Wikipedia talk:Primary Secondary and Tertiary Sources/Archive 2

That wasn't reorganization, that was a major rewrite
Any comments on Slrubenstein's version of the policy? In some ways it looks like an improvement, but (1) I don't like people using protection to shield major, disputed, edits (2) it makes the primary/secondary/tertiary distinction even more important, and different editors have different interpretations of the distinction (3) there are cases where primary sources are appropriate for (semi)-evaluative conclusions, e.g. citing The Origin of Species to show that Darwin concerned himself with certain problems in biology and believed that natural selection could resolve them. Jacob Haller 23:36, 27 August 2007 (UTC)

I think the section should give insight into good practice, recognizing exceptions, not ammo for wikilawyering, and for this reason, strongly prefer "should" to "must." Jacob Haller 23:36, 27 August 2007 (UTC)


 * I like the explanation of primary/secondary/tertiary, but I'll also second my preference for "should" over "must". But I strongly suggest striking "explanatory" from "never to make analytic, synthetic, interpretive, explanatory, or evaluative claims."  Explanatory is a good thing.  The other 4 adjectives indicate O.R. Squidfryerchef 03:26, 28 August 2007 (UTC)


 * I would suggest "to avoid making new analytic, synthetic, interpretive or evaluative claims, not found in the original sources." Jacob Haller 03:37, 28 August 2007 (UTC)


 * I wouldn't call it a major rewrite, it is mostly a reorganization, and it does present a more logical grouping of the concepts. The only significant change I see is including explanatory in the secondary sources section, which did add consistency. But I think we may need a definition of what explanatory means in this context. Dhaluza 09:44, 28 August 2007 (UTC)


 * Gee, so many places on this page this comment could go, but I'll just post in this one section. Like others, I agree with should over must. I also agree with removing the explanatory condition, as on occasions it helps the article to have some explanation if can be kept NPOV. I also tend to think that there should be something in there to the effect of "there are intended as general guidleines, not hard and fast rules, as there is always an exception to every rule" (as many points others on this page have pointed out time after time for either side of the argument). We need something that spells out guidelines, not rules, to keep some of the edit wars to a minimum. There are a lot of "rule nazis" or "wiki-lawyers" on here who interpret everything as a hard and fast rule with no leeway either way. wbfergus 14:15, 28 August 2007 (UTC)


 * I agree. I should also say I'm puzzled as to why there is such an intense desire to paint primary sources as being less reliable.   Primary sources can be reliable or unreliable, secondary sources can be reliable or unreliable.  The proponents of a rule that rejects primary sources assert that primary sources are less reliable - but as far as I recall they've never provided a reference for that claim.


 * I just looked at the Wikipedia article on the Jahn-Teller effect. Unless you look at the original articles (the primary source) you don't know how Jahn and Teller proved the theorem, don't know that the articles provide tables that make determining Γvib for a symmetric molecule almost trivial.  Perhaps somewhere there's a secondary source that explains that Jahn and Teller proved the theorem by exhaustive analysis of the degenerate symmetries in every molecular symmetry point group (that's partly why those tables are in the article) but I've never seen it.


 * I also think I recall that in some college history class I heard that historians were supposed to seek out primary sources. --Minasbeede 14:55, 28 August 2007 (UTC)


 * You're absolutely correct that professionals (and students learning the field) are encouraged to seek out primary sources to analyze and interpret them, especially in comparison to the existing work in the field. However, this policy prohibits that kind of analysis and is against the fundamental nature of the project. (See BirgitteSB's eloquent related comments.) Also, this policy prohibits the kind of interpretation you're forwarding. If the tables and raw data indeed lead to such a conclusion, a secondary source is needed to make that claim. If the claim cannot be found in independent sources, it isn't a notable one, no matter how crucial you think it may be. Those interested in primary sources and original research have many Foundation projects to choose from, such as Wikiversity, Wikibooks, Wikisource and Wikiquote, however Wikipedia is not such a project. It would also be even better if the editor is an expert and publishes a reliable source about the matter that we can cite here. Vassyana 16:12, 28 August 2007 (UTC)


 * If you are referring to what I said about the Jahn-Teller papers ("the kind of interpretation you're forwarding") I'm certain that the papers themselves state that their approach is to examine each point group (well, duh, if they did it they'd surely say so) and pretty certain that the papers also point out that the tables they use in that proof can also be used to create Γvib for a symmetric molecule. It's irritating to experience editors who are distressingly eager to find "original research" fault with simple statements instead of paying attention to the meat of what is being said.  My point is that secondary sources overlook key parts of those publications since the secondary sources concentrate on a single aspect of the papers (the theorem proved.)  You have to go to the original papers, the primary sources, to see the entire range of what they published.  (I could check on the Γvib statement - I have copies of the papers on my computer - but I'm tired.  The point remains that the primary sources are superior to the secondary sources, in this instance.  It should not be necessary to have to explain why the primary source is used when the facts are that the primary source has the superior information.  Wikipedia does not need and should not have a policy that makes a blanket condemnation or deprecation of primary sources.  "In this instance" is italicized in order to emphasize how foolish it would be to require justification of use of a primary source for every instance in which the primary source is superior.  Let the superiority of the result be adequate.  If it turns out in some cases that use of primary sources produces an inferior result then let that be adequate reason to switch to a better source.  You cannot, however, make an accurate blanket statement about the quality of sources on the basis of whether or not they are primary.)  --Minasbeede 03:59, 29 August 2007 (UTC)


 * If you're writing an academic thesis then you're actually supposed to be doing original research. You'll see that term on the mastheads of academic journals "...a publisher of original research".  We didn't make up that term just for Wikipedia.  That said, primary sources are a nice thing to have in a WP article for various reasons ( fact checking, NPOV, detail, providing resources for others to do research ). Squidfryerchef 02:55, 29 August 2007 (UTC)


