Archive 15Archive 18Archive 19Archive 20Archive 21Archive 22Archive 25

On sources, and the classification thereof

From all these discussions, I can't help feeling that there's some confusion, or a lack of clarity, as to what type of source (primary/secondary/tertiary) a given source is. As I understand the definitions, a primary source is a photograph, video, eyewitness statement (in the context of, say history), or a TV show itself, a script, a novel, etc. A secondary source is anything that isn't primary. A tertiary source is a special case of secondary sources which do not use any primary sources themselves.

Thus, if someone edited wikipedia with information they had gained first-hand, that would be unsourced, and the article would itself be a primary source. If they directly quote an eyewitness or a text, then that's sourced from a primary source, and the article is a secondary source. If the only directly used sources are secondary (including tertiary sources), then the article is a tertiary source (often seen as the ideal for wikipedia).

The principle here is original research - the composition of a primary source, or the synthesis of primary sources into a secondary (non-tertiary) source are research. If this is done for the article, then that is original research. Thus, if a person watches a TV show, reads a book, watches a film, or any similar activity and then writes a plot summary, that's original research. If they find one or more plot summaries elsewhere and reference them, then it's not original research, it's synthesis or composition of secondary sources.

Journal articles aren't a primary source, generally speaking - raw experimental data is a primary source. Primary sources are typically (but not always) devoid of interpretation. Interpretation is added through research in the synthesis of the secondary source. WP:NOR is saying that we don't do interpretation, we do synthesis compilation.

Does that make sense to people? SamBC(talk) 23:07, 14 August 2007 (UTC)

I think you might be on to something here. These terms are used in several different ways, and though I don't think your definitions are actually the most common scholarly ones, they may be the most useful ones for our purposes. We have to remember that the primary source-secondary source-tertiary source system was developed for the field of historiography, which is related to but distinct from the science of encyclopedia-writing. So, we might want to go with a definition more like this and less like this. I think we need to give these terms application-specific definitions for how they are employed in defining Wikipedia's sourcing policy.--Pharos 01:08, 15 August 2007 (UTC)
Sounds pretty sensible to me. Slrubenstein | Talk 01:09, 15 August 2007 (UTC)
Aye! ≈ jossi ≈ (talk)
Sounds about right to me. Vassyana 01:52, 15 August 2007 (UTC)
Yup! except for the very last sentence in the first comment, we don't synthesize sources to come up with original research. Other than that, it looks very reasonable. Dreadstar 02:02, 15 August 2007 (UTC)
The ambiguous interpretations of synthesis are also an issue, of course - synthesis can validly be used just to mean compilation with 'binding' text, while we prohibit synthesis used to generate new conclusions. If it helps, you can think of the problematic use of "synthesis" in my original comment as being replaced by "compilation". Here, I've done it now... SamBC(talk) 02:12, 15 August 2007 (UTC)
Now it's perfect..;) That's pretty much what I thought you meant, but I just wanted to clarify the use of the word 'synthesis'. Thanks! Dreadstar 02:18, 15 August 2007 (UTC)
I also think that the phrasing about being "created at or near the time under study" causes confusion. Just because Tacitus lived in ancient rome, that doesn't make his works primary sources. They're only primary sources if he was writing about things he directly experienced or documenting literally the reports of others. I've known history students at pretty much every level get confused by definitions using phrases like that. I agree that wikipedia should have its own (tighter) definitions of the terms, so as to define them unambiguously for use within policy, guidelines and discussions. Uses in articles should, of course, be based on the (looser) general definitions already in the encyclopedia. I think my explanations just now are also useful because there's no effective explanation I can see in WP:NOR that explains why we don't use primary sources, and why using them is original research. SamBC(talk) 02:09, 15 August 2007 (UTC)
Tacitus is most certainly a primary source in historical studies. That phrasing is correct. He is only secondary sources in a strict sense of the definition, outside the context of pragmatic application or the context of the source itself. In practice, such historical writings are treated as primary sources. Vassyana 02:57, 15 August 2007 (UTC)
Either we're at cross-purposes, or my historian friends disagree. It probably depends on the focus of the study in question, on the context. For example, studying the society and era of Tacitus in general, he'd be a primary source. For more specific matters that were already historical at the time (I'm given to understand that he wrote histories), then he'd be secondary. Of course, I might be getting Tacitus confused with other classical or ancient historians, in which case consider my point without the specific example. SamBC(talk) 03:06, 15 August 2007 (UTC)
Tacitus wrote about events before his time and during his time (though pragmatically stopping in the reign of the previous emperor). Ammianus Marcellinus did the same (mixing older material with eyewitness material). There are pragmatic reasons to group the ancient sources together and modern interpretations together. Howeber, the same ancient/modern split doesn't work with archaeological data, where the original context is the only pure primary source, and even the isolated artifacts, their catalogue records, published or unpublished descriptions, etc. are selections and interpretations thereof. Jacob Haller 03:45, 15 August 2007 (UTC)
That's contrary to common practice, as well as basic university teaching in the subject. For example in this syllabus, the historical texts (including Tacitus writing about what was history at his time) are primary sources.[1] Students of history are commonly required to divide their bibliography in primary and secondary source sections. Tacitus would be firmly in the former. Here's a few more examples.[2] [3] [4] [5] Vassyana 04:02, 15 August 2007 (UTC)
This misunderstanding is why I have a problem with us defining these terms in the most useful way for Wikipedia. Because the more common uses of these terms use them otherwise. It leads to confusion when Tactitus's works are widely known as "primary sources" but are not to be considered such by our definition. The issue is that if Tacitus writing about what was history at his time is to be a "primary source" than no "secondary source" can exist about September 11, 2001 attacks or Global warming. I have always thought we should get to the bottom of what make "primary sources" problematic and describe that in the policy. Get rid of the terminology which lead to misunderstanding and simply explain what the real problem is with using diarys and raw data as sources. I know that is not a popular opinion, but there is no way to make "primary source" not mean all the definitions that it commonly means so we can single out a few types of sources that are pitfalls in original research.--BirgitteSB 14:15, 15 August 2007 (UTC)
In particular I like the definition of primary sources in the UNR, that includes this often forgotten attribute: created or otherwise produced during the time under study. For example, a newspaper article for the Attack on Pearl Harbor, is certainly a primary source. In WP, some editors believe that newspaper articles about a current or recent event are secondary sources, when actually they are not. ≈ jossi ≈ (talk) 02:13, 15 August 2007 (UTC)
Actually, some are, most aren't. You do get news reports that are reviews and synthesis of other reports. It depends what you're taking it as a source of. For example, a report written by someone who wasn't there, that clearly bases all of its information and conclusions on sources who were there (and identifies the information as such) are secondary. For example, a lot of news coverage of the current foot and mouth disease "crisis" in the UK take the form of reviews and analysis of what has happened "so far". However, a news report saying "and nows there's another case, argh!" is primary. SamBC(talk) 02:20, 15 August 2007 (UTC)
Yes, a fine distinction that many people do not get. ≈ jossi ≈ (talk) 02:23, 15 August 2007 (UTC)
Well, if we are going to put together domain-specific definitions for wikipedia, then that clarification (written a bit better) should probably be part of it. It's looking like ending up as a "wikipedian's easy guide to classification of sources". Or maybe idiot's guide ;) SamBC(talk) 02:25, 15 August 2007 (UTC)
I just added another use of compilation to my original point, as an "or" with one instance of synthesis. Just for clarity. It's bolded so it can be seen easily even if you've already read it. SamBC(talk) 02:25, 15 August 2007 (UTC)

Further to all this, I've just started a new "proposal" at Wikipedia:Classification of sources. I don't know for sure that it'll be useful, it's certainly not in the right tone yet, and if it is useful I don't know if it should stay a separate page. However, that's useful for working on it. If you think we should have a clear internal definition or illustration, please join me and help to build it up and put it in the right tone. It's mostly copy-paste from my initial comment above. SamBC(talk) 02:44, 15 August 2007 (UTC)

Does this not risk creating undue weight for the concept? I think we fell down that whole discussing synthesis - getting so involved in the concept that you forget that all it is saying is "No OR" in a particularly obscure section. Spenny 14:54, 15 August 2007 (UTC)

Village pump

In case anyone's interested, there's a Village pump discussion on the pri/sec/tert sources issue. Dreadstar 21:23, 15 August 2007 (UTC)

Primary sources

"Original research that creates primary sources is not allowed. However, research that consists of collecting and organizing material from existing primary or secondary sources within the provisions of this policy is, of course, encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia.

Although most articles should rely predominantly on secondary sources, there are rare occasions when they may rely on primary sources (for example, legal cases). 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."

  • Shouldn't primary sources be discouraged as strongly as possible? When using only primary sources, it is very difficult (if not nearly impossible) to build an article without including original research by drawing conclusions from the information. Secondary sources allow the sourcing of conclusions about primary sources without engaging in original research.
  • Isn't "[o]riginal research that creates primary sources" nonsense? Original research can only prouce a secondary source. Writing an allegedly inspired religious text would be creating a primary source, but would hardly be original research. Drawing upon such a text for reference material would be a conflict of interest and the result would probably be original research.
  • Should we provide an encouragement to use primary sources with secondary sources?
  • Legal cases? Seriously? Legal cases should be among the least desirable situations to use primary sources. Legal scholars very often debate for decades what a case is about, in what fashion and what exact conclusions were drawn. If secondary sources find it so easy to debate such "simple" facts extensively, how can those primary sources possibly be a good example for the no original research policy?
  • The line between descriptive and analytic/interpretive claims is very thin, highly subjective and easily abused. As an example, Jesus says it is better to cut off a hand than to sin. This would be easily presented as Christian scripture encouraging severe self-mutilation if only the primary source (the Bible) was used. However, that is not an accurate presentation and not supported by reliable secondary sources. While many are prone to respond "that's silly", imagine it were an obscure religious group with whose beliefs you were not at all familiar. Now imagine it is a small obscure group that is often called a cult. The abuse potential is obvious.
  • Also, notability and neutral point of view would seem at least to preclude a heavy use of, or dependence on, primary sources. The policies form a coherent unit, mutually supporting each other, not providing contrary encouragement or loopholes.

Just some thoughts. Your thoughts? Vassyana 23:21, 11 July 2007 (UTC)

"Original research that creates primary sources is not allowed. Research that consists of collecting and organizing material from existing primary or secondary sources within the provisions of this policy this and other content policies is, of course, encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia.

Although most articles should rely predominantly on secondary sources, there are rare occasions when they may rely on primary sources (for example, legal cases). 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."

There. ≈ jossi ≈ (talk) 00:07, 12 July 2007 (UTC)

I strongly support these edits. In particular, I agree that the sentence "Original research that creates primary sources is not allowed." is junk - how can someone "create primary sources" by editing Wikipedia? Enchanter 00:56, 12 July 2007 (UTC)
If you witnessed something and added what you saw to Wikipedia, you'd be creating a primary source. SlimVirgin (talk)(contribs) 19:16, 21 July 2007 (UTC)

I agree with the edits in the guideline. And I understand the preference for reliable, secondary sources, since these may be the result of more fact-checking than the primary source could or did. However, when secondary sources are not (freely) available, I have no problem with using a primary source provided the use thereof is clearly indicated in the article.   — Xiutwel (talk) 08:08, 12 July 2007 (UTC)

Due to the supporting comments and lack of objection, I have changed the page to reflect Jossi's proposal. Vassyana 13:10, 18 July 2007 (UTC)

I can't understand the rationale for devaluing primary sources and privileging tertiary ones. It results in nightmarish edit wars, where a few people with dictionaries sometimes go around inserting misleading definitions, reclassifying ideologies, and rewriting history to fit. (For example, Pierre Proudhon and Benjamin Tucker identified themselves as socialists; Marx refers to Proudhon as a socialist, etc., but some dictionary definitions of Socialism have excluded them.) Devaluing primary sources means that one editor with a dictionary can trash a political article and no amount of facts can get in the way. Jacob Haller 21:19, 2 August 2007 (UTC)
A related problem is that the secondary sources may simply contradict the primary ones. Jacob Haller 01:17, 3 August 2007 (UTC)
Generally, politics is an exceedingly well-covered area, in terms of references. Beyond that, what you're worried about ("dictionaries") would probably be original research! Presenting an argument about whether or not X person fits the dictionary definition of Y label is certainly original research, unless that claim can be cited to a reliable secondary source. Even then, information should be presented in proportion to its appearance in reliable sources. Extraordinary claims require extraordinary proof and extreme minority views should probably be excluded. So, it seems that your concern is already covered. :o) Regarding secondary sources contradicting primary sources, stick to the secondary sources generally. In such cases, there is usually a significant discussion regarding the contradiction and why it exists, which usually provides a solid analysis of the material. If it's an unusual or extreme claim, it may require additional references or removal of that claim. Just some thoughts. Cheers! Vassyana 12:42, 7 August 2007 (UTC)
I don't have a poly-sci library. I have to make do, and the primary sources are much more easily available than most secondary sources. As far as politics goes, I am mostly involved in working on anarchism, socialism and their related pages. For anarchism, primary sources are readily available, but except for An Anarchist FAQ, decent secondary sources require serious cash. For socialism too, primary sources are readily available, but secondary sources are largely divided into those covering specific movements: some mostly cover Marxism, some mostly cover social-democracy, some also cover anarchism, etc. and many of these are partisan sources, dismissing the "wrong" movements as "ancient or medieval" (for movements which developed in the mid-19th century). In both cases, some scholarly works use special definitions remote from the movements, e.g. political scientists who use "anarchism" to mean "polycentric law" or economists who use "socialism" to mean "central planning."
And there are editors who insist on using dictionaries and rejecting primary sources, or who regard the use of one word to qualify another word as OR/IS unless I can find the same two-word combination. Jacob Haller 17:14, 7 August 2007 (UTC)
Public and university libraries offer a low to no cost solution. Additionally, there are many sites (such as Questia) that allow online research for relatively inexpensive rates. Anarchy, socialism and other such widely covered subjects hardly require any serious investment in cash to research. The sources are plentiful, available and easily had at no to minimal expense. Out-of-date sources and original research based on primary sources just aren't up to par. Most of us don't have personal libraries, but we "make do" and manage to find modern reliable sources just the same. Vassyana 10:30, 9 August 2007 (UTC)
The public libraries where I like have nothing in these topics. The university libraries are 90 minutes away (mostly waiting between buses) and $2.60 for the round trip, as well as closed for the summer. I can't afford to waste that much time or pay that much money for secondary sources. Jacob Haller 17:22, 9 August 2007 (UTC)
I'm sure the public libraries have something, even something reliable, on the topic. You're telling me they don't even have general history books that discuss anarchism or socialism? You're saying they do not have historical or biographical books that discuss Marx or Proudhon? I'm sorry but that's more than a little hard to swallow. Your fare is cheaper than mine and about the same time for the trip, so I've little sympathy on that count. There is a minimum expense and effort involved in doing research. Vassyana 23:14, 9 August 2007 (UTC)
In the natural sciences, the problem is the other way around. Chains of evidence are created through generation of data and the attempt to re-create said data. When erroneous claims stick, it is very often due to not reading the primary source. See for example Simkin & Roychowdhury, Complex Systrms 14: 269 for a discussion of the scope of the problem. I think the devaluation of primary sources is a bit of a bias. Consider for example Perrin's_Beaked_Whale#_note-1. To argue for or against the older (the specimen description, mistaken for another species) or the younger source (which is the original description of the species) being correct would be OR; as both vertebrates and invertebrates are known food in this genus it simply cannot be told what is correct. I am not sure which one is a misprint; the primary source is rather unavailable. I will try to check whether there has been an erratum but the species description not mentioning such suggests that the primary one indeed read "invertebrate". Any tertiary source relying on the original description would be expected to repeat what is apparently an error. In my experience, especially in the more specialist fields of the natural sciences, the preference of sources should be primary (if available) - secondary - tertiary. Especially pop-sci tertiary sources are usually full of errors; OR if you will on the reviewers' behalf. Wikipedia provides a novel means to prevent this (as no error must be stand uncorrected for more than a few seconds after its discovery), and this opportunity should be duly honored. Dysmorodrepanis 06:50, 19 August 2007 (UTC)

On the appropriate uses of primary sources

It occurs to me that there's a pretty good, relatively objective criterion that we could use to cover when primary sources are appropriate, and it ought to cover pretty much every case:

Primary source are acceptable when they require and receive no interpretation in the article, but are merely a source of an objective fact.

Thoughts? SamBC(talk) 02:16, 15 August 2007 (UTC)

Mmmmm, no... Primary sources are easily abused, by selective quoting, quoting out of context and other fallacies. That is why their use is deprecated, and secondary sources always preferred. Also, if a primary source has not been used in any secondary source, then its use in a WP article will violate WP:NOR. And if the primary source is cited in a secondary source, then use the secondary source. ≈ jossi ≈ (talk) 02:21, 15 August 2007 (UTC)
So we can't use, say, census data to describe populations of countries/cities, or similar primary sources for heights of buildings, or official match stats to provide sports information (if you think that records of individual sporting events should be in wikipedia, anyway)? SamBC(talk) 02:27, 15 August 2007 (UTC)
Census data, in particular its reliability and accuracy, is definitely an exception, not the rule. Additionally, it's general reliability and useful as a plain source of data is well-accepted and well-covered in reliable sources. We're quite permitted to apply common sense to determine if something is a notable exception to the rule. Heights of buildings and sports statistics are also both easily acquired from secondary and tertiary sources, when notable. Vassyana 02:52, 15 August 2007 (UTC)
Any source is readily abused by selective quoting, quoting out of context and other fallacies. Any source is no better than the interpretation given by the editor. A secondary source has the joy of this potentially being applied twice, once by the original author, and again by the WikiEditor - this is certainly true of journalistic sources. Spenny 02:28, 15 August 2007 (UTC)
Secondary (and tertiary) sources provide context, analysis and interpretation. Primary source abuse potential is much more significant. See below for an example. Vassyana 02:52, 15 August 2007 (UTC)
Whilst secondary sources can provide those things, to be able to validate them, it potentially requires some form of analysis: does this context show a bias of the contributor, is it a complete representation, is it accurate? We are dependent on the editor accurately representing information - distortions are inevitable in the required paraphrasing to wikify the information. This is true regardless of the source. I don't disagree with arguments such as the biblical distortion example below. I do disagree that assuming secondary sources are somehow more immune from the same issues and we have the extra issue of finding out whether the source is working to the same agenda as the editor. Spenny 09:08, 15 August 2007 (UTC)
(ec)I have to agree with Jossi. Such a plain statement is bursting with abuse potential. After all, it's not disputable that (according to the Bible Matthew 5:28–5:30) Jesus said:
"If your right eye causes you to sin, gouge it out and throw it away. It is better for you to lose one part of your body than for your whole body to be thrown into hell. And if your right hand causes you to sin, cut it off and throw it away. It is better for you to lose one part of your body than for your whole body to go into hell."
While most contributors to the English Wikipedia are familiar enough with Christianity to spot abuse, realize there are similar examples from religions and so-called "cults" you are probably not familiar with. Even without making explicit OR statements, material such as this can be juxtaposed and used to present a very negative (and inaccurate) POV. Selective quoting, especially quoting from religious texts and leaders, is a serious concern. Even worse, it could be cited for the plain claim: "Jesus said it was better to gouge out an eye or cut off a hand, than to sin." That's a "plain reporting" of what that primary source says. It's not explicit OR, but certainly in many cases it would be effectively the same. It's fairly self-apparent how such paraphrases could be badly abused. This kind of problem is part of why primary sources are discouraged in favour of secondary sources. Vassyana 02:46, 15 August 2007 (UTC)
But that "plain reporting" is plainly not "objective fact", at least to any sensible editor. Plus, the bible is a bit of a straw man in such discussions - it can be used to counter most things in terms of quoting and sourcing, and probably to support quite a few of them as well. It's something of a boundary case. However, my suggestion could be amended by a guiding example, such as "such as official figures". It's probably worth also saying that the source must be clear in the statement. SamBC(talk) 02:54, 15 August 2007 (UTC)
It's only obvious to "any sensible editor" because they are familiar with the religion and text, as I mentioned immediately in my comment above. Vassyana 04:04, 15 August 2007 (UTC)
Accepted, the first part of my answer was hasty. What about the suggestions of qualifying the statement with guiding examples and requiring clarity of source? SamBC(talk) 04:06, 15 August 2007 (UTC)
I've been chewing on that. I'm a fan of {{sofixit}}. This is my suggested revision:

"Primary sources may be used in articles in the context they are employed in secondary and tertiary reliable sources. Primary sources that are considered accurate by reliable sources may be used for citation of "pure fact", including such raw data as U.S. census statistics."

I think it would cover the most appropriate uses of primary sources, provide some reasonable level of protection against abuse and include your sentiments, unless I've misunderstood them. Vassyana 04:22, 15 August 2007 (UTC)
It seems a little awkward to say "in the context they are employed in..." as that requires them to have been used in those sources. Apart from that it's great, but I'm not sure how to overcome that aspect. SamBC(talk) 04:31, 15 August 2007 (UTC)
I came come up with many of examples not related to the Bible, Sam. As Vassyana said above, height of buildings of such other examples, are the exception and not the rule. For example, To suppose that the eye with all its inimitable contrivances for adjusting the focus to different distances, for admitting different amounts of light, and for the correction of spherical and chromatic aberration, could have been formed by natural selection, seems, I freely confess, absurd in the highest degree. Guess who said that. And guess what is missing... ≈ jossi ≈ (talk) 04:07, 15 August 2007 (UTC)

(outdent) A couple words and an order swap might do the trick!

"Primary sources that are considered accurate by reliable sources may be used for citation of "pure fact", including such raw data as census statistics. Primary sources may also be used in articles in context as they appear in secondary and tertiary reliable sources."

