Wikipedia talk:Primary Secondary and Tertiary Sources

This article describes both a proposed policy change to the discussion of PSTS in the NOR policy page plus the proposed move of some of the existing content to a guidelines page where there can be more examples and clarification.

The following discussion is copied from the archives of the talk pages for WP:NOR to give a history of the discussion regarding this issue and the recent discussion leading to the proposal to deal with PSTS mostly through a guideline and with a more limited version in policy, emphasizing the avoidance of SYTN.--SaraNoon (talk) 20:50, 24 August 2008 (UTC)

Request left on my talk page
SaraNoon left on my talk page a request to review this page. In response to that request, I've had a chance to look over the current state of this project page in its current form. as distinguished from where it was roughly nine or ten months ago. My brief assessment based on a brief review of it, is that it is a very useful contribution to the discussion about WP:PSTS. Taken by itself, however, it appears to me to take a somewhat contra view of PSTS, and doesn't factor in the numerous and widely distributed variables across the wiki that were factors in arriving at the current form of the PSTS section as presented on the WP:NOR policy page. That is to say, at present, it's not self-standing, but rather is to a very perceptible extent a reaction to aspects of the current PSTS section that are seen as problematic by WP users who are advocating this version as a replacement. ... You have my interest because to me this page has something valuable to say about the PSTS debate. How much time I'll be able to devote to the discussion is dependent on factors presently beyond my direct control, though. Kudos to those involved for the thought that quite clearly was put into arriving at the present version of this project page. IMO, at present the most valuable potential modification to PSTS is in the contention that the delineations between primary, secondary and tertiary shouldn't, we might say, be taken too strictly and that like other editorial parameters in WP, are readily subject to debate and to the consensus process in general. ... SaraNoon, thanks for the note-- this remains an interesting debate about this WP policy issue. ... Kenosis (talk) 01:29, 8 September 2008 (UTC)
 * I got the same note and 100% agree with Kenosis. This is not a general inclusive look at what primary and secondary sources are and how they are best used but is clearly a position paper regarding current debate within Wikipedia.  For example the discussion of how all sources on current events are more primary than future sources will be is omitted.  Ten years from now every source about the United States presidential election, 2008 currently used will be seen as a primary sources compared to the sources that are written with the perspective of knowing who the winner was or what turned out to be the key events of their presidency.  Imagine if we currently tried to use any sources written during the election of George H.W. Bush about his pledge of Read my lips: no new taxes in the article on the United States presidential election, 1988 as secondary sources when they do not even have the proper perspective of the fact that the pledge was broken during his presidency?  PSTS is solely about perspective, and it is entirely a relative judgement.  This is page is written as if these assessments are more conclusive and permanent than they truly are.  Original research is completely independent of the perspective of the sources. While it is more common to see honest mistakes about OR made when using sources with a closer perspective, this is not an inherent feature of the more primary source. Rather it is an interplay between the goal of a NPOV encyclopedia to take a very distant perspective and common human fallibility of trying to force a round peg in a square hole because a source with a much nearer perspective is most convenient.  I would not support this as policy.-- Birgitte  SB  02:16, 8 September 2008 (UTC)