 * Uh-huh, but the policy is too broad and is interpreted still more too broadly, to the detriment of the quality of Wikipedia. The policy is like rejecting all arithmetic division because a fallacious proof that 1 = 2 can be made by doing a forbidden division by zero in something that looks like a proof but is not (because of that fallacious division.) --Minasbeede 03:59, 29 August 2007 (UTC)

No, Squidfryerchef has it just right. Minasbeede writes, "My point is that secondary sources overlook key parts of those publications since the secondary sources concentrate on a single aspect of the papers (the theorem proved.) You have to go to the original papers, the primary sources, to see the entire range of what they published" and I agree completely. If all you want to do is quote a primary source for a fact, completely independent of any interpretation or argument, that is absolutely permitted by the policy. However, if your goal is to fill in a gap in the secondary literature by going to the primary sources and using something hithertoo unused to support or develop an argument, analysis, interpretation, explanation ... well, then what you want to be doing is write your own book or article. Nu? Feel free! Go right ahead! Write a book and submit it to a press. Write an article and submit it to a journal! If that is what you want to do, you are free to do it. There are many venues for publishing your own original research, and nothing is stopping you from submitting your own work to a publisher. It is just that, Wikipedia is not the appropriate venue for editor's own original research. Do it anywhere you like, just not here. Slrubenstein  |  Talk 13:49, 29 August 2007 (UTC)


 * Huh? I can quote any secondary source to "fill in a gap" and you're content.  But if I go to a primary source and do no more than quote it to "fill in a gap" then I'm doing "original research"?  That's absurd.  Either way I'm quoting a source.  If there was "original research" done then it was the author of the primary source who did it, not I.  If all secondary sources (or all secondary sources quoted in a particular article) overlook something it is incomprehensible how anyone could call retrieving that something from a primary source "original research."


 * There's a noxious "mission creep" occurring. The policy against "original research" is justified on the basis of a desire to forestall patently incorrect "original research" of the flat-earth, perpetual-motion, etc. variety.  Now you (and others) claim that --Minasbeede 14:42, 29 August 2007 (UTC)quoting a primary source falls into the same category as flat-earthism.  (Oh, yeah, you allow such quotation so long as it neither contradicts nor augments material from secondary sources.)   Physician, heal thyself.  It is you (plural) who are attempting to co-opt Wikipedia to put forth your own views. --Minasbeede 14:42, 29 August 2007 (UTC)

It is not mission-creep. It is based on the realization that there are topics where what one editor believes is a "flat-earth theory" another editor believes is a "flat-earth" theory. Maybe this is why we disagree: I am not a fundamentalist. i do not think that a text means just what it says it means. Texts are read and reading is interpretation. When a fundamentalist says "The Bible says the world was created in six days" I know that there are Bible scholars who, using comparative philology, reference to other works of ancient Near Eastern literature, and other techniques, have strong arguments that that is not the point the Biblical author was making. I think the Fundamentalist is really projecting his or her own belief into a document written by people of another culture a long, long time ago. But the Fundamentalist insists that the text means just what it says. The fundamentalist and I will never agree. And I have discovered that there are people who read speeches by Hitler or Moussolini, and articles on genetic research, that are in effect fundamentalists too. If this does not make sense to you, well, I am sorry, I have tried in good faith to explain my view to you as best I could. We will just have to agree to disagree - and hope Wikipedia has policies that can enable us to work together on an article without getting into a revert war. Slrubenstein  |  Talk 14:53, 29 August 2007 (UTC)


 * I think we are in very good agreement about fundamentalists: they claim to put "the Bible" (or something) first but what they really do is use the Bible (or something) to support their own views. Do we agree that just because some editor sincerely believes something Wikipedia does not need to give equal coverage to that something?   Do we believe the same, even if there's a vocal minority (not merely a single editor) who believe (or claim to believe) in that something?   Obviously I use "flat-Earth" as a stand-in for other disputes.  I do mean to indicate an essential similarity between "flat-Earth" belief and some other beliefs.


 * I probably was overboard in my wording ("noxious mission creep.") What I'm saying is that sources are sources, be they primary or secondary, and that it's counter-productive to assert as policy that secondary sources are superior to primary sources.  That assertion can't be proven (because it isn't true.)  Isn't it counter-productive and a waste of time and effort to argue over an untruth?  Perhaps one or more of the advocates of the superiority of secondary sources would like to claim and even attempt to prove that secondary sources are superior in 20%, 50%, 80%, or whatever% of the time.  To base a policy on that assertion would mean that the policy is flat out wrong in 80%, 50%, 20% or (100-whatever)% of cases and provides improper guidance.  That's detrimental, and such a policy is unneeded.  The policy should simple be to make appropriate use of appropriate sources.  There is no single general rule that can be used to distinguish appropriate sources from inappropriate sources.


 * I'm not at all convinced that scientists who read (and accept) genetic research articles are fundamentalists just because they are scientists and operate and think as scientists. That they don't allow creationists to force non-science into science (nor allow science to be made subordinate to sophistic reasoning) doesn't make them fundamentalists. --Minasbeede 17:24, 29 August 2007 (UTC)