Thoughts? Vassyana 05:17, 15 August 2007 (UTC)

  • Quotes: I read this as "if a secondary source used the same quote" which seems too restrictive to me. I would rather include a general warning about selective quotation, require solid citation, and encourage fv tags and/or outright removal of manipulative quotations.
  • Pure facts: I agree but think there may be some exceptional cases (look at J20 or J27).
  • Author's views: I would add another category; people can cite primary sources as sources for their authors' views.
Regardless of the type of source, context is as important as content. Jacob Haller 06:21, 15 August 2007 (UTC)
The proposed wording says nothing about quotes and really it should not be a primary concern. Quotes should be used sparingly and only when they add something to the article without effectively creating original research by leading the reader. Preferably, quotes should be employed as they are used in reliable sources. We're allowed to use common sense to determine exceptional cases. Your concern about "sourcing" an author's views are covered by statement permitting primary sources to be used in the same context as they appear in secondary and tertiary reliable sources. So, if an author cites a primary source, or a reliable source indicates it was a source for their views, we can include that appropriately. If such information is not contained in reliable sources, advancing the unpublished, or unreliably published, claim that X primary source was the source of Y views by Z author would be original research. Vassyana 13:42, 15 August 2007 (UTC)
Another approach would be to say that the primary source should not be used (to advance a position) but can be used as justification to remove citations of dubious nature (and therefore potentially unverified statements). In other words, we cannot base articles on them as that leads inevitably to analysis, however, we should be able to use them in discussion with considerable weight to back up assertions of NPOV etc. I feel there is a WikiLawyering approach that is used where if you present a primary source even in discussion, that does not count. You must have had the "this is an inviolable secondary source" "but it contradicts the primary sources it is based on" "Tough - your source is a primary source so it does not count. Verifiability over truth." "???" discussion? Yes, that countering is subject to the same distortions but let's AGF. Spenny 09:22, 15 August 2007 (UTC)
You are suggesting that primary sources be restricted to talk pages, and such a restriction is not likely to gain anything close to consensus. Your assertion that using primary sources "leads inevitably to analysis" is way off base. If it leads to analysis by the reader, that's a good thing. The problem is if it leads the editor to put their analysis in the article text. But this is not inevitable, it's just one of many ways editors may go astray. We can't make simplistic prescriptive guidelines to address every one of them, because if we did the net effect would destroy Wikipedia. Dhaluza 09:41, 15 August 2007 (UTC)
I'm trying to square a circle, and paraphrasing arguments put to me, and I quite agree, the issue is not the source, but the use of the source. Primary sources are deprecated, I am told they cannot safely be used and that using a primary source to disprove a quote or other analysis is de facto OR. Yet that leads to a policy that supports nonsense, which does not seem sensible nor in the spirit of policy which is to keep out nonsense and factional views (I think). Policy works against editorial judgement - verifiability not truth - yet I can verify fact by reference to a primary source of the highest repute. I don't think my brain is yet mangled enough for WikiPolicy. Spenny 12:43, 15 August 2007 (UTC)
You've been told about right, except that primary sources are not completely excluded. We are simply encouraged to use them with care and rarely. You heard right that using a primary source to try and disprove a secondary source is indeed original research, in the most basic sense. If you believe a fact is significantly wrong and the topic is notable, there should be enough third party sources to verify or disprove the claim. It's certainly good scholarship to perform original research, and I will not disagree with that. However, it is not good encyclopedia building. Wikipedia exists to summarize extant scholarship. The basic policies of Wikipedia are built on that assumption. Those interested in original research and disproving extant sources, should find a reputable publisher for their views and/or check out Wikiversity and Wikibooks. Vassyana 13:42, 15 August 2007 (UTC)
I am not convinced by the "You should be able to find a source" view. The reason is that if you go down the source based research route, you can end up putting together snippets. This seems even to be encouraged by citation, as writing well thought out sections is prone to being challenged as synthesis by the cite-happy. These individual snippets can be well-sourced in their snippetness, but are not necessarily the whole picture. We clearly discuss citing at sentence level. Good sourcing and writing is about the whole source and the whole context. Controversial statements are best sourced on the basis of complete discussions not quoting convenient Googled sentences. We can see that the whole is wrong, we can find sources that discuss the entirety from different aspects, but we cannot necessarily unpick a statement from a misleading source. When someone is seeking to "advance a position" the snippet approach is one technique used: the whole is advancing the position, yet is unassailably composed of cited elements and is not OR. (I think this also depends on subject areas and the likely nature of sources: history and science, things are more straightforward than popular culture or current affairs). Spenny 14:28, 15 August 2007 (UTC)
If there aren't enough sources available to make a complete encyclopedic article, the article should like head over this way and its notability is questionable. If there are no reliable sources for a claim, it cannot be verified and likely violates WP:NPOV, which demands that the articles reflect the current state of existing scholarship. On top of all that, if there's no source for a claim/POV/theory (particularly if the material in the article is arranged to build towards it, or support it), there's an assumption of potential original research. You're quite right about the snippets and Google-searching, but that's not a condemnation of source-based research. It's a condemnation of poor research. There's little replacement for getting to the library, paying for Factiva (or another service), or otherwise acquiring access to a wide range of published materials. You're also quite right that context is important, and it should be reflected in the article, which is why using reliable sources from third party publishers with a reputation for accuracy and fact-checking is so important. Just some thoughts. Vassyana 04:27, 16 August 2007 (UTC)
I totally agree with your comments. The reality of Wikipedia is that there is a difference between the aim of producing the ultimate reliable encyclopedia with a robust view on notability, and the de facto public view of it as the place for a comment on everything (within reason). I'm actually comfortable with the second place, but then we have to accept that the rules for the idealised end product don't quite fit, and there we are more reliant on good editing practice of a large number of editors. I subscribe to the view that it would be sensible to distinguish Wikipedia article quality, and therefore rules, though that would just end up being a bureaucratic nightmare to police, so I don't push it. Spenny 08:42, 16 August 2007 (UTC)
Um, why US Census data? Can't we assume that any census is pretty reliable (unless there's direct evidence to the contrary); I'd've thought most census around the world were pretty reliable. SamBC(talk) 13:31, 15 August 2007 (UTC)
Sure, it was just an example from the top of my head. I've altered it to be more appropriately non-centric. Vassyana 13:44, 15 August 2007 (UTC)

I have given this issue a lot of thought for some time and I will direct to something I wrote previously about it. Pulling out my thoughts on question of when these sources are appropriate: sources which should only be be used with extreme care, making sure they are only cited to document that such a source makes such a claim (direct quotes of the source, careful paraphrase of what the author of the source claims, facts and figures in infoboxes).--BirgitteSB 14:28, 15 August 2007 (UTC)

That's a good comment, and I was impressed by it the first time around. There are a lot of parallels with the problematic synthesis section too, which seems to be "over-ruling" on the basic don't make it up principle. Policy is trying to find a set of rules which essentially are about not adding your own twopenneth. I'm not sure the rule exists in any more satisfactory form than that. Spenny 14:48, 15 August 2007 (UTC)
BirgitteSB, I pretty strongly agree with the linked post. I'm particularly impressed by the solution of "first" and "second" class sources, avoiding the field-varied meaning of "primary" and "secondary" sources. I think it well-includes the principles expressed and implied in WP:V, WP:NPOV and WP:RS. I do wonder if editors will generally take some exception towards having notable historical texts lumped in with "unreliable" sources. However, I agree there is little effective distinction in appropriateness as reliable references. Certainly, historical texts are "useful", but their reliability is questionable and best left judged by historians and other experts. You've given me some serious food for thought. Thanks! Vassyana 07:30, 16 August 2007 (UTC)

Census data

The problem I have with census data is that it can be cherry picked in ways that might even appear to be benign. TableManners 03:22, 16 August 2007 (UTC)

I'm not sure what you are referring to. Census data is useful in WP for saying how many people live in an area, and it is the most reliable source for this data. In the U.S. the census bureau is also a geographic resource (TIGER data) which is also useful. So if you are suggesting that census data should be depreciated because of the potential for misuse, I would disagree because any data can be misused. Dhaluza 09:17, 20 August 2007 (UTC)

Assumption of bad faith in original research

This line referring to OR contains an unnecessary assumption of bad faith:

"...'original research' has no reliable source, and is therefore believed to be the original thought of the Wikipedian who added it."

I changed "therefore believed to be" to "could be" because we don't have to assume it's OR, the simple fact that it could be OR is enough. But the change was reverted. WP:AGF should be followed, not excepted in this case. Dhaluza 09:33, 15 August 2007 (UTC)

I recommend getting consensus prior to making this change, as was noted earlier. Dreadstar 09:38, 15 August 2007 (UTC)
And that is what I am doing here-seeking consensus for changing the bad faith assumption you challenged in your revert. The text clean-up that is not intended to change the meaning was restored. Dhaluza 09:49, 15 August 2007 (UTC)
As SlimVirgin noted, your edit inadvertently changed the meaning. Dreadstar 20:33, 15 August 2007 (UTC)
I would say that NOR is such a fundamental feature of wikipedia that we have to be very cautious about it, and that means assuming that something is OR unless there's evidence against it, which is part of the point of V and RS. An unfortunte corrolary of your argument would be that any challenge of uncited material questioning it as OR would be seen as a breach of AGF. SamBC(talk) 10:56, 15 August 2007 (UTC)
Not at all. We don't have to assume it is OR, we can say things that could be OR violate the policy without the unnecessary assumption of bad faith. I don't think the corollary necessarily follows from this, but I would not object to specifically addressing it, i.e. by saying that challenging potential OR is not in bad faith. Dhaluza 11:01, 15 August 2007 (UTC)
But I'm not sure that the original wording assumed bad faith, just failed to assume good faith. It describes assumptions made about things that appear to be original research. The response made by editors should then assume good faith where possible, by saying that something "appears to be" original research. SamBC(talk) 11:03, 15 August 2007 (UTC)
Assuming something is OR has nothing to do with assuming good/bad faith. Assuming good faith means you must assume other editors are doing what they believe will improve Wikipedia. It does not mean you must assume they are correct or that they are following policy.
Assuming good faith: You find a statement that appears to be OR. You contact the editor after removing or tagging it and explain this policy to him assuming that when he added the statement he did not properly understand this policy and mistakenly crossed the line. Or if you know the editor and that he understands the policy, you assume he can locate proper source to back it up and the statement is only in violation of this policy because he has not yet added the reference. In that case, you request a source from him stating that as the referencing currently stands the statement appears to be OR. Assuming bad faith: You find a statement that appears to be OR. You remove or tag the statement assuming the editor is using original research to push his POV. In discussion with the editor your comments make it clear you think the editor not only knew he was violating original research, but also that he was purposely working against the neutrality of the article.
Assumptions of faith have only to do with motivations, not whether actions are correct or not. It completely appropriate to confront someone while assuming they are wrong or have violated policy, as long as you attribute their motivations to misunderstanding rather than malice. I don't think we should make any changes to policy based on WP:AGF--BirgitteSB 14:52, 15 August 2007 (UTC)
The problem seems to be the underlying assumption that having a POV is identical to editing in bad faith. That amounts to saying that a person who seeks truth may have (probably does have) a flawed POV (the person should be more open and tolerant WRT untruths.) That's nonsense, and the difficulties with formulating a NOR policy appear to be as faltering as they are because it is nonsense. It is abundantly clear that those with an (untrue, incorrect, self-serving, whatever) POV can misuse anything to further that flawed POV. That such editors do such things and misuse concepts, methods, etc. does not expose as corrupt those concepts, methods, etc., it only establishes that those things can be used corruptly - just as simple citation or quotation can be similarly misused. If that happens in Wikipedia the apparent proper thing to be done is to undo what is incorrect. Forbidding entire classes of material because the material might be an abuse is a bad approach. The discussion seems to well-illustrate that it's a bad approach to forbid what is called OR, with OR being construed very widely. I assert you cannot create an encyclopedia based entirely on quoting other sources, unless very often those other sources are other encyclopedias. So-called OR can be a powerful tool used to expose untruth. If the so-called OR is on shaky logical grounds or is too much of a reach then removing it from Wikipedia would seem to be exactly the correct action. That's not the same as removing all OR, which is the current policy. --Minasbeede 19:21, 15 August 2007 (UTC)
I disagree completely, you are reading my example backwards. The underlying assumption was that editing in bad faith is when you assume the another editor is purposefully working against the neutrality of an article. That is when you are approaching that editor with bad faith. The two examples I gave are about good/bad faith, refer to faith in challenging the OR statement. The second person character is assuming good faith in the first example and bad faith in the second. I was not referring to the motivations of third person character, which are generally impossible to determine (which is why we must assume good faith).--BirgitteSB 19:47, 15 August 2007 (UTC)
Isn't what I say consistent with that, with "you assume" meaning "you assume that the other editor, who has (and shows) a POV, is editing in bad faith"? I do see that in your interpretation of your words it is the person described by "you assume" who is editing in bad faith. I agree with what you say - any part of what I said that seems contradictory can be lined out: what you say is good. (Thank you for your comments.) --Minasbeede 14:02, 16 August 2007 (UTC)
You said: The problem seems to be the underlying assumption that having a POV is identical to editing in bad faith. I would disagree with that and instead say: The problem seems to be the underlying assumption that working against the Wikipedia's goal of neutrality is identical to editing in bad faith. Having a POV, which I take to mean a bias or a pre-existing opinion on an issue, is natural and unavoidable. But we need to assume our fellow editors are working towards the same goals of creating a neutral encyclopedia. It is when we begin to believe that our fellows editors are trying to undermine the creation of neutral encyclopedia, we are assuming bad faith.--BirgitteSB 15:00, 16 August 2007 (UTC)
Yes, I did say that. I don't think it violates NPOV to assert that 2 + 2 = 4 when some other editor claims that 2 + 2 = 5. If that other editor removes my edit that says 2 + 2 = 4 it's hard to assume that editor is acting in good faith, no matter how sincere his belief in that incorrect sum. But, good faith or not, I think 2 + 2 = 4 should remain. This example is quite simple. What happens in practice is that someone makes a truthful and valid conclusion that is not in any way a reach beyond logic or the facts but because it is a conclusion and not a quote someone else who asserts an untruth removes the edit, preserving the untruth. The real problem, though, is the irrational prohibition on synthesis and "original research." I know the reaction used to make hydrogen on the trailers in Iraq releases a lot of heat because I've ended up with boiling lye when performing the same reaction. I can certainly find a source that gives the ΔH for the reaction but the fact that enough heat is released to boil the solution communicates better than a dry number that would have to be explained. There's no fraud of any sort involved in citing my own results and directly reporting personally observed phenomena communicates well. But that's forbidden. The trailers were used to make hydrogen, the reaction used to generate hydrogen releases heat, always has, always will. It's absolutely neutral to assert that the cooling unit on the trailers is an essential component of a hydrogen generation system even though that contradicts the completely unfounded original story (in the CIA/DIA white paper) that the cooling unit was added to the design after it was discovered that the heat of summer interfered with biological WMD culture. (That latter claim would never pass muster in Wikipedia: it's blatantly unreferenced, always will be.) NPOV can disprove non-neutral POV, and such disproof would seem to be a necessary feature of Wikipedia. Instead it's shunned. That shunning favors the propagation of untruth. --Minasbeede 18:04, 16 August 2007 (UTC)
Dhaluza, it has nothing to do with bad faith. We assume something is original research when it appears that no source exists for it, not simply when no source is supplied; that is, when the material isn't attributable. SlimVirgin (talk)(contribs) 20:21, 15 August 2007 (UTC)
I agree with SlimVirgin; I don't read the current version as an assumption of bad faith, rather an observation on OR. Dreadstar 20:38, 15 August 2007 (UTC)
Perhaps this is already understood, but just to try and clarify a bit; the phrase "The only way to show that your work is not original research is to produce a reliable published source that advances the same claims or makes the same argument as you.", is referring to the content that 'you' (the editor) added without a proper citation or reference. The content added is an argument in the sense that it is "an abstract or summary" or "the subject matter" being presented in relation to the subject of the article. "Argument" doesn't mean that you are trying to present a POV, but presenting an unsourced 'argument' as defined above, into the article. Dreadstar 20:49, 15 August 2007 (UTC)
Anything not attributable (not only arguments) is deemed original research. That is, material need not actually be attributed to a source, but it must be attributable, meaning there must be a source out there somewhere for it. If none can be found, we assume it's a Wikipedian's own opinion. SlimVirgin (talk)(contribs) 21:21, 15 August 2007 (UTC)
Thanks for clarifying, I guess 'claims' covers any non-argument content; and 'claims' doesn't mean your claims, but the claim being made by the content added. Did I get it all covered? Dreadstar 21:27, 15 August 2007 (UTC)
Yes, I think so. :-) SlimVirgin (talk)(contribs) 21:48, 15 August 2007 (UTC)
So, with that in mind, the phrase that started this section "material for which no reliable source can be found and which is therefore believed to be the original thought of the Wikipedian who added it. " is clearly not an assumption of bad faith; it is a statement that the material in question is believed to be OR because it is unsourced content - and not a bad faith attempt to add biased, pov or otherwise false information to Wikipedia. It's a reflection on the content rather than the Wikipedian who added it, OR or not. Dreadstar 05:16, 16 August 2007 (UTC)
Quite right. Someone could put in an "obvious" conclusion in good faith, and perhaps not even consciously or intentionally. Vassyana 05:24, 16 August 2007 (UTC)
I still think the discussion has missed the point. Going back to the original quote:
"...'original research' has no reliable source, and is therefore believed to be the original thought of the Wikipedian who added it."
"believed to be" is an unnecessary assumption. We don't have to assume something is bad, just because we can't prove it's good. We can simply say we don't accept anything unless we know it's good. That is the unnecessary assumption of bad faith I was referring to. We should not assume something that is not attributed is OR, because simply adding a citation would prove the assumption wrong. We simply need to say we don't accept the material that might be challenged without the citation. Dhaluza 09:54, 20 August 2007 (UTC)
The problem is that there is a false equality made between "know it's good" and "from some source." Requiring everything to be sourced in no way makes everything good, it just makes it attributable to someone else. Making it attributable to someone else doesn't remove the need to examine it for reasonableness or correctness, but the apparent underlying assumption made while discussing the NOR policy is that does. It may be that material from other sources is, on the whole, more reliable than material that arises from the use of simple logic or synthesis but both kinds of material can be correct and appropriate for Wikipedia and both kinds can be incorrect and inappropriate. (Material that arises from simple logic or synthesis should be as good as the sources used for the logic or synthesis so it's not at all certain that sourced material is of better quality. Simple logic is valid.) Part of the irksome nature of this discussion is the blithe assumption, for as long as it is needed, that sourced material is always good. Making this assumption allows the automatic discrediting of everything that is lumped under "OR" but in a discussion the automatic should be avoided since relying on "the automatic" is begging the question (which is irksome.) --Minasbeede 22:41, 20 August 2007 (UTC)


No, the assumption is that sourced material is source, or that it's not OR. One of the core principles of wikipedia is that it does not include original research. It could be seen as axiomatic. SamBC(talk) 03:35, 21 August 2007 (UTC)
"Know it's good" is a quote from Dhaluza's comment to which I was responding. You say the policy could be seen as "axiomatic." A possibly more correct term is "dogmatic." That it is a core principle doesn't make it sensible, well-reasoned, useful, correct, or possessing of any other virtue. Those qualities have to arise from the nature of the policy itself, and in the case of NOR the virtues aren't there. Being "core" does not anoint with virtue. NOR is anti-thought, anti-logic. It is true false material can be dressed to appear to be logical or thoughtful. That is not valid justification for throwing out thought and logic, is it? The goal and desire, it would seem, would to be to keep the valid and truthful and to remove the invalid and untruthful. Cited material can also be false material and need to be removed. It's invalid to make the special assumption (when the flaws in NOR are being discussed) that if something can be cited it's true and has merit above that of all "OR." Isn't that begging the question (perhaps "begging the question" approaches being synonymous with "axiomatic," at least in this discussion)? --Minasbeede 09:48, 21 August 2007 (UTC)

Scope of OR

Yes, yes. So make some examples. If i say 'today it's hot' what sources do i must give to you? The C° meaured by my termometer? But it's OR, right? And maybe someone don 't find 35° 'hot' enough to be called so.

Behind my house there is an old church i don't know how wide it is and then i decide to measure it. Oh, my God, there is still an OR, and perhaps also bad faith, to add insult to injury.

We cannot copy other sources, because it' Copyviol, we cannot even resume them because it's OR (no syntesis, right?). Do you have the clue? Where is the good sense? Tell me how one can be authorized by your policies to do something that is not simply copy texts older than 90 years. Feeling as potential criminals every time that we post something that is not a he said so. That's the meaning of OR, IMHO.

I assert you cannot create an encyclopedia based entirely on quoting other sources, unless very often those other sources are other encyclopedias. Agree.--Stefanomencarelli 15:58, 16 August 2007 (UTC)

Protection of wrong version

{{editprotected}} The edit war was successful at removing a long standing sentence:

Can we put this back in and then reprotect? TableManners 03:13, 16 August 2007 (UTC)

Declined. It's still there. Check again. --- RockMFR 03:26, 16 August 2007 (UTC)
It's still in there: Wikipedia:No original research/Primary, secondary, and tertiary sources, second paragraph, last bolded sentence. Dreadstar 03:28, 16 August 2007 (UTC)
Sorry, I must have searched the wrong version. TableManners 03:43, 16 August 2007 (UTC)
I'm not sure what your definition of "long standing" is, so I went back to this version in effect exactly one year ago. Diffs show the issue was in contention then, but at much different point. For example at that time it read "Original research that creates primary sources is not allowed. However, 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 published primary and secondary sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia." —The preceding unsigned comment was added by Dhaluza (talkcontribs) 10:06, August 20, 2007 (UTC).
The issue is not whether the edit is longstanding. The issue is whether the disputed language documents current Wikipedia practices and reflects the current consensus of the broader Wikipedia community. See Help:Modifying and creating policy. Here, it does not. Many times, a policy seems to have consensus within a small group of editors, but when exposed to the broader community, the policy lacks consensus. In this particular case, there is clearly no consensus, given (1) that several editors have expressed contrary opinion, and (2) that in practice, good Wikipedia articles, including almost all FAs, use a variety of primary, secondary, and tertiary sources. COGDEN 17:23, 21 August 2007 (UTC)
It seems like the issue would be easier to reach consensus on if there was some way of clearly describing when it's possible to collate material from primary sources without it becoming synthesis. SamBC(talk) 17:50, 21 August 2007 (UTC)
I think the issue is everyone agrees with current practice while at the same time believe thier preferred version best describes it. No one seems to think that articles should be written differently than current practice. The opinion opposing your own is not (2) that in practice, good Wikipedia articles, including almost all FAs, should not use a variety of primary, secondary, and tertiary sources, but rather (2) that in practice, good Wikipedia articles, including almost all FAs, rely on reliable, published secondary sources. Personally I think that you must believe "Wikipedia articles should rely on reliable, published secondary sources" means something differently than I do.--BirgitteSB 18:12, 21 August 2007 (UTC)
That, in conjunction with the statement, "Although most articles should rely predominantly on secondary sources, there are rare occasions when they may rely on primary sources." Although the first statement only implies that primary sources are disfavored as a rule, the second one makes it explicit. The fact is, in current Wikipedia practice, primary sources are not rare. COGDEN 18:21, 21 August 2007 (UTC)
Why can't we just use the same standard for all sources. Citations to secondary and tertiary sources, too, must (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." COGDEN 18:29, 21 August 2007 (UTC)
I disagree about that this policy advocates rare usage of primary sources. I think it is saying it is rare for an article to rely on primary sources, not to use a primary source at all. I think we do use the same standard for all sources. The standard of no original research. The reason primary sources are mentioned is alert editors to the increased danger with original research in using them.--BirgitteSB 19:50, 21 August 2007 (UTC)
Well, if it's confusing or ambiguous, there is no sense discussing what we think it says, we should be talking about how to make it clearer and unambiguous. Frankly, I think the language from one year ago quoted above was much clearer, and I agree with COGDEN that the same standard should apply to all sources. Dhaluza 09:30, 22 August 2007 (UTC)

Houston, we have a trouble

No original research can be added to a wiki. So the only choices is to grab a researched document and copy inside the wiki but it's illegal (copyright) or you just can publish material that was created by someelse and changed it bypassing the copyright protection (cheating) or using published material that the original author allow to publish in the web.

