully) Jossi, you're so protective of exact current form of the page. Don't you think it might be a good idea to isolate the contentious part so that the rest of the page can be free of dispute? I haven't seen that yet for a policy page but it often works to settle long disputes in the main space. Having a separate PSTS policy page could actually make PSTS more important, not less, as other pages like WP:V and WP:RS share credit for it.Wikidemo 01:22, 25 October 2007 (UTC)
That option has been already discussed, bur never followed up. There is nothing to stop anyone from starting a new page in which this can be explored by interested editors. By the way, I am not saying that the current wording is perfect, and it is not my intention to "protect it". Just that if some editors want to explore alternatives, they should so it without impinging on current policy. ≈ jossi ≈ (talk) 01:55, 25 October 2007 (UTC)
How would you feel about creating a new PSTS page (or perhaps a more neutral title, like "Sources and Original Research") where we move the PSTS material, then transclude it verbatim into the section where it now sits? That way this page would not change one bit initially. People can debate, revert, protect, whatever, to their heart's content on the new page and hopefully calm things down here enough to unprotect this page. If you think that's a good idea I'll propose it again.Wikidemo 02:18, 25 October 2007 (UTC)
Protection policy may not be "settled" at this point in time but there is an existing policy that advocates of a position are not to protect (because doing so is obviously a ploy to avoid the "three revert" rule.) That policy has been violated during this discussion. May I suggest that no such violation occur again in the future (that is, advocates of a position that creates a rule they want to wield actually follow a pertinent rule while pushing what they advocate)? --Minasbeede 13:52, 24 October 2007 (UTC)
Stability is particularly important for policies that all editors have to rely on, and don't expect to have to check every ten minutes to see what's changed. Proponents of changes have evidently failed to gain consensus for these changes on the talk page, and now appear to be proposing to edit war. On any article the correct course is to deal with issues on the talk page or take it to a wider forum, not start disrupting the process by playing with 3RR. Once again, proposals are needed that fully match the effectiveness of the current policy, and provide a clear explanation for editors used to the PSTS description. .. dave souza, talk 14:26, 24 October 2007 (UTC)
The policy as currently protected contains non-consensus claims about use of sources. There does not need to be a consensus to remove non-consensus claims. There needs to be a consensus to include claims in the first place. There has never been a consensus on what "primary source" means in the context of Wikipedia policy. Never. And because there is no consensus for even that, claims about "primary sources" in this policy lead to problems rather than solutions. There is consensus that "sources too close to a situation may have bias or other issues and should be used carefully with regard to that concern." There is also consensus that "sources from non-experts often have slight errors arising from misunderstanding technical details." Using "primary source" in the first case and "secondary source" in the second case is not helpful. WAS 4.250 15:37, 24 October 2007 (UTC)
I don't wish to reenter the debate. My request was a simple one: advocates of any position in this discussion who are also administrators not lock the page: Wikipedia policy forbids them to do so. It's a simple request: honor that policy. --Minasbeede 16:01, 24 October 2007 (UTC)
I fail to understand the logic for unprotecting: the proposed "clarification" language can best be worked out here, even with the page being locked. That some seem to be champing at the bit to unlock the page, while either failing to realise or acknowledge that the discussion must take place here, seem to me to bode ill. &#0149;Jim62sch&#0149; 20:25, 24 October 2007 (UTC)
The discussion here has petered out. Very few of us seem to be interested in advancing the cause of consensus. Besides, I think we have consensus language already. Nobody has disputed the truth of User:Vassyana's language, which is:
"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 to not "go beyond" what is expressed in the sources, use them in ways inconsistent with the intent of the source nor use the information out of context. In short, stick to the sources."
If someone can find fault with this language, please speak up. Otherwise, there is no reason not to put it on the page and see if it sticks. Anyone who reverts will have to justify that reversion by actually commenting on the substance of the language, which moves us toward consensus. I think we've spent enough time seeking ways to show why consensus does not exist: let's try building consensus this time. COGDEN 20:48, 24 October 2007 (UTC)
That reads like a heavy handed threat. Reaching consensus is a two way street, and unprotection is premature since your proposal is not exactly getting a welcome reception. As far finding fault with your proposed change, I certainly can: It ignores the distinction between primary, secondary, and tertiary sources and the qualitative differences often found therein. These are common distinctions and categories that any college student is expected to know and understand. Any change of the policy that replaces the current part describing primary, secondary, and tertiary sources with one that places all sources at the same level undermine this policy, not supports it. Odd nature 20:57, 24 October 2007 (UTC)
That it is bare-bones (at present), and does not include certain controversial items, is not a legitimate criticism. You can't negate consensus language by arguing it doesn't include your own controversial add-ons. If you agree with the above language, you are part of the consensus. If not, then point out something that is false or could be expressed better. Absent this, I don't treat the above comment as a good faith suggestion that would help us find common ground and build consensus. COGDEN 21:03, 24 October 2007 (UTC)
I would have to agree with COgden, regarding the comment from Odd nature. Since there are so many variances of what consititutes a 'primary source', the version from Vassyana is an attempt to come up with language that is still clear as to its meaning, without confusing the issue with terms that don't have clear, concise, and consistent meanings across all disciplines. If this is treated as merely an addition to the current policy, how does this weaken or confuse the policy any more than it already is? I think most people can clearly see and understand OR when they see it in an article, especially if questions are raised and brought here. The big problem people seem to have is comprehending exactly what constitutes a 'primary source' in regards to which kind of discipline (or topic) is being discussed. If somehow consensus could be reached on where and how to move PSTS out of this policy to a more appropriate place, I think that most edit wars and 'regular users' confusion would be greatly alleviated. wbfergus Talk 16:47, 25 October 2007 (UTC)

Would the Arbitration Committee hear this?

Just a hypothetical here, since I still believe we can build consensus through negotiation. However, supposing we can't, that after these three months of discussion and eight-some-odd archive pages, there are still well-meaning editors who just can't bring themselves to work toward consensus and will always revert any moves we make in that direction, would it be appropriate to bring this issue before the Arbitration Committee? I don't know of any instance where it has resolved an issue like this, but there is no reason why it's not within their jurisdiction. We've tried negotiation, we've tried mediation, we've requested comments. Do you think the Arbcom would hear this issue, and is it worth a request? COGDEN 21:00, 24 October 2007 (UTC)

Arbcom's job is to hear cases of conduct such as is being displayed on this page by people acting as owners of the page. Whether arbcom will take the case, or taking it, act on evidence is another question. Typically, they act in whatever way they think best helps wikipedia regardless of rules and evidence. So in addition to evidence of mis-conduct, a convincing case for what is best for wikipedia would be needed. WAS 4.250 21:53, 24 October 2007 (UTC)
ArbCom does not get involved in policy making. If there is no consensus for changes to established policy, it means that there is no consensus and the status quo remains. If there are disputes about the status quo being such, same applies. ≈ jossi ≈ (talk) 00:53, 25 October 2007 (UTC)
The fact that the page has been protected for so long indicates that there is not consensus for the policy as currently expressed. I think it would be appropriate to take a case to ArbCom to set reasonable ground rules for resuming editing without taking specific action for past conduct, but with the threat of immediate action for future misconduct. The extended protection of this critical policy page is a disgrace, and needs to end as soon as possible. Dhaluza 01:00, 25 October 2007 (UTC)
I disagree. The fact that a few editors disagree does not mean that there is no consensus for the policy as expressed. In any case, you can try to file an ArbCom case, but it would be unlikely it will be heard. ≈ jossi ≈ (talk) 01:15, 25 October 2007 (UTC)
There are more than a few editors. I'm keeping a rough tally, and something like half the editors making comments oppose the present PSTS formulation. You would actually expect, given how this issue has been publicized, that far more commentators would support the present language than would oppose it. Plus, you have to factor in the strength of people's opinions on the subject. If more than a handful of people have strong, reasonable and specific opinions on either side of an issue, you don't have consensus. COGDEN 19:15, 25 October 2007 (UTC)

