Wikipedia talk:No original research/Archive 22

Archive 15

←

Archive 20

Archive 21

Archive 22

Archive 23

Archive 24

Archive 25

Before we unprotect the page

We need to resolve certain disputes. Based on the past week's discussion and comments from vaious people, I am singling out two specific disputes, which is more than enough to work on for now. Once we resulve these, people are free to raise other issues which we can work on, one at a time, until we resolve them.

Based on discussion, I have added a third topic: if we have a distinction, what names should we use?" I really think we need to resolve these three disputes before moving on. Slrubenstein | Talk 21:29, 31 August 2007 (UTC)

• General Discussion of the Status of Wikipedia Policies is not appropriate here, but rather at the village pump or on the [wikipedia listserve] - if you are not a member it is easy to register, just follow the link.
• Opposition to NOR Anyone who is opposed to an NOR policy should develop, perhaps in consultation with like-minded editors (on your own talk pages or subpage), a specific proposal either to abolish the policy or replace it, and create a new proposal page in which you present your proposal and basic rationale, and invite discussion. You can do so by providing a link to your proposal page here, on the talk pages for NPOV and V, and at the administrator's bulletin board, and other places as well. E.g.: Wikipedia:Proposal to replace No Original Research - use this if you believe that there are forms o original research that should be allowed on Wikipedia.
• The quality of sources please discuss this at Wikipedia talk:Reliable sources, not here. As that guideline changes, we can discuss its impact on V and NOR on their respective talk pages.
• Other issues with the current text of this policy Please "Watchlist" this page, be patient, and introduce new issues to this page after we have resolved the current conflict.

Well, there are currently two "quality of sources" locations: the actual guideline article and the NOR policy page. Much of the current controversy arises from having both a guideline and a separate policy section. Apparently the guideline doesn't have enough compelling force (but that's a personal interpretation that I'd welcome having replaced with a better characterization.) That dual-location situation is in part what creates a "can of worms." --Minasbeede 14:51, 31 August 2007 (UTC)

Let me just add that those who entirely oppose any original research policy will ultimately fail, given that Jimbo has expressed support for such a policy several times, and I think there is a long-established consensus. Jimbo wouldn't allow that to be changed. COGDEN 20:02, 31 August 2007 (UTC)

The issue isn't as much one of absolutely opposing the policy as it is of seeing flaws in how it is applied. An analogy. A relative works for State Farm. The relative states that the instructions are to provide a settlement to the policy holder if it can be found to be justified by the policy. In other words, if the insurance policy allows the payment, make it. There seems to be a human tendency that goes the other way: if the Wikipedia NOR policy can be used to justify removal, remove. (I'd hazard a guess that some insurance companies also act in that manner: if they can find a way to avoid making a payment they use it.) Beyond that I have no comment - I truly do not want to discuss this topic at this time (and maybe never.) --Minasbeede 20:10, 31 August 2007 (UTC)

If someone is misapplying this policy it is more a matter for their personal talkpage than this one. The policy can only do so much. We are still obliged to educate other editors one-on-one to bring them to a more complete understanding of the goals of Wikipedia. I understand that this talkpage is often a place for such education. And most of the time that works out fine. However, while this policy is protected we must focus on resolving the issues at hand and redirect the secondary uses of this talk page elsewhere.--BirgitteSB 20:15, 31 August 2007 (UTC)

That speaks past what I was saying but it doesn't matter. I don't want to discuss that, I just wanted to point out that it is false to characterize my position as my being totally opposed to NOR. I don't want to discuss that. Nor does Slrubenstein want that discussed here, and he's correct at this time and has my full support. All mention of this issue could be archived, to great benefit. When "this time" has passed I still don't want to discuss it. You're safe from that. I did put a short essay over in the page he created. (Sorry, SLR, I hope this is the last time.) --Minasbeede 22:38, 1 September 2007 (UTC)

Should the NOR policy distinguish between primary and secondary sources?

I have now slept on it and feel even more strongly than before that the NOR policy must make a distinction between primary and secondary sources. I think Slim Virgin is right (and I do not always agree with her and long questioned her on this). NOR needs to explain what "original research" is, and what it isn't and I do not think there is any way to do this without discussing primary and secondary sources.