 * NOR policy is supposed to be about NOR. It is not supposed to be about defining PSTS.  The problem is that PSTS is relevant to discussing NOR, and the current "policy" on PSTS is oft abused by wikilawyers who miss the intent of NOR and try to use the discussion of PSTS in NOR policy to block inclusion of relevant material by arguing about PSTS instead of the material itself. Unless I'm misreading them, it appears that both Kenosis and Birgitte agree that the definition of PSTS (as currently in NOR) is less than fully precise and is subject to confusion, debate, misuse, etc.  I can't say I've seen anyone deny that this is the case.  My goal, and recommendation, therefore is to deemphasize PSTS as a criteria for judging OR and instead encourage education on PSTS via a guideline so that it can be a matter for discussion among editors instead of a excuse for "my secondary source trumps your (what I consider, and you deny) primary source."  The perfect example of this is when people try to use a newspaper article about a scientific study, as a secondary source, but then try to deny others the ability to quote from the conclusions of the peer reviewed study (arguably secondary source material in a primary publication).  It's that kind of abuse of this policy that I am trying to address.--SaraNoon (talk) 17:05, 8 September 2008 (UTC)
 * I disagree that PSTS is necessarily relevant to discussing NOR. It does not clarify NOR, but actually leads to more confusion and misunderstanding than pure treatment of Original Research would do.  Original Research is not appropriate when using any sort of source. PSTS is simply a shorthand oversimplification used to talk to the less clueful people about NOR. Like telling a kid that gravity  means that what goes up must come down and ignoring the fact that if you have enough velocity to escape earth's gravity then this is not true. Some people believe discussiong PSTS is a useful way to deal with editors breaching NOR, but as you describe above this simply leads to a different set of problems.  I would say a worse set of problems.  But given that it is unlikely that we will have consensus to remove the discussion of PSTS, it would still be worse to supplement the oversimplification at NOR with the expanded, detailed, and authoritative sounding oversimplification written here.  It is much harder to teach people that PSTS is poor characterization that sometimes does not hold up to scrutiny when it has it's own entire policy page.  So an expanded PSTS policy I might support would probably start be along the lines of "The description of cardinal sources at WP:NOR is a simplified rule of thumb useful to novice editors.  It does not represent a complete understanding of the complex issue of using the best sources available."-- Birgitte  SB  17:56, 8 September 2008 (UTC)

I too was notified by SaraNoon. I am Opposed to this "proposed Wikipedia policy, guideline" in its current form and will probably remain opposed as I think the current WP:PSTS is detailed enough. Some people arguing on these pages clearly do not appreciate how often Primary sources are of use and which might be in danger if the current wording was to be implemented. Here are three examples: --Philip Baird Shearer (talk) 17:12, 8 September 2008 (UTC)
 * See how the European Court of Human Rights judegement on Jorgic v. Germany  is used in the Bosnian Genocide article. Is it a primary source or a secondary source? By quoting it directly about the Bosnian Genocide case, we can side step that issue, but its analysis of the arguments for and against a general Bosnian Genocide are fundamental to analysis of the article.
 * A case were the primary source may be more desirable than just relying on secondary sources was high-lighted during the editing of the Bombing of Dresden in World War II in 2006 (a controversial subject if ever there was one). By the selective extraction of different parts from a primary source by secondary source authors, those authors of the secondary source could use it to to reinforce their POV. From the same RAF Group briefing paper:
 * Dresden, the seventh largest city in Germany and not much smaller than Manchester, is also [by] far the largest unbombed built-up area the enemy has got. In the midst of winter, with refugees pouring westwards and troops to be rested, roofs are at a premium, not only to give shelter... but to house the administrative services displaced from other areas... The intentions of the attack are to hit the enemy where he will feel it most...and to show the Russians when they arrive what Bomber Command can do. ( Michael Zezima From Dresden to Baghdad 58 Years of "Shock and Awe")
 * Second example from Norman Longmante in The Bombers page 333:
 * Dresden, the seventh largest city in Germany ... is also far the largest unbombed built-up area the enemy has got ... At one time well known for its china, Dresden has developed into a industrial city of first-class importance and ... is of major value for controlling the defence of that part of the front now threatened by marshal Koniev's breakthrough. ... The intentions of the attack are to hit the enemy where he will feel it most ... to prevent the use of the city in the way of a further advance, and incidentally to show the Russians when they arrive what Bomber Command can do.
 * So using a secondary source that quotes a primary source may not necessarily avoid the problems of a biased POV even if experts have the resources required for interpretation. (The full text is available in Churchill by Samantha Heywood pp. 118,119)
 * A third example the International Court of Justice use different English spelling in different cases. More recent cases seem ot follow the same principles as Wikipedia. Those which involve the US use US spelling while those involving the Commonwealth use Commonwealth English, but this is not always the case particularly with the older cases. One can not just look at law books because they tend to us the spelling of the country they are published in. So for the naming of articles and the English to use in them it is easiest to look at the original court judegments (primary sources) to decide on which spellings to use for a particular judgement to meet the content policies of Wikipedia (see List of International Court of Justice cases)
 * Philip, I support the use of primary sources when used appropriately. I encourage you to look at the proposed policy to tell us more specifically how it would obstruct your concerns.  The examples you raise might well be positively addressed in the proposed new guidelines.
 * For example, doesn't the following proposed policy statement satisfy your concern?
 * It is therefore important to remember that, according to policy, primary, secondary, and tertiary sources may all be acceptable if used appropriately. Therefore, how material is classified is far less important than making certain that the material cited from the source is accurately described without inserting interpretations which are not specifically present in the cited source. That is the essence of the "No Original Research" policy.