Don't bother ArbCom. Insist that existing rules (in particular the rule that forbids those with a position to lock a page) be strictly observed. That would mean that those with a position could not slap a lock on a policy page whenever they saw a threat to some part of the policy they favor. The rule is clear: if they are partisans in a debate they are not to lock a page. Period. That should be particularly stringently enforced for policy pages as it is ludicrous for a group of editors to on the one hand insist that their pet wording be part of some policy (e.g., NOR) while they blatantly violate an even more important policy (showing their actual disregard for policy, other than as a tool to suit their agendas/desires.)

Review the Wikipedia pages about consensus (this is for personal edification) and notice how one example of improper behavior is for there to be continued opposition to a particular wording, with the opposition being expressed by a changing set of editors. Wikipedia recognizes that as prima facie evidence of an "ownership" problem or offense. Look over the history of the page (don't neglect to review the protection history) and decide, for yourself, if the history of WP:NOR is one that shows full active compliance with the rules and spirit of Wikipedia or if it shows offenses of the type already recognized and described in the policy-related pages.

I'm an outsider. I can watch Wikipedia self-destruct (or be destroyed from within) in a rather detached way. "Destroyed from within" has to mean "by administrators or some group of administrators." Editing, even bad or foolish editing, is not destructive of Wikipedia: Wikipedia is designed to prevent destruction from that direction - that's one if its primary strenths. Wikipedia is open to all to edit, Wikipedia explicitly advocates IAR, Wikipedia admonishes editors to "Be bold." The fundamental nature of Wikipedia is largely evident in and included in those three things. Those three things can be first neutered and then destroyed - the attempt is underway. --Minasbeede 03:54, 25 October 2007 (UTC)

You are missing a very important point and distinction: this is not an article, but an official policy page. Yes, IAR applies here, and there is good reason to protecting of a core policy of the project. As for being bold, be that in articles, but not on policy pages. ≈ jossi ≈ (talk) 04:23, 25 October 2007 (UTC)
I am missing nothing. I am saying that the things I cite describe and pertain to a largely open Wikipedia, which they do. I'm not invoking them per se, just using them to illustrate the extent to which Wikipedia explicitly adheres to being a wiki, with all that means. Locking a policy page (or any page) for months is gratingly contrary to openness. You apparently are missing that administrators with a stake in a content issue are not to lock the page where the content issue arises - and there is nothing I see that says that administrators are exempt from that prohibition on policy pages. There can be, for a while, an assumption of good faith when that kind of locking occurs but when the locking becomes part of the strategy used to protect a non-consensus wording then it can be seen to be exactly the behavior the policy against partisan locking is intended to prevent. While locking as a tool to control is done advocates of the position that is the reason for the locking have the luxury of avoiding good-faith discussion: they have a trump card. I have seen many pertinent issues raised by those who wish to have the wording changed and those issues ignored (no response given) by the advocates of the current wording. The strategy is clear: keep the page locked until everyone tires and goes away, with the result that the whole cycle repeats, because the policy section in question is flawed and does engender legitimate opposition.
A fresh editor could be bold even on a policy page. The natural result in most cases would be that the policy page would be reverted and the fresh editor advised to start a discussion on the relevant talk page. In the case of NOR many seasoned editors have engaged in extensive good-faith discussion on this talk page. They are being shown and will apparently repeatedly be shown that all their good faith is to no avail: the partisans will keep the page locked for as long as it takes to defeat them. (Jossi, below: the standard isn't "agreement," it is "consensus." Some may never agree. The discussion is incomplete, but that is almost entirely because those who want the current wording do not deign to contribute other than to reiterate that they insist on the current wording.) --Minasbeede 13:21, 25 October 2007 (UTC)
Also, WP:OWN is a good thing in policy pages. Yes, I monitor policy pages as part of my duties as a ccommitted Wikipeedian (both as an editor and as an admin), to ensure that policy pages remain as a reflection of established practices, and to ensure that changes are discussed and agreed upon. ≈ jossi ≈ (talk) 04:26, 25 October 2007 (UTC)
That's not relevant to WP:OWN. It would be if you didn't recognise consensus to change policy on the talkpage when you saw it. (I'm not saying that's what happening here, just pointing out that you probably haven't violated WP:OWN, so try not to say it's a good thing. It never is.) Relata refero 07:08, 25 October 2007 (UTC)
I think IAR, BOLD, consensus, etc., apply nicely to the policy pages too, only the thresholds are different. There are walled garden and ownership problems too. What's bold on a policy page would be a very incremental improvement - changing a word here and there, whereas on an article page boldness might mean a complete rewrite or deleting a section. An abrupt change to the way we source material, or worse yet an unstable policy on the subject, would affect millions of articles and Wikipedians. So even changing one word here is a little bold. There's a tension, present in articles but much more crucial in policy pages, between the notion that anyone can participate, and the need for competent, intelligent work by people who know what they're doing and who respect the history and broader goals of the project. It's the usual "Wikipedia is not a democracy" problem. It's not a dictatorship either, nor is it a meritocracy completely. Whatever it is, it's done an extraordinarily good job dealing with a very difficult, complex issue of how to administer a large project in a novel way. Having watched this for the past few months I see that Jossi has been a fair, intelligent, and capable watchdog. There are a few key places like policy pages, arbitration committees, bot approvals, and article deletions, where there's considerable danger if the people guarding the process are not as thoughtful as he is. I think the possibility of Wikipedia collapsing from within due to bad decisions made by an unaccountable and increasingly aggressive administrator class is very real, but not imminent. However, I don't see what that has to do with this page and secondary sources, which I think are a good thing and should not be lightly changed. Wikidemo 04:56, 25 October 2007 (UTC)
Whether a general explanation of primary/secondary/tertiary sources is necessary (or not) is not the issue. The specific issue is whether an explanation of primary/secondary/tertiary sources is necessary *here* in WP:NOR. This is effectively about wanting/needing another WP:SOURCES under WP:NOR.
However, "No Original Research" applies to *every* kind of source. Inversely, a violation of NOR is a violation of NOR irrespective of what kind of source is being used.
"Original research" has nothing whatsoever to do with the quality of sources, and everything to do with an author's presentation of them. Reading->Thinking->Writing. Sources affects the first step, OR affects the second and third.
-- Fullstop 07:51, 25 October 2007 (UTC)

Codgen: What did you have in mind that the arbitrators might be able to decide on? -- Fullstop 07:51, 25 October 2007 (UTC)