Some people do not understand the distinction or think it opens up a can of worms but I now believe that is because they think too literally. What I mean is, some people think that things are what they are. I do not mean to go into a long philosophical debate but Wittgenstien and Pierce and Dewey - generally acknowledged to be among the most important philosophers ever in the US and Europe - argued that the meaning of things including words depended on how they are used. To a large degree, what makes something a primary or secondary source is, how it is used. Primary sources can be used to establish facts without violating NOR, but they cannot be used to make analytic, synthetic, interpretive, or explanatory or evaluative claims without violating NOR. Secondary sources by contrast are the only sources we can use to make analytic, synthetic, interpretive, explanatory, or evaluative claims (as long as they are the claims explicitly made by the source itself).

We can, did, and I think still should provide examples of primary and secondary sources, but they are just examples. The essence is in how they can or cannot be used.

This by the way means it is conceivable for a text to be either a primary or secondary source - depending on how it is used. Let's take On the Origin of Species for example - I grant this may not be the best example, but I am just looking for a handy hypothetical. In it Darwin makes all sorts of analytic, synthetic, and explanatory claims about natural history. As a document concerning natural history, it is a secondary source ("According to Darwin, variation in finches on the Galapagos is best explained by ..."). But in it he also expresses his own beliefs and judgements. As a document about Darwin, it is a primary source ("According to Janet Browne, when Darwin wrote On the Origins of Species he was responding to ...."). In the article on Evolution we can use it freely as a secondary source (although evolutionary theory has gone far beyond Darwin, evolutionary scientists still consider the case he built to support his theory valid). In the article on Charles Darwin it is a primary source and while we can use it for facts, if we want to make arguments about what Darwin was like or what he meant, we should draw on appropriate secondary sources by historians of science, intellectual historians, and biographers of Darwin. I do not see this as a can of worms. I just see this as a more sophisticated way of thinking that understand things in terms of the ways they are used, and understands that appropriate uses depend on the context. Slrubenstein | Talk 10:43, 31 August 2007 (UTC)

I think distinguishing between the three main classification of sources is useful, but I also wonder about how they are included on the main NOR "Policy" page. It is nice and very useful to have them wiki-linked to the appropriate page with the full-blown description, but I question the "nutshell" approach to the descriptions on the "Policy" page. I looked over the "nutshell" on the "policy" page and the full-blown linked articles, and there are differences, though at this point in time they are minor. If those linked pages ever changed, even just a little bit more, there could be enough of a difference between the "official" definition and the "nutshell" definition on the "Policy" page, that it seems that someone could easily claim the "nutshell" definition was itself "original research", making an inference that wasn't present in the "definition" page. wbfergus 12:25, 31 August 2007 (UTC)

This is one of the reasons that a few of us felt that it might be useful to have a seperate, possibly "policy", classification of sources in project-space. For the purposes of official policy and guidelines, stable definitions that are consistent across disciplines is more useful than a reference to the general definitions that are either restricted in application or change meaning unpredictably across different domains. I would suggest further development and adoption of that page, and then referring to that from guidelines and policies (including NOR). SamBC(talk) 12:33, 31 August 2007 (UTC)

This Darwin example seems clearly wrong to me. Darwin would be a terrible source to use for an explanation of evolution in the Evolution article. On the Origin of Species is a primary source for all practical purposes. It is a seminal book written 150 years ago, which no longer summarizes the state of scientific understanding of evolution. It should of course be discussed in the evolution article, but in the context of discussing the history of the concept of evolution, not to explain evolution today. Similarly, in the Charles Darwin article, Origin should be used as a source to explain Darwin's views on evolution. I'm not sure I understand the distinction you are making here - in both cases, Origin is a source for explaining Darwin's views. john k 17:30, 31 August 2007 (UTC)