--SaraNoon (talk) 18:25, 8 September 2008 (UTC)


 * Yes, I agree with your first paragraph. No to the last, because with that one may as well throw out the prohibitions on using primary sources in WP:PSTS. The reasons for the prohibitions on primary sources is to stop OR on Wikipedia. A good example of where the current rule has worked well this is River Teme and navigation, where two opposing views have been arguing the point over a number of years. Some of the exchanges on the talk page are perfect examples of the WP:PSTS policy working as it should. I would be interested to see an example of an article where WP:PSTS in its current form is an impediment to the development of an article. --Philip Baird Shearer (talk) 08:14, 9 September 2008 (UTC)


 * I briefly looked at the River Teme discussion as you suggested. It did not appear to be as contentious of an issue as disputes over primary source vs secondary source occur in more contentious issues which draw on scientific studies, such as homosexuality, pornography, evolution, global warming, abortion, evolution, AIDs, chastity education, birth control pill, etc.  In these areas, it is not uncommon for one side or the other to rigidly insist on the right to delete both the material and citations to peer reviewed studies (decried as primary sources), even if the material cited is precisely just a properly attributed quote to the authors.  In such cases, there is no question of it being original research, but the policy regarding secondary vs primary sources in NOR is being misused to block inclusion of material and opinions (the peer reviewed opinions of experts in the field) because one side of a contentious issue wants to mute those opinions.
 * I am by profession an academic and scientist who knows the difference between an original interpretation and simply objectively summarizing the findings of A or B.  Like many academics with access to many relevent journal articles in a field, I've found that when I try to simply expand on an article's bibliography and add citations and brief summaries, even those easily confirmed by looking at an online abstract, if I attempt to add the material to a contentious article it will get pounced on and deleted and specious arguments against use of "primary sources" will be raised.  This is very frustrating and discouraging to people with expertise in a field who want to contribute to Wikipedia but are driven away by "blockers" of content.
 * My editing preference is toward inclusion of as many citations as editors from both sides of a contentious article believe are necessary to document "their side." Even if the citation is only briefly discussed or is blended with similar citations that make the same point, the inclusion of citations adds tremendous value to articles in that they provide those starting down the research path with an better bibliography to start their studies. But as stated, some POV-pushers are so anxious to push their POV that they want to deny readers access to a bibliography of citations which includes research they "don't believe is credible" -- even if published in noteworthy journals!  So that is my gripe.  I'd like to see a policy that you can never delete a verified citation (even if you cut nearly all the text), but that is unlikely.  So I am starting with an effort to just reduce the foothold of the wikilawyers who are abusing PSTS as a means of deleting material without actually bothering to discuss how it can be included and given proper weight.--SaraNoon (talk) 01:38, 10 September 2008 (UTC)
 * I chose commercial navigation the Teme river article as an example precisely because for those involved it is a contentious issue, but it is unlikely to be an issue that excites much controversy here. For those deeply involved this issue the only recourse is visits to the local libraries to look up unpublished records. The types of unpublished papers are described in more detail on the articles talk page in the section Commercial navigation. If it were not for NOR I suspect that the proponents of the two view would by now have created an article of very original research. I don't full understand what you are saying above. This is probably because I deal with historical subjects and international law. The wording of PSTS seems clear enough to me and few people who argue about controversial issues on articles I watch -- such as Historical revisionism (negationism), Unlawful combatants, Genocide, Bombing of Dresden, or even Depleted uranium (my interest is the section legal status in weapons) -- seem to disagree over what is a primary and what is a secondary source. Can you give an example in an article history and the talk page of the an article where you "add the material to a contentious article it will get pounced on and deleted and specious arguments against use of "primary sources" will be raised." --Philip Baird Shearer (talk) 12:50, 10 September 2008 (UTC)