I don't really see an issue for arbitration. While frustrating at times, I like the fact that it is very very difficult to change even one word of a policy page. You all know that I have concerns about the PSTS section, and that I am open to suggestions on how to correct this problem. I also think that we are nowhere near the point where we should actually impliment any changes; it should take a LOT more discussion before anything is actually changed. In other words, the process is working... it is working slowly, but it is working. There isn't really anything to arbitrate. Blueboar 13:04, 25 October 2007 (UTC)
I'm not saying we're quite ready yet, either. I think we should try unprotection at least one more time, and see if we can at least take one small step toward consensus. Three months with zero to show for it is not good. Some of us have come a long way toward compromise, but others appear to have no interest at all in obtaining consensus. From how it appears to me, there are one or two editors (and at least one admin) who are absolutely intransigent, and are likely to revert any move, however tiny, toward consensus. It's this phenomenon, really, that would be before the ArbCom. The ArbCom would basically rule that the PSTS section should be reset to a consensus position, and editors and admins should not block the natural movement toward building consensus replacement language. COGDEN 19:02, 25 October 2007 (UTC)
Go ahead and file an ArbCom case if that is what you want. As I said, I do not think that they ArbCom will hear the case. ≈ jossi ≈ (talk) 19:14, 25 October 2007 (UTC)

OR and translation

I run into an interesting case recently (Operation Wilno). This is a translation of a Polish historiography term operacja wileńska. The related event doesn't however seem to have an estabilished name in English historiography; the article was moved to Vilna offensive (as Vilna is more "politically correct" and the article describes a little more than just the Operation Wilno). A disambig at OW was created, but is now considered for deletion, with one of the main arguments being that the very term operation wilno is OR in English (even if it exists in Polish and is a disambig between 1919 and 1944 events). I wonder what's the stance on OR on such translations?

On a related note, I have run into an argument that WP:UE prefers that we translate names instead of using the original non-English title (that commonly comes up when I am writing articles on tiny Polish organizations that don't have an estabilished English name). Should they be translated into English but possibly invented names, or kept under non-English but certainly verifiable Polish names? -- Piotr Konieczny aka Prokonsul Piotrus | talk  13:15, 25 October 2007 (UTC)

Having looked at the AfD comments, I don't think the problem is really with the translation... but the notability of the name itself... even in the Polish original of "operacja wileńska". I think the objection is that the very name "Operation Wilno" (whether in Polish or English) is OR... that it isn't even supported by Polish sources. I am not going to get into whether it actually is OR or not (I am not qualified to judge)... but that is the nature of the objection as I see it.
On the broader question, I don't think simply translating something from one lanuage into English is OR by definition. I suppose that a given translation could be OR... if there were more than one way to translate the word or phraise, and an editor chose a particularly odd way to translate it... but I think that this is probably fairly rare, and would have to be dealt with on an individual basis, when and if someone objects. Blueboar 14:11, 25 October 2007 (UTC)
The Polish name, as used in at least 32 printed books ([12]), is not OR - despite the claims of several editors who refuse to look at evidence to the contrary :) -- Piotr Konieczny aka Prokonsul Piotrus | talk  14:18, 25 October 2007 (UTC)
Indeed accuracy of translation is of secondary importance in this case. If this term is not used in English disambig page fails WP:N my mile. M0RD00R 16:30, 25 October 2007 (UTC)

The translation of Polish: operacja wileńska into English would be Operation Vilnius. So, we are not dealing with the unquestionable translation. Operation Wilno could have been used if the pushers of this name have demonstrated that it is indeed used with this obsolete English name for the city in this context. Not a single English language source is given with such name yet. While exotic and obsolete English names may be used if academic sources use them too, no evidence of the usage of Operation Wilno in English has been provided. --Irpen 20:00, 25 October 2007 (UTC)

But, if the name operacja wileńska is used by Polish sources, then it is not OR to simply translate that name and use it here in the English Wikipedia. Nor is it OR to translate Wileńska as Wilno (both Wilno and Vilnius are common varients of the city's name.) There may be other problems with this title, but I think it is off base to object on NOR grounds. Blueboar 20:18, 25 October 2007 (UTC)

Press as a reliable source wording needs fixing

In general, the most reliable sources are books, journals, magazines, and mainstream newspapers; published by university presses or known publishing houses.

It looks to me like somewhere along the line someone shoe-horned in mainstream newspapers - that sentence does not run properly and confuses magazines in general with something which I think was meant to be far more specific: The Beano after all is from a known publishing house.

Although this is really the subject for reliable sources, as it is written here, it does not describe reliable sources. My only problem is how to fix, as I don't want another essay on RS and also we need to demonstrate context: the press as a reliable source for gossip and scandal but not for science. Spenny 19:54, 25 October 2007 (UTC)

I would say that mainstream magazines and newspapers are reliable sources for the fact that something happened and that it was news worthy. To use your example... the NY Times is reliable for news about a scientific discovery, but not for the science itself (ie the science behind the discovery). In this context, citeing the press is not OR. Blueboar 20:08, 25 October 2007 (UTC)
The Sun, the Mirror, and the Mail are mainstream British newspapers - the wording is meaningless as a test when you consider their deliberate biases and venom expressed in their pages. As we have debated before, there are different types of writing within them. I suppose as a test of OR, any source is proof of lack of originality but that is not what the paragraph is talking about, it is talking about reliability. Spenny 23:22, 25 October 2007 (UTC)

Alternative proposals: straw poll.

I'm writing a couple of proposals based on some of the recent discussion. The purpose here is not to "vote in" one version or the other. I just want to see which version is better, so that we can focus our efforts on achieving consensus. Keep in mind that these proposals are not perfect, and are just a starting point, not a final product. If you feel that both proposals are unacceptable, please don't say "neither". Instead, please create a "proposal 4" or a "proposal 5", etc., or suggest something calculated to move us toward consensus. This period of lack of consensus and flux has gone on far too long, and we need to start making positive proposals and positive efforts to move toward consensus.

Source-usage proposal 1: no primary/secondary source distinction

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 them in novel ways.

Within Wikipedia articles we will find statements of fact and statements of interpretation, analysis or conclusion. It is important to cite the right sources to back those different types of statements. Statements of fact should be cited to reliable sources that clearly demonstrate that fact. Statements of interpretation, analysis or conclusion should be cited to reliable sources that contain the same interpretation, analysis or conclusion.

Editors cannot include their own interpretations of previously-published facts, unless that interpretation is either 1) an obvious and non-controversial consequence of the facts or 2) can be attributed to a reliable source. Nor can editors expand on an author's interpretations of fact, unless that expanded interpretation is also found in a reliable source.

Sources should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources. In short, the policy of "No original research" requires that Wikipedia users stick to the sources.

Source-usage proposal 2: yes primary/secondary source distinction

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 them in novel ways.

One useful way to ensure that a Wikipedia passage does not "go beyond" the sources is to consider the difference between what are often called primary and secondary sources. A primary source is a document or person with original information concerning the matter being written about. Primary sources are usually close to the matter at issue, and present information that is not filtered through a third-party. A secondary source is a document or person who obtains the relevant information from another source. Secondary sources may repeat or comment on primary sources, but do not provide first-hand information.

