Wikipedia talk:Primary Secondary and Tertiary Sources/Archive 3

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Issues at hand

I believe most of those supporting the emphasis on reliable sources want to see:

• A clear emphasis on relevant sources (i.e. specifically written about the topic at hand)
• A clear requirement for up-to-date sources (e.g. sources that reflect current knowledge and views about a topic)
• A retention and clarification of the focus on reliable secondary sources.
• Deemphasis or deprecation of the distinction between secondary and tertiary sources.

I believe most of those opposing tight restriction on primary sources want to see:

• A clear allowance for peer-reviewed articles.
• A clear exception for especially reliable primary sources (i.e. census data)
• A common sense permission for using primary sources to provide illustrating or supporting facts and quotes.
• Acknowledgment that reliable secondary sources may be used in an unreliable fashion.

Some advocates of primary sources want to see:

• An allowance permitting primary sources to be used to counter secondary sources.
• Deprecation of source type distinction.
• Deprecation of no original research.

(Caveat) This may or may not be an accurate recounting of positions and issues, as I am subjective and biased.

I believe the first two groups of concerns are completely reconcilable and compatible, and in combination could represent a broad practical consensus. The third group of concerns is most problematic. Allowing editors to use primary sources to counter secondary sources would violate no original research, and is part of the abuse that lead towards policy reform in regards to source types. That would plainly permitting the advancement of unpublished positions and allowing Wikipedia to be used to "right great wrongs". Deprecation of the source type distinction would require a very broad revision of Wikipedia policy and guidelines, since the reliance on "reliable secondary sources" and "reliable third-party publications" is very well-ingrained throughout Wikipedia (and as such, such a discussion cannot be resolved here in a single policy). Additionally, since the distinction was further elaborated and discussed (in this policy) due to widespread abuse, removing such distinctions without a replacement to address the underlying issues would be little more than an invitation to acute abuse. Deprecation of NOR is not practical since it's one of the "non-negotiable three", so to speak. That last point is especially of concern to me, as while we should assume the best, efforts to weaken the policy from such a POV are (at the least) suspicious. Vassyana 18:08, 30 August 2007 (UTC)

I may be inclined to consider engaging in discussions about fine-tuning the policy by discussing group 1 and 2. Regarding group 3, I do not see these proposals having any chance of gathering momentum. The policy of NOR is one of the core policies and it is well established. We can fine tune, sure, but it would be a waste of time to challenge the principles upon which this policy was crafted. ≈ jossi ≈ (talk) 18:18, 30 August 2007 (UTC)

I agree completely with Jossi. In effect, people who would fall under group 3 are opposed to NOR - not just the text as it exists or existed a week or a month ago, but the core idea. Those people should not be involved in this discussion, as talk pages are for discussing improvements. Of course they have a right to express themselves, but - if you are right that there are people who fall into group three - the thing to do is create a new policy proposal page which is in effect a proposal to abolosh the NOR policy, and they can have a full discussion on that page. It is a different kind of discussion and should be kept separate from this one. Also, it is the kind of discussion that would require the input of hundreds of Wikipedians - abolishing a policy, even if it involves replacing it with something new (in fact, all the moreso) requires the active consideration of a huge chunk of the wikipedia community. (A discussion of how to improve the wording of a policy without changing the policy itself of course does not require the same scale of involvement and support). Slrubenstein | Talk 18:27, 30 August 2007 (UTC)

I also concur with Jossi. Group 3 goes to the heart of the policy, a policy that is well established and should not be challenged in those regards, since it is critical to the content of Wikipedia. Dreadstar † 18:40, 30 August 2007 (UTC)

Aren't there only two issues: (a) promoting secondary sources over primary sources and (b) the over-zealous application of NOR? I'm quite content to focus on (a) alone even though (b) is an issue with me. These aren't linked in my mind. I don't want primary sources put under a cloud and beyond that I don't think it's even proper for a policy page to ever do such a thing. The policy forbids certain types of editing. While I could argue with that I don't and won't, at least not now (possibly never.) The part of the policy that forbids OR in all its various aspects is sufficient. There is no need, in the policy, to single out any class of source for particular attention. I'll assume that there's been some rash of bad edits, all of which involved misuse of primary sources (that seems to be the claim.) The raw policy, without any mention of the class of source, is fully adequate to cover those. That there has been this (assumed) rash of events does not mean that the next person who uses a primary source will commit any wrong. I think it is improper to put that next person under a cloud. It also has to be obvious to all that Wiki lawyers will over-apply anything in the policy. It would put an unfair burden on those who use primary sources properly for the policy to single out primary sources, thus giving Wiki lawyers another club. Those who use primary sources improperly are hit by the general non-source-type-specific policy rules. That's enough. If guidance on the use of different types of sources is needed provide that outside the policy.