Can we describe good Wikipedia research practice on another page?
I suggest supplementing this page (instructions on what not to do) with another page, possibly Source-based research on what to do, providing guidance for common situations: what sources to look for, etc., etc. Jacob Haller 05:27, 29 August 2007 (UTC)
 * Not a bad idea. Were you thinking it should be an essay, or a guideline? Were you looking to have WP:PSTS (or my proposal, or some other replacement) moved or placed there? I'd also add it might not be a bad idea, if we take this approach, to specifically solicit input from the WP:V and WP:MOS talk pages. Cheers! Vassyana 18:23, 29 August 2007 (UTC)
 * Not really sure. I suppose it should start as an essay, but try to achieve enough consensus that it could later be upped to a guideline, and start with links from here to there as well. But my own thoughts on the subject are very disorganized, a matter of one question at a time. Jacob Haller 18:39, 29 August 2007 (UTC)
 * personally, I think it should be a personal essay. Wikipedia has always been rule averse in part because there are perhaps an infinite number of ways to do someting right and also because of the quasi-anarchic ideals of a wiki-community. Se we have always strived to have as few policies about content as poissible - NPOV, V, and NOR have worked for a long time and while I feel passionately about these policies I would really be loathe to add much more to them. Plus, we have peer review, and Good or Featured articles, that can act as community-agreed models of what we should be striving for (maybe what you want to do is write an essay on "How to write an article that will become FA" or something like that?). If anyone cares I explain my view in greater detail here The role of policies in collaborative anarchy. hee is another possibility: there are some superb books precisely on how to do research, that are used all the time in college and graduate school courses. perhaps a series of articles on these books might help provide editors unused to high-level research with some useful tools. We have lots of article on various books, so why not collaborate on some articles calling attention to books on research? Slrubenstein  |  Talk 18:45, 29 August 2007 (UTC)
 * How would you feel about deprecating WP:PSTS (or an appropriate replacement) to a guideline? This could potentially resolve some objections voiced and allow the topic to be more fully addressed, while avoiding policy bloat. It could also more appropriately represent a consensus where there are known exceptions and examples provided. Just chewing some thoughts. Vassyana 18:53, 29 August 2007 (UTC)


 * Guidelines (as distinguished from policies) are fine and surely would be useful. The problem I see is that for no good reason primary sources are being deprecated, with much of the deprecation being of the form "this kind of abuse can be committed when primary sources are used" with no recognition of the equally possible abuse that could arise if a secondary source is used nor of the possibility that the secondary source itself has made a misrepresentation.  It's fake rationality: all the good points of secondary sources are listed, all the possible bad points of primary sources are listed (or asserted as true), and then the forced conclusion is reached that secondary sources are better.  It's selective application of stringency: primary sources are regarded with distrust and suspicion and have to be frisked, metal-detected and X-rayed, secondary sources get through the security gate with a nod.  That's further compounded by assertions that editors cannot use a primary source to correct an  assertion made by a secondary source that misrepresents that primary source: the primary source is permanently on a no-fly list.  I'm doing some simple genealogical research.  I can see where ancestors lived in census reports: primary sources.  I can't see that where they lived would ever be notable in Wikipedia but where someone else lived (Willard Gibbs lived in New Haven - no surprise) might be notable.  If the census record says person X lived in New Haven and a secondary source says the person lived in Miami why is the secondary source given favored treatment?  (Yes, I know that something can be composed that adds a lot of additional things to the question and thereby puts the secondary source in a favored light.  The problem is that the policy in effect assumes that there is always something or might be something that makes the secondary source more reliable.  I can see no justification for that assumption.)


 * All of this discussion is unnecessary, but while there is a move to have policies that favor one class of source over another then there will and should be conflict. If, however, there were simply guidelines on how to use sources and on constructs to avoid when any source is used that would be fine.  If there's notable features to different kinds of source go ahead, note them. --Minasbeede 21:42, 29 August 2007 (UTC)


 * I like Vassyana's suggestion. One way to do it would be to merge Classification of sources with WP:PSTS to form a guideline.  --Coppertwig 22:50, 29 August 2007 (UTC)


 * It's not a move towards that, it's a well-established principle. Look over Verifiability and take note of how third-party reliable sources are discussed. Vassyana 23:11, 29 August 2007 (UTC)

Straw poll to test for consensus to remove edit protection
This policy was edit protected because edit warring over language restricting the use of primary sources. This subject has been discussed extensively on this talk page, and was also the subject of a previous straw poll. The question posed by this straw poll asks a very specific question intended to see if the edit war can be ended and edit protection can be removed to allow normal editing again:


 * '''Do you agree or disagree that language added to this guideline during the past year or so, replacing language treating primary and secondary sources equally with language devaluing primary sources in favor of secondary sources, regardless of the merits of the arguments or your personal views on it:
 * '''a) did not represent a consensus view of the Wikipedia community as a whole,
 * '''b) is unlikely to gain broad consensus in the near term, and
 * c) any such language added to this policy without first obtaining consensus should be removed.
 * -- Dhaluza 23:41, 28 August 2007 (UTC)

Comments
Comment. This is another biased poll, if better than the last one. The default implied position is that the change was made without consensus, an issue that has been repeatedly addressed and refuted. (Particularly the ludicrous assertion it was "snuck" in.) Most objections (by far) to primary source limitations originate in a rejection of the fundamental principle of no original research, rejection of other fundamental policies (such as NPOV), rejection of the requirements of reliable sourcing (such as WP:V) or outright distortion of how the policy reads (as it does not prohibit primary sources). Also, primary sources have been deprecated since before the most recent version established approximately a year ago. (Note that even in that earlier version, primary source reliance was still considered an exception.) The use of primary sources was simply tightened in response to gross abuse of the more permissive (but still highly restrictive) language. So long as the above misrepresentations and misunderstandings persist, it will be exceedingly difficult to achieve a consensus. Vassyana 02:17, 29 August 2007 (UTC)
 * There is no assertion in the question that it was "snuck" in (over a year's time, no less), it only asks if the language was developed through a consensus process, and that it reflects (or can attract) consensus. Dhaluza 10:31, 29 August 2007 (UTC)
 * The editor who crafted the last poll insisted it was snuck in, or otherwise inserted through dubious means. My refutation of that claim occured in the context of refuting his claim that there was no consensus/real discussion for the change. I apologize for the lack of clarity. Vassyana 15:00, 29 August 2007 (UTC)
 * The issue wasn't that it was "snuck in" by improper means. The issue was that it wasn't looked at carefully by the broader Wikipedia community, much like the WP:ATT proposal, which initially had a narrow "consensus" by a few editors, but failed after it was subject to broader discussion. Because of their importance and potential to harm the entire Wikipedia, policy articles, unlike content-domain articles, aren't subject to "inertia", and must always reflect widespread practice and consensus. CO GDEN  18:03, 29 August 2007 (UTC)
 * According to you, the change "was inserted into the policy in late 2006 without much comment or notice". I can't believe you continue to insist the issue was not broadly and vigorously discussed. Please stop repeating variations of the same disproved claim. Vassyana 20:48, 29 August 2007 (UTC)