Make Sure Theory is Correct & Policy Lead In to Guideline
I generally agree with Kenosis, Birgitte, and Philip Baird Shearer above, but I think an article such as this could be valuable, and it has some good parts now that are helpful. As a beginning, though, we must make sure that the theory is correct and well-supported by non-Wikipedia sources. There is a great deal of confusion in terminology within Wikipedia, and people have been talking past each other, and as a result, some of the normative parts are likely to confuse things. For example, while the statement that "interpretations" may only be supported by secondary sources might be absolutely true in the sense commonly used by experts outside Wikipedia, it may not be true under a lot of definitions floating around Wikipedia, which are often based on quick-and-dirty definitions that obscure the concepts of primariness and secondariness.

For example, if I am a casual Wikipedian who is not well acquainted with source classification theory outside of Wikipedia, I might look at some list of typical "primary sources" and see "peer-reviewed journal" listed there, and assume that I cannot cite a peer reviewed journal for an interpretation. Yet, under non-Wikipedia scholarly standards, the peer reviewed journal is certainly primary--because it almost always presents new ideas and/or new data--but is also almost always secondary--because it interprets those new ideas and/or new data--and under the normative "secondary only" rule above would be entirely appropriate for citation. But I might not know that unless I understand the theoretical difference between primariness and secondariness. This article captures some of the complexity of these necessary determinations, but also mixes in some apparently-original research by Wikipedians that I don't think is well-supported by the literature in this field, and will only confuse people. A first step might be to take away the normative statements for now, and only keep statements that are verifiable, as if this article were a mainspace article.

Or, we can just rely on the primary source article, which has already found a lot of these sources, and is presumably NPOV, or on its way to becoming NPOV. Looking at this article for the first time in a while, I was surprised that a lot of good work has gone into it, and we ought to re-use as much as possible, rather than re-inventing the wheel. CO GDEN  02:00, 9 September 2008 (UTC)


 * I think the primary source article is very good. In my view, any and all such information can and should be included in Guidelines regarding PSTS.  At a guideline level, it is also great to try to flesh out the theory and make it correspond with the prevailing definitions in academia.


 * The crux of the problem and dispute, however, is clearly at the policy level, where any discussion can and should be brief yet at the same time clear enough that it helps to encourage article development in a way that minimized content disputes. But the current policy statement on PSTS fails in that regard and in some minor ways admits that it is inadequate for the purpose of defining primary and secondary sources in an objective manner that can actually serves as a binding principle for NOR.


 * Personally, I find the idea of NOR very easy to understand. My guess is that most other editors do as well.  But including this PSTS "standard" and the preference for secondary of primary sources into NOR, creates an "objective standard" (which is not really objective or standard across all disciplines) that becomes a focal point for creating rather than resolving conflicts over sources.


 * Such conflicts arise when I run into people who don't bother asking for evidence that my cited contribution doesn't include my own spin or original research, but instead insist that since I cited a peer reviewed medical journal that fact alone justifies their deletion of my contribution (and the cite), because science journal articles are "a primary source and not allowed," even when I bend over backward to merely quote the authors so as to ensure that I am not injecting any POV. Admittedly, this only happens in articles related to contentious issues, but those are also exactly the ones where policy should help to reduce the areas for conflict (by supporting inclusion of verifiable material) rather than provide footholds for wikilawyering effort to exclude material.


 * If the PSTS were not already a part of NOR, I think it would be better not to have included it and this potential confusion of classifying sources. But I don't know the history and I can see why it can be helpful to point people in the direction of understanding these classifications so they can be discussed as needed in sorting out both the quality and appropriate use of the sources, as in the River Temes article discussed above.  But more importantly, as this PSTS section has been in Wikipedia for a long time, ripping it completely out might do more damage than good and would clearly only ignite far more debate.  So I don't support removing it, either.