Even more, there are a lot of "not so important material", that never was researched by a authority or a compentent entity, this kind of material cannot be showed in Wikipedia or even worst can be used by wikipedia only by a biased point of view, for example if we talk about Microsoft, there are little room for any outside to Microsoft to talk about it but Microsoft have almost all the rights and authority to talk about himself.

Anyways empiric talking, the "no original research" rule in wikipedia is not widely applied, common sense overule any no practical law. No original research is used currently to bash someelse, asking for any tiny details about a specific text, it's not anymore (and never was) a way to keep clean wikipedia. --Magallanes 18:58, 16 August 2007 (UTC)

A few comments:
There is a very big middle ground between purely original research ( as in doing your own experiments ) and copying other people's works into the Wikipedia ( which is usually copyright infringement ). What we do allow are brief quotes and paraphrases, which are used in citations and links back to the original sources.
However, yes I do think some interpretations of NOR make it difficult to write about popular culture, highly technical topics, subcultures, underground topics, and topics of local interest. There are certain issues that tend to get ignored by the academic world and by the mainstream media, and I support giving editors some leeway in the use of "primary sources" as long as they don't introduce any new facts of their own. Squidfryerchef 23:05, 16 August 2007 (UTC)
Indeed. I happen across the problem very often. In the things I do (bird phylogenetics mainly) we have a frequent problem that geoscience data (geography, paleontology) is often ignored by molecular studies, because due to the little personal and research overlap between the fields and because of the fundamental incompatibility of GIS and phylogenetics software. Molecular phylogeneticists and paleontological phylogeneticists are more often than not blissfully unawares of each others' research. Molecular clock models are often being used uncalibrated or weakly calibrated, leading to secondary sources that shine on the outside but are rotten at the core. This leads to results in which molecular phylogenies are published which are flat-out contradicted by the material evidence (the fossil record), but that this discrepancy is often not discussed in follow-up publications.
For example, we have recently had 2 publications about parrots, a molecular phylogeny that proposed they evolved in the Turonian (I think) and a review of the fossil evidence that argues that with almost 100% certainty (the absence-of-evidence problem, hence not full certainty) they did not. Neither papers' authors were aware of each others' research (which was conducted at the same time) and therefore do not discuss it; even their references have but little overlap.
A Wikipedia article would have to address this discrepancy, but this cannot be done in a way that would not be considered OR by at least some interpretations - there is at present no secondary source that explicity states that the "molecular" hypothesis is probably wrong. What we are left with are 2 highly valuable sources that utterly contradict each other in a specific point, and one sticks to "good practice" and the other (in this respect, its other findings seem to be good) doesn't. It is obvious which scenario is more likely to be correct, but it is hard to put this on WP in a way that may not be considered OR by at least some interpretations.
As interdisciplinary research cooperation is more of a desire than a reality, these things happen not too rarely. The careful WP editor, bent on sourcing properly, will often find discrepancies between secondary sources that have never before been addressed, because the sheer scope that is technically possible on WP is barely reached by even the most comprehensive reviews. The question how to deal with contradicting secondary sources in the absence of a citeable discussion of this discrepancy (in which case the discrepancy is usually resolved anyway) - especially when one of the conflicting claims is much more likely to be correct than the other - needs to be addressed, as the scope of the problem will grow along with WP. For the time being, WP's own definition of Original research as "...not exclusively based on a summary, review or synthesis of earlier publications..." [emphasis added] might serve as a stopgap. Dysmorodrepanis 07:49, 19 August 2007 (UTC)
It really is as simple as presenting both of the contradicting views without trying to reconcile them. It may be best to mention that in A field they say X and in B field they say Y, in the example you mention. Or you could say fnordtologists* such as (author cited) posit X and gimbletologists* such as (author cited) state Y. It's not very hard at all to avoid OR in this instance. Wikipedia rules do address that some care might need to be taken to avoid undue weight. If one of those sources does not stick to good practice, then peer review or an article countering the "poor" practice should exist. I realize this is not universally true, but determining who uses "good practice" or "reliable science" is a subjective judgment and best left to the professionals. If a wiki editor is a professional, then they should seek a reputable publisher to submit their contrary findings or critique. Vassyana 08:07, 19 August 2007 (UTC) * Nonce.
Well said Vassyana, and I couldn't agree more. Where we don't have widely accepted truth, we simply present the facts, and let the reader decide. I agree that this is not OR, but many editors take an extreme view and argue that since we don't have an authoritative secondary source to resolve the issue, putting primary sources in juxtaposition is OR, and so we must delete the article, which is of course nonsense, but there's plenty of that to go around here on WP. Dhaluza 10:18, 20 August 2007 (UTC)

While you continue to discuss without problems about the santity of NOR and its stricth interepretations, in Wiki there are stuff like this: Domination of the skies, superior seamanship, and prudent, timely maneuvers helped to nullify the overwhelming odds. In the highest tradition of naval service, the finest qualities of the American sailor became commonplace during the heroic fight. Devotion to duty, daring courage, uncommon bravery, and an indomitable spirit were part and parcel of this victory.

So don't cause me LOL in so indecent manner. This piece is what could be called shameless agiography and pubblicity for US Navy, and there are thousands of this stuff. While i try to reduce every optimistic overclaiming made by italian aviators with the cross controls with several sources (it's unuseful to post that x has downed 12 enemy aircraft when are available sources that considering both sides, dismissed all this claimings without any problem), there are thousands of pages like this, written directly by J. Wayne. Obviousely you are not worried about, your goal is forbid brain activity. I am, instead. Babbling about NOR when gratuitus agiographies are so well spread is a thing that shows how these discussions and policies don't guarantee nothing. Sorry,but it so.--Stefanomencarelli 13:59, 20 August 2007 (UTC)

This whole issue over the "articles should rely on secondary sources" does raise a good point though...

I just noticed that the article has been protected due to edit warring, so I thought I'd take a look as to why. I think there's some merit to the "compromised" rewording of the sections in question.

For most articles common sense reasonably dictates that secondary sources are going to be the ideal sources to base an article off of. That gets thrown out the window, however, when we start talking about works of fiction. In that case you are usually writing an article about the primary source itself, in which case the bulk of your information will come from the primary source. A good chunk of it should also come from secondary sources, true, (otherwise it probably fails WP:N) but in that case I would not say that the article relies on secondary sources so much as it uses secondary sources.

Now obviously this situation only applies to a certain subset of articles, far from a majority, but also far from being "rare" as the current revision of the policy states. (Anybody care to try and count the number of articles on films, just for a warmup?) The basic idea behind the sentences in question is sound, but it needs to be slightly reworded to reflect the fact that some articles by nature rely on primary sources, not secondary sources. -- Y|yukichigai (ramble argue check) 12:20, 18 August 2007 (UTC)

Hi Yukichigai, most articles about films rely on reviews, not on primary-source material. Similarly, most articles about novels rely on sources who've written about the novel. It is, indeed, unusual to find a legitimate article that relies entirely on primary sources. If you can find one, please let us know. SlimVirgin (talk)(contribs) 12:23, 18 August 2007 (UTC)
Umm, we may have gotten some wires of communication crossed. I'm not suggesting any articles ONLY need primary sources (it's pretty hard to satisfy WP:N with just primary sources), I'm merely pointing out that the primary basis for information contained within articles about works of fiction is primary sources. Works of fiction rely on primary sources more than they do secondary sources. -- Y|yukichigai (ramble argue check) 21:57, 18 August 2007 (UTC)
What about an article about a railroad, where there isn't going to be a lot of controversy or incorrect information, and, some might argue, it's inherently notable. Most of the information from the article, about where the stations are and what kinds of trains they have, would come from the railroad's web site, a primary source. There might be a few citations to newspapers that ran opinion pieces about the railroad, but this is one kind of article where the primary source has the best information. Squidfryerchef 14:25, 18 August 2007 (UTC)
P.S. We need to clarify whether we're talking about articles with only primary sources versus articles that cite mostly primary sources. I agree that except in articles where inherent notability implies, we need at minimum one secondary source to establish notability. Squidfryerchef 14:31, 18 August 2007 (UTC)
Yeah, that I don't disagree with. Virtally all articles (all that I can think off off the top of my head) need secondary sources at the very least for WP:N purposes. The point I'm trying to make is that the wording seems to indicate that all articles, no matter what, should be based on secondary sources over primary sources. For works of fiction this is quite the opposite; information about the work of fiction is far less reliable when based off of secondary sources. -- Y|yukichigai (ramble argue check) 21:57, 18 August 2007 (UTC)
While that's true for basic factual information, it's imperative (as long as there's a policy of "no original research") that no interpretation be put on the text/film/whatever that aren't from a secondary source. Well, I suppose an author or similar making a statement about how things are supposed to be interpreted is technically a primary source, but that would be an exception in this case (provided it's stated as being the author's own interpretation). SamBC(talk) 22:22, 18 August 2007 (UTC)
That's sort of tangential though; a separate section addresses that problem completely. Like you said, information about the interpretation of the work of fiction can come from the primary source still. It's not really relevant to my point, regardless. -- Y|yukichigai (ramble argue check) 22:38, 18 August 2007 (UTC)
The point is that people shouldn't be writing a Film Studies/Literature/Theatre Studies/whatever essay. They should not say, "character Y is cruel, as shown in chapter umpteen when they fnordle character X's brip" (excuse nonsense words) or similar. Plot summaries are inherently original research unless they're based on other published plot summaries. There's a line to tread between OR and copyvio, but it's not so thin a line. SamBC(talk) 00:47, 19 August 2007 (UTC)
I disagree. It's quite easy to write a plot summary which simply lists the events of the film without any sort of interpretation or other type of OR. It may not be the most inspiring and articulate plot summary, but numerous plot summaries rely on no sources outside of the primary source for the information they contain. -- Y|yukichigai (ramble argue check) 04:18, 19 August 2007 (UTC)
FWIW, I have found myself messing with some cryptozoology articles recently, and here we have a case where it is the secondary sources that most often violate WP:V. As CZ is basically dealing with animals too cryptic, or apocryphal or probably fictional for classical zoology to bother with, secondary sources are prone to outlandish claims and go at great lengths to POV-push their pet interpretation. The only really reliable sources are indeed the original observations, and these are often misrepresented, misquoted etc in the secondary literature. This is the case why most cryptozoology websites are, from the point of a professional zoologist, so abysmal: because they rely on the twisted version of a primary source, twisted even further to serve some particular end (such as that there are surviving pterosaurs). Dysmorodrepanis 08:04, 19 August 2007 (UTC)
There are plenty of sources that "debunk" cryptozoology, so this is not as difficult of a situation as you would make it appear. If you have trouble finding such sources, you could for example contact the Center for Inquiry, who would be quite happy to point you towards opposing sources. Be aware that secular humanist debunking material is no less biased. However, that's not really a problem, since we don't try to avoid the bias of the available sources here, but rather represent all notable POVs appropriately. Vassyana 15:25, 19 August 2007 (UTC)
Again, at WP:N an attempt to make secondary sources an absolute requirement lead to a similar edit war and edit-protection, which was finally worked out by making it strongly favored, but not absolutely necessary. This dispute has many of the same underpinnings. One notable example of using primary sources was the thousands of articles created for geographic places in the U,S. based on U.S. Census Bureau data. I have also noted below that most articles on aircraft and airports rely heavily on primary source data because it is the most reliable source in these areas. Dhaluza 09:42, 20 August 2007 (UTC)
At WP:N, the consensus was not that secondary sources should be the favored sources. The consensus was that notability is best evidenced by the presence of secondary sources, which is entirely true and in line with current Wikipedia practice and the consensus of the community. If somebody has written about something, it's notable. However, for purposes of WP:NOR, once the notability criterion is passed, the best Wikipedia articles use a combination of primary, secondary, and tertiary sources. Current practice is to use all three, and this policy statement must reflect current practice, not lead it. COGDEN 17:33, 21 August 2007 (UTC)
Yes, I should have said that in the current compromise language of WP:N, secondary sources are strongly favored for judging notability. And you are correct that many editors objected to the attempt to modify WP:N to make secondary sourcing a prescriptive requirement there, because it did not reflect community consensus or practice. I suspect the same thing is happening here. Dhaluza 09:24, 22 August 2007 (UTC)

Time chain

The NOR policy should reflect the dynamics of the subject matter at hand. An article about a breeding edge issue such will require different sources than an article about the French Revolution.Arebenti 13:19, 18 August 2007 (UTC)

No original research is a core policy. It doesn't matter what the subject is, wikipedia only includes information and analysis that are directly based on reliable third-party (generally secondary) sources, in order to ensure that there is no original research or invalid synthesis. SamBC(talk) 13:29, 18 August 2007 (UTC)
Much content in many articles on TV shows are based on primary source material. Much content in many "wrestling" articles is based on first person claims (as is natural, cause lots is made up fiction). Much song information is from those with a direct financial interest. Many articles on small companies rely on those companies web sites. Should we run wild deleting all this? No. It should be sourced better, but we need to get there in increments. Absolutist terminology is generally not helpful for this. WAS 4.250 16:44, 18 August 2007 (UTC)
That's a point, I should have said that it's wikipedia policy to only include information blah blah blah... the point is, we should resist the addition of such material beyond a reasonable start, and we should resist attempts to change policy to say that all that stuff is okay. It's currently tolerated because such policies describe a goal, and that is still the goal AIUI. Some of those articles probably should be deleted, and some kept. If anyone wants to start going through it all and tagging stuff, then that's cool. For me, the effort isn't worth the return. SamBC(talk) 17:15, 18 August 2007 (UTC)
To further the point by WAS 4.250, aircraft articles are based on primary source manufacturer data, which is the most reliable source. Secondary source "tests" are not reliable because of the cost of doing a proper flight test, and the engineering data needed. For airports, we also rely on primary source data from the aviation authorities. Secondary source data for basic airport characteristics usually just copies the primary source data. So primary sourcing is critical to our coverage of aviation as well. Dhaluza 09:29, 20 August 2007 (UTC)
Unless Jimbo speaks, Wikipedia policies are not prescriptive. They only reflect current consensus practice, or at least a consensus aspiration. For a good explanation of this fact, see Help:Modifying and creating policy, as well as Wikipedia:Policies and guidelines. A "policy" that's just a goal by some editors is not a policy, but a proposal. If we can convince the Wikipedia community to make that "goal" a predominant practice or aspiration, then it can become a policy. COGDEN 17:38, 21 August 2007 (UTC)

Is listing a subset from a published government document original research?

Hi All, Requesting feedback on this: The National Institutes of Health posts a yearly Excel spreadsheet of medical grant awards. About 56,000 entries, $20 billion. Does this fall under WP:NOR: "Figure 2 lists entries from the NIH's 2006 Grant Database containing the words Giardia." On the one hand, it involves some kind of research activity by the author. On the other hand anyone can verify the statement by downloading the Excel file and repeating the search. Thx, Gastro guy 22:49, 18 August 2007 (UTC)

In general, any process, with the possible exception of complex mathematical calculations, that can be perfectly reproduced is not original research . The only potential issue is with the interpretation of data ... one must be careful to state only that which is unquestionably supported by the data. For instance, interpreting "country with the highest GDP per capita" as "wealthiest country" is original research. — Black Falcon (Talk) 22:58, 18 August 2007 (UTC)
Thanks! Gastro guy 06:25, 19 August 2007 (UTC)
Your NIH example is not original research. However, cherry picking information might produce NPOV issues, which is a different beast entirely. —Kanodin 07:26, 19 August 2007 (UTC)

What about information where sources cannot be found at all?

I am interested in putting up entries in WP on the Cairo metro system. No revolutionary theories, just the names of the stations, their general location in Cairo, prominent landmarks nearby, etc. This kind of content is common for other cities such as London, Paris, Moscow, etc. Most metro station pages I have read cite no sources anyway. Since no sources are cited, what is this referred to, if not original research? 41.196.184.19 10:40, 19 August 2007 (UTC)

For starters, there should be an official map from the operator which you can cite. You can translate what the cartographer depicted into plain English without crossing into OR, as long as you don't interpret it or draw any non-obvious conclusions. So you could capture the names of the stations, and their order on each line. This would probably only be enough info for a summary article, and if you tried to create stubs for each station, some editors would object to using a single primary source and probably nominate for deletion. So start off with the content merged into one article, and split off sub-articles as you find more sources. Dhaluza 11:19, 19 August 2007 (UTC)
For most major cities, there should be abundantly available travel guides that describe major points about the city, including the public transportation system. There are plenty of sources available for articles like this, especially for "world centers" like Cairo and London. Just be cautious using such sources, since Wikipedia is not a travel guide. Vassyana 15:21, 19 August 2007 (UTC)

Okay, another question posed

I just spent somewhere around an hour reading all of the above (and several other pages), and I'm more confused now than I was before I started reading everything. So, without any further ado, here's a particular case in point, hopefully which this group can help solve.

I've been primarily interested in several articles where I was personally invloved in the subject matter, Axe Murder Incident and Joint Security Area (JSA), though I've also made numerous edits to a slew of other articles as well. So, I am an eyewitness to many of the events/articles that I have working on. This makes me a primary source, correct? As a primary source on these articles, "part" of what I edit may fall under original research, which I understand. I try to comply with this rule as much as possible by finding other references that I can cite which support my edits.

However there are several things which I either have problems with, or completely disagree with. First, let me start with a quote from this article that I vehemently disagree with, followed by (what I consider) a very good example. The quote is "The threshold for inclusion in Wikipedia is verifiability, not truth." So, in essence any lies (falsehoods) which have been published or otherwise made publicly available are perfectly legitimate Wikipedia inclusions since they can be verified from the published source, but somebody who was an eyewitness who hasn't published a "document" of some sort can't rebut the falsehood, therefore perpetuating the falsehood into eternity. Now for the example.

Regarding the Axe Murder Incident, the US Army released a statement the following day that one of the reasons for a slow reponse was that the Quick Reaction Force (QRF) was over a mile away from the DMZ at the time of the incident. This is blatantly false to anybody that has ever been there simply due to the fact that at NO TIME is the QRF ever 1 mile away from the DMZ. The farthest the QRF ever gets from the DMZ is approximately 100 meters, which is where the Advance Camp is located, anf this is usually just for a very short period of time. On this particular day and time however, the QRF was actually sitting at the entrance to the JSA which is right smack dab in the middle of the DMZ. Here is part of an email I received from the guy who was actually working the checkpoint at the entrance to the JSA that day.

"Thirdly, to answer the question about who recovered the USA soldiers bodies? It was 3rd Platoon. The information below, "In actuallity, the QRF was sitting in their trucks at CP#2, at the entrance to the JSA, at the beginning of the fight with the 3rd Plt. Ldr awaiting authorization from Capt. Bonifas (who was already dead) for authorization to go in, right smack dab in the middle of the DMZ," is a very accurate statement. Moreover, it took the actual JSA CDR, on this day, to order the 3rd Platoon Leader to go in to recover the bodies.

Lastly, Mike, as you may recall, the above quoted statement truly supports what I have been telling you, M.S. (from the History Channel) and others all along. And, that is, that I was assigned to UNC CP#2 at the time of the August 18, 1976, Axe-Murder Incident. When I initially saw the QRF platoon approaching my guardpost, I was somewhat happy that help was coming. Contrary to what was stated by the 3rd Platoon Leader during the December 28, 2004, airing of the documentary, it was he in the lead vehicle (and not an E-7 Platoon SGT) that I vigorously tried to wave into the JSA -- but my request fell on death ears. Instead, the foregoing platoon leader, subsequently, got out of his vehicle, went into my checkpoint, and then called the JSA CDR, whom, upon his arrival to UNC Check Point #2, ordered him to go in. You can imagine how much time had elapsed! I was saddened by his actions.

I specifically remember this aspect of the incident because due to my assignment at CP2, the LTC ordered me to assure that the soldiers congregating around to view CPT Bonifas' body did not, in fact, see him. I did, and I have been living with the residuals of it ever since. That's why I am telling this story, as it actually occurred."

BTW, my name is Bill, and the reference to Mike above is to another buddy of mine who was also included on the email. I also changed a person's name to just their initials in case they don't want their name known. So, since the US Army issued a statement almost immediately that was wrong in so many ways, it is allowed to stand since the other eyewitnesses and/or myself haven't published anything to contradict it, and the lie is allowed to florish forever? And seriously, if I was to issue to press relase staing the actual events that day, would any news organization really run it or care? No, because it would contradict the "official" version as published by the US Army.

I am very careful about my edits, trying to remain as nuetral as possible and report the actual events as accurately as possible, however this came to a head when somebody wanted to change the word "Murder" to "Killing", claiming that the term murder is POV and trying to diminish the actual events of that day Talk:Axe Murder Incident. I do have some references to a web page I have where I placed some of my recollections from Korea, and also from another buddy who was there with me (Mike, from the quoted part above). Both of our references are listed as "eyewitness account...". Also regarding the term "Murder" for the above immediate problem, a Google search for +"Axe Murder" +korea returns 1100 hits, while the same search that replaces "killing" for "murder" only returns 209 hits. Both sets of hits return some entries for other events, but more inaccurate results occured with the word "killing". So, the term "murder" is also more associated with the events of this day than the term "killing" is as well, adding prevalent thought and opinion to the term.

In conclusion, the history of the world would be in sad, sad shape if it was solely relegated to whichever side expended the most effort in promoting their version of events. I do see where this leads to edit wars, etc., but in many cases, I strongly feel that they need to be decided on a case-by-case basis, where a "group" of educated, non-partisan parties, preferably with both an interest and some background on the subject matter, can decide what should be "allowed" or not. Otherwise, as others have said above (and paraphrased by me), all Wikipedia is, is just a regurgitation of what's already available, so where's the added "benefit" of even using Wikipedia, if all you're are going to see is what's already available everywhere else? Has anybody ever tried (and had success) writing Encylopedia Britannica telling them a cited "fact" is incorrect, offering proof, and actually gotten them to change it? I haven't, and I seriously doubt if anybody else has had any luck in that regard either. That's what makes Wikipedia unique and adds value to many of the articles.