Well it was chosen in haste and perhaps a bad choice, although I think you misunderstand. If someone wanted to make a claim about variation among finches, I would argue that Darwin's book is a valid secondary source. But there is no need to belabor this one example. Can you think of any texts that could be a primary source in one context, and a secondary source in another? That is the issue. If you can't think of any, I would just take that as a sign that the boundary between primary and secondary source is even clearer than I thought! What is your answer to the question that heads this section? You do not have to take a position on what I say; the question is your position on the current policy. Slrubenstein | Talk 17:37, 31 August 2007 (UTC)

Folks, for some this is a very contentious issue. I think we first need to resolve the basic question of whether the policy must or should make this distinction first. Then discuss editorial issues. Sambc, if you want to, creat a policy proposal page and invite discussion of your idea there. But here, the question is simple: should NOR distinguish between primary and seconday sources? That's all! Slrubenstein | Talk 12:38, 31 August 2007 (UTC)

To a certain extent, people seem to react differently to "whether the policy … make this distinction" depending on what the distinction is. I would say that yes, it should me the distinction, purely because avoiding original research is best helped by different strategies in each case. However, the definition needs to be appropriate and stable, wikipedia-specific but based on real-world definitions. Is that acceptable supplementary material in this discussion, or is that still being too specific? SamBC(talk) 12:52, 31 August 2007 (UTC)

In regards to MinasBeede's point about the actual "terms" primary and secondary sources differing from the non-Wiki world, how about re-phrasing them to "Primary Wiki Sources" and "Secondary Wiki Sources", etc, so that a distinction is more readily made? I agree that the types of sources need to (or at least should) be included in the "policy", but the definitions themselves should not be re-created on the policy page, unless somebody can devise a way so that whatever the "official" definition pages get edited to is always automatically included verbatim. This would really clutter up the policy page though, and I really see no benefit that the simple wiki-linked article doesn't already cover. In short, yes make the distinguishment between types of sources with just a link to what the term is defined as, for consistency. Offer examples of OR, why OR is bad, etc., but do not try to re-create a definition here when it can easily change from the linked article/definition. This page is ONLY about NOR (or what OR is), not about what type a source is or what a "primary source" is. Keep it simple, keep it consistent, and there will be far fewer interpretations of what it's "supposed" to mean or how it "should" apply. This is all very similar to what I do in real life, designing databases and getting just one simple description/rule for each and every field in every table. Everything is defined only once, though it may be linked or cross-linked in numerous places. wbfergus 13:58, 31 August 2007 (UTC)

• While have I doubts about whether the terms "primary" and "secondary" are wisest choice to describe the distinction, I have no doubts that such a distinction between types of sources should be made. Although in reading the intro to the question above, I do like the idea of simply defining these terms by how they can be used rather than trying to define them by example and then dictating how they can be used. It clarifies alot of the things I find problematic about the terms.--BirgitteSB 14:30, 31 August 2007 (UTC)

But that's another aspect of the "can of worms." Either there's a good, useful description of what makes a source "primary" or "secondary" (in the Wikipedia sense) or there isn't. If there is a good, useful description then that description alone is sufficient. If there isn't a good, useful description then every source must be analyzed (by someone) to determine if it's "primary" or "secondary" (in the Wikipedia sense.) That's hard to distinguish from "arbitrary" - which "someone" is allowed to make the distinction? In any real case there will be a Wikipedia editor who has introduced some material and another who wants to alter or remove it because it is OR. If it's OR because a ""primary" (in the Wikipedia sense) source was used then how does the first editor distinguish between policy and prejudice? --Minasbeede 15:12, 31 August 2007 (UTC)

I think the basic issue of NOR is that it's difficult to explain. Personally, I feel like it's kind of a "I can't really define it, but I know it when I see it," kind of thing. I've previously indicated that I think the idea that novel synthesis is banned by NOR is pretty ridiculous - an encyclopedia article is by definition a synthesis, and doing things like summarizing books, or trying to describe the state of the literature on a subject, are bound to involve some degree of original synthesis. It is novel analysis that we need to be careful about, and this is what we shouldn't be able to use primary sources for. The classic OR using primary sources is essentially to go to, say, the text of some primary source and then to quote that text, out of context, to support some new argument which is not explicitly made by the primary source. But I don't know that this is something that is unique to primary sources - one can easily do the same thing with secondary sources. I think the key issue here is not that primary sources are forbidden for any particular purpose. It's that, in evaluating primary sources, we need to stand by the judgments of reliable secondary sources. john k 17:36, 31 August 2007 (UTC)