 * That said, however, I do think there is (a) wide recognition of the fact that the current PSTS policy cannot possibly define PST source in a brief fashion that is adequate and therefore opens the door to confusion, dispute, and edit warring which is counterproductive and (b) a growing consensus that some of these issues are better dealt with at a guideline level rather than a policy level.


 * If I am correct in (a) and (b) above, my goal is then to replace PSTS in policy with something briefer which deemphasizes editorial efforts to define sources as primary or secondary as the means of determining whether the material in question can be included. Instead, recognizing that both can be appropriately used, we should encourage editors who may benefit from and have a need to explore and discuss PSTS in working on an article to study a more expansive guideline linked to policy.  The advantage of a guideline is that while not binding, it should help to clarify the issues and keep people focused on the core issue of carefully reporting verifiable sources without original research. And also, because it is a non-binding guideline, it is makes it a bit harder for wikilawyers to play "policy" as a trump card for deleting material and helps to keep discussions more focused on the key question: is the source being accurately represented without the insertion of original research on the part of the contributor? --SaraNoon (talk) 02:14, 10 September 2008 (UTC)
 * SaraNoon, you write: "since I cited a peer reviewed medical journal that fact alone justifies their deletion of my contribution (and the cite), because science journal articles are "a primary source and not allowed,"" please could you provided the best archetypical case you can, because I can not see how such a line is argued under the current WP:PSTS by the person wishing to delete the fact unless it is more complicated than you indicate. --Philip Baird Shearer (talk) 12:59, 10 September 2008 (UTC)


 * I'm not saying they were justified in deleting the material according to the current policy, of course. This editor was a rather senior editor and I suspect knew better, but was wikilawyering to block material that undermined the POV he was trying to push.  Because it was a controversial issue, he was also backed by a number of like minded editors, and they were looking for any excuse to block the material.  Their claim boiled down to (1) science journal articles are primary sources, (2) secondary sources, such as magazine or newspaper articles synthesizing complex scientific issues, are preferred and even mandated by PSTS.  It was rubbish, but consumed weeks of debate which eventually yielded allowing me to cite one peer reviewed article only.


 * In short, the issue I'm addressing is abusive attempts to misuse this policy, not reasoned and balanced attempts. By deemphasizing PSTS I hope to help others like myself to at least have the peer reviewed science articles we wish to use as source material treated as well (and, hopefully better) than "secondary soruce" popular magazine articles.  In other works, I want to take a blunt a tool used by obstructionists, especially against newbie scholars who have solid material to contribute but will run into this kind of abuse of policy.--SaraNoon (talk) 22:38, 10 September 2008 (UTC)


 * Could you point us to the relevant article discussion page where the problem occurred? Jayen 466 23:37, 10 September 2008 (UTC)
 * I agree, it would be useful if you could link to an example where the policy was misused. I think you may be trying to change too much policy to solve one local difficulty. Colin°Talk 13:46, 3 October 2008 (UTC)

COGDEN suggested we review non-Wikipedia sources to ensure we aren't confusing things with our own OR into PSTS. Here are some wrt science and medicine:
 * Evolution of Scientific Information - Primary sources - Secondary sources - Tertiary sources Timken Science Library; Guide to Library Research in Science.
 * Primary, Secondary & Tertiary Sources James Cook University; Library Guides.
 * primary source - secondary source ODLIS — Online Dictionary for Library and Information Science
 * Primary source - Wikipedia; The Free Encyclopaedia (OK, that ours it would be odd if our policies disagreed with our articles!)
 * Types of nursing information University of Connecticut; University Libraries.
 * Guidelines for Assessing Professional Information Washington State University; Department of Veterinary Clinical Sciences.
 * Primary, Secondary and Tertiary Sources in the Health Sciences (PDF) University of Minnesota; Bio-Medical Library.