Wikipedia is not a primary source, and may never contain primary be the original source of information. Rather, Wikipedia is a secondary (or even tertiary) source. Wikipedia may repeat and quote sources, but may not present undocumented information or theories concerning those sources. To the extent Wikipedia comments on a source or analyzes it, such commentary or analysis must not be original, and must be verifiable. Conducting background research that consists of collecting and organizing information from existing primary and/or secondary sources is, of course, strongly encouraged. All articles on Wikipedia should be based on information collected from reliable published sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. However, in presenting such information in a Wikipedia article, great care should be taken not to take the information contained in the sources out of context or present it in a manner that violates the provisions of this policy.

Sources should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources. In short, the policy of "No original research" requires that Wikipedia users stick to the sources.

Source-usage proposal 3: minimal say nothing about sources

{nothing: the general OR policy is sufficient, and we need not make specific reference to sources}

Source-usage proposal 4: clarified intro, revised primary/secondary source section

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" what is expressed in the sources or use them in ways inconsistent with the intent of the source.

Wikipedia articles include statements of fact, as well as statements of interpretation, analysis or conclusion about these facts. Statements of fact should be cited to reliable sources that clearly demonstrate that fact. Statements of interpretation, analysis or conclusion, including the notability of facts, should be cited to reliable sources that contain the same interpretation, analysis or conclusion. Editors must avoid including their own interpretations of facts.

In short, the policy of "No original research" requires that Wikipedia users stick to the sources.

Primary, secondary, and tertiary sources

In taking care to avoid "going beyond" the sources or using them in novel ways, sources can be considered in terms of three categories:

  • Primary sources are documents or other sources written by people very close to the situation being written about. Primary sources that have been published and meet the reliable source criteria may be used in Wikipedia, but only with care, because it is easy to misuse them. For that reason, anyone-without specialist knowledge-who reads the primary source should be able to verify that the Wikipedia passage agrees with the primary source. Any interpretation of primary source material requires a secondary source.
Examples of primary sources include archeological artifacts; photographs; United Nations Security Council resolutions; 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 notebook|experiments or observations; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs; sources which are the subject of the article or are closely involved in promoting the subject of the article.

An article or section of an article that relies on a primary source 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 sources should be careful to comply with both conditions.

In some cases, such as peer-reviewed scientific papers, primary sources can provide the best and most accurate sources of fact. However, articles must be based on reliable third party sources, and care should be taken in considering the reliability of sources which have a close interest in the subject of the article.

  • Secondary sources draw on primary sources to make generalizations or interpretive, analytical, or synthetic claims. A journalist's story about a traffic accident or 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, is a secondary source.
  • Tertiary sources are publications such as encyclopedias and other compendia that sum up other secondary sources and primary sources. Some tertiary sources are more reliable than others, and within any given tertiary source, some articles may be more reliable than others. For example, articles signed by experts in a general or specialized encyclopedia can be regarded as reliable sources.

Secondary and tertiary sources can provide both statements of fact, and statements of interpretation, analysis or conclusion about the facts which are the subject of the article or section. They should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources.

Source-usage proposal 5: clarified intro, primary/secondary source essentials

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" what is expressed in the sources or use them in ways inconsistent with the intent of the source.

Wikipedia articles include statements of fact, as well as statements of interpretation, analysis or conclusion about these facts. Statements of fact should be cited to reliable sources that clearly demonstrate that fact. Statements of interpretation, analysis or conclusion, including the notability of facts, should be cited to reliable sources that contain the same interpretation, analysis or conclusion. Editors must avoid including their own interpretations of facts.

In short, the policy of "No original research" requires that Wikipedia users stick to the sources.

Primary, secondary, and tertiary sources

See Primary, secondary, and tertiary sources FAQ for guidance and examples

In taking care to avoid "going beyond" the sources or using them in novel ways, sources can be considered in terms of three categories:

  • Primary sources are documents or other sources written by people very close to the situation being written about. Primary sources that have been published and meet the reliable source criteria may be used in Wikipedia, but only with care, because it is easy to misuse them. For that reason, an article or section of an article that relies on a primary source 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 sources should be careful to comply with both conditions.

In some cases, such as peer-reviewed scientific papers, primary sources can provide the best and most accurate sources of fact. However, articles must be based on reliable third party sources, and care should be taken in considering the reliability of sources which have a close interest in the subject of the article.

  • Secondary sources draw on primary sources to make generalizations or interpretive, analytical, or synthetic claims.
  • Tertiary sources are publications such as encyclopedias and other compendia that sum up other secondary sources and primary sources. Some tertiary sources are more reliable than others, and within any given tertiary source, some articles may be more reliable than others. For example, articles signed by experts in a general or specialized encyclopedia can be regarded as reliable sources.

Secondary and tertiary sources can provide both statements of fact, and statements of interpretation, analysis or conclusion about the facts which are the subject of the article or section. They should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources.

Source-usage proposal 6: less-controversial compromise revision of #5

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" what is expressed in the sources or use them in ways inconsistent with the intent of the source.

Wikipedia articles include statements of fact, as well as statements of interpretation, analysis or conclusion about these facts. Statements of fact should be cited to reliable sources that clearly demonstrate that fact. Statements of interpretation, analysis or conclusion, including the notability of facts, should be cited to reliable sources that contain the same interpretation, analysis or conclusion. Editors must avoid including their own interpretations of facts.

In short, the policy of "No original research" requires that Wikipedia users stick to the sources.

Primary and secondary sources

See Primary and secondary sources FAQ for guidance and examples

In taking care to avoid "going beyond" the sources or using them in novel ways, sources can be considered in terms of two categories:

  • Primary sources are documents or other sources written by people very close to the situation being written about. Primary sources that have been published and meet the reliable source criteria may be used in Wikipedia, but only with care, because it is easy to misuse them. For that reason, an article or section of an article that relies on a primary source should not interpret the facts stated in the source, unless that interpretation is either (1) an obvious and non-controversial consequence of the facts stated in the source or (2) can be attributed to the primary source itself or another reliable source. If a primary source makes its own interpretations of fact, editors cannot expand on that interpretation unless that expanded interpretation is verifiable by reference to a reliable source

In some cases, such as peer-reviewed scientific papers, primary sources can provide the best and most accurate sources of fact. However, articles must be based on reliable sources, and care should be taken in considering the reliability of sources that have a personal interest in the subject of the article.

  • Secondary sources repeat the original information found in primary sources, and often make further interpretive, analytical, or synthetic claims.

Secondary sources can provide statements of fact, as well as statements of interpretation, analysis or conclusion about the facts which are the subject of the article or section. They should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not already present in the sources.


Do you have a "proposal 7"? If so, place it here. Please make sure it is fundamentally different from the above proposals. If it is a relatively non-controversial an edit or suggestion, you may mark up the above proposals

Which is better: 1, 2, 3, 4, 5, or 6?

Please indicate which of the above proposals is better (or best), and a brief explanation why. Do not say "neither" unless you have a positive proposal to add, or something calculated to move us toward consensus. Marking up any of the above proposals is allowed. If you are okay with both, say that, so we know if there is a consensus to move in that direction.