Go ahead and discriminate between primary and secondary and see where it takes you, but when you start juxtaposing types of sources and drawing judgments as to which is superior, you start moving toward original research. If you find contradicting evidence, you should cite all the relevant sources, then let the reader choose which one is right. Wikipedians can analyze whether a source is reliable, but once we deem a source reliable, we have no business comparing the relative reliability of two reliable sources. We can only represent perspectives based on the prominence of their views. If you simply find two sources with contradicting evidence, cite both (including the one that you think is wrong). If the contradiction is basic, like a number difference, go ahead and acknowledge that. If the contradiction requires complex reasoning to reveal, do not articulate the nature of the unrevealed contradiction (unless, of course, a source reveals the contradiction). —Kanodin 18:41, 31 August 2007 (UTC)

If we have strong evidence that one is wrong, we should not present it. This should really all be common sense, not some kind of mystical activity. As writers and editors we must make judgments of our own about what sources are better than others. That's fine. We just need to be careful to be able to source it, and if someone comes along challenging it, the challenge needs to be either accommodated or rejected on the basis of what reliable sources say. I think the basic issue is that we shouldn't draw our own inferences from texts. This tends to be done more with primary sources, but can be done with anything. john k 21:39, 31 August 2007 (UTC)

Summary so far, for this section/question

Well, so far it appears that everybody is in agreement that a distinction should be made in regards to the types of sources. I didn't see any objections above that a distinction should not be made. The only objections I see (maybe I'm biased or blind), is in the wording of how to label the types of sources, and whether or not the "Policy" page should attempt to offer a definition of the types of sources.

Personally, I feel that the details of what the various "source types" are should be defined on the separate pages already created, and this "Policy" page merely point out that different types exist and here's the link for the definition. This page is solely about NOR and should not expand (rightly or wrongly) into the realms of other subjects, no matter how contentious or how badly needed. When it's necessary to delve into those realms, link to that subject appropriately and say something like "For more clarification or discussion, see ...". Keep each "Policy" and/or "Guideline" focused only on that subject without blurring the lines of distinction. wbfergus 17:49, 31 August 2007 (UTC)

Watch out for that word "everybody." I don't agree. But if it's to be done I believe the first thing that needs to be done is to clean up the terminology. "Primary" is defined in the Wiki policy differently than it is elsewhere. That being the case (and whether or not the word is or isn't used absolutely consistently everywhere else) I think some change in terminology needs to be made. The difficulty which seems to be continually unaddressed is that the Wiki definitions of "primary" and "secondary" don't seem to be solid enough so that the definitions themselves could be used in place of partial enumerations. (That's a "can of worms," but some people clearly want to have that "can of worms.")

I also favor moving source type definitions elsewhere and having a compact, cogent policy. There are clearly those who do not agree and will block any such move. Well, OK. But please clean up the language even if you never deal with the real problems - and soon. Maybe someone could re-edit the existing policy and substitute "purple" for "primary," "sienna" for "secondary," and "turquoise" for "tertiary." That's nonsensical, but at least it helps eliminate the conflict/confusion that comes from using terms that mean other things elsewhere - with confusing overlap of meaning from Wikipedia to the rest of the world.—Preceding unsigned comment added by Minasbeede (talk • contribs)

You keep saying you object to the words primary and secondary but so far I have not noticed your suggesting an alternative. If there is no alternative, we stick with primary and secondary. Slrubenstein | Talk 21:18, 31 August 2007 (UTC)

I'd take the material completely out of the policy. If the terms are to be kept as is then here's a draft of text that might be added to the policy:

Wikipedia distinguishes between different types of source and uses that distinction as a starting point for determining if a source is appropriate for citation by Wikipedia articles. Wikipedia uses the words "primary," "secondary," and "tertiary" (see below) to categorize different types of source. Please note that while these words are also words used outside Wikipedia for the distinction of different types of source the meanings of the words within Wikipedia are not precisely the same (as detailed below) for use within this policy.