 * Bad material is bad material and if it arises through any method or pathway it is bad. That it is bad or a misuse of something (here the "something" is "primary sources") does not poison the entire well.  It's as though you are claiming that secondary sources inherently cannot be in themselves an example of misuse nor the grist for misuse, or that they somehow mysteriously possess qualities that make them superior to primary sources.


 * The issue is (or should be) good over bad. When some editors seem absolutely dead set on tarring an entire class of potential source material it's hard to not think that those editors have some scheme in mind. --Minasbeede 02:28, 29 August 2007 (UTC)

I see no reason nor justification for a priori denigrating primary sources. Were I to speculate on what the consensus here is I think I'd violate all sorts of policies - and be doing something useless. If you want Wikipedia to be good you want it to be based on good sources. Those may be primary, may be secondary. I see no useful purpose served by instituting a policy that denigrates one type of source. What that would most likely accomplish would be to give more ammunition to "Wiki Lawyers." Someone relies on a primary source, the "Wiki lawyer" removes the material and then both insists that the use of that source be justified AND repeatedly rejects all such justification using arguments the "Wiki lawyer" would never use against secondary sources, even though that argument (or an even better one) might be even better applied to some other material that is or is said to be from secondary sources. Why add another hoop through which editors (when challenged, and challenges may most often occur for material for which there are editors trying to favor their own POV) must jump? If it's good material from a good source and is used appropriately that would seem to be enough. If it turns out that overall Wiki sources are 10% primary and 90% secondary or the other way around what does it matter? What you want is good material from good sources. You can't honestly make any kind of blanket statement about the reliability or utility of primary sources nor of secondary sources. (You can make sophistic arguments. That's always possible.  It would seem that Wikipedia would intentionally shun sophism, everywhere.)  The negative claims made about primary sources are themselves unsourced, they're just claims. They smack of OR - and of the worst kind. --Minasbeede 02:22, 29 August 2007 (UTC)

A, B, and C, and I'd like to take it a step further and get rid of the Smith/Jones example. Squidfryerchef 02:46, 29 August 2007 (UTC)


 * Comment, I agree with Vassyana on the quality of this straw poll. My position is that the policy needs to be returned to its pre-edit-war state, and the changes proposed by Cogden and the others need to have consensus before they are implemented.  The edit-war has specifically targeted removing or changing the following two sentences, which previously had consensus and have been included in the policy for almost a year prior to the edit-war:
 * "Wikipedia articles should rely on reliable, published secondary sources."
 * "Although most articles should rely predominantly on secondary sources, there are rare occasions when they may rely on primary sources.'


 * Those two sentences, and the rest of the policy that I linked to above, should remain in place until consensus is reached for changing or removing them. This is the process according to Wikipeida policy and guideline.  Dreadstar  †  04:47, 29 August 2007 (UTC)


 * Why should they remain? They embody an unfounded blanket denigration of the quality of primary sources.  That they were added at some point in time does not remove the fact that the basic premise for them is not supported by fact.  According to the policy embodied in those two sentences, for example, one can't quote anyone, one can only repeat a secondary source's quotation of someone.  Unless B says A said X A didn't say X. That is ludicrous.  It has not been shown that favoring secondary sources over primary sources in any way enhances the quality of Wikipedia.  Anecdotal or isolated examples in which it is claimed that reliance on primary sources has led to lower quality don't suffice to justify a blanket condemnation of primary sources.  I can't really even understand why anyone is eager to reject use of primary sources: they are, just as the name implies, primary sources and considerable weight should attach both to "primary" and to "sources."  I certainly do make objection to the extreme nature of the NOR policy (which rejects logic, rejects perfectly valid synthesis) but that's separate from my objection to the denigration of the use of primary sources.  It would be false to characterize my objections to denigration of primary sources as being linked to disagreement with the NOR policy.--Minasbeede 12:58, 29 August 2007 (UTC)


 * I believe this is misstatement of how WP:Consensus works, specifically "policies and guidelines document communal consensus rather than creating it." The question properly asks whether these changes document existing consensus or try to create a new one. Like verifiability, it is the burden of those wishing to include the language to show it represents consensus, not the other way around. Dhaluza 11:19, 29 August 2007 (UTC)
 * And Cogden's proposed wording does not reflect current consensus, so it should be removed until a new consensus is found to change the pre-edit-war wording. There is proof that the previous wording that stood for almost a year had consensus.  We don't remove or reword standing policy just because someone comes along and disputes it and claims no consensus for the current wording.  It would be chaos to do so.  A new consensus has to be reached to change the wording. Dreadstar  †  16:53, 29 August 2007 (UTC)
 * If there is no consensus (which appears evident already), the simple solution is just to delete the controversial language entirely, replacing it with nothing. Essentially, there would be no official policy regarding primary or secondary sources, until some consensus is reached for some positive language. CO GDEN  18:10, 29 August 2007 (UTC)
 * That's exactly backwards, there was consensus for the pre-edit war wording. You need consensus to change the wording. Dreadstar  †  20:37, 29 August 2007 (UTC)
 * It doesn't matter if there "was" a perceived consensus at one point. There's clearly no consensus on the controversial language now. Wikipedia policies, because of their importance and their potential to do damage to the Wikipedia institution, must reflect present consensus and must describe current practice. The whole WP:ATT promotion issue showed that sometimes, a perceived policy consensus among a group of editors isn't an actual consensus within the Wikipedia community. CO GDEN  18:58, 30 August 2007 (UTC)