  • 1 or 2: I think both are essentially equivalent. COGDEN 07:30, 9 October 2007 (UTC).
    • I think version #4 is problematic, particularly the statement "Any interpretation of primary source material requires a secondary source." Actually, primary sources can interpret themselves or other primary sources. #4 is too complicated, and much of it has nothing to do with OR, and belongs in the NPOV policy or its own article.COGDEN 17:42, 10 October 2007 (UTC)
    • As to #5, I think that's much better than #4, but it still contains serious inaccuracies. For example, the statement that articles relying on primary sources can "make no analytic, synthetic, interpretive, explanatory, or evaluative claims" is not true. Primary sources may themselves contain analytical and explanatory claims. There is no Wikipedia consensus that you can't quote a primary source for its own analytical claims. COGDEN 17:42, 10 October 2007 (UTC)
    • 6: I'm willing to live with it, though I don't think it's ideal. At least it's a roughly accurate description of current Wikipedia convention and practice. COGDEN 18:04, 12 October 2007 (UTC)
  • 4 has a lightly copyedited version of proposal 1 as an overall statement, then a revised PSTS subsection related to that statement, explicitly stating the requirement for third party sources and the preferred use for fact of, for example, peer reviewed scientific papers. .. dave souza, talk 22:04, 9 October 2007 (UTC)
  • 4 Seems to be both the best written and keeps in the needed disinction between primary, secondary and tertiary sources which is necessary for NOR to have teeth. JoshuaZ 22:45, 9 October 2007 (UTC)
    • How does adding a primary/secondary/tertiary sources add "teeth" to the NOR policy? To me, it seems to make it gummy, by adding confusion and ambiguity. You want to add your own crackpot creationist theory? Don't cite a primary source thermodynamics paper, just cite an evangelical author's interpretation of that paper that supports your theory. You can't even argue that approach with version #1. I favor #1 because nothing in it is controversial that I can see, and it has the advantage of being clear, concise, and offers a bright-line rule, which adds considerable teeth to the policy, in my view. COGDEN 20:22, 10 October 2007 (UTC)
  • 1 or intro to 4. My personal belief is that detailed discussion of types of sources should be presented in a guideline, not a policy. I believe policies should be limited to clear, non-negotiable principles, while matters that involve implementing and executing principles in ways subject to complications and exceptions that require judgment is what guidelines are for. I think discussion of source categories is the latter type. Best, --Shirahadasha 23:30, 9 October 2007 (UTC)
    • This was something I gave thought to, and while it's good to have explanations including examples included in self-contained policy documents to avoid excessive cross-referencing, it's attractive to strip the policy to essentials as I've done in proposal 5 above. .. dave souza, talk 11:48, 10 October 2007 (UTC)
  • I tend to lean towards #4. I really would like to see something like "most introductory-level textbooks" included as an example of a class of tertiary source. ... Kenosis 23:40, 9 October 2007 (UTC)
  • 1 is best because it avoids the unnecessary diversion into source typing that only causes more arguments than it solves. WP is supposed to be a general purpose encyclopedia anyone can edit, but the search for a general purpose definition of primary/secondary sources that anyone can understand has proved fruitless. There is no need to continue searching for that holy grail, when we have perfectly usable cups on the shelf. Dhaluza 00:28, 10 October 2007 (UTC)
    • However, that significantly weakens policy directly related to the use of sources for fact or analysis, and thus at the heart of NOR. .. dave souza, talk 11:48, 10 October 2007 (UTC)
      • How so? By explaining the NOR policy without unnecessary complexities, #1 actually strengthens the NOR policy. The more I see this assertion as fact, the more it looks like PSTS here to benefit PSTS, not explain NOR. Dhaluza 23:25, 10 October 2007 (UTC)
  • 1 seems best. Most of the sources for most of the articles I've been working on fall through the cracks of PSTS. Jacob Haller 01:43, 10 October 2007 (UTC)
    • Please note that 4 and 5 include new language to make explicit that in some circumstances primary sources provide the best source of fact. Does that meet your concerns? ... dave souza, talk 11:48, 10 October 2007 (UTC)
      • Nope. Some of the anomalous sources are political or theological treatises by one or more of the subjects of the articles. These are neither primary, nor secondary, nor remotely tertiary by the definitions in #4 and #5. These are sources for their authors' views. As written #5 requires non-primary sources (these are neither primary nor secondary; expect edit wars) and favors third-party sources (these are first party, but their critics are third-party. who's more reliable?). Jacob Haller 20:41, 10 October 2007 (UTC)
    • Still 1 - 6 is an improvement over 5 but the source-typing still mishandles interpretive primary sources. We could distinguish fact-heavy primary sources from interpretation-heavy primary sources - lab notes would be the former, political essays would be the latter, and most research papers would be in between. Finally we would need historical sources, as in the issue of the 1880 newspaper. Jacob Haller 21:48, 12 October 2007 (UTC)
  • 5 I think it is a good rewording of existing policy, that is not so controversial as to be damned, but rebalances the wording to address a number of concerns. A useful step forward. Spenny 12:22, 10 October 2007 (UTC)
  • 5 seems best to me. It still enforces the distinction between types of sources without getting bogged down in what type a source is, let that get handled on the FAQ, guideline, whatever it turns out to be. I could live with 4 if I had to, but perhaps instead of third-party, maybe second-hand might be better, though that still can have a conotation of second-rate. wbfergus Talk 13:30, 10 October 2007 (UTC)
  • 4 (first choice) & 5 (second choice) – we need good explanation of types of sources in official policy -- Vision Thing -- 18:23, 10 October 2007 (UTC)
Take a look at Wikipedia:No original research/Sandbox/Various examples and then take a look at primary sources (not to mention secondary sources and tertiary sources). Do we (or anybody), really want that much 'stuff' in this policy? All those definitions of various source types have nothing to do with OR or NOR. It may be useful during some discussions, but has no bearing really on if something is OR or not. It's how the source is used, not the source itself. Or, probably more accurately, it's how the 'data' from the source is used, and in what context, not what 'type' the source is. wbfergus Talk 18:51, 10 October 2007 (UTC)
  • None of the above. I believe this draft is far superior to the above examples 4 and 5. It is far more clear about what is being discussed and why. However, I will still note my opposition to a secondary/tertiary distinction, which I find to be superfluous (and is noted even in academia as an unclear distinction) and therefore an additional useless layer of complexity. I believe both 1 and 2 contain terribly unclear language that will lead to far more confusion and conflict that already exists. I am not unsympathetic to the intent of those proposals, however the form they take is utterly unacceptable to me. In particular, the third paragraph of 1 is rife for abuse. Opening it from the POV of exceptions for an editor's opinion is particularly poor form. Also, "an obvious and non-controversial consequence of the facts" is far too broad, inviting arguments along the line of "it's obvious and not contradicted by the sources". 2 seems to mix without clarity primary sources as the original source of facts & claims and primary sources as non-independent references, or those close to the subject. Vassyana 05:55, 11 October 2007 (UTC)