Thanks for any and all opinions on this, regardless of whether there's agreement or disagreement from my POV. wbfergus 15:57, 22 August 2007 (UTC)


I think your biggest misunderstanding is the role of an encyclopedia. An encyclopedia is not supposed to be the place where ground is broken, discoveries are made, or truth uncovered. An encyclopedia is the place all the currently acknowledged information on a subject is laid out comprehensively. When someone reads an encyclopedia, they should come away with knowledge of "things as they are known to be". Encyclopedias are a starting off point to research, not the final answer. You start research by finding out what is already said on the subject. The threshold for inclusion in Wikipedia is verifiability, not truth does not mean choosing falsehood over truth, but disavowing both as irrelevant. Encyclopedias are a summation of existing knowledge, what has already been printed and then acknowledged and/or disputed. Your difficulty with WP policies is because you wish Wikipedia to be something it is not. It is designed to be "just a regurgitation of what's already available". The benefit is that it is a free-content regurgitation of what's already available. That is the revolutionary part.--BirgitteSB 16:29, 22 August 2007 (UTC)
"Truth" is highly variable, particularly when getting into differences between eyewitness accounts. Any editor on Wikipedia could claim to have been an eyewitness and try to present the "real" version of events. In the absence of information from a reliable source, there is no way to verify such a claim. No matter how "true" someone insists it is, with no way to verify such a claim we cannot in good conscience present it in an encyclopedic article. This is especially true of claims that contradict the conventional wisdom regarding a topic. If someone has vital information that contradicts the current consensus about a topic, I would strongly encourage that person to seek a reputable publisher for their claims. Wikipedia collects and paraphrases existing published claims. That supports the central purpose of building the encyclopedia. It is not the place for novel research and claims that cannot be verified in published literature. Vassyana 16:41, 22 August 2007 (UTC)
The benefit is that it is a free-content regurgitation of what's already available. That is the revolutionary part. Unfortunately, it's not revolutionary. A simple Google search will result in numerous hits on an item, almost all of which are free. It is then up the reader to compile the various results into a cohesive "train of thought". What makes Wikipedia "different" or unique is that various readers can do this themselves (oftentimes with input from a multitude of other readers who usually have an interest in the subject), to create a new version of what an encyclopedia may already have, but with far better references (due mainly to constant bickering back and forth), and the ability to also allow for eyewitness accounts in applicable context.
I can see the points being made above regarding something like religion or some new science-related technique and it's results, etc., but as a general blanket policy, it will in the long run, severely restrict any benefit of Wikipedia. Why use it if all it is going to do is repeat what's already considered "common knowledge" by a few monied sources? Most people that "publish" do so for some sort of gain, usually either monetary or because it's required for their doctoral thesis. They rarely will seek to contradict the established "concensus of opinion" on a subject unless they are seeking to make a name for themselves, delve soley into "original research" themselves (regardless of whether can can actually prove it), or merely as some do, to create controversy. Once they have done so, then all of a sudden it's established as fact, because some wacko actually went to the effort to get something published? Or, in the example I cited above, the US Army's version will always be the version repeated through the ages (to protect a fellow officer), rather than the accounts of eyewitnesses who contradict the publish statement, since in this case, 1.) there simply isn't enough information to warrant more than a few pages at most, far from what would be required for a "book", and 2.) as I stated in my post above, even if the other eywitnesses and myself all banded together and issued a "press release", would any of the news organizations actually publish it? No. So therefore, regardless of the truth, the lie prevails, simply due to the power or monied source of publishing might? Why use Wikipedia then? There is plenty of free information out there already, much of which is BS, but at least the user can also have access to additional sources of information which Wikipedia's "core group" won't allow because it's not published as a hard-copy book or magazine/newspaper article? Much (most?) of what is in Wikipedia is already freely available through the web with a simple Google search.
Regarding "Any editor on Wikipedia could claim to have been an eyewitness and try to present the "real" version of events.", I backed up my claim as an eyewitness (just to avoid this possibility), with the UNC Certificate of Appreciation I received from Gen. Vessey, the UNC CINC at the time. So, without picking a fight or arguement, the above statement was written "off-the-cuff" without looking at any of the linked materials to determine the true and full scope of the question, which in itself is one of the problems with many Wikpedia articles. Many editors read one or two "articles" on a subject and then deem themselves expert enough to judge the acurracy of all subsequent edits to that Wiki article, even if they read the "fringe theory" versions of that subject. Or, as is the case on here lately, when "editors" come here to push the POV of their country, even in the face of overwhelming evidence to the contrary. I at least have presented a credential verifying my status as an eyewitness.
Anyway, my main point of contention is that since I was an eyewitness, my revert of the word "killing" back to "murder" is being questioned as OR, whereas if anybody else had done the same thing, it wouldn't have, since the overwhelming preponderance of opinion and stated publications all call it "murder" (more than 5:1 use the term murder). I guess I should have made that point a bit more clear. Thanks again for the feedback though. wbfergus 18:14, 22 August 2007 (UTC)
"Free-content" means free as in freedom, not free as in beer. There are plenty of google hits that are free as in beer while being tied up by restrictive copyright. That content cannot be derived from with obtaining a license from the copyright holder, likely for a fee and with restrictions. Wikipedia is available under a free-content license (GFDL), which allows anyone to make derivative of it's content, for any purpose, so long as they they license it under the GFDL as well. For example this means that the plentiful (in english and some other fortunate languages) information you would find by google is available for you to read and learn from, but you are prohibited from translating it into Swahili. It may seem that we are regurgitating what is already available to you. But it is not widely available to everyone, and in the end Wikipedia is making information available in places it was not accessible before. In twenty years you will understand exactly how revolutionary this is.--BirgitteSB 19:06, 22 August 2007 (UTC)
Unfortunately, there is a lot on here that more or less is copied verbatim from other sources (even copyrighted ones), then a few minor changes are made, fixing a typo, rearranging sentence structures, maybe deleteing a sentence or two, etc. Does that automatically make it legal? By removing some of the context, if that context applied to the entire paragraph and laid the groundwork for an assumption in the last sentence (or another paragraph the editor didn't agree with), doesn't that constitute OR? If someone edits something thats been published to change the meaning, so they can skirt potential copyright issues, and winds up (even inadvertently) changing a meaning of something, doesn't that in itself constitue OR? I'd maintain that almost everything would be OR, even it was copied verbatim and then a couple small changes were made. All of a sudden it's no longer a secondary source, as it is now different and "interpreted" by somebody else. That suddenly makes it "original". If it wasn't "original", and was based even remotely on copyrighted work, then it's a copyright violation, since they (almost) all state something like "cannot be copied in whole or in part".
Sure would help make all of this unique and revolutionary though if there was free beer. wbfergus 15:45, 23 August 2007 (UTC)
You misunderstand some of the concepts behind original research and copyright. Ideas cannot be copyrighted only the creative expression of an idea is copyrighted. Original research involves a novel idea or synthesis of ideas. A new expression of a documented idea is not original research, nor does it infringe on the copyrights of the documented expression. Editors rewriting the ideas expressed in existing sources in their own words is exactly what is supposed to be happening here.--BirgitteSB 17:29, 23 August 2007 (UTC)
Okay, then my eyewitness accounts (the very few I have added, about two total), meet these criteria of not being OR, and my revert of the word "killing" back to "murder" is not OR either. The few occasions where my personal eyewitness account came into play was in clarifying a few vague points made by secondary sources. This is the point I was trying to get clarification on, though I guess I muddied the waters with my example above, which hasn't been used in any article. I've only used the example on talk pages, as the Lt. that the Army report covers up, has already had to live with his bad decision these last 31 years. I know that already would be a great burden to bear for most people, without making it more blatantly obvious who the person was.
Perhaps there could be an award of free beer for those who get an article to "FA" status? wbfergus 18:00, 23 August 2007 (UTC)
Has your account been published in a reliable source? From what I understood above using your own account would be original research. Whether the word "murder" is original research depends on what reliable sources have published about the incident. If there are reliable sources that describe it as murder you can cite them and it would not be original research. If reliable sources disagree, explain all viewpoints and who holds them.--BirgitteSB 18:35, 23 August 2007 (UTC)
The example I've used above has only been used on talk pages, not in any article. The article itself, "Axe Murder Incident" however is titled after not only the facts, but far more numerous published sources than any variation terming it "killing". Several of the references even call it "Murder", and I also have in my possesion probably one of the first booklets ever done (within three days of the incident), title "Axe-wielding Murder at Panmunjom", which several of the pictures I've added were taken from (the booklet itself says everything may be used, no copyright). wbfergus 19:18, 23 August 2007 (UTC)
The content of that article can only be settled on the talk page, not here. But I hope you have a better understanding of this policy now.--BirgitteSB 19:31, 23 August 2007 (UTC)

Is the article Jennings Rutter Battle original research?

The article describes a "battle" between Ken Jennings and Brad Rutter, two US game-show contestants. They have appeared on various game shows, sometimes head-to-head and sometimes not. The article attempts to compare their achievements and declare a "winner" of each "round". I believe that editors have invented the structure of this "battle" themselves, although each specific fact in the article is referenced to a source. Is this OR? --Cinematical 16:27, 22 August 2007 (UTC)

It could quite easily be synthesis, a form of OR, but I don't have time right now to check the sources. Notability might be an easier thing to challenge, though. SamBC(talk) 16:55, 22 August 2007 (UTC)

COGDEN's edit (Two-part test as applied to secondary sources)

Cogden, could you make sure you have clear consensus for any changes before making them, please? This is a policy page and it has to be stable.

The change you made didn't make sense to me. You say all sources should be handled the same way. You then say "the source should not be used to support or imply any analytic, synthetic, interpretive, explanatory, or evaluative claims, unless those claims are supported by some other reliable source."

That means that a secondary source cannot be used to make any analytic (etc) claim, unless supported by some other source. So you're now saying that multiple sources must be available for each edit, which is a major change.

It's also not true that all sources should be handled the same way. Primary and secondary sources are, as a matter of fact, handled very differently. The policy has to reflect that. SlimVirgin (talk)(contribs) 18:59, 22 August 2007 (UTC)

Looking it over, I agree with what you're saying. It should have been phrased differently:
"the source should not be used to support or imply any analytic, synthetic, interpretive, explanatory, or evaluative claims outside the source itself, unless those claims are supported by some other reliable source."
With this addition, I think this represents current practice and consensus with respect to primary, secondary, and tertiary sources. For example, if I have a secondary source that makes one conclusion (e.g., "Thomas Jefferson slept with slaves"), I can't use that source to make a second, somewhat different analytical conclusion (e.g., "Thomas Jefferson wasn't a racist"). This is actually a slightly stronger formulation of the ban on original research, but I think it represents the practice of the Wikipedia community. COGDEN 19:29, 22 August 2007 (UTC)
But that just says "don't go outside the source," and don't violation SYNT, which we say already. SlimVirgin (talk)(contribs) 21:12, 22 August 2007 (UTC)
We already say it, but we could say it more concisely as a two-part test. Also, by saying it's a requirement for primary sources, but conspicuously omitting it as a requirement for secondary sources, we kind of imply that the requirement is not as important for secondary sources. COGDEN 17:24, 23 August 2007 (UTC)
As to the first part of the test:
"it should be clear to any reasonable, educated person without specialist knowledge that the source is cited accurately,"
I think this is quite clearly Wikipedia policy, and it probably would have the support of Jimbo, who said that one reason for the original research policy is the fact that lay editors on Wikipedia "aren't really equipped" to verify new scientific or historical theories. [6][7] Therefore, any citation to a source (primary, secondary, or tertiary) should not require specialized knowledge in order to verify that the citation is found in a reliable source. It can be a citation to a highly-technical source such as a journal article or a physics textbook, but a lay person must be able to look at the source and determine that what is cited matches what is stated in the article. It doesn't really matter in this sense whether the source is primary, secondary, or tertiary. COGDEN 20:13, 22 August 2007 (UTC)

Based on use

I'm not certain that changing the wording to say "based on how they are used" for the Primary, Secondary and Tertiary source use is a purely stylistic change. I think it needs to be discussed and a consensus reached before making this change. Dreadstar 20:03, 22 August 2007 (UTC)

I removed the challenged sentence regarding "based on use". Do you actually disagree with this, or are you just protecting the interests of those whom you think might disagree. Are you saying that primary sources are absolutely primary sources, regardless of how they are used? Based on discussions here and elsewhere, I thought it was non-controversial that the same reference can be either a primary source or secondary source, depending on what is being cited. COGDEN 20:17, 22 August 2007 (UTC)
Thank you for removing it. I think I've made myself clear in the above discussions that I do not believe use of a source necessarily affects its status as primary, secondary or tertiary. If I missed something that shows this to be the incorrect view, please let me know. Thanks! Dreadstar 20:22, 22 August 2007 (UTC)
I agree that use doesn't affect primary/secondary/tertiary status in most fields; however, I thought that "based on use" had general support. A citation is not there for the source, it is there for the claim. Jacob Haller 20:27, 22 August 2007 (UTC)
(ec) Materials that are generally considered primary sources (such as historical works) are always primary sources. However, some secondary works may be used as primary sources. For example, the work of a historian may be a perfectly reliable secondary source for the historical events that the book addresses. However, the foreword from the same book, describing the scholar's approach and general opinion would be generally considered a primary source in an article about the historian herself. Vassyana 20:31, 22 August 2007 (UTC)
(ec)The clearest way the use of the source matters is this example. An notable historian writes a book about WWII. The book can be used as secondary source in WWII article, and as a primary source in the author biography. --BirgitteSB 20:34, 22 August 2007 (UTC)
So, there is consensus for adding that single sentence about use? I may have missed that part in all the action. If there is support, it's a pretty complex issue that would require more than a single sentence of explanation, similar to what Vassyana and Birgitte just had to describe. Is the concept of 'use' explained in that detail anywhere? Dreadstar 20:37, 22 August 2007 (UTC)
If you disagree, that's evidence there is not consensus, at least a Wikipedia consensus. But since the distinction between primary and secondary sources is used outside Wikipedia, we should back up whatever we put in the policy statement with reliable sources. I don't happen to know any, though. Most sources discussing this issue don't make that fine a distinction. COGDEN 21:05, 22 August 2007 (UTC)
I think we should leave out "based on use." Most people know what primary and secondary sources are, if only roughly. Those who don't know can look it up. We should keep it simple here. SlimVirgin (talk)(contribs) 21:14, 22 August 2007 (UTC)
(ec)This section of the policy has always tended towards the often (rather than always) accurate interpretation. Although I think the 'use' issue has agreement among the Wikipedians that have examined the issue, it is only explained on various talk pages. As yet, no one has written an explanation of the issue that is simple enough for newbies without containing inaccuracies through oversimplification. The current explanation is correct most of the time. The benefit of having it available to point editors to in without an argument has always been decided to be worth it's inaccuracy on other occasions. Several times the talkpages have been over this and come to that conclusion. This is one reason people will not speak up here, but will revert changes. They likely think there is nothing new to be said about the issue, and the decision will be the same as always. However I think it is only a matter of time before the person who is capable of writing this issue clearly enough for a newbie without any inaccuracies comes along. Or instead a person with some unforeseen out-of-the box solution to current stalemate.--BirgitteSB 21:25, 22 August 2007 (UTC)
If we put it in, we'd need to explain it; and the detail needed to describe those rare situations where it would be useful seems to be a bit much for the small impact it would have, so I'm all for keeping it simple and leaving it out for now. Thanks for the excellent explanations everyone! Dreadstar 00:47, 23 August 2007 (UTC)

I did a brief review of who defines primary and secondary sources based on use, and I found some definitions that I think are particularly germane to this policy page, because their "use-based" distinction has a stronger connection to the idea of "original research". Here are a few definitions of primary sources:

  • "Manuscripts, records, or documents providing original research or documentation." [8]
  • "Primary sources contain original information and are usually the place where the original information first appears. Examples of primary sources include interviews, diaries, letters, speeches, results of experiments or original research, literary works, autobiographies, original theories, and other materials. Compare to secondary source."[9]
  • "An original work such as a book, manuscript, or document produced by an author. A book can also be a secondary source. A scientific or scholarly journal article is usually a primary source."[10]
  • "The main source used to defend a research question. For example, critical essays, documented studies, scholarly or technical journals, or interviews with experts."[11]
  • "Information which has not been interpreted by another person, ie, original articulation of an idea or concept."[12]

Here are a few definitions of secondary sources:

  • "second-hand report or review of original research that is written by someone other than the original researcher"[13]
  • "A document which interprets or analyzes a primary source. It is something written or reported about someone else's work."[14]
  • "A source that contains information that other people have gathered and interpreted, extended, analyzed, or evaluated, such as newspaper articles, a documentary on television, a website, a science text, and an encyclopedia entry."[15]
  • "materials or sources that contain information that has been cited, translated, or based upon another primary or original source."[16]
  • "Materials that are not original manuscripts, contemporary records, or documents associated with an event, but which critique, comment on, or build upon primary sources."[17]

Thus, whether something is a primary or secondary source really depends upon what aspect of the work is being cited. If I'm citing some original research (such as a new scientific theory or a new interpretation of history), it's a primary source, even if the original research was derived from evaluating and analyzing other primary sources (such as raw scientific data or diary entries). In other words, all original research begins its life as a primary source. If someone else adds an additional new idea, they too can be cited as a primary source as the originator of that new idea. Secondary sources don't add any new theories or interpretations, they just comment on primary sources.

According to this framework, sources can be both primary and secondary sources. A biography is a secondary source to the extent it describes, critiques, or analyses primary source material in a non-novel way, but a primary source to the extent it introduces novel theories, analyses, or interpretations about the subject's life. This is not the only way out there to distinguish primary and secondary sources, but given the focus here on the meaning and use of "original research", does adopting this type of framework make sense? COGDEN 21:53, 22 August 2007 (UTC)

There's been some discussion and attempt at clarification of this at the essay/talk page Wikipedia:Classification of sources. Might be worth a look, and if there's consensus on what it says (or is made to say) it can be linked and/or used as basis of new wording/section? SamBC(talk) 22:40, 22 August 2007 (UTC)
Wikipedia:Classification of sources is just starting out as an essay, and it considers the three types of sources with respect to WP:NOR, WP:Notability, and WP:NPOV. We'll have more time to work that out, but here, since this is a policy and it has to reflect Wikipedia-wide practice and consensus, we need to make a change quickly. We can't let it remain much longer in its non-consensus prescriptive state. I'm just thinking that focusing more specifically on how the primary/secondary distinction relates to original research, we can avoid the larger more controversial issues. COGDEN 17:29, 23 August 2007 (UTC)
In that direction, here's few thoughts:
  • Aggregation of primary source information, sorting it, and doing basic maths on it isn't original research.
  • Interpreting it without reference to another source making that interpretation is.
  • Similarly, putting together multiple interpretation to make a new interpretation may be original research.
  • There are interpretations that wouldn't be considered original research, but there's probably no way to delineate them besides examples and common sense.
Do these make sense? Can they be worked into/reflected in the policy? SamBC(talk) 18:07, 23 August 2007 (UTC)

Primary v. secondary sources in the scientific literature

This is somewhat related to the above thread, but with a slightly different focus. I would like the cited examples of primary and secondary sources to include something about the application of these terms to the scientific literature. To me, primary sources in this context include journal articles reporting on novel research conducted by the authors of the article. Examples of a secondary source, in a scientific/medical context, would be a review article synthesizing available primary sources, a textbook chapter, or a statement from a large/respected organization such as the WHO/CDC/NIH/etc. The basis for this? Editors may claim journal articles as "secondary sources" and selectively cite specific articles, out of context, to advance their point. Editors should not be in the business of deciding which, of the thousands of peer-reviewed journal articles published every week, are the most scientifically significant ones. Instead, such determinations should be made by experts in the field, as indicated in the sort of secondary sources I've described above. This dovetails with WP:WEIGHT. What I'm suggesting would be, in fact, a very minor change (proposed changes are in italics):

  • "Examples of primary sources include archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; scientific journal articles describing novel research conducted by the article authors; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs."
  • "An historian's interpretation of the decline of the Roman Empire, or analysis of the historical Jesus, is a secondary source. Scientific review articles, chapters from widely used textbooks, or position statements from well-known scientific bodies are secondary sources. Wikipedia articles should rely on reliable, published secondary sources."

I don't want to open another can of worms about SPOV and so forth; these stipulations would exist to clear up an issue of what is primary vs. secondary within the scientific literature, and would not in any way prioritize scientific sources over non-scientific ones. MastCell Talk 22:35, 22 August 2007 (UTC)

We've tried to discuss and address this at the essay Wikipedia:Classification of sources — have a look and see what you think. Unfortunately, we're not too sure about the wider acceptance of this (either on wikipedia or in academia). SamBC(talk) 22:42, 22 August 2007 (UTC)
I think it's an interesting essay. I wish it wasn't necessary to spell this out, but I've seen it become an issue from time to time. I guess I'm wondering if there is any objection to adding the examples I've listed above to the list of examples in the current policy, or for that matter any objection to the way I've delineated primary vs. secondary in this context. I want to keep this particular proposed change separate from the "use-dependent classification" and "third-party" issues. MastCell Talk 22:59, 22 August 2007 (UTC)
I would support the change. I think you're right, and we ought to make this clear. COGDEN 17:35, 23 August 2007 (UTC)
I have added the proposed examples to the policy page. MastCell Talk 18:28, 23 August 2007 (UTC)
Don't agree with this idea. Many important journal articles aren't expounded on in "review" or "meta-analysis" articles, or in the mainstream media. This could lead to a lot of edit-warring and removal of good cites. I say that this kind of journal article may be both a primary _and_ a secondary source. If it's peer-reviewed it is a secondary source. Squidfryerchef 19:37, 23 August 2007 (UTC)
I also disagree with the change, as I've stated elsewhere on this talk page. (Sorry for the repetition -- should have posted here in the first place.) --Coppertwig 19:43, 23 August 2007 (UTC)

Could you repeat your objections for my sake? No one is saying that an important journal article not covered in a review or the mainstream media needs to be removed. However, such articles are generally handled as primary sources de facto (see WP:MEDRS for example). They can certainly be cited, but when it comes to determining WP:WEIGHT, secondary sources like those I listed are essential. The idea that peer-reviewed == secondary source is one I strongly object to, and one that does not mirror current practice on Wikipedia. There are thousands of peer-reviewed articles published every week; calling them all "secondary sources" puts individual Wikipedians in the business of deciding which of these thousands are most relevant. That determination should be made not by editors here, but by reliable secondary sources such as reviews, textbooks, scientific bodies, etc. MastCell Talk 20:49, 23 August 2007 (UTC)

In a scientific research article, the actual data (such as filled-out surveys) is a primary source; the published report is a secondary source and normally contains interpretation of the results by the scientific authors. It says clearly in the policy that a journalist's report of a traffic accident is a secondary source. There are huge numbers of such newspaper articles. Part of the task of Wikipedians (who are intelligent human beings, not robots) is to discern what is interesting and relevant.
The wording in bold type on this page makes it sound as if primary sources are not to be used or are rarely to be used. Defining scientific articles as primary sources would lead to large amounts of good data being cut out of articles.
Scientific articles normally contain commentary on other scientific articles, (often in their introduction), quite similar to what one finds in review articles but on a smaller scale. I see no reason not to treat this part of scientific articles as a secondary source. Scientific articles also usually/often contain interpretation, commentary, opinion, speculation and/or recommendations based on their results. Again, I see no reason not to put this into the same category as a journalist's account of a traffic accident.
Common sense, the wiki editing process and the basic idea of "no original research" is normally enough to keep people from building up new interpretations from a collection of scientific articles. Special words that make it sound as if scientific articles should rarely be used are not needed and could lead to a lot of harmful clearcutting of articles and loss of information, or at least a lot of wasted time arguing over and over again that scientific articles are OK to use as sources. --Coppertwig 21:25, 23 August 2007 (UTC)

Unprotected?