"The law, in its majestic equality, forbids the rich as well as the poor to sleep under bridges, to beg in the streets, and to steal bread." -- Anatole France

Here it seems that some want "the law" to single out one class for the prohibition of certain behaviors. --Minasbeede 18:24, 30 August 2007 (UTC)

I will have to disagree with you. Primary sources are inherently problematic in most, if not all instances when the subject is controversial. The potential for abuse of such sources is way to great to dilute the currently wording or the spirit of the policy. If you use a primary source such as Census data, in an article that is not controversial, no one will challenge that source on the basis of NOR... so there is no need to dilute the policy to "allow" these uses. ≈ jossi ≈ (talk) 18:35, 30 August 2007 (UTC)

OK, in some cases they are problematic. Do the non-source-type-specific rules cover the problems or do they not?

You do not actually have the power to declare what will or won't be challenged, do you? At best can't you just identify what it is that you won't challenge?

I'm not talking about any "dilution" of the policy, I'm talking about making it uniform. If some parts of the policy hit use of primary sources harder let it happen: we're both content. In reality I don't see anything useful about the current source-targeting language: it's like "if you drive a Corvette then you really, really, need to observe the speed limit." Perhaps some Corvette drivers do tend to speed more than other drivers. The speed limit is the speed limit and the law specifying the speed limit doesn't need to mention Corvette drivers. Attempting that makes writing the law harder: do you then also have to mention Mercedes drivers, Volvo drivers, etc.? --Minasbeede 18:48, 30 August 2007 (UTC)

Policies and not "laws" per se. They are designed to foster an understanding, and that understanding is related to what WP is and what WP is not. The current formulation of this page captures that understanding quite well, and yes, it can be fine-tuned, but not at the expense of making is easier to introduce OR via the improper use of sources, in particular primary ones. ≈ jossi ≈ (talk) 19:08, 30 August 2007 (UTC)

I understand the difference between Wiki policies and actual laws. I think my meaning was clear enough, and it should be obvious to all reasonable persons that I was not trying to force a complete equation of Wiki policies with laws, I was just giving an example and using that example to illustrate, outside the current issue, my impression of source-specific language in policies.

It is apparently your claim that what I propose (removing source-type-specific language) somehow makes it "easier" to introduce OR, "in particular primary ones." There's no issue at all (it's not a straw man, it's a vapor man) about the introduction of anything, since almost anyone in the entire world can at any time introduce whatever that person desires (I've reverted a few pages to remove garbage myself.) I have said, many times, that the policies, applied to all sources, can remain just as they are (or as they develop.) I'm not opening any door to anything and given the actual nature of Wikipedia there's no door at all anyway. The policies come in when they're used to justify removal of material that came in through that 1000-semi-wide entryway. I again assert: apply the policies. I continue to object to any blanket policy statements about any type of source. Let the policies be applied uniformly, let improper material of whatever sort be promptly dealt with (probably removed.)

In practice the policies aren't used in any sense that involves understanding, they're used as justification for removing material. --Minasbeede 19:53, 30 August 2007 (UTC)

As mentioned above, if you want to deprecate the emphasis on reliable secondary/third-party sources, that is a issue much larger than this single policy. That preference is well-ingrained throughout Wikipedia policy. Continuing to discuss that issue here will only distract from the issue we need to address for this policy. If you wish to make changes overall to that emphasis that is endemic to Wikipedia, you will need to address that on a community wide basis and get consensus to change multiple Wikipedia policies and guidelines. Vassyana 19:41, 30 August 2007 (UTC)