 * It is also a misrepresentation of the policy. As written, there is nothing in the policy that prevents an editor from adding "A said X" as long as there is a verifiable source (e.g. A's autobiography, a published transcript of a speech or whatever) The NOR policy does not reject the use of primary sources.  The NOR does not reject logic or synthesis either, it just rejects an editor's logic or synthesis - whether the editor is me or Minasbeede.  That is because Wikipedia is not a soapbox or a forum for advancing one's own views.  If that is what you want to do the world wide web has LOTS of opportunities for you.  This is just not one of them. Slrubenstein   |  Talk 13:41, 29 August 2007 (UTC)


 * My understanding of logic may be flawed. If "all men are mortal" and "Socrates is a man" doesn't include by logic that "Socrates is mortal" then I'm really confused.  Concluding that "Socrates is mortal" does not seem to me to be my "personal" logic or synthesis: that conclusion flows from the laws of logic.  If an editor takes information from source A and other information from source B and combines these sourced, verifiable pieces of information to synthesize C, C being factual and correct, it's very hard to understand why that is forbidden, other than that "it's the policy (dogma)."  To me the test should be whether or not all reasonable people could synthesize A and B to reach C, making C verifiable as not being "an editor's logic or synthesis" since it can be seen that the synthesis is valid.  I don't anticipate that the policy will soon be changed and this response isn't concerned with such change.  My point is that it seems deceptive and extreme to blithely characterize all synthesis as representing an attempt to use Wikipedia as "a soapbox or a forum for advancing one's own views."  If the view advanced is clearly inherent in the source material by the application of valid synthetic logic then to me that removes it from being "one's own views."   I'd certainly agree if the synthetic logic is strained or non-obvious or if the sources are dubious or stretched beyond what the material properly indicates then the material is inappropriate and ought (a) to never have been edited in to begin with and (b) should be promptly removed.  It isn't NOR as a whole that rejects primary sources, it's the two disputed sentences.  My interpretation of those two sentences is perfectly valid: they do imply that a misquotation of a person's words has priority and prominence over the person's words themselves. Nor did I state that an editor could not, under the control of those two words, report that A said X.  What I said was that, according to the policy, an editor is forbidden to refute a secondary source misquotation by referring to the primary source.  That is the tenor of the two sentences: secondary trumps primary.  The effect is that Wikipedia becomes a soapbox for the views of the secondary source which misquotes the primary source and the policy (the two sentences) favor the editors/accomplices of the dishonest secondary source who wish to misrepresent the primary source to favor their non-neutral POV. --Minasbeede 14:19, 29 August 2007 (UTC)
 * This is the problem: in my opinion, there are many editors here who are unreasonable and who make utterly unreasonable inferences from primary source material - and no doubt they think the same of me. We will never reach agreement.  But Wikipedia asks us to take each other at good faith.  How then can we work together on an article and produce something half-way decent (or even better!) that satisfies us both?  Only - I argue - because we have policies that provide a framework for making enabling not just editors with opposing views but editors with opposing views of what is a strained or non-obvious inference.  You can say I am just soap-boxing and I will say no I am not, this is obviously what the primary source means (or vice versa).  This policy protects us from that.  You are right that there are bad (unsophisticated or ill-willed) secondary sources out there.  But for every topic I know of there are good sources and it is not hard to find them, and to edit an article that makes it clear which sources represent the mainstream opinion ... and I do not necessarily mean mainstream opinion of all human beings, which would sink the evolution article. In the evolution article, we favor secondary sources and a standard of what is mainstream in terms of biologists; the article on the planets relies on the dominant views of astronomers.  NPOV may require us to say that some people still believe the sun orbits the earth, or that creationists reject evolution.  The evolution article has a brief summary acknowleding these other views, and links to articles that address them in full.  There are ways of representing scholarly debates accurately without relying on OR. Slrubenstein   |  Talk 14:33, 29 August 2007 (UTC)

[I dispute that creationism falls within scholarly debate (and in particular within scientific debate) but that's an aside. I'll accept that Wikipedia handles that appropriately, or at least adequately.] I can see no way that any OR I might attempt would ever be useful or necessary in establishing that the theory of evolution is valid: there's a plethora of existing evidence that is tremendously over-adequate for that task. I do not think that it is in any way valid to characterize using an accurate quotation from Darwin to refute an inaccurate quotation from Darwin as "original research." As I read the two sentences at issue the exact implication of those sentences is that using primary source material to refute dishonest/incorrect/misleading secondary material is deprecated/forbidden. --Minasbeede 14:51, 29 August 2007 (UTC)

Comment, I didn't follow the debate at the time, and don't know what happened. However, I see no general agreement on (1) what the present policy involves (2) what good practice involves or (3) whether the present policy describes good practice. Jacob Haller 05:26, 29 August 2007 (UTC)