Not disagreeing. What I thought about this was that it would be healthy to try some incremental changes in the right direction, seeing as getting progress was rather fraught. I did a comparison with what was there now and what was proposed and thought it was a step in the right direction. I had a theory that we should get used to policy changes being changed occasionally :) Spenny 16:01, 11 October 2007 (UTC)
I don't disagree. However, I think the current proposals are as unclear, or moreso, than the current text. As I tried to emphasize, I believe clarity is a central concern. It's no improvement to move towards some equally or more unclear. I think the previous draft I link (rlrubenstein's?) is quite similar to the current version in policy, but significantly clearer in its presentation and intent. Vassyana 23:38, 11 October 2007 (UTC)
  • None of the abovem 1 and 2 are a terrible way to address this issue, with the potential of negative unintended consequences down the line. 3 is not an option. 4 and 5 need work, maybe along the lines of previous proposals as per Vassyana's above. ≈ jossi ≈ (talk) 16:29, 11 October 2007 (UTC)
      1. 3 is an option. It's not a good one, but it's the one that will be forced upon us if we fail to reach any consensus. So you disagree with anything in particular stated in #1 or #2? Or is it just that you so badly want to include the language you favor that you will not accept any option that doesn't include your favored language? COGDEN 22:42, 11 October 2007 (UTC)
  • 1 looks good as it avoids introducing counterproductive complexity. However, in view of so many people who think that splitting sources into categories is useful, I can also go along with:
  • 2 which may be a good start as it provides a compact sketch of source categories without taking the reader for a ride. But what should be mentioned, is the distinctions that have practical use for NOR:
1) Data facts should be backed up by references to the original (primary) sources that give that data, just as is done in scientific literature (not second-hand info!).
2) Interpretation of such (primary) sources should in turn be backed up by references to (secondary) sources that make that interpretation.
It could be useful to give more detailed explanation about primary, secondary etc. sources in Wikipedia but not on the policy page. A manual similar to WP:NPOVT would be appropriate for such explanations. Harald88 11:44, 13 October 2007 (UTC)
The interpretation of primary source data need not come from a secondary source. It can also come from a the primary source itself. Many primary sources (particularly journal articles) contain both primary information and an analysis of that primary information. Citing that analysis is fine, and is not OR. COGDEN 18:16, 15 October 2007 (UTC)
Note also that peer-reviewed scientific publications are more appropriately called secondary sources - which shows again that not too much attention should be drawn to such categories! - A good example of a reliable primary source would be an authenticated letter of one scientist to another in which a result is communicated. 83.77.88.88 12:26, 13 October 2007 (UTC)
Some disagree, but usually peer-reviewed scientific journals are classified as primary sources, since they contain original (primary) information. They are the primary source of the new scientific information, idea, or theory presented therein, which is usually why you cite them. COGDEN 18:16, 15 October 2007 (UTC)
  • None of the above, or rather Only the intro from 4/5/6 or a tersified version of 1. Discussions of primary/secondary sources are not pertinent to NOR (they may be to RS). In this context, original research refers to any paraphrasing (or "selective" quotation) to make a source appear to say something that it does not explicitly or obviously say, or to suppress qualifications/reservations that the source makes. This applies to all sources, irrespective of whether they are primary, secondary or tertiary. -- Fullstop 13:39, 14 October 2007 (UTC)
  • 4 or 5. I feel strongly that the types of sources need to be included. Regarding OR, WP is quite different to what contributors are used to when it comes to academic writing. When writing your thesis, you are encouraged to produce OR -but not here. It's a different beast. <<-armon->> 22:50, 14 October 2007 (UTC)
Even in a thesis, though, citing a primary source is not original research. When writing a thesis, you cite primary sources because they are the best source of old research. Then you add your new ideas, which are indeed OR. With Wikipedia, we just omit the last step. However, we still need to cite primary sources to describe the old research and previously-published ideas. When we can't get primary sources for a theory, idea, or data, we use secondary sources. (E.g., when we can't cite Einstein's thoughts on Relativity, we cite Brian Greene's thoughts on what Einstein was thinking). COGDEN 18:16, 15 October 2007 (UTC)
I think you're missing the point. In an article on Relativity, we don't cite a secondary source on Einstein's thoughts simply because we can't mind-read, but because we need an expert to tell us what thoughts were relevant, and how and why they are. Even if all Einstein's thoughts were recorded, we still need a secondary source -otherwise, anyone could create any number of OR narratives simply by selection what to highlight. This is precisely why primary sources are problematic for our purposes. As these policy pages should a way for good faith editors to understand what we're doing, and serve as a way to avoid arguments, it's important to include it. <<-armon->> 03:49, 16 October 2007 (UTC)
By "thoughts", I mean Einstein's theories and ideas about Relativity, which are highly citable. It might be useful to have a second "expert" filter the ideas of a first expert. But that doesn't mean that citing the first expert is original research. Likewise, citing irrelevant material is not original research. It's irrelevant research.
The problem with this line of thinking that says editors can't be trusted to cite information directly from the horses mouth shows its ugly head most prominently when scientist #2 is actually a crackpot scientist. Is what a crackpot scientist says about what Einstein says more important than what Einstein says himself? How about a creationist, a flat-earther, or an evangelical? True, secondary sources filter irrelevant information, but filters are often ideological, too. And sometimes, secondary sources filter out the truth, by introducing mistakes and misinterpretations. And none of this has to do with original research, anyway. COGDEN 00:45, 17 October 2007 (UTC)
  • 2b below or 4. The types of sources need to be included. The missing here is distinction between scholarly/scientific sources and others. For example, an article in a peer-reviewed scientific journal (primary source) could be better than a newspaper report on a scientific discovery written by a journalist without proper background (such reports are often laughable for a specialist). The scholarly sources can be retrieved by Google Scholar, for example.Biophys 23:48, 14 October 2007 (UTC)
That's an issue for WP:RS, not Original Research. If it's old, published research, even if it's completely unreliable crackpot research, it's not original. It's unreliable, just not original. We need the OR page to stick to the point. COGDEN 18:16, 15 October 2007 (UTC)
Likewise, I think there should be some discussion of reliable sources. It's not off topic because these issues are interrelated. Because WP is a encyclopedia that anyone can edit, in order for it to work, contributors must understand why the prohibition on OR is in place. <<-armon->> 03:49, 16 October 2007 (UTC)
It confuses people to say that "X is original research" when actually, it isn't. I don't see how confusing "reliable" with "original" adds any clarity to the policy. COGDEN 00:45, 17 October 2007 (UTC)