While I agree with Jossi, I have a different response to Minasbeede, which I think expands on Vassyanha's point: yes, some cars are more dangerous to drive than others, and it makes sense to act appropriately. This is not something the founders of Wikipedia gave much thought to, it is something we learned by experience. At first NPOV was a short statement, and there were few if any other policies. Over the years, experience has taught us that that short version of NPOV was not enough. NPOV itself became longer, more developed. And V and NOR were created. It is crucial to recognize that all of them not only have to be consistent with NPOV, they really are the offspring of NPOV. But we realized that inherent in NPOV was a distinct idea and that became NOR. And part of the experience that led to the development of NOR was the recognition that the vast majoriuty of NPOV violations involved the use of primary sources, and that encouraging the use of secondary sources could help protect NPOV. Now, does this mean there is a flat out ban on primary sources? No! Several years ago NOR had this clause, and I would not object if it or something like it were put back in the policy: "In some cases, where an article (1) makes descriptive claims that are easily verifiable by any reasonable adult, and (2) makes no analytic, synthetic, interpretive, or evaluative claims, a Wikipedia article may be based entirely on primary sources (examples would include apple pie or current events)." But you are right that there is a difference between primary sources and secondary, and that the former are riskier and therefore require greater care. Now, what is the best way to express this in a policy? I concede this is an open question and maybe we can improve the text of the policy in this regard. But this remains at the heart of the policy. Slrubenstein | Talk 19:56, 30 August 2007 (UTC)

You overlook the point: the speed limit is the same for all cars. The standards actually are the same for all editing, the source-type-language in the policy is superfluous, inoperative, and unnecessary. But the language does create a huge opportunity for "Wiki lawyers" to annoy those who happen to use primary sources in a fully appropriate way. Maybe some feel very strongly about abuses that have occurred by the improper use of primary sources. That's not a reason to clog the policy.

The answer to the question "Now, what is the best way to express this in a policy?" is: You don't. It's not necessary. --Minasbeede 20:24, 30 August 2007 (UTC)

I pretty much agree with how this section started. I'm going to think on it tonight and post something tomorrow, as maybe I have a different perspective on this, being a relative "newbie". The way I'm currently thinking, all of this makes perfect sense to all of you who have been involved in this "policy" for quite awhile, but it is awfully confusing for someone such as myself, and probably others. And then when we (newbie's) get hammered by the "wiki-lawyers" who take NOR to the extreme, we get even more confused and defensive. I noticed a part on the policy page that is confusing in itself (to me at least), especially when taken in the context of Wikipedia:About#Using Wikipedia as a research tool and other spots on the same page. wbfergus 19:32, 30 August 2007 (UTC)

There is another reason for treating primary sources differently that I want to mention, and, to keep the metaphor, this has to do with the kind of driver as much as the kind of car. Above (in threads that really ought to be archived, when this discussion is refactored so conversations are not going on simultanously in in thirty sections) COGDEN says correctly that one can write an encyclopedia based on primary sources. But those encyclopedias, e.g. Encyclopedia Brittanica, are written by PhDs or MAs supervised by PhDs, meaning, people who are specifically trained to do original research using primary sources. Anyone who wants to work with primary sources can do so, and do what others who do so do: submit an article to a peer-reviewed journal or to a project like Encyclopedia Brittanica, where a panel of PhDs ensures it is up to their standards. Wikipedia does not propose to abolish EB. But it does present itself as a radical alternative to EB in that it is not written by people trained to do original research, and articles do not have to be approved by a panel of experts trained in original research. Wikipedia is open to all. And the price we pay is certain content policies that you won't find explained in EB. This makes perfect sense to me: if Wikipedia is to say to the world, there is another way to write an encyclopedia, not by PhDs and not edited and supervised by PhDs, there ought to be something to compensate for the lack of credentials. What compensates is our policies. When someone says "How can you trust an article written by people who don't have the proper credentials and .. what?... there is not even any editorial board or peer-review?" our response is, instead we have our policies. And in this particular matter, NOR is essential. Slrubenstein | Talk 20:06, 30 August 2007 (UTC)

So you simple equate using a primary source with OR. Apparently you also claim that using secondary sources forestalls OR. I agree with neither. (In doing so I feel I must point out that I am not defending OR, even though very clearly my words here in no way defend OR.) --Minasbeede 20:19, 30 August 2007 (UTC)

Exactly! We do not have our own stable of experts, so we must rely on outside experts (by relying on reliable third-party references). Vassyana 20:40, 30 August 2007 (UTC)