It's a draft, I'm not wed to any of it. --Minasbeede 21:58, 31 August 2007 (UTC)

That makes sense, although I continue to advocate the addition of a separate (short) policy defining primary, secondary, tertiary, for ease of reference from other policies/guidelines. I might be suffering from a misconception regarding the need to refer to those terms elsewhere, though. SamBC(talk) 01:24, 2 September 2007 (UTC)

Me, too. I think the existing "Wikipedia:Reliable Sources" already covers and somewhat overlaps (and it shouldn't "somewhat overlap," the two should be parallel) the source-classification parts of the NOR policy page. So I'd advocate making "Wikipedia:Reliable Sources" say whatever it is that some desire to say about sources, be structured so that the "bottom-line" parts are easily referenced from within the NOR policy, and then reference them. A huge bonus would be that all the discussion about sources would then be away from NOR and instead on the "Reliable Sources" talk page. Look at the history of this talk page for a few thousand posts back and you'll appreciate what I'm saying. If all the source classification stuff is in one place ("Reliable Sources") then the issue of being parallel disappears - another bonus. --Minasbeede 01:34, 2 September 2007 (UTC)

We should make a distinction between two types of sources, but use different names

Can the (Wikipedia) definitions of "primary" and "secondary" be made complete enough and thorough enough that they suffice to define "primary" and "secondary," as used/needed by the NOR policy? --Minasbeede 20:29, 31 August 2007 (UTC)

(copy-paste of my own previous response to implication of same question above) Well, specific words to use as terminology aside, what do people think of the definitions given at Wikipedia:Classification of sources? Could that be used to try to develop a consensus-based wikipedia-specific definition, while leaving the question of what words to use seperate, for now? SamBC(talk) 20:42, 31 August 2007 (UTC)

That depends on whether primary sources are required to be "rare", and whether secondary sources are required to be "primarily relied upon". We shouldn't constrain Wikipedia practice unless we fully define what the terms of the constraint. A better idea would be just to delete the section on primary/secondary sources, and work out definitions and distinctions at Wikipedia:Classification of sources. The only thing we have a consensus on now is that sources not be used to support or create original research. If we can (1) agree what primary and secondary sources mean, and (2) agree on specific instances where primary sources are preferable, or where secondary sources are preferable, then we can work towards a guideline to that effect. COGDEN 22:29, 31 August 2007 (UTC)

Wikipedia:Classification of sources uses nonstandard definitions of primary and secondary sources, particularly with regard to ancient historical sources, most of which are secondary according to the essay, and primary according to scholarly practice. I like its discussion of the POV issues concerning both primary and secondary sources, and role thereof, but fear its dismissal of the POV issues embedded in many tertiary sources. I think the emphasis on third-party sources is misleading; for describing beliefs, though not many other purposes, internal sources are better. Jacob Haller 22:42, 31 August 2007 (UTC)

Egad. Another question: just how many places (and what are they) in Wikipedia is there an attempt made to classify sources? We have NOR, the Reliable Sources guidance page, and Classification of sources. Any more? (I'd ask if the definitions in "Classification of Sources" agree with those in NOR but as a person able to read that would be sort of wrong of me.) (See my talk page.) --Minasbeede 22:49, 31 August 2007 (UTC)

WP:SCLASS isn't finished, although contributions have slowed down (to say the least). If you think it needs work, but could be useful otherwise, then contribute. Discrepancies with real-world uses seem, however, to be necessary because real-world uses vary and we need a consistent use, a consistent definition, that can be applied to everything wikipedia covers, more or less. SamBC(talk) 23:56, 31 August 2007 (UTC)

Discrepancies with real-world usage my be necessary but I was asking about internal consistency within Wikipedia (which I could have found out for myself and which, true to form, I haven't yet done. Now I'm thoroughly shamed into doing it.) If discrepancies are necessary (with real-world usage) I think it is at least as necessary (as having such discrepancies exist) to point these out in order to lessen confusion. If terms were used that were not exactly the same as real-world terms the problem would not exist and I advocate that the terms used be different if the definitions are different. --Minasbeede 00:13, 1 September 2007 (UTC)

I think the first order of business would be to stop using the words "primary" and "secondary" in a way different from how they are used outside Wikipedia. Choose/develop a different way to distinguish the usages from the Wikipedia point of view.