Discussion

"Wikipedia is not a primary source, and may never contain primary information." That is ambiguous as it can be read as, "we cannot copy information from a primary source and use it here." What it means to say is more like "Wikipedia is not a primary source, and may never be used to publish original, or primary, information." I wonder if that is the source of some of the confused debates in the past? Spenny 08:03, 9 October 2007 (UTC)

I agree with the ambiguity, and I'm making a minor edit. COGDEN 08:17, 9 October 2007 (UTC)

Would a sensible other version be to combine the two: talk about facts, then note that there is a useful tool that can be used to help the thinking about sensible sourcing. Spenny 08:05, 9 October 2007 (UTC)

I'm sure that someone believes that these options are an improvement, but I'll be damned if I can see how. Anything that does not differentiate between primary, secondary and tertiary is a non-starter. And by the way, Wikipedia is a tertiary source -- no, really, it is. &#0149;Jim62sch&#0149; 17:30, 9 October 2007 (UTC)
I don't think your comment is a constructive one. We need to arrive at consensus here, and if you don't agree, please offer your own version. If everything is a "non-starter", then we'll never get consensus, and the default will be option #3 (when consensus cannot be achieved, we say nothing). Starting from scratch, I'd like to know which version, #1 or #2, moves us in the right direction toward consensus. Surely one of them is an improvement over saying nothing, right? If we chose #1 or #2 (or both), would you block our progress toward consensus in that direction? COGDEN 20:16, 9 October 2007 (UTC)

I simply cannot contribute to wikipedia and follow these neverending, boring and depressive discussions, so sorry, but i'll post this intervent to make clear how many things are seen in a real wrong way:

I, again, have to say that Wikipedia policies cannot grant the reliability nor search about truth. This example above written:

Caesar's Gallic War is considered a pinnacle of ancient reporting. It however is also considered a fine example of propaganda writing and is known to contain numerous errors of hearsay. We cannot use it to support facts independently of reliable secondary sources without violating the basic principles of no original research and neutral point of view.

In other words, if you post something that is from secondary source it's OK, no thinking is need, on the contrary, activate own mind is forbidden. If you post something that is from primary source it must be discrimined from propaganda claims and facts.

Well, just tell me, Wikipedia searchs the truth only when primary sources are involved? So i can post every newspaper he said/she said and confutate or support every thing, even if it's Murdoch news talking about GWB? Hitler talking about Jews? Mussolini about fascism? No thanks, there's something really wrong with this discrimination. You cannot 'search the truth' only when a primary source is involved, forbid an analisis that MUST be used if you want 'discriminate' between propaganda claims and 'facts' (yes this involved brain activities) on every source, included that famous british racist encycl that almost started wikipedia.

In substance, Wiki forbids 'personal POV' and 'Original Research' but on the other hand, suggest 'analysis' on facts and claimings, but just from primary sources. This seems quite absurd to me. One wikipedian cannot seriously 'made' discriminations between facts and propaganda 'relativized' just to the kind of sources. And what about thertiary sources? Wikipedia.en, is true that boosted thanks to a 1911 british enciclopedy? And so why this 'disprace' for other enciclopedies (written by professionists, moreover)?

Botton line. I think, and i am not alone, that burocracy and tecnocracy, without clue of their limits, tends to de-humanize and worsening the society, treating humans as machines or 'problems'. The simple fact that Wikipedia cannot recognize that every enciclopedy written in the history has both 'OR' and 'NNPOV' and pretend to be 'different' with two zillions not-professional-and generally anonimous-workers is ,frankly speaking' laughable and out of reality.

Claiming the proibition of 'OR' with secondary sources and the 'syntesis' while at the same time is requested to discriminate on the type of sources (primary, sec, third, reliable, unreliable and so on...) is cleary an abstract construction, that gives only to wikilawyers the arrows to take out unliked statements, even if quite trivials.

Another thing, all this display also how the trust on wikipedians is really low because if not, it would be not necessary all this arrays of 'policies'. This bring also to much hipocrisy, because you cannot call for wikilove when you see the 'other' just a potential lyer and troll. And moreover, you can't say seriosly that my (or someone else) brain has no right to think while you asks for 'sintetize' with 'other words' available (and often in conflict) sources, re-writing and merging in a single article several book, magazine, newpapers articles.

Wikipedia calls for the 'use of brain' (for copyright), so it's laughable that then order to 'not express own opinions' during the process. Psicologically, biologically and logically it's simply impossible. And this will been until Wiki will be made directly by bot programs (if properly programmed).

All the words in this and other discussions means only that is pratically impossible to accept the only possible conclusion about OR and other stuff: if you run an opera like Wikipedia you cannot (even if this is the final dream for burocrats/tecnocrats) avoid lies, false statements and many other things, and the only way to improve Wikipedia is simply the partecipation and the discussion about this and that article. Let's let pruned by selection and logic the articles instead to fall in Maccartysm and hunt for witches and eretics. No to zealots, publicans and phariseis, but yes for a community capable to facing one each other work. --Stefanomencarelli 21:40, 9 October 2007 (UTC)

I'm not really following your point, but maybe you can answer this question: starting from scratch, which of the above proposals, #1 or #2 or both, most leads us in the general direction toward consensus, and would you or would you not block progress toward consensus if we headed in that direction? Or do you have a fundamentally different proposal you'd like to add? Or any suggestions to the above proposals? COGDEN 21:54, 9 October 2007 (UTC)
I can clarify his point (I think): It's misleading to state that one should be careful with using "primary" sources. Instead, one should take care with the correct use of any source. If you (artificially!) categorize sources, there are different risks in any category. Harald88 11:19, 13 October 2007 (UTC)

Discussion about version #5

It seems that thus far, versions #1 and #5 are getting the most attention. I don't think anything in #1 is controversial, so I'm not commenting on that, but I wanted to address the controversial aspects of #5 and suggest some possible changes that would move us toward consensus and actually reflecting Wikipedia practice, should we choose to base a consensus draft on #5 rather than #1.

  • The introduction to #5 is great.
  • I have no fundamental problem with saying that it is "easy to misuse" primary sources. I happen to think it's just as easy to misuse secondary sources, and I think this statement is unnecessary and creepy fluffage, but I don't really care.
  • The statement that sections relying on primary sources can "only make descriptive claims" is not accurate. Primary sources themselves may make statements that are not merely descriptive (e.g., an editor can write, "according to Plato (a primary source), Socrates was the greatest philosopher who ever lived"). We can cite non-descriptive, conclusory, or analytical statements if they are in the primary source.
  • The statement that sections relying on primary sources can "make no analytic, synthetic, interpretive, explanatory, or evaluative claims" is demonstrably false. Primary sources may themselves contain analytical, synthetic, interpretave, explanatory, or evaluative claims. For example, in the Diary of Anne Frank, a prototypical primary source, she often makes commentary about the world around her, explains things, and evaluates people and situations. The proposed language would restrict us to quote Frank only regarding cold, hard facts. Such a restriction is not a good description of current Wikipedia practice.
  • I would suggest replacing the above language with the following:
    "an article or section of an article that relies on a primary source should not interpret the facts stated in the source, unless that interpretation is either (1) an obvious and non-controversial consequence of the facts stated in the source or 2) can be attributed to the primary source itself or another reliable source. If a primary source makes its own interpretations of fact, editors cannot expand on that interpretation unless that expanded interpretation is verifiable by reference to a reliable source"
  • The statement that "articles must be based on reliable third party sources" has no meaning to me. What is a "third party source"? Who are the first and second parties? If this is about WP:BLP and self-published sources, than we should say this. But that has nothing to do with OR, so we should take this statement out. It's adequately covered in WP:RS. Let's avoid creep.
  • The statement that "Secondary sources draw on primary sources to make generalizations or interpretive, analytical, or synthetic claims" is not exactly true. Secondary sources do not have to make analytical claims. They can just repeat the primary source without comment. I would suggest replacing this with:
    "Secondary sources repeat the original information found in primary sources, and often make further interpretive, analytical, or synthetic claims."
  • Why do we need to mention tertiary sources? Does this really add anything? Historiographers don't necessarily use this category. It's unnecessary, so why don't we just get rid of it in the interest of avoiding creep?