I'm probably violating something here, but I wanted others to readily see this. Why (or how) did the "Policy" page get unprotected? I though it was being protected until everybody could reach concensus and stop the edit wars on it, and it definately appears that there is no concensus yet. Does someone more vested in this care to revert it back to the protected version? wbfergus 17:26, 31 August 2007 (UTC)

I will, for a week. If we resolve things before then, anyone can unprotect it. Slrubenstein | Talk 17:29, 31 August 2007 (UTC)
Just in case anyone is confused. The old protection was set-up to automatically expire after a week. No one explicitly unprotected it, it just happened automatically.--BirgitteSB 18:24, 31 August 2007 (UTC)

Compromise on archiving

I suggest that if anyone wants to continue a particular thread from the archive they should be welcome to bring that particular thread back in to activity with their new comment. We need not activate the entire recent history, nor refuse to discuss anything that is of particular interest to someone here.--BirgitteSB 19:44, 31 August 2007 (UTC)

Let me point out that this page, as is, is already almost long enough to archive. We cannot resolve all disputes at once, and an attempt to do so will in my opinion simply leave the policy permanently protected. If we can reach a consensus on the two issues on the table, we can then archive this page and open discussion on two more. But no matter what - the purpose of this page is to discuss improvements to the policy page. It is not a soab-box or a blog. Discussion should be limited to concrete proposals to improve the article, discussion of such proposals, including criticisms and alternatives. Slrubenstein | Talk 19:48, 31 August 2007 (UTC)
I was pleased to see the earlier material archived and a new start attempted. I think what BirgitteSB suggests is fair, proper, and almost unavoidable but if nobody does what she suggests that would be quite satisfying. --Minasbeede 19:51, 31 August 2007 (UTC)
I'm not concerned about the length of the talk page. We can handle having multiple simultaneous topics on the same page. I just don't want anybody's comments to be shuffled to the archive page before others have a chance to respond to them. But Birgitte's suggestion sounds workable, as long as everybody is aware that there are active discussions and un-answered back-and-forth in the archives, and that anybody is welcome to resurrect any of the archived topics if they want to respond to something. COGDEN 19:54, 31 August 2007 (UTC)
I would agree with BirgitteSB's suggestion, it minimizes the length and simultaneous discussion without stamping on active discussions. SamBC(talk) 19:55, 31 August 2007 (UTC)
I agree with the archiving and with BirgitteSB's suggestion. There was way too much soap, with the lather removed the page's readability has increased a thousand-fold without compromising the integrity of the discussion. Dreadstar 20:04, 31 August 2007 (UTC)
I go to sleep, and then, the next day, in the afternoon, get back to this discussion and all of the old discussion, and much of the intervening commentary, has been archived. It's going to take some time to figure out what people have said. Jacob Haller 20:17, 31 August 2007 (UTC)
On the other hand, the policy finished out it's week of protection without any success from the week of discussion. Sometimes trying any new idea is better than continuing to do what is known not to work. But please feel free to bring any thread out of the archive that you feel would be helpful to resolving this.--BirgitteSB 20:25, 31 August 2007 (UTC)

This page is now something like 70 kb meaning half of it should be archived - yet we have made no progress! For starts, can someone refactor the discussion in the sections below this one, moving what belongs into the sections above, and anything else off the page (perhaps to anyother policy or guideline talk page?) Slrubenstein | Talk 09:40, 1 September 2007 (UTC)


How I see the source issues

I believe the sources section should have two purposes:

  • It must enumerate types of sources, their appropriate uses, and certain inappropriate uses.
  • It may also provide guidance on research practice for Wikipedia, but that's not its core task.

I believe that when discussing sources:

  • It is better to use plain language, or at least common terms with their common uses.
    • Hence my opposition to ad-hoc redefinitions of "secondary" and "primary."
  • It is better to define more types/cases than try to fit every type/case into primary-secondary-tertiary without subcategories.
  • It should distinguish different sections and/or different uses within one work.

I can't really figure out where the discussion is going right now, or why. I can't see why we can't just start with "category 1," discuss whether these are appropriate sources, these are inappropriate sources, these are appropriate sources in certain contexts, or the category should be divided into two or more parts, and then move on to "category 2." I'd suggest starting with tertiary sources, where we can probably reach the widest consensus, and then move "backwards." Jacob Haller 22:26, 31 August 2007 (UTC)

I think what you are suggesting is a good idea, but we really should be having this discussion with respect to an essay, with the goal of creating a guideline. This type of fine-grained discussion isn't really appropriate in a pillar policy, I don't think. COGDEN 22:32, 31 August 2007 (UTC)
I agree. This is the discussion page for the "No Original Research Policy", not for defining what category the various types of source information fall in to. Those are completely different things. All this "page" needs to do (or actually should do) is reference the different types available (preferably an official "Guideline") and note something along the lines of "Original research (OR) can easily enter an article when using 'primary source data' (whatever it getes defined as), so more care must be taken to prevent OR than is usually taken with 'secondary source data', though both types can be misused to create OR". wbfergus 23:14, 31 August 2007 (UTC)
How about looking at it from a different (I hope useful) point of view. Material with no supporting citation, no matter what it says, may be and be be suspected of being OR. Material that cites sources but that has words such as "therefore," "I believe," "I conclude," "we can see," "it can be seen," "X is therefore seen to be wrong/lying/misinformed" etc. (not an exhaustive list) are probably OR unless the indicated words come directly from the cited source and are embedded in the proper content from that source. --Minasbeede 23:25, 31 August 2007 (UTC)
Add to the list "however", "nevertheless", "notwithstanding", "apparently", and many others. OR is easy to spot. ≈ jossi ≈ (talk) 23:44, 31 August 2007 (UTC)
Thank you. And if an editor finds a source that disagrees with another source that has been cited in Wikipedia it is never (hardly ever?) proper to add to the relevant article anything similar to "Aha! X can be shown to be wrong because of (blah-blah-blah) from this source (citation Y.)" That's not a proper Wikipedia construct unless the "Aha" part is a quotation from the cited source. Either replace or remove the improper existing material (if it is really improper.) Note that the spirit/philosophy of Wikipedia is such that before removing material an editor should try to see the material objectively as the original editor would have seen it. (That last sentence may need work, or burial.) Even if citation Y does make that conclusion it is still necessary to determine if the material belongs in the article in question or if the better solution is to simple remove the offending words. In other words, is it pertinent to the article in question to indicate that X is in error? --Minasbeede 00:05, 1 September 2007 (UTC)
Not really. Words like "however" are the easiest to spot but the least likely to distort the sources. I've run across passages which are cited, and which completely contradict the source. In that case FV-tagging or removal is appropriate. I've come across passages which led to disputes about their interpretation; in this case, removal is also generally appropriate. I've inserted passages which reflected one very reasonable reading of the source, and then gone over them, and found there were other possible readings, so I've removed those passages.
One of the most common OR practices involves taking "the definition of x-ism" and "the definition of y-ism" and showing either that "x-ism" is a form of "y-ism" or "x-ism" is incompatible with "y-ism." The problem here is that "x-ism" and "y-ism" invariably have multiple meanings. For example, several times per year, someone or another goes through Libertarian socialism or, for that matter, Anarcho-capitalism and declares the term is an oxymoron, usually but not always on the talk page. Jacob Haller 00:16, 1 September 2007 (UTC)
There are of course three solutions for contradictory sources.
  • You can list the sources, and what each source says
  • You can try to determine the most reliable source
  • You can remove all the sources and their statements
Of course, it depends on undue weight, and so on. The really nasty problems come when the claim is important, but several sources make slightly different versions of the same claim, and listing each version would take forever. Jacob Haller 00:16, 1 September 2007 (UTC)
Thanks. The question "In other words, is it pertinent to the article in question to indicate that X is in error?" was really meant as a question an editor should ask himself in the type of situation described. I can accept that the default answer is "Not really." That sits well. --Minasbeede 00:21, 1 September 2007 (UTC)
If X is indeed in error, we will surely find a source that makes that assertion. Easy. And if we cannot find a source that makes such assertion, it is not up to WP editors to make it. ≈ jossi ≈ (talk) 04:32, 1 September 2007 (UTC)
The classic fairy-tale example that comes to mind is the small child who pointed out the emperor had no clothes - without having to find someone else to quote and unable to do so, since the "common knowledge" was that the emperor wore very fine clothing, so fine it could not be seen. Note that above I indicated that the default answer "Not really" sits well with me but there's clearly a strong tendency among some in Wiki-land to make that the only answer. No matter how clearly a Wiki editor sees that the common knowledge (or the common spin, or the prevalent spin) is wrong he's forbidden to say so. There's no place in Wiki-land for anyone to say "The emperor has no clothes" on his own, he has to find someone else who has said it. Once the child in the story spoke the truth there were a plethora of people who said the same thing. Prior to that nobody did.
Another example (possibly obscure) just occurred to me. It's the Gary Larsen cartoon of a bunch of geese walking in a V formation. One looks up and sees other geese flying in the same direction in a V formation and says "Hey, look what they're doing." We are, according to some, supposed to only walk as we migrate south. --Minasbeede 11:04, 1 September 2007 (UTC)

This page is now too long and must be archived, and we are no closer to any resolution. Please put all talk that is simply critical of NOR in another page. Nothing is stoping you from creating a page to propose replacing NOR with a different policy. But stop bitching about the policy here It is not the place. Be constructive. Discuss how to make the NOR policy stronger, or discuss on another page an alternative to NOR. You are free to do either one but they belong on separate talk pages.

It is pertinent to NOR but I quite agree this particular discussion can be deferred/archived/whatever - and I'll be glad of it. What I wrote was in response to Jossi. Please, in the future, aim fairly and aim at him (or whoever it is) the same as you aim at me. Otherwise you would seem to be prejudicing the discussion to favor your point of view, which tendency might also apply in the case of the insistence on giving special preference (with some nebulous justification) to secondary (in the unclear and apparently indescribable Wiki sense) sources. .--Minasbeede 11:41, 1 September 2007 (UTC)

My comment was meant to address many people which is why I did not use any names. In any event, talk that questions the value of an NOR policy itself can now go here: Wikipedia talk:Proposal to replace No Original Research Slrubenstein | Talk 12:51, 1 September 2007 (UTC)

Notice of draft revision

My impression of the current policy (not based on any factual checking of the edits) is that it grew over time to include the terms and definitions that now appear contentious. Those same issues really don't belong on the policy page though, as they "really" have nothing to do with the NOR policy, their presence seems to have been added for clarity, but have now become a point of much discussion. So, if the policy was "rewritten" so that it only pertained to NOR, and then had very brief explanations of the other "terms" with links to those pages/guidelines, the policy would be easier to understand for other newbies like myself, and any problems people had with "Sources" or anything else, could be discussed and handled there, where they should be, instead of here. NOR has has nothing to do with defining sources or synthesis, though those are areas where OR usually appears. Keep them separate.

What I created in my Sandbox is an attempt (probably poorly done) to show how some minor changes could help improve the policy and minimize future disruptions of the policy, without actually changing the policy. In previous discussions (now archived) it was asked several times "...then how would you propose to reword ...", and people rarely did. I think Vassyana was the only one who took any effort, but that was solely towards the "Sources" problem. My attempt was to create something to minimize future disruptions of the policy, that others can see and edit without an edit war on the policy page (which looks really, really bad). Others who have problems with the current policy are invited to make their proposed changes there as well. Additionally, if anybody else agrees that removing the "contentious" sections from the policy and having those discussed and maintained separately would benefit the overall policy, I would invite those people to help with rewriting those and appropriately linking them. I don't think this can all be done (and agreed upon) very fast, but I see it as a better use of people's time than just constant "discussing" here. After all, more than a week has gone by with the policy locked, and no progress has been made.

I have two variations. One (using primary, secondary and tertiary source info) is at Sample NOR Policy.wbfergus 10:31, 2 September 2007 (UTC)

The other, using Vassyana's sources proposal is at Sample NOR Policy #2.wbfergus 14:27, 3 September 2007 (UTC)


Thank you. What you are doing is a good step in the right direction. --Minasbeede 17:13, 3 September 2007 (UTC)

Over a year

I've been looking, and the discussion about primary/secondary has been going on for over a year in this talk page. One can look at a long effort like this and see either of two things:

(1) a diligent effort to reach consensus on something

(2) a diligent effort to escape consensus on something

Those who see (1) probably feel that to remove the something now as destructive of a long period of good, hard work. Those who see (2) probably feel that to remove the something now ends a discussion that should have ended far earlier.

At one point in the history I see complaints about a secondary discussion page having been created. I'm not one who asserted "stealth" but the discussion at that time sure resembles an objection to stealth. Sadly, I didn't keep track of when that was, but it was less than a year ago (as I have only looked a little more than one year back in the history.) When I see the date 1 September 2006 on primary/secondary discussion (that doesn't look all that different) I can only shake my head in wonder. --Minasbeede 14:30, 2 September 2007 (UTC)

Why not just recommend source-checking?

I think we should just advise that editors examine existing claims, and note claims which the sources given do not support and vice-versa. Ideally we should have ways to mark citations as fv, the source supports some but not all of the indicated claims; better sources required, the source can be interpreted to support the claim; better sources required, the source was not immediately available; I plan to check it later, and the source was not available; can other editors please check it?. We could even say that some kinds of sources require greater scrutiny than other kinds. Jacob Haller 21:00, 8 September 2007 (UTC)

There are already tags for that purpose. :) Vassyana 06:29, 9 September 2007 (UTC)
Those are pretty cool. I've been on here roughly a year and half and didn't know those existed. Stuff like that needs to be in a "Guideline" on sources, so all that pertinent information is in one place and easily found. That would go a long way towards solving some edit wars I've seen. Thanks for the link Vassyana! wbfergus 11:33, 9 September 2007 (UTC)
Thanks. I'd forgotten the verification needed tag. I agree with wbfergus that these should be linked and/or shown on the page. Jacob Haller 17:50, 9 September 2007 (UTC)


Which policy is available for situations in which is given a source simply ignored by other contributors? And i not talk about strange and unproof things. I talk about guys that simply not rated at all the sources, even if Wiki should not have the purpuse to prune the sources, expecially if those sources simply not are liked from some guys not exacty NPOV. I have even posted links to sites from the info are extrapolated, but those guys are simply rollbacked accusing me to not have post any source at all. These are situations, sorry to say but that's in which no policy can do enough. Simply helps some very POV guys to act pruning. But still i cannot understand how they can say 'unsourced' when i post sources.--Stefanomencarelli 18:39, 9 September 2007 (UTC)

Without specific I cannot say for sure, but it sounds like they are saying the online source is not a reliable source. Vassyana 23:56, 9 September 2007 (UTC)
That is what we have a dispute resolution process in WP, Stefano. ≈ jossi ≈ (talk) 11:37, 11 September 2007 (UTC)

peer reviewed academic articles are primary sources ?

If so could we add them to the list of examples of primary sources to make this clear ? Rod57 09:43, 12 September 2007 (UTC)

The whole primary/secondary sources issue is under strenuous debate atm, so it'd be hard to get consensus on such a specific right now. SamBC(talk) 11:15, 12 September 2007 (UTC)
I agree with Sam that consensus is in flux on whether to list examples or even discuss the differences between primary and secondary sources ... but this question can probably be settled regardless of the primary/secondary debates we are having. I don't think peer reviewed academic articles should be classified as primary sources. I suppose a few might qualify as primary (although I can not think of an example), but the vast majority should be classified as secondary sources - having based what they say on primary data and sources. Blueboar 12:34, 12 September 2007 (UTC)
I think the simple position is that the primary source issue is not relevant (with the qualification as Sambc notes that it is debatable). If they are peer reviewed in respected journals then that satisfies the original research test. Whether the concepts within are novel is not relevant as the peer review is assumed to give credibility and it can be shown that Wikipedia is not being used to spawn an original idea (read: wacky theory) that has no credibility outside the mainstream. There is an issue of being careful to give due weight to the topic - and one research paper does not necessarily make a topic notable, but we really shouldn't be worried about this type of source. Spenny 13:28, 12 September 2007 (UTC)
With respect to what was known up to the time of publication of a scientific article the material in the article (at least part of it) is supposed to be "novel." That's the whole point of scientific publication: the publication expands knowledge. That it is/was novel within it's field is no bar to its being cited in Wikipedia: as far as Wikipedia is concerned the publication of the material made it no longer "novel" in Wikipedia's terms. In this regard Wikipedia doesn't care about primary or secondary as applied to sources: they're supposed to be published (pretty clear distinction) and reliable (less clear.) It almost appears that for some "primary" amounts to "first-person historical" or "first-person philosophical," that there's an onus applied to these, that publication doesn't remove the onus, and that the onus grows as the date of publication falls further into the past. --Minasbeede 15:08, 12 September 2007 (UTC)

It is impossible to classify scientific articles as either primary or secondary sources. Take this 2004 paper as an example A trypanothione-dependent glyoxalase I with a prokaryotic ancestry in Leishmania major.

In the fist paragraph of the Results section I quote some dissociation constants I measured. If a Wikipedia article cited these results, it would be using the paper as a "primary source" - a source that presents novel, previously-unpublished data.

In the first paragraph of the Discussion on the other hand, I discuss previously-published data on Methylglyoxal synthase in the light of my results and bring these data together and interpret them. Here, if a Wikipedia article cited the paper on the possible problems caused by methylglyoxal synthase, it would be using the paper as a secondary source. Scientific papers are neither primary nor secondary sources, they contain elements of both. Tim Vickers 17:16, 12 September 2007 (UTC)

"It is impossible to classify scientific articles as either primary or secondary sources" and other than for the purposes of current wording in NOR we really don't care within Wikipedia. For the purposes of the wording in NOR it can't be done: sometimes the material is primary, sometimes it is secondary. We aren't going to change, the nature of scientific articles isn't going to change. If there's a conflict there's only one place change can be made to resolve it. If there's not a conflict why is there such a long history of discussion of the primary/secondary/both nature of scientific articles on this talk page (including in the archives)? --Minasbeede 17:37, 12 September 2007 (UTC)
In science, I would think that the primary material would be the data upon which the article is based, not the article itself. Again, if we focus on the core intent of this policy, we are saying that editors should not discuss ideas or draw conclusions of their own, but should simply report what others have said about the ideas and what those others have concluded. The reason for the caution (and that is an important word... it is a caution and not an outright ban) about using primary material is that it is very difficult to use it without forming conclusions of one's own. Thus, in an article on some scientific theory we should not cite primary "evidence" that supports or refutes the theory, we should use scientific reports and peer reviewed journal essays, etc. that discuss such evidence, and discuss the author's conclusions. For example, in an article on Global Warming we should not say "the polar ice caps are melting <cite to report of ice caps melting> thus global warming is occuring" we should instead say "Scientist note that the polar ice caps are melting and concludes that global warming is occuring <cite to report about ice caps melting that concludes that global warming is occuring>" Blueboar 18:08, 12 September 2007 (UTC)
OK, but it also says in "global warming" that homonuclear diatomic molecules (in particular, N2 and O2) don't absorb in the infrared. That's pertinent to the discussion but it isn't a conclusion (in the sense used above) and it hardly matters whether the source for the statement is primary or secondary: that's how it is. (This does explain why the major atmospheric gases don't play a part in the warming, making trace gases the dominant factors.) A policy applies to everything. If the policy causes harm or isn't inappropriate in some cases then it needs to be altered to remove the harm or inappropriateness. It's gratuitous to examine the sources for the statement that homonuclear diatomic molecules don't absorb in the infrared to determine whether the sources are primary or secondary. Plus apparently the real application of this policy wording that favors secondary over primary seems to be almost exclusively outside the realm of science. Science only gets dragged in because there is also in science a notion of "primary" and "secondary" sources - another clue that the policy wording has a flaw.--Minasbeede 18:18, 12 September 2007 (UTC)
Tim Vickers, if I cited your novel, original work on the interpretation of previous Methylglyoxal synthase data, I'd be citing you as both a primary and a secondary source. You would be the primary source of your novel conclusions based on the combined data, but the secondary source of the data and findings of previous researchers. You can never be the secondary source of your own original insights. You're the one that came up with them, so you are the primary source. COGDEN 18:31, 12 September 2007 (UTC)
No... this misunderstands what we are talking about. Just because a work makes new conclusions, that does not make it a primary source. Most accademic works make new conclusions. It is the data upon which those new conclusions are made that is the primary material... In science it would be the observations and experiments upon which the conclusions are drawn. In history that data would be eye-witness accounts of an event, or the original documents that the historian consulted to reach their conclusions. In any case, once it has passed through the mind of the publishing scientist or historian we get away from OR. Blueboar 18:45, 12 September 2007 (UTC)
Since these terms are so hard to define, and a single source can be argued to be either primary, secondary or both primary and secondary, does including this tenuous set of classifications really make the policy clearer? I don't think it does. Tim Vickers 18:55, 12 September 2007 (UTC)
The terms are not hard to define... we spell them out quite well:
  • Primary sources are documents or people very close to the situation being written about. An eyewitness account of a traffic accident is a primary source. United Nations Security Council resolutions are primary sources. 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. Examples of primary sources include archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.
  • Secondary 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. Wikipedia articles should rely on reliable, published primary and secondary sources.
I have highlighted the parts that deal with science issues... note that Cogden's and Tim's example does not fit our definition of Primary... and does fit the definition of secondary. It is the data that is primary, not the report analyzing it or reaching conclusions from it. Blueboar 19:25, 12 September 2007 (UTC)

It would seem from the title of Tim Vicker's article that he is presenting data that is of evolutionary significance (for a single species.) He's saying (I think) the evidence indicates something about the evolution of a particular species, about how it should be classified. Is that conclusion primary or secondary? It is Tim Vicker's OR, published. His conclusion: primary or secondary?