Using primary sources clearly lends itself to original research. Caesar's Gallic War is considered an example of the heights of ancient reporting. However, it is also known as a masterwork of propaganda and known to contain inaccurate hearsay. Without modern reliable references to tell us what is accurate, what is propaganda and what is hearsay, we engage in original research by evaluating such claims for ourselves (and therefore using our own theories/judgments, instead of those of outside experts). If that is the case for a particularly notable, reliable and well-regarded historical source, how much more so for less reliable source material? Nearly all examples of truly (or exceedingly) reliable primary sources (such as census data) are widely regarded and noted as accurate in reliable third-party publications (in fact, that's how we know they're so reliable). Even then, such sources lack the context needed to craft an encyclopedic article, and are therefore best used in a supplementary fashion, rather than being relied upon for the article. Vassyana 20:42, 30 August 2007 (UTC)

Minasbeede, I think you are acting in good faith, but I think you are oversimplifying my position. I am making analogies and pointsl not simple equations. I admit, a beter analogy would have been the difference between a car and a truck (meaning, semi) (rather than say comparing a volvo to an SUV). I was trying to give you some insight into what motivated the creation of the NOR policy, I wish you would take it in that spirit. Slrubenstein | Talk 20:31, 30 August 2007 (UTC)

You don't need to do original research to use primary sources, or even to base an entire article on primary sources. If a lay person can understand the technical conclusions of a peer-reviewed journal article, then there's no reason not to cite the author's relevant conclusions directly from the primary source. There's no need to wait for those same conclusions to appear in Discover magazine or USA Today. There's also no reason not to quote—directly from the primary source—the words of some celebrity in a newspaper interview, or a novel, or a movie, or a non-technical government report, or an autobiography. You don't need a PhD for any of that. Of course, if you do need a PhD to understand any source—primary or secondary—there's a danger that by "dumbing-down" the source you are introducing original research. COGDEN 20:24, 30 August 2007 (UTC)

I'm afraid Vassyanna's description accidently misstates my position and may misstate others' positions as well, but I think this may be the first step to a broader agreement. To borrow his/her phrasing with minimal changes, I would like to see:

• A global warning about adding interpretations, etc., etc. but allowing explanations if needed and recommending citation of any explanations
• A clear emphasis on relevant sources
• A clear emphasis on up-to-date sources among secondary sources
• A clear allowance for peer-reviewed articles
• A common sense permission for using primary sources to provide illustrating or supporting facts and quotes
• Acknowledgment that reliable secondary sources may be used in an unreliable fashion.
• An allowance permitting the use of certain primary sources for certain topics. This may include:
  • peer-reviewed articles (as above)
  • especially reliable primary sources (e.g. census data)
  • the use of an individual or movement's own works (but not sacred texts) to describe an individual or movement's beliefs
  • An allowance permitting these types of primary sources to be used to counter secondary sources
• Either use more widespread definitions of primary/secondary, or use other terms than primary/secondary, or deprecate the source type distinction

As you can see, this includes elements of all three groups, as well as other concerns. I think all of us want better source-based research, but we disagree about which are the most common or most dangerous types of OR. Jacob Haller 20:34, 30 August 2007 (UTC)

We seem to be more or less on the same page. I've been putting some thought into your previous suggestions, as well as the primary/secondary distinction mess. Vassyana 20:40, 30 August 2007 (UTC)

Also, if I misrepresented your position, that was not my intention. I only meant to draw together the broad points that editors have expressed. Vassyana 20:42, 30 August 2007 (UTC)

I meant that your description didn't match my main concerns, or, I suspect, Minasbeede's or COGDEN's. Once we know that we have been talking past each other, and missing certain concerns, then we can address these concerns.

I often run across differences between how one movement describes its beliefs, how its opponents describe its beliefs, and how secondary sources describe its beliefs. In general, I'd prefer to rely on the movements' descriptions and avoid their opponents' descriptions; the secondary sources may support the movement, oppose the movement, be nonbiased, be nonbiased but depend on one side's sources, etc. so the primary sources are still useful.

I think that "primary countering secondary" should be allowed for (and only for) exceptionally reliable primary sources, special cases like authors and movements on their beliefs (for movements it may reduce statements about "all x" to "some x"), and misquotes, and requires absolute clarity in the primary source. Jacob Haller 21:16, 30 August 2007 (UTC)

We have an exception available to accommodate self-published sources (which are by default primary sources), in our verifiability policy. See WP:SELFPUB. This will address your specific concern about movements. ≈ jossi ≈ (talk) 22:55, 30 August 2007 (UTC)

Self-publication, movement ties, and primary/secondariness are three different issues; the first two depend on the source, and the last on both the source and its use. Certain edit wars have come up regarding the use of anarchist writings on the anarchism page and related pages. But the same issues affect any political philosophy.