COGDEN 18:20, 10 October 2007 (UTC)

If it's a primary source in the sense of being raw data, it won't have analytical claims, and even if it's republished as part of another source, we can still use it as a primary source. If it's a third party source making such claims about the subject, then it's a secondary source as defined. If the subject or an author close to the subject makes such claims about the subject, then we can report the fact that they make the claims as in WP:NPOV, but can't take them as a third party analysis of the subject: it might be worth repeating that option here. WP:V states "Articles should rely on reliable, third-party published sources". Were you thinking of something else? ... dave souza, talk 18:51, 10 October 2007 (UTC)
Not as defined in the above drafts. Primary sources are "documents or other sources written by people very close to the situation being written about". That doesn't mean just raw data. People very close to the situation being written about can make analytical and interpretive claims about that situation. Your reference to "third party analysis" seems to indicate you'd like to introduce a first-party/third-party distinction on top of the primary/secondary distinction. Seems a bit too complicated. We'd end up with a 2X2 grid, including: (1) first-party primary source analyses = bad (e.g., Anne Frank analyzing her situation), (2) third-party primary source analyses = good (e.g., Anne Frank recording her father's analysis of their situation), (3) first-party secondary source analyses = good? or is it bad? (Adolf Hitler commenting on Mein Kampf), and (4) third-party secondary source analyses = good (the Ku Klux Klan commenting on Mein Kampf). That sounds way too creepy and complicated. COGDEN 19:37, 10 October 2007 (UTC)

Your Plato example is a good example of what's permissable. That is a descriptive claim. You are stating that Plato said "X", not supporting X by reference to Plato. Regarding your Frank example, the same holds true. Noting her observations as her own is one thing, citing facts based on them is a whole other ball of wax. Your suggested replacement text is a wide-open door to abuse of this policy, practically inviting original research. (I can further clarify this point if it does not follow to you.) "Third party", as used throughout policy, adheres to its normal definition of "independent", or a party not directly connected to the topic being discussed. It may be better for clarity to replace such uses with "independent" to make the meaning clear. Could you provide an example of a secondary source that simply regurgitates a primary source without generalization, analysis, synthesis, explanation or other accompanying claims? I agree on tertiary sources, as I've noted in the past. I think it is superfluous to what we are discussing here and even in academia the distinction between secondary and tertiary is very poorly defined (and openly noted as such). Vassyana 05:15, 11 October 2007 (UTC)

That's a very fine distinction between "making analytical claims" and "citing analytical claims" (that in my experience already goes over the heads of a lot of editors), but it doesn't avoid the problem. Let me alter the examples to illustrate the point: The first example is taken directly from the Socrates article, which says, "Socrates refused to escape for several reasons. 1. He believed that such a flight would indicate a fear of death… etc." (describing two other reasons, and citing Plato's Crito). Here's a non-descriptive claim citing Plato as a primary source, which relates Plato's synthetic and analytical conclusions about Socrates' actions. It was Plato who made the analysis and synthesis.
From today's featured article, The Smashing Pumpkins, we have the statement, "During the tour, Iha and Wretzky went through a messy breakup, Chamberlin became addicted to narcotics and alcohol, and Corgan entered a deep depression" (citing a primary source interview with Corgan). A few sentences later, the article states, "The recording environment was very difficult, and the band fought constantly. The contemporary music press portrayed Corgan as a tyrant during the recording sessions" (citing another primary source interview). These are analytical and synthetic claims, but the analyses and syntheses are okay because in addition to being made by the Wikipedia editor, they were originally made by the primary source journalist, who is cited. I could probably find an example in practically any featured article where a synthetic, interpretive, or analytical claim is made, and a primary source is cited that contains that claim, but it's not cited in a "merely descriptive" way.
Regarding the "wide-open door to abuse of this policy, practically inviting original research", please explain how the draft would encourage OR abuse. It's just a restatement of WP:V which doesn't even require a source at all if a statement in the article is obvious or non-controversial. We don't need to worry about abusing reliable sources when a reliable source isn't even required.
Examples of secondary sources that simply regurgitate a primary source without generalization, analysis, synthesis, explanation or other accompanying claims: (1) in the 1800s, newspapers used to copy original articles from newspapers in different cities, (2) a reference work, such as a book of engineering tables, (3) a copyright violation such as straighforward plagiarism, (4) legal case reporters, (5) video news clips replayed in a news program, (6) re-compilations and anthologies, etc.
COGDEN 09:46, 11 October 2007 (UTC)
An 1880 newspaper is not a secondary source, but a primary source. A plagiarized source should not be used in WP. Media articles, and video clips of newsreels that do not provide commentary or analysis are primary source (only media that analyzes and synthesizes and comments on primary source material is a secondary source). Compilations and anthologies will be secondary sources only if what they compile are such. ≈ jossi ≈ (talk) 16:38, 11 October 2007 (UTC)
An 1880 newspaper is not a secondary source, but a primary source. Fine example of how nobody can agree on what "primary source" means when applied to Wikipedia's sourcing policies. WAS 4.250 16:47, 11 October 2007 (UTC)
Jossi, if you are contending that an 1880 newspaper is by definition a primary source, I would have to disagree. Oh, I suppose that they could be primary sources for some articles (an article on Yellow Journalism, say) but for many articles I would call such a document an out of date secondary source. Think about it... if your contention is correct, at what point does a Newpaper shift from being Secondary to being Primary?... after one year? Five years? Fifty? WAS has it right... a prime example of how nobody can agree. Blueboar 19:09, 11 October 2007 (UTC)
A newspaper article on the Attack on Pearl Harbor, is a primary source for the article on that subject. That is basic scholarship. ≈ jossi ≈ (talk) 01:41, 17 October 2007 (UTC)
That really depends on what you're citing the newspaper article for. If you are citing it for its original (primary) information, it's a primary source. If you're citing it for its repetition of news reports and documents published in other places, it's a secondary source. COGDEN 18:51, 17 October 2007 (UTC)
Even based on the terms of the definitions on #5, these things and people are not primary sources because they are not "very close" to the situation being written about. They are removed by one step, the step of quotation or incorporation by a separate, distinct author who is not at all close to the original situation. COGDEN 22:34, 11 October 2007 (UTC)

I incorporated my comments above into a new compromise version #6, which hopefully should be less controversial than #5, and maybe more likely to lead to a consensus draft. COGDEN 18:06, 12 October 2007 (UTC)

drop the inane primary/secondary/tertiary source stuff (in #5) and you have my vote. :) As I see it, that is the crux of why this NOR discussion is going nowhere. -- Fullstop 22:22, 24 October 2007 (UTC)


Hmm "Research that consists of collecting and organizing material" is rewriting a topic based on articles; in other words copycat researchers( the types who get banned at schools ); what i sometimes miss are original ideas who explain. Rather then claiming to be a publisher or having it from another source, it should be possible to explain to others in ones own words. Probaply the "Math" and "Physiscs" parts could realy bennefit from such a policy, they are realy hard to read for normal people like me. Peter-art 11:23, 26 October 2007 (UTC)