 * I don't believe that I could fairly vote in this poll, as this has (evidently) been going on for longer than I've been looking at the policy, even though I've tried to read and understand what all has been argued back and forth. Anyway, I had a thought that may help some, though it seems like it would be a pain to implement, especially with the dearth of existing articles. How about something that distinguishes between the types of sources cites, either different "reference" sections like "primary source", "secondary source", etc., or a different "template" that may accomplish that, where the reference is cited as "ref=primary" or "ref=secondary". This way, on those rare occasions where a primary source is used, it becomes more evident which "piece" of the article is using the primary source vs. a secondary or tertiary source. Just an idea while trying to find a way out of the quagmire. Personally, I think there are occasions where primary source information is just as good (if not better, see the section/question below) than most secondary sources, and any policy regarding primary sources should be worded so that it explicitly states something along the lines of "primary sources can be used in special occasions...blah, blah, blah" with a few "good examples". There are to many "wiki-lawyers" on here who are abusing the policy as it is currently written. I tend to agree with User:Minasbeede wbfergus 13:36, 29 August 2007 (UTC)
 * I've put a few proposals out there defining primary and secondary sources, and I've concluded that there is not even a consensus as to what those terms actually mean. So I don't think any policy that distinguishes primary sources from secondary sources, and treats them differently, is going to be successful. There will always be disputes as to what source fits into what category. For example, I might argue that Darwin's Origin of Species is a primary source because it is where the idea of evolution originates, but some other editors would apparently argue that it is actually a secondary source, because the book analyzes, interprets, and comments on raw data. If we can't agree on that, we shouldn't have a policy that distinguishes between primary and secondary sources. CO GDEN  18:19, 29 August 2007 (UTC)


 * It appears that the easiest way to finally garner concensus on these issues is to simply prolong any arguments either way until so many people are burned out on the subject the the few reaining people can finally push through whatever definition they want at that time. It seems prety sad that with all of educated, intelligent people on here, that they can't agree to how to properly word a "policy" that is misused to dictate how articles are written or what should (or could) be in them. This is starting to remind me of watching little kids on a playground arguing about how to play tetherball. Admittedly, it's a broad policy open to many different interpretations, and User:Vassyana seems to be the only person so far who has made an effort to get something written out for people to comment on (right or wrong). Instead all that happens is incesstant bickering back and forth, one camp wants to rely solely on interpreted secondary sources, with the occasional tertiary source, and the other camp wants to allow primary sources as viable input to an article, and neither side wants to back down or compromise. wbfergus 19:17, 29 August 2007 (UTC)
 * I don't think any side wants to "rely solely on interpreted secondary sources with the occasional tertiary source." No one has said that, have they?  Both sides want to allow primary sources as input, it's the extent to which those primary sources can be relied upon.  That's what I understand the discussion to be about.  On side wants some limits on Primary sources, the other wants unrestricted utilization of primary soruces, some even to the point of including our own opinions on what primary sources say.  I believe use of primary sources in integral to the 'pedia, but not in a completely unrestricted manner.  Dreadstar  †  20:48, 29 August 2007 (UTC)
 * You must have missed my multiple language proposals, which failed because we can't agree on the definition of primary and secondary sources. That tells me that consensus isn't in the works right now for any policy distinction based on the difference between primary and secondary sources. What we need to do is just delete the controversial sentences, and work on the Classification of sources essay or some other place where we can hammer out a consensus definition of primary and secondary sources. Since WP:NOR is already a policy, and a core policy at that, we should be very conservative about what is allowed to remain there. CO GDEN  18:58, 30 August 2007 (UTC)


 * I'd agree with Dreadstar, in his judgment of the relative positions. I specifically attempted to incorporate most of the concerns expressed about primary source restriction, such as peer-reviewed articles, particularly reliable primary sources and their use for citing supporting or illustrating facts and quotes. On the other side, I incorporated the emphasis on reliable secondary sources and specifically defined such sources based on existing consensus (extant policy) and concerns expressed about primary and dubious sources. While I have a clear position, the draft was an attempt to find some compromise between the two opposing views, based on current policy overall and the concerns expressed on both sides in the history of this talk page. To be blunt, some opposing responses are based on misunderstandings and/or distortions. Also, some opposing responses are exactly the kind of concerns I've already considered and integrated into the proposal. If only one side is willing to compromise, compromise is impossible. Vassyana 22:03, 29 August 2007 (UTC)
 * I don't know where you've found a consensus in the definition of primary and secondary sources. I've seen on this talk page numerous conceptions of what primary and secondary sources are, and people have strong opinions. Unless and until we get broad consensus on definitions, it will be impossible to reach consensus on policy distinctions based on those definitions. For example, many editors here (myself included) think that peer-reviewed journal articles, novels, movies, interviews, and autobiographies are almost always primary sources. If that's so, the a policy that they should be "rare" or not "primarily relied upon" would de-legitimize most of Wikipedia's best featured articles. Simply defining everything that's an acceptable source as "secondary" doesn't help either: we shouldn't confuse people by using terms like "secondary sources" in non-standard ways. CO GDEN  18:58, 30 August 2007 (UTC)