• Proudhon's General Idea of the Revolution in the 19th Century is a multiply-published, (anarchist) movement, mostly-primary source, as are Bakunin's God and the State, Tucker's State Socialism and Anarchism, etc.
• Kropotkin's article on anarchism in the Encyclopedia Brittanica was a non-self-published, (anarchist) movement, largely-secondary source.
• Carson's Studies in Mutualist Political Economy is a self-published, (anarchist) movement, largely-secondary source.
• Most encyclopedia articles are non-self-published, non-movement, secondary sources.

One edit war had concerned the intro, and rival definitions, and this edit war had been repeated several times in the past couple years. I rechecked my sources, and tried to reach a compromise which the sources would support, and the minority objected to my use of primary sources, without disputing the content or relevance of the sources.

There were two concerns raised. One was the use of primary sources, which one or two editors insisted was absolutely forbidden, with most editors ignoring this. I had found several primary sources which contradicted the contested definition; I tried to craft an explanation of the primary sources' objections, which could fit in the intro without interrupting its flow, and the second objection was that this amounted to OR and/or IS. I asked people to list their concerns, and to propose various solutions, figuring that with a half-dozen major sources making similar arguments, we could reach something which summed up the arguments and which had as little OR and/or IS as possible.

I concede that cramping several essays into a couple words did involve interpretation. Still, it would have been better to work everything out than to have one side insisting that "primary sources are forbidden, therefore, we won't discuss ways to minimize interpretation," (without anyone else agreeing on their interpretation of NOR). That's how I wound up here. I had regarded my position as common sense and consensus until then; I still regard it is common sense and widespread practice.

Of course, this is how I remember the debate; not how the other parties may remember it. Jacob Haller 00:17, 31 August 2007 (UTC)

I believe that using specific content disputes to drive policy, is a bad idea. I have seen this again and again in policy discussions, and it does not work. Policy pages need to provide a framework that editors can use to collaborate and find common ground: How to apply policy is something that editors need to work out within the context of specific articles. Note that there is no mention on NOR about primary sources being forbidden. ≈ jossi ≈ (talk) 00:48, 31 August 2007 (UTC)

A good effort, but I still see problems. Books aren't peer-reviewed. Is "Origin of Species" excluded? Other scientific books? Columns in "Scientific American"? Also, what is happening here? Supposedly Wikipedia doesn't care about truth, only about citability. Why limit articles to those that are peer-reviewed? That's a limit placed on primary sources but not on secondary sources. When did the magic fairy tap every secondary source author with her wand and make them all reliable?

Part of the motive seems to be to avoid getting in to deep. The shore is in the opposite direction from how you are headed. --Minasbeede 20:54, 30 August 2007 (UTC)

I agree. There are some good ideas floating, but they should be hammered out in an essay first (maybe Wikipedia:Classification of sources), and then maybe promoted to a guideline. COGDEN 21:30, 30 August 2007 (UTC)

(ec)Does the policy actually limit articles to those that have been peer-reviewed? I didn't think it did, and I would agree it shouldn't. There is or should be a diffefrence too between NOR and V, which are policies, and RS, which is a guideline. In many wikipedia articles we ought to favor peer-reviewed articles, but certainly not exclusively. The difference between a primary source and a secondary source is not that one is peer-reviewed and the other is not. It is that they are used different ways. Slrubenstein | Talk 21:19, 30 August 2007 (UTC)

Okay, how would you identify reliable sources? Jacob Haller 21:16, 30 August 2007 (UTC)

This talk page is not about reliable sources; that is a separate guideline with its own talk page. We will never resolve any disputes if we keep bringing up separate discussions. Slrubenstein | Talk 21:19, 30 August 2007 (UTC)

WHAT? There's ALREADY a guideline page for "reliable sources"? Why, then, clutter up this policy at all with any discussion of sources? --Minasbeede 21:43, 30 August 2007 (UTC)

I think the rationale is that some people think that primary sources pose more of a "danger" to original research than secondary sources, and that it belongs here. I think the issue of whether a source is "reliable" is different from the question of whether new or non-obvious conclusions are drawn from the source. Unfortunately, the proposed reasons for treating primary and secondary sources differently have more to do with WP:NPOV and WP:Notability than they do with WP:OR. COGDEN 22:53, 30 August 2007 (UTC)

If it's what they think then that smacks of OR, doesn't it? In any case, that they think it and are well-intentioned doesn't make the language work. Does the language work?