IMO, they align with the current definition at WP:NOR. I'm opposed to the suggested replacement text for WP:PRIMARY, and think the supplementary text is merely an essay at present and is so far off being a suitable guideline that officially "proposing" it is premature.

Sara mentions difficulties where editors have sought to replace statements cited to reliable peer-reviewed journal articles with statements cited to newspapers. This is an issue for WP:V and WP:RS and help with specific problems can be requested at WP:RSN. But please don't try to change the definition of primary and secondary such that all peer-reviewed literature is regarded as secondary. The idea that "Wikipedia articles should rely mainly on published reliable secondary sources" and that "Primary sources that have been published by a reliable source may be used in Wikipedia, but only with care" is not usually an issue of reliability or verifiability. The reason that PSTS is discussed at WP:NOR and not WP:V is that the primary/secondary distinction mostly affects how you can use the source, not that you can use the source. That said, one area where PSTS affects whether a fact from a primary source should be mentioned/cited is WP:WEIGHT -- quality secondary sources are your friend when it comes to that judgement (e.g. a literature review in a journal or textbook, not a newspaper article). Colin°Talk 13:46, 3 October 2008 (UTC)

Why is this page not marked Essay?
I'm concerned about why this page is marked as a proposed guideline. I now see (from What links here) that some editors are referring to this page even though it deviates from WP:NOR and WP:V (see Colin's post above), and doesn't have wide acceptance (these talk archives were copied over from NOR). Can we please get this page moved back to Saranoon's userspace, or marked essay? There seems to be some real misunderstanding (see Colin's post) here, yet "What links here" reveals that editors are using this page for guidance. Sandy Georgia (Talk) 18:28, 4 October 2008 (UTC)

secondary source material are the only acceptable source
This was rejected as an interpretation in a very looooooong discussion over WP:PSTS it came about over the wording of PSTS (See this edit on 28 November 2007, that was put in for clarification and also Wikipedia_talk:No original research/Archive 28 and Wikipedia_talk:No original research/Archive 29 the talk archives on this addition).

Also see this article Richard Debaufre Guyon where I recently added a comment about the man from a tertiary source: "The 1663 Chambers Encyclopaedia states "Indomitable courage, and an incessant care for the comfort of the troops under his command, were the chief features in Guyon's character"." -- this would also be out under this interpretation unless it is intended to get rid of tertiary as a definition. --Philip Baird Shearer (talk) 13:31, 10 September 2008 (UTC)


 * In fact that bit makes little sense at all. If the given source actually documents the interperation, it is acceptable per WP:OR. If the given source is published and reliable it is acceptable per WP:V If the interpration is of significance to the topic it is acceptable per WP:NPOV. What is a policy-based reason for adding any requirement beyond those three for adding an interpratation to Wikipedia? —Preceding unsigned comment added by  BirgitteSB (talk • contribs) 18:26, 10 September 2008


 * The restrictions in WP:PSTS on primary sources was added to reinforce the rest of WP:NOR as one of the main methods of original research is a novel interpretation of primary sources -- as is the prohibition of using primary sources that have not been published in a reliable secondary source as another primary method of OR is the use of unpublished facts and figures. --Philip Baird Shearer (talk) 07:40, 11 September 2008 (UTC)
 * That passage appears to have been composed with a weak understanding of tertiary material. Durova Charge! 08:20, 11 September 2008 (UTC)


 * Right, but if the source documents the interpretation it cannot be a novel interpretation. No matter what kind of source you are using. The condition of "when documenting interpretations" makes the OR angle on what sources to use in the next clause irrelevant. -- Birgitte  SB  15:21, 11 September 2008 (UTC)
 * In fact, it has it backwards, when documenting an interpretation, the source to use is the primary source, which in this case is the published interpretation. I think the intention is when making an interpretation, or in interpreting material, .... DGG (talk) 13:56, 4 October 2008 (UTC)

Essay
Why is this essay, with a talk page copied from elsewhere, marked as a proposed guideline? It should be moved back to the user space or marked as an essay. Sandy Georgia (Talk) 02:02, 12 October 2008 (UTC)