Quality vs type
Unindent - I see part of the problem is that the emphasis should first be on quality of source, then the viewpoint. As it stands, it feels more like my secondary source, warts and all, can trump your primary source. Secondary sources may not contradict the primary source, but they may be a superficial interpretation, which in turn can be misread. It is then difficult to use the primary source to correct the misinterpretations in the face of this policy (especially as this may require some obvious synthesis to align the original wording to the secondary interpretation). So I want wording that puts the primacy of quality sources first. Spenny 14:41, 11 August 2007 (UTC)
 * I view those as two separate issues. Quality of a source is different than the type of source (primary, secondary or tertiary).  I don't think you can put one in front of the other, they need to be looked at in parallel, and not in isolation from one another.  – Dreadstar  †  19:24, 11 August 2007 (UTC)
 * Where would you put a BBC solid news item vs. the report it was based upon? For example, say we have a Government White Paper (a primary source) which is summarised and commented on by the BBC news item (secondary source). Having read both, we find that there is an element of over-simplification which distorts what is actually said. Clearly the primary source is of the highest quality, it is what is being discussed, and the facts of its statement must outrank the summarisation of the reporter, even if this is done by the simple expedient of the direct quote for the reader to interpret by themselves. Obvious as this may be, it appears that is not accepted as reasonable by some, and is undermined by this statement of policy. You may see this as a discussion on real research vs. articles in New Scientist where I would be far more accepting of your position, but often this debate is about a much lower test of qualities. I agree with that the concepts need to be considered in parallel, but that is not what policy appears to say, it appears to say it is right to disregard primary sources, regardless of quality. Spenny 22:10, 11 August 2007 (UTC)
 * In the example you provide, you should not use a primary source in that fashion, even in the version "lenient" in relation to primary sources. It is fundamentally a violation of this policy, since you would be employing an "unpublished analysis ... of published material that appears to advance a position". We cannot put forth our own interpretation or analysis of primary sources without violating the most basic principle this policy is based upon. Particularly in regards to government matters, what you find to be an oversimplification, another may find to be a perfectly fine overview. Secondary sources are advantageous because they generally include analysis and commentary, avoiding the pitfalls of trying to employ primary sources without giving an analysis (explicit or implied) and advancing a position, which is a very difficult task to be generous. The policy reflects those issues. Vassyana 05:34, 12 August 2007 (UTC)
 * That misrepresents my point. My point was that I could quote a primary source without an analysis to show that a secondary source was flawed, if needs be leaving it to the reader to make that analysis. The problem with the primacy of secondary sources approach is that it leaves the door open to quoting news items, which may be giving a controversial aspect undue prominence. In the context of verifiability vs. truth, if I can verify that a secondary source is misleading, or unreliable, then I can have both. We are not talking about hard stuff here, we are talking about statements of fact. Yes, on a Government paper, we can see that what is said is not always what is meant, but we should not assume that a third party interpretation outside the notable journals is necessarily any more accurate than the original words.
 * I have a particular example in mind (not that it is in Wikipedia as far as I am aware). During the floods in the UK last month, someone leaked an about to be published white paper on changes to planning policy. One short paragraph discussed the fact that it was inappropriate to stop building on flood plains as in fact a lot of areas that have been built on for centuries fell into this definition, including the centre of London, where it is considered that the Thames Barrier is a suitable solution to the problem. Someone tried to make a story that the Government was reckless as they were making policy to ignore the flooding risk (which also ignored the fact that the flooding affected areas that were not even considered flood plains). Some of the news media ran with that slant, making the story by selective quoting and taking the comments out of context. We could write an article based on the newspaper reporting, and find a number of articles taking that line, thus asserting the correctness of publishing the distortion as verifiable. We can correct this by simply quoting the original short paragraph in context which requires no analysis. Spenny 11:08, 12 August 2007 (UTC)
 * I don't think it misrepresents the point at all. If the obvious point of the edit is to put forth the claim that the newspaper was wrong, at the very least that is highly questionable. There's far more harm potential in that kind of example (particularly of cherrypicking quotes and POV pushing, two very common inappropriate uses for primary sources), than for gain. notability and NPOV require multiple third party sources. In general, if there are enough third party reliable sources to establish notability, present a balanced view and write a complete encyclopedic article, primary sources should not be necessary except in rare cases, as reflected by the current policy formulation. Vassyana 18:07, 12 August 2007 (UTC)


 * Whilst I understand this viewpoint as a statement of principles, it is a great way to give credibility to urban myths. Someone who is knowledgeable in their subject area will understand the misrepresentation of the media and can readily correct it without unreasonable synthesis. I still think you misrepresent what I am saying to justify your position: "quoting the original short paragraph in context" is quite different from "particularly of cherrypicking quotes and POV pushing" and whilst I understand what you are afraid of, it should not be used to justify the inclusion of the misrepresentations of the media. You seem to have been lucky and not had that sort of editing used against you. Unfortunately, the media has a tendency to use press releases, the same news agencies, pundit of the day and so on, so giving the impression of wide consensus, whereas it may well be just the blind reproduction of one particular press officer's summary. Secondary source does not mean disengage critical faculties.
 * The other effect I have been subjected to is the one line summary in a general news article used to assert a viewpoint, the headline quote used where we cannot see behind this source to understand the context to validate this secondary source itself. Often, that one line quote will be out of context. Much as you fear primary sources cherry picking and POV pushing, this happens just as much in secondary sources of the media, where so often these are a collection of cherry-picked items of interest. It is leaving Wikipedia open to just the same sort of spin-doctoring that the press rails against in the UK, when they were party to the culture that begot it. If you have ever been party to an incident where the media is involved, you will know that even the most innocuous of happenings can get incredibly distorted, even to the point where they will contradict the pictures being shown.
 * Policy that fits well with scientific credibility does not transfer well onto topics of general interest where even the so-called reliable sources need to be tested for veracity. Spenny 22:51, 12 August 2007 (UTC)
 * The statement "I could quote a primary source without an analysis to show that a secondary source was flawed" is the very definition of Original Research. Dreadstar  †  23:49, 12 August 2007 (UTC)
 * What on earth do you mean? In any case, checking the secondary sources against the primary sources improves the articles. At times secondary sources misquote primary sources, contradict the primary sources, etc. If the secondary source says "the author said A" and the author says "not A," and does not say "A," then the secondary source is ... wrong. Jacob Haller 00:17, 13 August 2007 (UTC)
 * I explained it in more detail in this posting. Dreadstar †  00:24, 13 August 2007 (UTC)
 * Ultimately, what is research? In Wikipedia, there appears to be a distinction between research: as in inventing things, histories etc. and publishing that work; and research: looking things up. Looking things up is not banned - a news article mentions a statement, it appears fishy, we go to the primary text and see it is wrong. We do not research, we simply observe that the article in question is of questionable validity. If the citation is not of sufficient quality, then the statement it supports needs to be challenged. There is nothing inappropriate in that. Spenny 14:35, 14 August 2007 (UTC)

I've expressed my position above (1) quality is important (2) most claims can use either primary or secondary sources (3) some types of claims require certain types of sources (we could describe these in more detail). Jacob Haller 05:06, 12 August 2007 (UTC)


 * Absolutely agree. Let's say you have Anna Frank diary or De Bellum Gallicum. So we should trow them in WC because are 'primary sources'? Then let's say that we have Mein Kampf or a diary written by a mad reclused in a neuropsichiatric hospital that thinks to be Napoleon. Are these latter sources, still primary, as the same level of 'reliability' of the firts two? What about 'The second world war' written by Churchill, that awarded even, if i remember well, with a Nobel?? And what about the immense lies that we can read in 'secondary sources' in wich, expecially if politic-ecomic interests are involved, Wiki should pose exclusive faith?