I see below that "primary" and "secondary" here don't mean what they mean in the real world and that that may be the source of the problem. While the proposed archiving of the whole thing hasn't been done I'll leave off commenting on that last in adherence to my pledge of brevity. --Minasbeede 23:32, 30 August 2007 (UTC)

If this is just a gesture at levity, it is not that funny. And if you are serious I would have to question your good faith as you seem to wilfully ignore good faith attempts to explain the policy to you. "Reliable Sources" is one thing. "NOR," which depends on a distinction between primary and secondary sources which are different concepts from that of reliable sources, is another thing. If you truly do not understand that, you do not understand the policy and should request mentoring rather than engage in editing the policy. Slrubenstein | Talk 21:48, 30 August 2007 (UTC)

You didn't answer either question, and they were serious questions, if not asked in an utterly serious way. I have NEVER edited the policy and have no intention of ever doing so. Go look. I'm discussing, not editing.

I repeat: is there already such a guideline page? I repeat: if there is such a guideline page why clutter up the NOR policy page with any discussion of sources (because there already is a guideline page)?

"I don't care" may seem like a flippant answer but it does represent my view. If the language doesn't belong in the policy page then the motivations for putting it there, no matter how well-intentioned, make no difference. There is the NOR policy: this is the discussion page for that policy. I assert that the policy should be applied, I assert that it should be applied to material that arises from all types of sources. (I have removed my own edits because I realized they were OR.) I rather imagine that what I assert should be done is actually, most of the time, exactly what is done. (Perhaps I should have said that sooner. I apologize for the delay.) Even if primary sources are what is most often abused when the NOR policy is violated (for which my request for evidence remains unanswered: that claim continues to be blatantly unsourced: it's apparently OR, or worse) I don't think that any language in the NOR policy page is going to help the situation or make any difference - again, no matter how well-intentioned the advocates of such language are. Such language will, I think, cause grief for those who properly use reliable, published, primary sources. For that reason I oppose the language.

I am not trying to dilute anything, I'm not trying to open a back door for OR. Not mine, not that of anyone else. I'm trying to make the policy settle back and just be policy. If I wanted to use a primary source (let's say "Origin of Species," just for an example) I'd go ahead and do it. I'd attempt to use a reliable part of that source, I'd attempt to follow all policies.

Could I at least have an answer to my two questions? Is there any possibility that I might have a bonus: some actual examples of how primary sources have been abused? --Minasbeede 22:23, 30 August 2007 (UTC)

Slrubenstein answered the questions on my talk page, for which I thank him.

Answer (1): see Wikipedia:Reliable sources

There is a Wikipedia page on reliable sources, it is a guideline, not a policy. Shouldn't a great deal of the primary-secondary material be there (and only there)? --Minasbeede 23:40, 30 August 2007 (UTC)

Oops, saved too soon.

""NOR," which depends on a distinction between primary and secondary sources which are different concepts from that of reliable sources, is another thing."

I dispute that statement. I don't see that NOR in any way depends on any such distinction. The test for NOR is whether or not it's the editors own work. The test does not and can not extend to the sources the editor uses (except for the trivial case where the source is the work of the editor. There's some further quibble language, but it doesn't affect what I say.) Original research is forbidden, citation is both allowed and demanded. Citing a primary source is not OR. (Again, with quibbles, but I assert all such quibbles are fully covered by NOR without any source-type-specific language.)

Another request: if there's ever been any case in which the source-type-specific language was essential to the enforcement of NOR please show it to me. --Minasbeede 22:33, 30 August 2007 (UTC)

You may not agree with that, but NOR depends primarily on such distinctions between primary and secondary sources. ≈ jossi ≈ (talk) 22:57, 30 August 2007 (UTC)

I don't see how you can say that, unless you define primary source as anything which, when cited, produces original research. Using traditional distinctions between primary and secondary sources, either one can lead to original research if interpreted in a way that nobody has ever done before. It's not the type of source that matters, it's how you use it. If I cite Discover magazine as a secondary source summarizing the Theory of Relativity, and then use that citation to support my personal theory of antigravity, I've committed the act of original research. It can happen just as surely with a secondary source as it can with a primary—sometimes even more easily. COGDEN 23:06, 30 August 2007 (UTC)