I should point out that if/when others cite Tim Vicker's paper used here as an example they far more likely will be citing it for its conclusion than for its raw data or its review of other work. The conclusion is the part that constitutes the major part of the advancement of knowledge that the article presents. --Minasbeede 20:05, 12 September 2007 (UTC)

The problem is that for some, the policy has broken down sources into their constituent elements, rather than a cohesive whole. I don't think that was the intention. Surely the intention was to take a rational view on the reliability of the source and how it has already been evaluated is part of that. I am not convinced that policy makers would have thought that there would be an encyclopaedic issue that would delve into the source data of a scientific paper, but more rationally it must be assumed that we are interested in the analysis, and more likely the summary conclusions of the paper rather than any detailed discussion. Therefore, it is reasonable to assume that in this context, the primary/secondary issue of the form Is this the first occurrence of a concept, the primary source for that discrete idea? was not considered relevant when the formulation was considered. Spenny 20:26, 12 September 2007 (UTC)

One of the biggest concerns of the referees of a paper is whether or not it is novel, whether or not it is the primary source for a discrete (and novel) idea. If it just rehashes old material the paper ought to be rejected and usually is. The referees don't act to restrain publication to fall within what is already known/believed, they act to require that whatever is published extends what is known/believed. (Review articles and the like excepted.)

The referees may not know whether the author's conclusions are correct. If they see obvious incorrectness they will alert the author in their referee statements and exhort the author to correct errors. The editor may refuse to allow publication until the author does so. If they're not certain they may say that in their comments but recommend publication. (I think there's a story about Physics Review Letters giving priority to all the papers that the referees didn't understand. The referees did not want to impede the spread of new, good ideas and since they didn't understand they couldn't judge "good," so by default they said "publish.")

I agree that probably not all aspects of the issue were considered when the formulation was considered. We're doing some considering now. --Minasbeede 20:41, 12 September 2007 (UTC)

Many peer reviewed articles on theoretical subjects should be considered primary sources. E.g., the article in which supersymmetry was first proposed, string theory, special relativity, General relativity, inflation theory, conformal symmetry, etc. etc. etc. Count Iblis 20:51, 12 September 2007 (UTC)

Folks... we obviously are not talking the same meaning of "primary" here... if you read the policy, it is clear that entire papers and articles are not to be considered primary sources... raw data and notes on experiments are listed as being the primary source material. Now, it may be that the policy uses the wrong word to discribe that concept (if the term "primary source" is regularly used by scientists to mean something other than the raw data) but that does not mean the concept is flawed... just the use of the term. We can keep the concept and change terminology (although given the fact that we do define how we are using the term, I don't see the need). Blueboar 21:03, 12 September 2007 (UTC)
Minasbeede... why are we considering it now? Is it really suddenly relevant? Blueboar 21:06, 12 September 2007 (UTC)
Any answer I'd give would surely be biased. It's a nice bias (I like it) but bias isn't what's needed here. I fear that the best answer is "review the recent history, as seen on this page and in its archives."
Responding to your previous paragraph, it would appear that the policy isn't all that clear - as you yourself seem to recognize by your first question. (I did propose changing the terminology a while back.) --Minasbeede 22:21, 12 September 2007 (UTC)
I'll quibble and say that the policy is perfectly clear... the only thing murky is that a few people seem to define the terminology in a unique way. I'll go back through the archives and try to figure out why this has suddenly become an issue. Blueboar 23:10, 12 September 2007 (UTC)
Addendum... well, having reviewed the issue, I still don't quite see why (or even how) this became a such "hot" issue now. No one seems to be pointing to an article where this is an issue or anything. Is it a concern based on reality, or is it just a few people worrying that "someone might not understand"? Pardon me for asking such blunt questions... I came into this debate late in the game and am trying to understand why it is even being discussed. Blueboar 23:33, 12 September 2007 (UTC)
People really are operating on mutually-exclusive definitions of primary/secondary. Some use a very strict dictionary definition, contrary to practice. Others draw upon their field familiarity in describing and defining them, which is some I myself am guilty of, drawing upon my studies in humanities for my own understanding of the distinctions. Due to their well-established varying definitions, I believe no matter how well we define them, there will continue to be misconceptions and misgivings. Vassyana 01:27, 13 September 2007 (UTC)
I called that the "First Order of Business" back on August 31. [18] --Minasbeede 01:43, 13 September 2007 (UTC)
Quite right. I made a significant revision to my proposal much to that effect that failed to generate enough support. I address some of the salient points regarding what we're trying to address below in a new section. If you have some suggestions on what potential distinctions we should use, or what they should rely upon, it would be sincerely appreciated. At this point, I think that it is important to focus on what we're trying to distinguish here, including why and how. Exact terminology is really of a secondary concern (no pun). Reaching a consensus on what we should be addressing in this policy is more important, and contentious, than what exact wording we should use. However, it seems clear that the primary/secondary distinction is confusing and contentious to many. Vassyana 01:53, 13 September 2007 (UTC)

Ahead, not after

I don't want to mess up the section below (leave it clean for better responses) but this is a response to it. I've done a recent personal review myself and have recognized that my reason to be paying attention to WP:NOR had nothing to do with source typing: I'm troubled by the synthesis section (which I'll not discuss now: there are issues enough.) So I came into the source typing discussion more because I saw it and recognized I had thoughts and an opinion than because I had really strong feelings about it. (Or if I have strong feelings I don't have any agenda behind those feelings.) It looks to me a lot like source typing is inherently too cumbersome to save. The bottom line in source typing is "primary bad" or "primary suspicious" or "bogeymen lurk in primary sources" or something of the sort. The problem is that yeah, sometimes the problem is there - but sometimes it isn't. A policy applies to all. There are cases where I wouldn't hesitate to use 17th century sources. That's "use," not "misuse." No matter how the source typing section is written I'll pay very scant attention to it. There are overriding principles that existed before any source typing was done and those seem to me to be fully adequate. While there are 17th century sources I'd willingly use there are sources from this month I'd never even consider. I recognize that the issue is framed primarily as one of source type but there's the undercurrent of "that-primary-stuff-wasn't-checked-out" and "that-primary-stuff-is-from-long-ago-and-may-not-correspond-to-what-modern-scholars-say." The problem is that these are being made a universal (BEWARE ALL PRIMARY SOURCES) and that's not needed. On top of that is the poor choice of words and on top of that is the confusion that still exists over the Tim Vickers sample article, as an example. There's at least three aspects of the article that have been identified: the experimental data, the historical review (if that's what Tim Vickers called it) and the conclusions. If the article persists and is cited it will mostly be the conclusions that are the reason for the citation. Those conclusions are of the bugaboo "primary" source variety. I think the conclusions are not outrageous and almost certainly will survive every test of time (not having read the article I'm not basing this on what it says but on a feeling.) It's primary, it will (probably) remain always citeable. I think it is plain wrong to put all such articles under any sort of cloud. I don't think the cloud is justified. Yeah, yeah, the article was refereed but if it were a chapter in a book it wouldn't be: books, which on the one hand Wikipedia anoints as "good" sources, often are also primary sources and would fall under the source-typing cloud. Or there'd be a clause explaining why books, though primary sources, aren't really supposed to be watched (or, for all I know, the policy might end up saying "books in particular have to be treated with caution.")

Besides which an editor could be quite aware of many sources that agree with each other and select the primary source because it's the best selection for what the editor is doing. Or he could know that the source he chose is the only source available. He's supposed to use verifiable sources. As yet he's not required to exhaustively explain why he chose any particular primary source - but that could be coming, for all I know. Just like primary soures, the editor who chooses to use one is under a cloud.

Besides which it's rather absurd to make any general statement about source types ahead of their use. If 99.9% of some type of source is bad (whatever "bad" means) then probably that source type will be used infrequently. I can't see how anyone can feel that it's correct to a priori cast aspersions on an editors choice of source, including that 0.1% of the time when it's "good."

I can't help tighten the wording: my heart isn't in it. I'd rather chuck the whole thing. I can live with its being a guideline. --Minasbeede 02:28, 13 September 2007 (UTC)

Template

I propose the creation of a template for new editors that explains to them why their additions were reverted if they are original research. Unless there is one that already exists that I am not aware of? SefringleTalk 18:15, 16 September 2007 (UTC)

There is {{uw-unsor1}} for unsourced contributions and {{uw-biog1}} for unreferenced contentious edits about living persons. There is no specific tag for original research. At least, there are none listed on the Big Page o' User Talk Templates. Vassyana 20:20, 16 September 2007 (UTC)

Use of an example to test

I suggest that an example is agreed on to test in actual practice what we are talking about. Perhaps the article Gillian McKeith would serve. The use of Ben Goldacre's words from his blog versus the use of his words from a newspaper is one issue. The use of sources that refute the "poo lady"'s claims that do not use her name is another issue. As BLP's represent our strictest use standards, and the data and issues are small in number, and there is no ongoing major conflict at that article, perhaps it is a good choice. WAS 4.250 01:08, 17 September 2007 (UTC)

Squaring the circle

Let me be clear: I think this is a generally commendable policy, but one that can often be used in bad faith by "wikilawyers" or by people who just like to waste other people's time (like if someone says, "There is no citation here for the claim that these two words rhyme", and while that example is hypothetical, it's not exaggerated.)

I think two major things need to be added to the policy page:

  1. A clear statement near the outset of the intent of this policy, which I would summarize roughly as:
    1. Keep Wikipedia free of non-notable pseudo-science and crackpottery.
    2. Help to assure that articles that provide interpretation of literature, history, etc. present the thinking of those who are generally considered authorities in the relevant field, not those of random people.
    3. Help to assure that factual claims are reliable.
  2. Link explicitly to Wikipedia:These are not Original Research, which may need to be renamed and/or become a more balanced discussion, but which is a necessary corrective to the over-literal mentality that prevails among a lot of our contributors, perhaps especially those who are more familiar with areas in which strict rules must be followed (games-playing, citation style in college term papers, for two examples) than with those that require some judgment in weighing conflicting imperatives.

- Jmabel | Talk 19:02, 16 September 2007 (UTC)

Blueboar is working on a revision of the wording of the policy. Can you help with that, perhaps, at the appropriate time, suggest wording changes like those suggested above?
This discussion has been going on for some time. The main policy page has been locked for close to 24 days already and is probably going to soon be locked again. It would be good to concentrate effort on the particular part of the policy that is the source of contention. My solution would be other than rewording the policy but I favor an effort to improve the wording. That (improving the wording) appears to be productive. --Minasbeede 19:09, 16 September 2007 (UTC)
The essay is a good idea and makes some valid points. However, it does need a serious revision. Any assistance in revising the essay would be appreciated. Vassyana 20:30, 16 September 2007 (UTC)
I've made some significant revisions to the essay. Feedback and assistance would be appreciated. Vassyana 21:16, 16 September 2007 (UTC)
Your edits were good; I made one further one, which I hope will be uncontroversial. What do you think of the idea of pointing explicitly to that essay from the "see also" section of the policy page? - Jmabel | Talk 21:37, 16 September 2007 (UTC)
Thanks. :) Your clarification was excellent and exactly in line with how I was trying to tighten and rework the essay. I would definitely support linking the essay from the main NOR page, as it would help mitigate the wikilawyering that you and others have pointed out occurs. Vassyana 00:19, 17 September 2007 (UTC)
I reverted one change you made (about how to deal with typos & misspellings) because your change ignored or contradicted the intent of the essay -- which is to argue that intelligent use of material should never be confused with original research. And IMHO, slavishly repeating every letter of a text is not intelligent use -- which is why I wrote it how it reads. This is an essay -- the expression of an opinion -- not a statement of policy. It should be quoted when someone agrees with it & believes it expresses why they took some action; it should not be quoted as if it were policy. If you don't agree with it then write your own essay -- don't rewrite mine. -- llywrch 01:43, 17 September 2007 (UTC)
I've responded to you on the essay's talk page, but please remember you don't own articles in the Wikipedia namespace. If you wish it to reflect solely your own views, please keep it to your userspace. Vassyana 19:19, 17 September 2007 (UTC)

An alternative proposal

This is an alternative to Blueboar's proposal. I would have no objection to changing the word "sources" to "materials." it also may be possible to incorporate elements of Blurboar's proposal and mine. Slrubenstein | Talk 20:05, 16 September 2007 (UTC)

I will again object to the distinction between secondary and tertiary sources. Even experts admit the distinction is often a big grey line and difficult to determine. Review literature and textbooks are among the most reliable sources, and as such should not be divorced from reliable secondary sources in general. The main need for any such distinction is the desire to not be derivative of other encyclopedias, but they are a very limited subset of tertiary sources. The concern about being derivative of other encyclopedias is a valid one, but I don't think it has a whit to do with original research. For our purposes here, distinguishing between primary and secondary material is quite sufficient. Vassyana 20:25, 16 September 2007 (UTC)

First, let me know Blueoar's version makes the same distinctions. Second, I tried to respond to your (and others) concerns in my rewording. Slrubenstein | Talk 20:37, 16 September 2007 (UTC)

I see the caveat at the end, but I think it's a bit insufficient considering the block of text preceding it prejudices against tertiary sources. As I said, I also think it's completely superfluous to a discussion of original research. To be clear, beyond that, I think this is an excellent draft proposal. A single sentence at the end of secondary sources mentioning the problem with being derivative of other encyclopedias would suffice without delving into the secondary/tertiary divide. (I am aware Blueboar's draft makes the same distinction and I have similarly objected.) Vassyana 21:07, 16 September 2007 (UTC)
Being clear, my objection is based on the observation that encyclopedias (the real "problem") are a very limited and small subset, while the majority of such sources are particularly reliable (such as review literature and textbooks). Vassyana 21:13, 16 September 2007 (UTC)
Vassyana, I have no problem with what you are saying... why not take the relevant sections from the the two drafts and work up a phrasing that incoporates your ideas. In other words let's see some proposed language :>) Blueboar 22:23, 16 September 2007 (UTC)
I was trying to avoid adding another draft, in whole or part, so as not to dilute the discussion further. However, I will work on an example of how secondary and tertiary can be combined, if you feel such a draft would be helpful instead of distracting. Vassyana 00:27, 17 September 2007 (UTC)

I think Vassyana is right which is why in my version I wrote "However, some encyclopedias and other tertiary sources, such as Annual Reviews, have signed articles, and often articles that explicitly promote the author's own views. In this sense, a tertiary source can also be viewed and treated as a secondary source." perhaps Vassyana thinks this needs to be stated more strongly or clearly, or would suggest a specific change to my proposal? If it is in line with the sentence I just quoted, I am pretty confident I will hav no objection. If it diverges from what I wrote, I am certainly open to suggestions and will try to accommodate or dialogue until we agree. If you think the caveate at the end is simply insufficient, offhand I see two suggestions: (1) beaking it off into a new paragraph exclusively for when tertiary sources are encouraged i.e. one paragra[h that is proscriptive, and another that is prescriptive. OR, (2) divide "tertiaty sources" into two kinds, anonymous and signed. Other ideas? Slrubenstein | Talk 12:07, 17 September 2007 (UTC)

I've made two drafts. One combines the secondary/tertiary sections (User:Vassyana/PSTS draft 1). The other alters the focus/thrust of the tertiary sections (User:Vassyana/PSTS draft 2). Cheers! Vassyana 16:00, 17 September 2007 (UTC)
Either of these would work for me. If I had to pick one, I would go for draft 2... but that is not a set in stone choice. Blueboar 16:12, 17 September 2007 (UTC)

protection

I think we are rapidly moving towards consensus. I have no objection to changing "sources" to "materials" in my proposal, which I think would please many other editors here. I have no objection to reworking the section on tertiary materials, in line with Vassyana's concerns, and I have suggested two possible solutions and invited other suggestions.

There has been little criticism of the section on the history of the proposal/its origins that I proposed; but I hope people will at least go over it for style and clarity.

I extended protection another week just so we can focus on these proposals and keep moving towards consensus. Hopefully all current conflicts will be resolved before the week is out. As we move forward any admin can move material into the policy even in its protected state. My intention is that the continued protection give us space to finish the forward movement to consensus on these issues.

By the way, I know Llywrch, Vassyana and others have been working on an essay on what is not original research. I suggest this: that after we achieve agreement on the two proposals (one on originas of the policy, one on materials) we move that essay to this page and consider working on it as a proposed new section of the policy next. Slrubenstein | Talk 12:21, 17 September 2007 (UTC)

The consensus being reached is, at least as far as I am concerned, a consensus among those who believe this section is proper. I stopped commenting to allow such consensus to be developed.
The test for the source typing section, which it has failed to pass in the past, is for it to exist and to not be the source of repeated objections. By that I mean objections by others, not myself. I remind all editors of WP:consensus, where it says "A good sign that you have not demonstrated a change in consensus, so much as a change in the people showing up, is if few or none of the people involved in the previous discussion show up for the new one." There has been a long line of objections to source typing in the past, as can be seen in the history of the talk page.
If you reach real consensus, wonderful. If not it will show.
My somewhat detached language may seem to indicate I am withdrawing from this discussion. That is true. I'm still a party to the mediation request but with 5 holdouts it's highly unlikely mediation will occur so my my absence will not have an impact. In the unlikely event that all agree to mediation I will participate enough to allow the process to work.
There are good Wikipedia reasons to withdraw and good personal reasons. I bow to reason, thankfully. --Minasbeede 13:35, 17 September 2007 (UTC)
I think the essay is better left as an essay. It isn't policy, and shouldn't be hammered into looking like policy. It's a discussion of how to interpret policy sanely. Policy as such should stay simple, but there need to be side discussions (guidelines and essays) explaining the subtleties. - Jmabel | Talk 18:49, 17 September 2007 (UTC)
I agree with Jmabel about the essay. I'm flattered that some of the ideas I proposed will be incorporated into policy, but in general it's meant as a warning to people who insist on pushing the language of this policy beyond its logical limits -- honest to God, I remember encountering one person who insisted that photographs a Wikipedian take are original research -- while at the same time encouraging people to intelligently follow this policy when writing an article, & not worry about the Wikilawyers. Other points not covered (e.g. Jacob Haller's comment about Neologisms here), or other opinions, are important but would be better discussed in their own essays. -- llywrch 19:53, 17 September 2007 (UTC)

A Different Approach

OK, here is the Spenny re-write, which paraphrases the thinking, but adds in a couple of points. I've bolded the key phrase which is the root of what all this source jiggery-pokery is about. It is this key concept that we are battling with, but it is not explicitly stated. Spenny 23:59, 17 September 2007 (UTC)

Recognising the Right Source Materials

Within source material we will find a mixture of facts and interpretations. Some sources will be basic observations that offer little in the way of analysis; other sources will offer analysis of information and draw conclusions. When adapting material for use in Wikipedia, it is important to only introduce facts and observations that already exist.

A Wikipedia article or section of an article can employ statements of facts only if the material is used (1) only to make descriptive claims, the accuracy of which is easily verifiable by any reasonable, educated person without specialist knowledge, and (2) never to make analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on statements of facts should be careful to comply with both conditions.

Typically, statements of fact will tend to be found in materials such as historical documents; personal diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; raw tabulations of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works (such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs).

Wikipedia articles can include analytic, synthetic, interpretive, explanatory, or evaluative claims (1) only if such claims come from a reliable, verifiable source and (2) the point of view is clearly identified and accurately represented. In general, it is this type of source that is of most use to Wikipedia as an encyclopaedic article is a summary analysis of a topic and using raw facts to build an article will often require too much original work that could be challenged.

Further, it is most likely that works which are a comprehensive analysis of the topic will be most suited to providing a comprehensive demonstration of a particular viewpoint. Cherry picking individual concepts from different sources may suggest editing to support a point of view, and this can be especially so in the case of using passing references. (This is really NPOV, and I haven't worked this up well)

Useful analysis is typically found in academic journals, Government Inquiries, and media sources such as serious documentary programs and a few respected newspapers. {Yes, I know this bit is weak, but its past my bedtime).

Wikipedia strives to be a superb encyclopedia in its own right. Since Wikipedia does not want to be derivative, materials found in encyclopedias or similar reviews are often of limited value for Wikipedia research. Annual Reviews and Encyclopedia Brittanica articles often provide extensive bibliographies that are valuable tools for identifying important materials, and therefore of great use to Wikipedia editors. Nevertheless, these materials do not necessarily have the same content policies as Wikipedia and for this reason should not be viewed as authoritative. However, some encyclopedias and other material, such as Annual Reviews, have signed articles, and often articles that explicitly promote the author's own views which can be viewed as source material in their own right.

Proposal

Ok... I think we are at the stage where we agree on the basic principals (1) we want to talk about material and not sources (2) the issue we want to adress is the misuse of primary materials - eitors taking those materials out of context and making analytical, interpretive, or conclusionary statements from them. (3) All but the most basic analytical, interpretive or conclusionary statements included in our articles should be backed by reliable sources (defined as secondary material) that contains the same analysis, interpretation or conclusion.