 * The problem is clear; but it don't lies on 'primary or secondary sources' but in the manners of guys like Vassyana, that are acting with only one fixes idea: i am right, you not. This is problematic for me; the reverting to the changes made, (not without reasons) to this (quite absurd itself) policy. That's a problem.--Stefanomencarelli 14:15, 12 August 2007 (UTC)


 * The Gallic War is a masterpiece of propaganda, so probably a poor choice even if it is the main historical reference for the events. It is up to scholars to determine what is propaganda, what is likely true and what is likely false. Thus, we should be citing reliable secondary sources regarding the reliability of such sources, at the very least. The Diary of Anne Frank is a first person/eyewitness account, which as a whole are notoriously unreliable. For both of those examples there are a plethora of reliable secondary and tertiary sources that are easily available, including in-depth analysis of the sources themselves, so more likely we should avoid the primary sources all together except as referenced in reliable third party sources. Also, please refrain from personal attacks. There's no reason this conversation cannot be polite. Vassyana 18:17, 12 August 2007 (UTC)

This is hypocrisy. As long as the Gallic War can be cited it would appear it fully meets all Wikipedia criteria, which, let me remind you, specifically exclude considerations of truth. --Minasbeede 20:26, 30 August 2007 (UTC)


 * "... which as a whole are notoriously unreliable." You got a source for that?  Among other things it is humorous to see such flagrant use of personal opinion and OR (and worse) in defense of policies supposedly designed to prevent the use of personal opinion and OR.  Some first-person accounts are unreliable in some ways, some aren't.  If you're including material in an encyclopedia (or in anything that attempts to be accurate and fair) you have to use judgment.  --Minasbeede 15:04, 23 August 2007 (UTC)


 * Most of these appear to be straw man arguments, the policy does not prohibit the use of Primary sources, it clearly establishes that great care must be taken when using them becuase they can be easily misused: "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".  A good example of misuse is the above statement about using a primary source to show that a secondary source is wrong - that's an unpublished analysis. If a primary source is misrepresented by a secondary source, then - if it's important - another secondary source should have reported the discrepancy.  It's not our job to provide such an analysis.  Dreadstar  †  23:56, 12 August 2007 (UTC)


 * More jaw-dropping absurdity ("A good example of misuse...")  A (primary) says X.  B (secondary) says A said Y (which is materially different from X.)  According to the above it is "original research" to point out that A said X, not Y.  That makes Wikipedia in favor of propagating lies.  What is going on here?  This seems to say that if a creationist misquotes Darwin it is forbidden to show that he is lying because actually checking what Darwin did say is so-called "original research": B's lie has, by policy, greater weight than A's original statement and cannot be refuted by pointing out that what B says is unsupported by and contrary to the (primary source) facts.  --Minasbeede 13:30, 29 August 2007 (UTC)


 * Ah, I now see where I am wrong. If B misquotes A and we can see that by referring to A's words in A's primary source the refutation would take the form of removing the reference to what B has said that misquotes A.  What A said surely takes precedence over what B erroneously claims A said.  If what A said fits in whatever article it was that used B's words then A's statement can replace B's misstatement about what A said.  It is not our job to provide an analysis but it is our job to do an analysis of the material incorporated into Wikipedia and surely it is correct to remove or replace something in error if it can clearly be seen that the something is in error.  Possibly there could be a discussion about whether any of this would be appropriate in an article that for some reason touched on B's accuracy but there's no immediate need I see for that to be discussed (and further complicating an already complicated talk page.)  Specifically, I agree that it is not our job to point out B's apparent inaccuracy where B has been used a source and that use includes an inaccuracy.  Just undo it. --Minasbeede 17:48, 30 August 2007 (UTC)

Minasbeede is mischaracterizing this example. We do not know whether it is true that A said Y and to claim so would be wrong, which is why our policy prohibits it. What we do know is that according to B, A said Y, to say so would be to identify the POV (B's pov) and the source (B) and so we can most definitely add that to an article, complying with both NPOV and V. Minasbeede is upset at something if her jaw is dropping, but I fail to see what and I fail to see any reasonable meaningful argument for abandoning our policies. Slrubenstein  |  Talk 13:36, 29 August 2007 (UTC)


 * (it's his jaw.) The example given is "using a primary source to show that a secondary source is wrong."  A says X, that's in the primary source.  B says A said Y.  According to the claim that makes my jaw drop it is against Wikipedia policy to go to the primary source where A said X and use that to show that what B claims (that A said Y) is forbidden.  I guess it should have been made more clear that Y is incompatible with X.  If A says X and B claims that A said Y, where Y is incompatible with X, I see no flaw in using the primary source to refute B's claim.  It is absurd to require that C has to quote A to refute B  before what A actually said can be referenced in Wikipedia to refute B's false statement.  A said what A said.  If B claims otherwise then A's original (primary source) words surely can be used to refute B.  --Minasbeede 19:38, 29 August 2007 (UTC)

A recurring problem. Suppose A said "labor notes will never work" and B said "A proposed a system of labor notes." Later, B became an influence on several political parties and governments; academics devoted intensive study to B's work, and not to A's, and used A-ism as a shorthand for labor notes. Jacob Haller 17:57, 29 August 2007 (UTC)