It is obvious that someone can do OR with secondary sources. The emphasis is on the fact WP:OR has specific language to address the necessary caution needed when using primary sources, in particular when secondary sources are available. ≈ jossi ≈ (talk) 23:10, 30 August 2007 (UTC)

Regarding peer reviewed articles, my understanding is that such sources are not considered primary sources in Wikipedia, and are well accepted as sources for articles providing that claims made are attributed to the scholar/author, and providing that WP:NPOV#Undue weight is not breached. New editors sometimes forget to appreciate that our co0re content policies do not stand-alone. We seek compliance with V, NOR, NPOV simultaneously. ≈ jossi ≈ (talk) 23:04, 30 August 2007 (UTC)

You are taking the fact that peer-reviewed articles are acceptable, and inferring that "therefore, they must not be primary sources". But you are assuming that primary sources are, by definition, unacceptable. The fact is, journal articles are almost always primary sources under traditional usages of the term (because they are the first published source of new ideas and original research). They happen to be primary sources that Wikipedia accepts and encourages, and which aren't "rare". COGDEN 23:11, 30 August 2007 (UTC)

In Wikipedia we do not use the "traditional" usage of these terms, and that may be the core of the problem in this discussion. ≈ jossi ≈ (talk) 23:19, 30 August 2007 (UTC)

Proposal

By my count, we are fairly evenly split on keeping the current version in place or reverting back to the version I've pointed out. Since there is no consensus for keeping the disputed recent changes, I think we should revert back to the version I've suggested and move forward with discussion from there. At that point we can archive most of this talk page, unprotect the article and have a fruitful discussion as to changes. Right now, we're still in an edit-war state which is totally unacceptable. Edit warring is not the right method for making changes to any article, much less policy.

I'd like to clear all this up, so we can possibly set up an RfC to draw other eyes, and this page is entirely too convoluted and argumentative - mainly due to the method used to make the changes, edit-warring. That needs to be stopped.

I propose we unprotect the article, revert back to the pre-edit-war version, archive this talk page and then discuss the changes proposed, as is the standard and best practice for dispute resolution. Anyone disagree? Dreadstar † 16:26, 30 August 2007 (UTC)

I don't think anyone disagrees, but the concern seems to be how to "organize" this discussion page so it doesn't get as convoluted again with 2 or 3 more days. If that can be organized and presented in a fashion acceptable to others, then I think that everybody would go along with it. But just reverting the main policy page and allowing this page to remain as is is just asking for more confusion and trouble. wbfergus 16:46, 30 August 2007 (UTC)

I think it's been clearly shown that the previous version that stood for several months had consensus, so it is a good starting point. However, it's also clear that people on both sides of the debate have varying issues with the current formulation, though these concerns are not mutually exclusive. I feel the willingness to compromise has been very one-sided.[1] Unprotecting the page will not be problematic unless people continue to insist on pushing through major changes without discussion and consensus. Vassyana 17:25, 30 August 2007 (UTC)

You are treating this policy article as if it were a Wikipedia content article. It's not, and there is no policy "inertia" for statements that at one point seemed to have consensus, but clearly do not after wider exposure (see WP:ATT for case in point). Policies must describe, not lead, Wikipedia practice WP:Policies and guidelines. There is no place in a policy for controversial statements, and the fact that things are "evenly split" means that there is no consensus and the controversial statements must be removed. OR is one of the pillar policies of Wikipedia, with the potential to do wide-spread damage to the Wikipedia institution if we get it wrong. If any part of OR is clearly shown not to have a present consensus, or not to describe Wikipedia practice, it should be removed immediately. COGDEN 20:05, 30 August 2007 (UTC)

In that case, the proposal fails, since it seems the edit war would only be continued if the previous version were to be put back into place. Further warring is an unacceptable outcome.Dreadstar † 22:15, 30 August 2007 (UTC)

COGDEN is quite wrong about inertia. For obvious reasons, articles change several times a day. For equally obvious reasons, policies should change at a much slower rate. And discussion on this talk page should be about ways to improve the text of the policy. Anyone who wants to change the policy need to present a proposal to the whole community by creating a new proposal page and publicizing it. Slrubenstein | Talk 10:12, 31 August 2007 (UTC)