Now to the hard part... drafting language. As a starting point, I am going to actually propose the language I drafted above (in my "what if we just swapped words" thread). We probably need to edit it further (for example: dropping the tertiary material section?). But I think it may get us off to a good start. Anyone disagree? Blueboar 12:33, 16 September 2007 (UTC)

Nope. wbfergus Talk 12:47, 16 September 2007 (UTC)
Is that "nope" as in "I reject the proposal"... or "nope" as in "no problems here"? Blueboar 12:50, 16 September 2007 (UTC)
That's a "Nope" as in "Nope, I don't disagree. Go ahead and try your hand at it. The worst that can happen is after a lively debate it's not accepted. wbfergus Talk 12:56, 16 September 2007 (UTC)
The issue the policy as a whole must address is the misuse of all materials, not only primary. Maybe you mean in this particular discussion about the policy we're only addressing the misuse of primary material. Anyway, I see no problem so far in how you're proceeding. Thanks again. --Coppertwig 17:08, 16 September 2007 (UTC)
Yes, continue. Since you've said it I'll point out that you also now have to clarify "all but the most basic." That's 4 more words than "all," and the added four words make a difference. The words can't be left vague in a policy. I'm not trying to be difficult and I don't ask you to change the order in which you do things, just want to say that at some point that need has to be addressed - unless before then you're convinced by others to take out those 4 words. --Minasbeede 18:10, 16 September 2007 (UTC)
Actually, I would be comfortable saying just "All statements of analysis, interpretation or conclusion..." Blueboar 18:42, 16 September 2007 (UTC)
Understood. I agree with wbfergus, above. "Go ahead and try your hand at it. The worst that can happen is after a lively debate it's not accepted." As you will be debating language with which I don't agree I'd rather let you just debate it until it's as you think it needs to be. It's not an easy task and I don't want to make it harder. --Minasbeede 18:48, 16 September 2007 (UTC)
Well, if no one objects, then can we get an admin to unlock the page and export the draft? Or do we need more drafting? 19:37, 16 September 2007 (UTC) —Preceding unsigned comment added by Blueboar (talkcontribs)
I object to the wording of the draft but I don't object to the change. I can say now what it is I object to or I can wait. If even a few would be happy with making the change that would be something. --Minasbeede 19:45, 16 September 2007 (UTC)
I disagree, for the reasons I stated above. Defining "primary materials" (the presumptively bad one) as anything that doesn't contain an analysis doesn't address the whole point of this article, which is Original Research. Citing "primary materials", as you have defined them, is not original research. Even if citing them is bad (which I don't think is true), citing them is not original research, and has nothing to do with original research. Indeed, it is impossible to commit original research when citing "primary material" alone. Actually, original research can only happen when you cite "secondary materials", which according to this definition would include an editor's original, unpublished interpretation of the "primary material". My own pet interpretation of the Gospels is "secondary material". It's not reliable of course, but I think the revised definitions don't get us anywhere, and are likely to lead to more confusion, since we refer to reliable sources as "sources", and these definitions contradict the established academic definitions of primary and secondary sources. COGDEN 19:07, 18 September 2007 (UTC)
Cogden... keep up... you are disagreeing to a proposal that has already been set aside as other proposals have been put forward. :>) ... as for the rest of your comments... While the point is somewhat moot (as we seem to be moving away from even "primary material" etc.), I very much disagree that it is imposible to commit original research when citing primary material alone. Take the Constitution of the US... that is primary material (in fact it is a text book example of a primary source as well). If an editor writes something like: "It is fact that US citizens have a constitutional right to own handguns" and cites the constitution to back that statement, the editor is indeed committing original research. He/she is inserting his/her own interpretation the constitution. To make it not OR, he/she needs to cite some other source that includes this interpretation. Blueboar 19:25, 18 September 2007 (UTC)
Other reasons why the proposal is confusing and probably counter-productive:
  • All "secondary material" is by definition a primary source, because it contains original analyses and conclusions.
  • Most primary sources, if they are published, are "secondary material", because to get something published you usually have to say something new, and make conclusions.
  • Some secondary sources are "primary materials", because they simply convey previously-published information without further analysis.
  • Some "material" is neither "primary" or "secondary". For example, a source that contains conclusions, but does not draw on "primary material" to make those conclusions, falls into this limbo.
COGDEN 19:23, 18 September 2007 (UTC)

Comments on the alternative proposal

Well, I don't have a problem with the alternative... but I suspect that others may. For one thing, it does not solve the problem that not everyone agrees with our definition of what a "primary source" is. This version certainly defines it in a much clearer way, but it is a definition that is unique to Wikipedia. Personally, I have never had a problem with that... I understand that in Science the term means one thing, in History it means another... while on Wikipedia it means yet something else. When I apply the NOR policy on Wikipedia, I use Wikipedia's definition, just as I would use the Science definition if I were writing a scientific paper. But, based on the comments that keep being raised here... I am not sure if others can make that distinction. Was there a reason you wanted to keep the word "source"? Blueboar 23:38, 16 September 2007 (UTC)

To your last question, habit, on and off Wikipedia. I have no objection to changing it to "materials" if the change makes a diffeence to anyone. As to definitions unique to Wikipedia: I think this is unavoidable because Wikipedia policies address articles covering professional topics (law, architecture, medicine, business), life sciences, physical sciences, social sciences, and humanities. Our policies must be flexible and inclusive, and the result will be that we will have to rely on terms whose definitions may be unique to Wikipedia. I see no peoblem with that as long as (1) the terms are clearly defined and (2) a given term, or set of terms, is inclusive enough to guide people working on very different kinds of articles.
Blueboar, I think my alternative is more developed than yours so I propose that absent your objections we use it as a working model. However, I also know that you have been noodling around with your proposal in dialogue with a variety of other editors - I think my alternative has value only if you will edit it to incorporate whatever in your proposal you think is an improvement. Slrubenstein | Talk 12:04, 17 September 2007 (UTC)
Absolutely no problem with that... your version is more succinct and much less verbose. Good work. Blueboar 13:32, 17 September 2007 (UTC)

I've had a look at the proposal above and also Vassyana's drafts. Sorry to be a dog with a bone, but I am still struggling to see what the differentiation of sources brings to the party. The list of primary sources still has the issue that as a generalisation it doesn't work. Simple example, TV documentary could well be a sound secondary source. The list does not explain what makes it a source for fact as opposed to a source for opinion and so this simply raises the query - is Pepys' Diary a source of fact or opinion? It also strikes me that the policy statements made aren't really dependent on this source typing. I guess the other observation is that the differentiation on fact and opinion doesn't really hold water with distinguishing with source typing by example, though it is clear that the distinction between fact and opinion is the issue that is being explored by the policy wording (which I think stands up pretty well on its own). Spenny 16:56, 17 September 2007 (UTC)

Ian, this is why we changed from "source" to "material"... we are no longer differentiating different types of sources, we are differentiating different types of source material... the stuff IN the sources. To look at your examples: In most cases, a TV docuementary will contain both primary and secondary material... the parts of it that are primary material (say, for example, raw footage of Vietnam War soldiers carrying wounded men to a helicopter) should not be used to support a conclusion (say: that the number of wounded caused Americans to turn against continuing the war)... but the parts of it that are secondary material (say, the commentary of the narrator, or a shot of a historian talking about the war) can be. Pepy's Diary on the other hand is clealy both a primary source and primary material, and caution should be used when citing it. His diary amounts to being an eyewitness account and should not be used as a source for analytical, interpretive or conclusionary statements about the events he discusses. We can (and I think probably should) quote him in our articles ... but as essentially an eyewitness... for analysis or interpretation of his comments we need to cite to an historian that has performed such analysis and interpretation, and not interpret it ourselves. It is important to remember that this policy does NOT say "don't use primary material" (or even primary sources)... in simply says to be aware that some uses are considered OR. Blueboar 17:31, 17 September 2007 (UTC)
OK, I understand, but then that is not what the examples are, they do not do what you have just done, picking out stuff in sources, it is still a list of primary sources. (That is something to consider as a solution). I am commenting on what I am actually reading. However, that does bring out the point, do we want to clearly and explicitly talk about the individual concept under discussion? As it stands that doesn't come out. (I nearly thought about spawning that out as a separate thread, but I am trying to go step by step on this). Spenny 17:48, 17 September 2007 (UTC)

Two points - and if both of you agree, I would ask you to suggest appropriate changes. First, any list is of illustrative examples. By this I mean two things: first, it is not inclusive; second, it is not exclusive by which I mean these things are not always examples of (primary, secondary, etc.) (sources, materials, etc.) This shouldn't be too surprising: an apple pie could be (1) an example of my mom's cooking (2) an example of things you can do with fruit (3) an example of a desert (4) an example of symbols for America etc. This leads to my second point: whether something is a primary or secondary whatever depends on how it is used, the context. perhaps this gets closer to Ian's attmpt to get at the concept; i am suggesting that the concept involves how something is used and in what context. Does this make sense to you guys? If so, what language could more effectively communicate this? Slrubenstein | Talk 17:57, 17 September 2007 (UTC)

First... I have done some more tweeking which might solve this (let me know)... but to answer Slrubenstein's question on primary, secondary being based on use... no. Some things are primary by definition and not by use. The US Constitution (for example) is an example of something that is always primary material by definition ... it does not contain any analysis or interpretation... it just IS. We can quote it in our articles, but we can not interpret it or draw our own conclusions from it. For that we need to cite a reliable source such as legal scholar or Supreme Court justice, etc. Blueboar 18:12, 17 September 2007 (UTC)

Some observations:

  • I'm really uncomfortable with making up definitions like "primary materials" and "secondary materials" out of thin air. For one thing, it's confusing. In normal conversation between academics, if you said "primary materials", everyone would assume you meant "primary sources".
  • A policy is supposed to document current, previously-consisting Wikipedia practice. Here, we're making up categories that have never been explained or explicitly used before on Wikipedia.
  • Since these categories are made up from scratch, rather than growing organically from Wikipedia practice, there are bound to be logical inconsistencies, and there are:
  1. The policy says you can never use primary materials in support of interpretive claims. However, diaries, historical documents, census results, interviews, laboratory and field observations, artistic works, and works of fiction may, and usually do, contain interpretive and explanatory claims by the original author. But wait! That means they're also "secondary material", so you can include them. Or can you? I'm confused.
  2. The article says, "The conditions that apply to the use of primary materials also applies to the use of primary source material included in secondary sources", but the "primary" section says "never to make analytic, synthetic, interpretive, explanatory, or evaluative claims". So material that makes analytic claims should never make analytic claims? I'm confused again.

COGDEN 19:57, 18 September 2007 (UTC)

How to promote instruction creep

It's easy. Whenever there's a dispute between a more restrictive and a more flexible version of the policy, just find an example of something that would be allowed under one version but not the other, and claim that anyone who supports allowing that is against the policy and should therefore be excluded from the discussion. It's not necessary to actually succeed in excluding anyone; it's enough to put a chill on the making of certain types of assertions. --Coppertwig 16:43, 17 September 2007 (UTC)

Are you saying this is being done here? If so, I don't think it is intentional. Blueboar 18:29, 17 September 2007 (UTC)
It seems to me that everyone involved here is trying in good faith to improve the policy. My message was not intended to imply otherwise. --Coppertwig 21:34, 17 September 2007 (UTC)
I am worried about instruction creep here for another reason, because the new proposals use a whole lot of words to say very little, while introducing a lot of nuances and logical inconsistencies that will provide fodder for the next generation of edit wars. In User:Slrubenstein's proposal, for example, when all is said and done, "The conditions that apply to the use of primary materials also applies to the use of primary source material included in secondary sources". Why, then, do we need to make a distinction? COGDEN 20:07, 18 September 2007 (UTC)

Blueboar's re-draft of Speny's approach

Taking what Ian wrote and tweeking.... mostly adding 1) that facts need to be cited to reliable sources as well as opinions... and 2) stressing that analysis etc should be the same as in the source. I could live with this approach. Blueboar 01:15, 18 September 2007 (UTC)

I think this is a good approach that will help lay to rest a lot of the disagreement. I could support this as well. Vassyana 01:18, 18 September 2007 (UTC)

Recognising the Right Source Materials

Within source material we will find a mixture of facts and interpretations. Some sources will be basic observations that offer little in the way of analysis; other sources will offer analysis of information and draw conclusions. When adapting material for use in Wikipedia, it is important to only introduce facts and observations that already exist.

A Wikipedia article or section of an article can employ statements of facts only if the material (1) comes from a reliable, verifiable source, (2) is used only to make descriptive claims and (3) never to make analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on statements of facts should be careful to comply with all three conditions.

Typically, statements of fact will tend to be found in materials such as historical documents; personal diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; raw tabulations of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works (such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs).

Wikipedia articles can include analytic, synthetic, interpretive, explanatory, or evaluative claims (1) only if such claims come from a reliable, verifiable source, (2) the point of view is clearly identified and accurately represented, and (3) the analysis, syntethis, interpretation, explanation or evaluation matches that in the source. In general, it is this type of source that is of most use to Wikipedia as an encyclopaedic article is a summary analysis of a topic and using raw facts to build an article will often require too much original work that could be challenged.

Further, it is most likely that works which are a comprehensive analysis of the topic will be most suited to providing a comprehensive demonstration of a particular viewpoint. Cherry picking individual concepts from different sources may suggest editing to support a point of view, and this can be especially so in the case of using passing references.

Useful analysis is typically found in academic journals, Government Inquiries, and media sources such as serious documentary programs and a few respected newspapers.

Wikipedia strives to be a superb encyclopedia in its own right. Since Wikipedia does not want to be derivative, materials found in encyclopedias or similar reviews are often of limited value for Wikipedia research. Annual Reviews and Encyclopedia Brittanica articles often provide extensive bibliographies that are valuable tools for identifying important materials, and therefore of great use to Wikipedia editors. Nevertheless, these materials do not necessarily have the same content policies as Wikipedia and for this reason should not be viewed as authoritative. However, some encyclopedias and other material, such as Annual Reviews, have signed articles, and often articles that explicitly promote the author's own views which can be viewed as source material in their own right.

Comments

I have been following this debate only intermittently, so some of the issues I mention below may have already been discussed. If so please simply point me to the correct section. (for easy labelling I'll continue to use PS, SS and TS terminology in my comments below)

  • "Typically, statements of fact will tend to be found in materials such as ..." : This should be the other way around. PS typically contain statements of facts; it is not true that statements of facts are only or typically, found in PS. Statement of facts are also found in secondary sources, and it is in fact preferable to use SS for sourcing facts too. Editors will perhaps interpret the above version of the policy to mean that they should look for PS, such as historical documents, diaries etc to state raw facts (such as when a war began; when Hamlet was written etc.)
  • Signed article in EB are good sources for material on Wikipedia; and Annual Reviews (or other review articles in general) are in fact ideal sources for writing wikipedia articles on subjects they cover (since they are "works which are a comprehensive analysis of the topic" and much better than individual research papers that only present a narrow and novel POV). As currently worded, it is not clear if using Annual review articles as source is being deprecated, or if these are being listed as exceptions to the general rule about using TS.
  • "Wikipedia strives to be a superb encyclopedia in its own right. Since Wikipedia does not want to be derivative ..." I don't see what this has to be with quality of sources, as opposed to copyright issues. If copyright didn't prevent it, wouldn't we have liked to download the complete (at least signed) content of Britannica and other standard encyclopedias onto wikipedia as the starting point for further revisions and improvements ?

Abecedare 02:47, 18 September 2007 (UTC)

Good points, and I agree with them. With regard to the last, I was trying to get away from the tertiary source whilst respecting the spirit. Some of the reason is the issue of copyright. I'd be tempted to simply delete it as it is not useful policy - the policy is in the section above. I've got to save the planet today, so let's give this a little time to brew (a bit busy saving the planet today). I hope what I have demonstrated is that in principle, even through my garbled editing, we can get the spirit of what is required without getting into technical terms. Note to others: feel free to hack away, I don't want to own this, even in talk. Spenny 07:25, 18 September 2007 (UTC)
Spenny's draft as modified by Abecedare's concerns is moving entirely in the right direction. WAS 4.250 09:42, 18 September 2007 (UTC)

I like this because it is a fresh approach, but I don't think it reflects Wikipedia practice, maybe in part because the language is not precise or comprehensive enough. "Statements of facts" are frequently used to support analytical, synthetic, etc. claims, particularly when the statements of fact are found in the same source as the analytical claims. Usually, an editor will cite the fact, cite the author's interpretation of the fact from the same source. The fact is used by the author to make the analytical claims, not the editor, it's not original research. Such a use of "statements of facts" would barred by this draft.

Also, since this changes the primary/secondary dichotomy to a fact/interpretation dichotomy, we have the additional problem of defining what is a "fact" and what is an "interpretation". This seems a bit too metaphysical for a pillar Wikipedia policy. For example, if Author X says "Joe was drunk", is that a fact, or an interpretation? And does it matter? Both facts and interpretations can be cited by Wikipedia editors, so long as neither the facts nor the interpretations originate with the editor.

As a corollary to the above, how can we really say that analytical sources are "of most use to Wikipedia"? Sources with pure, raw facts are useful too, maybe even more useful, since without the raw facts, what is there to interpret? COGDEN 20:27, 18 September 2007 (UTC)

Cutting to the chase

The more I think about all of this, the more I think that we were approaching this back-asswards. The point of this policy is that we have to answer the question: "Oh yeah? Sez who?" with... "This reliable source over here sez so." If the answer is "I say so" or "the facts say so", or some thing like that, you are venturing into OR. The type of source being cited does not realy matter, the type of statement that it is being cited for does. For a statement of fact, we need to cite a reliable source that states that it is a fact. For a statement of opinion, we need to cite a reliable source that states that opinion. And for a statement of analysis, interpretation, synthesis, conclusion, etc., we need to cite a reliable source that contains the same analysis, interpretation, synthesis, conclusion, etc. Ian's approach is good... it does essentially approach the issue based on what the source is being used for, instead of what the source is. But it is a relatively long winded way of saying it. Can we be more concise? Blueboar 13:47, 18 September 2007 (UTC)

Precisely what myself and several others were arguing for several weeks ago. This policy does not need to define or really care what the types of sources/material are. That's all material for a different page, whether policy, guideline, whatever. All this policy needs to say is here are further definitions of these 'materials', but if you edit an article, be prepared to cite your edits with reliable 'materials', in the proper context as the 'material' you cite. wbfergus Talk 14:17, 18 September 2007 (UTC)
Yeah... I know... I just had to come around to what you have been saying in my own unique way and in my own time. I thank you for being patient with me while I underwent this process. I guess it is what they call "consensus building". Blueboar 15:11, 18 September 2007 (UTC)
No problem. I work with a bunch of scientists, and it's extremely aggravating to me on how long they can argue over a point for months or years only to finally come to a conclusion that what a non-scientist said was the easiest way to address the problem. So, I'm used to it by now. wbfergus Talk 15:27, 18 September 2007 (UTC)
I think you're right. That's probably a pretty good way to say it, too. That could be a good basis for a draft. COGDEN 20:32, 18 September 2007 (UTC)

<undent> Good point that it's what the source is being used for rather than what the source is, it ties in with thoughts I've been trying to pull together. Primary material can be raw data, but of course it can also be an original synthesis. Thus something that's a secondary source becomes primary material when it's the subject of the article or of the part of an article. An example which includes no raw research is Icons of Evolution. This book is a secondary analysis of primary sources, in a synthesis which is notable only as primary material about the ideas it promotes. The book has been thoroughly analysed by various secondary sources which form the basis of the article.
Perhaps another way of looking at the issue is that editors have to be careful to summarise accurately both the primary material providing the facts that are the subject of the article, and secondary analysis, description or summaries of those facts. Particular care has to be taken to avoid inadvertently summarising primary material in a way that presents our own opinion or understanding about the facts, so it's advantageous to use secondary material which shows attributable analysis.
In my view the primary/secondary distinction is useful as a way of assessing material, and while some examples are useful for explanation, we don't want a prescriptive list of what fits in each category. .. dave souza, talk 20:21, 18 September 2007 (UTC)

I'm becoming rather concerned with the view, largely espoused by Cogden, but seemingly being repeated by others now, that sources that produce analysis based on primary source/materials/whatever are, in themselves, primary sources for that analysis. That is not how the terms primary/secondary sources are used at all. Presenting new analysis, synthesis, commentary, etc etc etc is exactly what secondary sources do, and they are still secondary sources. Don't forget that these terms originate in history/historiography. SamBC(talk) 20:35, 18 September 2007 (UTC)

"Primary sources are documents or people very close to the situation being written about." When the document is an analysis (in the case above an unreliable one) it's a primary source for the situation being written about. In history, The Origin of Species is a primary source for Darwin's ideas and his analysis of the sources he sets out in that book. As I understand it... ... dave souza, talk 21:35, 18 September 2007 (UTC)
"Origin of Species" is perhaps a bad example, as it's complicated by a number of considerations; that might also make it a good example if we look at it closely. If you are studying Darwin, or the hitory of the theory of evolution, then it's a primary source. If you're studying evolution itself it's a secondary source that is also the first publication of much of its own primary data (all the sketches and so on). Its usefulnesss as a source in that case is, however, somewhat compromised by the fact that it's utterly out of date. In any case, the interpretation of basic facts, analsysis and commentary thereof, does not constitute a primary source. It's secondary material (or a secondary source, or whatever) because that's what a secondary source is. A document that largely reports and compares, dispassionately, the content of various secondary sources is, I believe, described as a tertiary source. SamBC(talk) 22:14, 18 September 2007 (UTC)
Sam and Dave... this is exactly why I came to the conclusion that we need to completely stop addressing this issue from the point of view of what kind of source/material is being used, and shift to what kind of statement you are using the source for. If you start with classifying the statement (statement of fact, statement of opinion, statement of conclusion/analysis/interpretation) it no longer matters what classification of source you are using. Primary, secondary and tertiary becomes irrelevant. The policy simply states that a statement has to be backed by a reliable source containing the same sort of statement.Blueboar 22:29, 18 September 2007 (UTC)
Or indeed, to avoid lawyering abuse, that contains the actual same statement, albeit in other words, or written longer, or so on and so forth. The main point about sourcing such analysis and conclusion is that we can cut the verbiage and reasoning, and just report the overall view, unless the reasoning itself is of interest. I was trying to make the points I made, however, more to try and correct some misunderstandings people seem to have. They aren't entirely relevant anymore, or shouldn't be, so I apologise. SamBC(talk) 22:37, 18 September 2007 (UTC)
Blue Boar: No, the differences don't diappear. The difference is larger in historical articles, for which the terminology was invented; but even in Darwin, there is a difference between the same statement in the Origin of Species and in a modern biology textbook. Darwin was making an argument (now largely accepted); the textbook is reporting consensus among biologists. This is why secondary sources are valuable: Darwin got several things wrong. Septentrionalis PMAnderson 22:45, 18 September 2007 (UTC)
OK, the differences don't disappear... my point was that it no longer matters as far as editing this policy is concerned. If we don't talk about the different types of sources in the policy anymore, we don't have to argue over our definitions of the terms (at least not on this page). Blueboar 22:54, 18 September 2007 (UTC)
It's significant in that sources very close to the subject don't give the analysis or opinion that make "secondary sources" valuable, since we must avoid doing that ourselves. As you say, we can avoid this classification problem by defining it by use – WP:V#Self-published and questionable sources in articles about themselves effectively applies to even the most reputable sources when they're the subject of the article, and in that case the aim is to find reliable sources summarising or commenting on the source. This is what I was trying to point out about Icons of Evolution, which happens to be very unreliable and so although it's a "secondary source" in that it analyses other "primary sources", to us it's only useful as a source of facts about the philosophy it presents, and for analysis of that philosophy we need to turn to reputable expert opinion on the book. ... dave souza, talk 13:26, 19 September 2007 (UTC)

Perhaps something like this

OK... on the theory of puting my money where my mouth is... here is a very rough draft of a short, concise... use oriented paragraph:

Citing the Right Source Materials

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 should take care not to confuse a statement of fact with one of interpretation analysis or conclusion. To say that "The Constitution of the US guarantees the right to own a sub-machine gun" sounds like a statement of fact, but is in fact a statement of interpretation (the interpretation being that The 2nd Amendment applies to sub-machine guns). Thus we should not cite the Constitution itself for this statement, instead we should cite an article by a constitutional scholar that contains this interpretation.


comments and criticism

The example may not be the best... but it was worth a try. In any case, is this more in the right direction ... or am I off base again? :>) Blueboar 20:23, 18 September 2007 (UTC)

I think this is in the right general direction. I especially like the first four sentences. I'm not sure what to make of the distinction between fact and interpretation, however, because sometimes it gets a bit metaphysical. Plus, I don't think we need to make that distinction. Both facts and interpretations are not allowed if they originate with the editor. If that's true, it's not really important for the editor to know the difference, other than the fact that interpretations are derived from facts.
But I agree with the intent of the last part of the draft. I think we can do it without explicitly distinguishing between fact and interpretation. Maybe we can say something like this:
Editors cannot include their own interpretations of previously-published facts, unless that interpretation is either an obvious and non-controversial consequence of the facts, or it 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.
For example, that the Second Amendment to the United States Constitution includes a 'right of the people to keep and bear arms' is a verifiable fact that may be cited in a Wikipedia article. I might also wish to also say that "the Second Amendment may protect a person's right to carry a sub-machine gun into a courthouse". However, if this statement is derived from my own observation, the statement is impermissible unless that conclusion can be found in a reliable source. The Second Amendment itself cannot be the source for my statement, since it is not an obvious or non-controversial consequence of the Second Amendment.
COGDEN 21:01, 18 September 2007 (UTC)
I really like the first paragraph, as it very succinctly encompasses what I think we're all trying to say. However, I think this starts to move off-base in the second paragraph. A large part of the (quite vigorous) debate about the second amendment is whether the "right of the people to keep and bear arms" is heavily modified (and substantially changed) by the preceding clause "A well regulated militia being necessary to the security of a free State". (Personally, my interpretation says the right of the people is the right of the people is the right of the people. However, constitutional scholars and judges have long debated how this hashes out.) It's more of a good example of the dangers of quote-mining/cherry-picking, than a good example of raw use, since it would imply something that is decidedly not settled. Vassyana 21:37, 18 September 2007 (UTC)
Good point, Vassyana... another example then. Cogden, I am concerned with your "obvious and uncontrovercial" exception. There is bound to be a lot of disagreement over what constitutes "obvious and uncontrovercial"... again I go back to "sez who?" ... if an interpretation/analysis/conclusion is an "obvious and uncontrovercial consequence of the facts", there should be some reliable source that makes the connection and says so. I understand what you are trying to say, and I don't disagree... but I think we need stronger language, and perhaps an example, on this point. Blueboar 22:08, 18 September 2007 (UTC)
We can use different language than "obvious and uncontroversial". I just want to make sure the policy doesn't lead editors to be overly-literal, for example requiring a reference for the idea that "Joe was a Native American", when the reference only says "Joe was a Cherokee". We can use a different example, too. Yes, probably something less controversial than the 2nd Amendment. COGDEN 20:04, 19 September 2007 (UTC)

For a better example, how about:

For example, that Franklin D. Roosevelt survived polio and that he became president of the United States are both verifiable facts that may be cited in a Wikipedia article. The dates of each event are easily verified, and it would be perfectly admissible to say that he "survived polio, and later became president of the United States" without finding a source that specifically uses the word "later", or otherwise explicitly compares the dates. However, any statement about the effect of the disease on his career would require separate citation, since it is not an obvious or non-controversial consequence of these easily verified facts.

(If that's too U.S.-centric, feel free to substitute something else.) I think this has an advantage in indicating both the type of extrapolation that is allowed and the type that is not. -- Jmabel | Talk 23:18, 18 September 2007 (UTC)

Draft as of 9/20/07 (15:48)

Removed for clarity Spenny 17:24, 20 September 2007 (UTC)

Comments on 9/20 (15:48) draft of proposal

Ok... I think this is the most recent version of the proposal, incorporating those comments that seem to have consensus. I would ask that people not make major edits to it until they have been fully discussed and have reached some degree of consensus. If you think something should be added/changed/deleted from this... post it as a new suggested draft with a new date/time heading. The idea is to keep the different versions clear so people can compare them and comment on them. Blueboar 16:05, 20 September 2007 (UTC)

New draft, incorporating comments

Citing the Right Source Materials

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.

For example, that Franklin D. Roosevelt survived polio and that he became president of the United States are both verifiable facts that may be cited in a Wikipedia article. The dates of each event are easily verified, and it would be perfectly admissible to say that he "survived polio, and later became president of the United States" without finding a source that specifically uses the word "later", or otherwise explicitly compares the dates. However, any statement about the effect of the disease on his career would require separate citation, since it is not an obvious or non-controversial consequence of these easily verified facts.

Proposed addition to the draft, added by ..dave souza, talk 12:19, 20 September 2007 (UTC) – see discussion at #Proposed addition to the draft below... dave souza, talk 12:40, 20 September 2007 (UTC)
Dave, I have taken your addition out... all it does is re-add the exact same material that everyone has been objecting to. The point of the new approach is to NOT talk about the different types of sources. Blueboar 12:59, 20 September 2007 (UTC)
comments on this draft

I like this... the FDR example is a great help, although I know that there will still be people who wikilawyer this exception to death. Blueboar 23:35, 18 September 2007 (UTC)

I could support this. It clearly illustrates the point. Vassyana 02:08, 19 September 2007 (UTC)

I could support this. It clearly illustrates the point. WAS 4.250 08:26, 19 September 2007 (UTC)

Good. Very good. However, is "editors cannot" the best description? We know that editors can, until someone reverts them. Would "editors cannot," "editors should not," "editors must not," or something else work best? Jacob Haller 04:51, 19 September 2007 (UTC)

Agree. "Must not" is the word I would suggest (cannot is famously ambiguous). We understand the development evolution and we understand that there must be a pragmatic interpretation of the rules in this context: I think that is the "should" exception. It is at this point where we step back and say that if there are issues caused during article development, then we have other mechanisms to deal with that, including "under construction" banners which give the clear indication that rules do not yet apply.
Would have preferred an example based on Winston Churchill myself :) Seriously though, it is clear enough that we do not need a USA frame of reference to understand the point, so I have no concerns.
We now need to bite the bullet, assuming we have a consensus starting to form. This proposal really walks away from source typing - indeed the title no longer fits the words, and it strikes me that it is really just some really good words to put in the policy and the sources section withers away (I am quite happy to see a link to the discussion essay). To risk being contentious, can we agree that the synthesis discussion becomes redundant in this context - it is essentially the final sentence of the second paragraph.
I think what may be helpful is to take a copy of the main page and put it together so people can judge the page as a whole before implementing it on a trial basis. Spenny 08:22, 19 September 2007 (UTC)
As a rule I am opposed to examples of how to apply/enforce the policy. In my experience they invariably become sources of confusion, or invitations for people to add other examples that end up making it confusing. here is why: a policy has to be clear, but general enough to apply to situations we cannot foresee. We have to rely on the good sense of editors acting in good faith to apply the spirit as well as the letter of the policy when controvery arises (when there is no controversy people seldom refer to poliies and that is okay). A policy will end up being applied in various ways. It is inevitable that many times people will have to apply it in a way different from the example. When that hapens, some may find the example confusing and change it with what they think is a better example, or think it incomplete and add another example .... after a while the policy will be overwrought and people will need to spend weeks rewriting it again. I propose instead that we try to be a simple and direct as possible and perhaps add that the actual application depends on the circumstances and editors need to use good sense. Slrubenstein | Talk 08:37, 19 September 2007 (UTC)
In the specific case of distinguishing between data/fact and conclusion/analysis and some people's concerns that our editors will make claims that are the one while citing a source that in fact contains the other one; I think that providing an example is very useful and should be done. WAS 4.250 08:48, 19 September 2007 (UTC)
In general, I like the draft of this section. However, I'm also torn between the other points made above. Spenny makes good points that I have been arguing for, for awhile now. The most difficult for me is the example. Like Slrubenstein, I think that just having one example, no matter how good, will eventually lead to certain cases in the future where somebody will add another example, then another, etc. However, I wouldn't be oppsed to having the example if there was also some way we could gaurantee that future examples wouldn't be included here, but on another page instead, like one that specifically details the various types of sources/materials and their proper usage in Wikipedia articles. These changes make the policy much clearer and by far easier to understand, but as Spenny stated, having these areas (sources/materials and their usage and synthesis) on separate pages will be a lot of work. Personally, I'd rather see them as offical guidelines appropriately linked from the other policies, but then to make them official guidelines for the the other policies will probably increase the difficulty in creating the them, as more people will have comments on what should be included, or the wording, etc. In the long run though, it will make for easier interpretation of the 'rules' (both policy as guidelines), by having everything more standardized, consistent, and modular. A change in one area won't neccessarily mean a change is also need in the other areas (like how each policy has it's own variation of the source issue). If the 'interested' editors from the other policies all agree on a change to the sources page, then the change to the sources page can be changed without similar changes being (possibly) made on the three main policies, so they are all 'in sync'. So, to wrap up, I feel that is good progress in the right direction, but there are still a couple of stumbling stones in our path to be cleared up, and it will be a long and probably contentious process to overcome, though ultimately it should succeed and make life in the future much easier, it's just going to be difficult getting there. wbfergus Talk 10:39, 19 September 2007 (UTC)
I agree that examples can pose problems, but they can work in policy pages, if done properly. A good non-Wikipedia example of where examples like this work effectively is in the American Restatements of Law, which look at country-wide caselaw and derive basic rules (restatements), and then follow them up with several model examples applying the rules. The examples are usually taken from actual caselaw. So if there's an issue with the FDR example (which I think is good, but maybe there's not a consensus on that), perhaps we can find an actual example where this issue has arisen on a Wikipedia page, and been resolved by a consensus effectively applying these principles, leading to a stable result. COGDEN 20:27, 19 September 2007 (UTC)

See, here's why sourcing is so important, and why examples in policy are dangerous: FDR never had polio. See Franklin D. Roosevelt's paralytic illness. – Quadell (talk) (random) 12:07, 19 September 2007 (UTC)

I agree with the draft, with a couple of tweaks: We can call it a "paralytic illness" instead of polio, and there's a minor grammar issue. I think the example is a good one. I know examples are difficult in policy pages, but I think one would be very helpful here. Model examples can work effectively to elucidate policy (I'm thinking about the American Restatements of the Law as a non-Wikipedia example). COGDEN 20:19, 19 September 2007 (UTC)

Re: Spenny's question on biting the bullet and deleting the synthesis section... No... for two reasons:
1) Dropping the "Primary, Secondary and Tertiary Sources" section, and replacing it with this new approach is going to be something of a surprise to many editors. It may even be contentious. We may have reached agreement and concensus, but that does not equate to the broader Wikipedia community agreeing with us (for an example of how policy page editors can have consensus, only to find that the broader community does not agree... see the WP:ATT proposal). I really think we need to go slow and take things one step at a time here. Let's get our current proposal set and approved before we move on to other things.
2) I dissagree with cutting it. Synthesis is a discrete form of OR, one that Jimbo has singled out as being something to watch out for (in various comments). it is also a very complicated issue. As such It needs to be discussed seperately in the policy. We can work on the wording if people have issues with it, but the concept needs to be discussed seperately. Blueboar 12:37, 19 September 2007 (UTC)
I am content to go one step at a time. I thought it would be useful to test the water on synthesis, but am realistic. I would take issue with you on synthesis being a discrete form. I am content with it being a special case, but in the end it is that same fundamental problem: adding in an opinion that is not supported in the source documents - and it encourages technical sounding complaints. Spenny 12:50, 19 September 2007 (UTC)
Agreed... "Special case" is a better way to put it.
As to the proposed change... Does anyone disagree with the basic thrust of what we are proposing: Replacing the "Primary, Secondary, and Tertiary Sources" section with "Recognising the Right Source Materials" (understanding that we probably need to work on the exact wording some more)?
If not, I think it is time to start advertising the proposal... posting at the Village Pump and notifying the larger community. We can expect some resistance to the proposal... something this fundamental is bound to generate resistance from those who are used to the current language. We will need to convice the broader community that change in and of itself is not a bad thing, and that our proposed change is for the better. Blueboar 13:13, 19 September 2007 (UTC)
I don't disagree. However, when the proposal is 'advertised', I would state that what this policy is doing with the current 'Sources' section and this proposed version, may be the first step in also getting similar changes done on WP:NPOV, WP:V, and possibly WP:BLP as well, so the main policies can become more standardized in this area, using common criteria, definitions, and examples, eventually leading to less confusion in interpreting the policies and their use. Add something else about creating a "guideline" for the source issues (probably incorporating WP:RS, which is still somehow different from WP:BLP#Reliable_sources, which in turn is linked from WP:V). This would help highlight the current convolution of the various policies referring back and forth to different 'things', whther they are the same in context or not, it leads to a lot of confusion. If they all linked to the same "guideline", much of the confusion of interpretation would be alleviated. wbfergus Talk 13:38, 19 September 2007 (UTC)
I'm on board. We can deal with any possible redundancies in the Synthesis section later. COGDEN 20:41, 19 September 2007 (UTC)
D'oh...Too late! - I was BOLD and already "advertised" the proposal at the Village Pump (Policy)... you might want to add your comments after mine. Blueboar 13:46, 19 September 2007 (UTC)
LOL... Okay, I'll head over there and make the comment. Thanks for being bold. wbfergus Talk 14:04, 19 September 2007 (UTC)
Looks good. Thanks again for your good work, Blueboar. Suggestions re the polio problem pointed out by Quadell: First of all, I suggest never stating in a policy that "X is a verifiable fact" or making statements of any sort of fact. This can be fixed by saying "Suppose it is a verifiable fact that ...". Inserting "suppose" in no way detracts from the value of the example in my opinion. Secondly, since a problem has been pointed out with this particular example, I think it's better to fix it in addition to inserting "suppose". Possibilities include "polio or a polio-like illness", or "a paralytic illness, long thought to be polio" or "a paralytic illness". If it weren't for that detail, it's a really excellent example in my opinion. Another way to fix it could be to make it anonymous: Suppose it is verifable fact that someone named Smith survived polio, and also that the person became president of the United States, and that the dates of each event are easily verifiable. It would be perfectly admissible to say that Smith "survived polio, and later became president of the United States" without finding a source that specifically uses the word "later", or otherwise explicitly compares the dates. However, any statement about the effect of the disease on Smith's career would require separate citation, since it is not an obvious or non-controversial consequence of these easily verified facts. I'm not sure whether one would also have to insert something like "and suppose the date of the polio is earlier than the date of becoming president"; I think that could just be left as implied. --Coppertwig 16:30, 19 September 2007 (UTC)
Good suggestions. I like the idea of saying "suppose...", and referring to "a paralytic illness". COGDEN 20:41, 19 September 2007 (UTC)
How about "Sticking to the sources" as the title? It strikes me that what we are trying to encourage people to do is not so much to go and find the "right" sources to footnote their ideas, but rather not to introduce ones own ideas in the first place if they go beyond the sources currently available. Also, I agree that examples can be helpful and support their use. Policy documents can suffer from the disease of being so precise and general as to be completely incomprehensible; examples help avoid that disease. Best, --Shirahadasha 21:47, 19 September 2007 (UTC)
Agree that sounds like a better title – this is about "care should be taken not to "go beyond" the sources or use them in novel ways", and restates that well. Minor point: "2) can be attributed to a reliable source." should surely read "is attributed...." As to what this replaces, the next section discusses that.. .. dave souza, talk 09:46, 20 September 2007 (UTC)
  • Approve in principle. I've sat out this one but noticed that it's getting near a consensus. The change is useful and it looks good, whichever exact version and wording people settle on. Once it's implemented we can review WP:RS to make sure it agrees. Question - is this what became of the earlier primary/secondary/tertiary sources discussion and is it supposed to replace or complement that part of the policy page? If we're getting rid of those definitions we ought to put them somewhere else. Wikidemo 02:04, 20 September 2007 (UTC)
  • As far as I can see, this is essentially a clarification of "care should be taken not to "go beyond" the sources or use them in novel ways", and appears to be intended to replace much of the Primary, secondary, and tertiary sources section. The "Research that consists of collecting and organizing material from existing sources within the provisions of this and other content policies is encouraged.." bit has been lost, this is often cited when arguing against over-restrictive interpretations. As I've commented at #Cutting to the chase above, the earlier primary/secondary/tertiary sources view has advantages when considering whether sources very close to the subject can give useful analysis or opinion, and as you note these definitions are needed, not least to minimise disruption to the understanding of editors familiar with the policy. A third paragraph could concisely discuss the point, possibly under a subheading to ease navigation. .. dave souza, talk 10:04, 20 September 2007 (UTC)
What about using the proposal I made, above, concerning sources, followed by this clarification? Slrubenstein | Talk 10:34, 20 September 2007 (UTC)
The point of my reworking of your suggestion was to show that it was possible to address these issues without tying ourselves in knots over classification. I do share your concern that we have drifted too far from understanding that there are different qualities to the sources that editors should be conscious of (but not governed by). Pragmatically, summaries and reviews are most likely to produce the sorts of information that we can readily use to write a summary style article without being prone to our own interpretations. We can discuss this without classification of sources. I think that by removing that red herring we are seeing discussion here which is about clarity of NOR which is healthy. Spenny 11:00, 20 September 2007 (UTC)

Proposed addition to the draft

As discussed above, it seems to me that the draft would work well with a subsection concisely explaining the usefulness of primary/secondary/tertiary sources. I've based this proposal (as shown at #Primary, secondary, and tertiary sources above) on the existing policy with some changes to bring it into line with the draft, and the examples etc. commented out. Slrubenstein's proposal may form a better basis for this subsection, but it seemed to me to be useful to put a proposal up for discussion about the principle. ... dave souza, talk 12:24, 20 September 2007 (UTC).. add link 12:38, 20 September 2007 (UTC)

Dave, I get what you are saying, and I see how such a subsection would be useful ... but I am not sure it can be done. In the new approach we are trying to avoid using the terms "primary/secondary/tertiary source"... because they cause so much confusion. Different accademic disciplines have different criteria for what these terms mean. Wouldn't such a subsection simply re-add the same problems the current proposals now solve? I would have to see some suggested language to get a good sense of what you are proposing... but I am sceptical. Blueboar 12:55, 20 September 2007 (UTC)
Addendum... I now see that you did give a draft... by ammending Spenny's draft above. Unfortunately, all that you did was to paste in the same material that we have been trying to move away from. The whole point of our new approach is to NOT include a definition of what the different types of sources are, since that seems to be the crux of the confusion and objections. Blueboar 13:03, 20 September 2007 (UTC)
I think there is a middle ground here and I'd ask Dave to take a look at the section Blueboar's re-draft. The point of that was to see if we could get the sense of policy without the contentious terminology. My feeling was that there was something to work with, and that by taking such an approach it brought out other issues (hence the emboldening). Spenny 13:08, 20 September 2007 (UTC)

We are moving fact and furious folks... and I am a bit confused as to which draft we are now discussing... would someone please post a "proposed" version that incorporates the latest thinking? (perhaps we should entitle them with: Proposed draft as of date/time and periodically update?) Blueboar 13:58, 20 September 2007 (UTC)

Excellent idea. I was thinking somewhat obliquely along the same lines. It would help anybody coming here from the Village pump to easily see what is being discussed, without having to read through mountains of other 'stuff'. wbfergus Talk 14:09, 20 September 2007 (UTC)

I'm opposed to any changes to the long-standing formulation of this key policy. The proposed changes open the door to a long and slippery slope, and for that reason alone I'd objection to their addition, much less their specifics. I appreciate the effort here, but once we start making exceptions, it will never end. FeloniousMonk 15:30, 20 September 2007 (UTC)

I don't think posts that allude to some undefined fear uncertainty and doubt are very constructive, perhaps you could be more explicit on your concerns. The position that some accept here is that the source typing is not relevant to policy, is contentious and is not required to support a robust policy of NOR. The discussion has been persuasive enough that some have changed their position so I think it is reasonable that you should be explicit in your reasoning as to why you do not accept this as we cannot debate FUD. Spenny 15:41, 20 September 2007 (UTC)
Monk... Are you objecting to change simply because it is change... or do you see a flaw in our approach? How does the new version "open the door to a long and slippery slope"?... what "exceptions" do you see us making? To most of us it simply restates what the policy has always said in a clearer less confusing form. Can you give us a better idea of what you are worried will happen if we adopt the proposal? Blueboar 15:43, 20 September 2007 (UTC)
Monk, the problem is that the present formulation does not have a consensus. The question is not about whether we should change it. I think everybody has moved past that. The issue is what should we change it to, so that it will reflect longstanding Wikipedia practice, and everybody will be happy with it. This has been a good opportunity to review what the underlying principles of this policy are, and how to best express them. I think most people here, if presented with a questionable article, would agree on whether or not it is original research in most cases. We are just trying to find the best way to express that consensus, but the devil is in the details. COGDEN 20:34, 20 September 2007 (UTC)
I agree that the WP:PSTS section in this policy is confusing, contentious, unnecessary, misplaced, and not supported by consensus, so it must go. The real issue boils down to making sure facts are backed by factual sources, preferably original or close sources, and that all analysis, interpretation, etc. use reliable sources to do the analysis interpretation, etc. See, I said it without using primary, secondary, tertiary, or even quaternary. Dhaluza 01:56, 21 September 2007 (UTC)
Dhaluza, I am going to quibble with your comment... sometimes an original or close source is appropiate... sometimes it is not. By focusing on the source you confuse the issue. This policy isn't about sources (or it shouldn't be)... it should be about statements made in wikipedia articles. When there is an NOR problem in an article, it is the statements that are OR, not the sources. Blueboar 02:05, 21 September 2007 (UTC)
I think we are basically saying the same thing in a different way. I agree that focusing on the source has been a distraction, and recognizing that is a breakthrough. I think I was focusing on the use of the source to back the statement, but distinguishing between factual and non-factual statements. The statements are only OR if they are not backed by reliable sources. Also, when is an original or close source not preferable for factual statements (other than when they are corrected by newer sources)? Dhaluza 11:04, 21 September 2007 (UTC)
When it is incomplete, tendentious, inaccurate, or out of context, which (in historical articles, where this distinction is most commonly made) it usually is. The chief problem with primary sources is that they are not intended to be history; quoting them directly and uncritically, as most editors would do even if WP:SYNTH didn't exist, produces very bad history. Septentrionalis PMAnderson 16:24, 21 September 2007 (UTC)