(Something about NOR I decided to move to my talk page, where it may languish and die, was here.) Minasbeede 13:32, 31 August 2007 (UTC)

I agree (and I've seen discussions on other users talk pages as well), that this is an issue. However, let's try to tackle just a couple things at a time. Once the above points have finally been agreed upon (regardless of which terms are used), we can open it up to deide on which terms to use. This may also neccessitate the "moving" of the current articles on primary, secondary, and tertiary sources as well. There several different issues that need addressed before the page (policy) can (or should) be unprotected, so let's take it in small, manageable chunks at a time. wbfergus 17:06, 31 August 2007 (UTC)

It's not entirely different to how it's used elsewhere, and the fact is that its use elsewhere isn't entirely consistent (see recent discussions on the use of the term referring to 'hard' science material. I think it's an awkward question. Anything else would seem neologistic or just plain awkward (such as the suggested wiki-primary-source or primary-wiki-source). However, using the familiar times is liable to cause confusion as people won't realise we have more precise wikipedia-specific meanings. My gut says to use the normal terms, and have a policy defining them in a way that isn't completely different to any of the real-world uses. Link it from every policy and guideline to which it is appropriate, and no-one will stay confused as to the meaning for very long. SamBC(talk) 17:10, 31 August 2007 (UTC)

It may well be different terminology would be awkward - but the current situation is awkward. If there's external confusion about what "primary" and "secondary" mean that doesn't license the use of the terms differently in Wikipedia and it also confounds the confusion. If Wikipedia has special meanings then it would seem Wikipedia should have special terms. In a matter of policy I don't think there need be an absolute bar to a neologism, although avoiding neologisms would be best, if it can be done. The goal is (I'd think) to communicate accurately and succinctly, with as little room for confusion as possible. Either it is possible to accurately define what (Wikipedia) "primary" and (Wikipedia) "secondary" mean or it is not. I stuck the "(Wikipedia)" on just to identify the terms to which I refer. So far things seem to be such that it is not possible. If so, this whole topic is dead in the water. (Gladly proven wrong.) --Minasbeede 19:59, 31 August 2007 (UTC)

Well, specific words to use as terminology aside, what do people think of the definitions given at Wikipedia:Classification of sources? Could that be used to try to develop a consensus-based wikipedia-specific definition, while leaving the question of what words to use seperate, for now? SamBC(talk) 20:38, 31 August 2007 (UTC)

I like Reliable Source and Auxiliary Source, per Vassyana; however, auxiliary sources should be reliable too, so I suggest Core Source and Auxiliary Source. Jacob Haller 20:37, 31 August 2007 (UTC)

I think those words are simply too loaded, and feel like they refer to the role of the source within the article rather than the nature of the source or the usage thereof. SamBC(talk) 20:39, 31 August 2007 (UTC)

OK, then the zeroth order if business is to decide whether the terms, as used in Wikipedia, are meant to apply to the role of the source in an article or to the nature of the source or to the usage thereof (if that's different from the others), something else, or some combination of these. --Minasbeede 20:45, 31 August 2007 (UTC)

I had suggested moving from source classification to claim classification but that would require too much work. Anyway, I like "Core/Auxiliary" because it defines two categories of appropriate use, and then we can debate what (traditionally-defined) primary sources go in which categories (for given topics) and which secondary or tertiary sources go in which categories. It distinguishes wiki-suitability from secondariness. Jacob Haller 21:04, 31 August 2007 (UTC)

In that case, core/auxilliary is a seperate axis to primary/secondary and can't really be used instead, although their use may help to clarify different standards of suitability. It seems to me that they distinguish which sources are being 'relied upon'. SamBC(talk) 21:17, 31 August 2007 (UTC)

Well, we can't talk about whether the secondary/primary distinction matches, or more-or-less matches, the core/auxiliary distinction (i.e. whatever does qualify sources) unless we use two sets of words. I think the current policy reflects appropriate considerations for core sources and auxiliary sources, but describes them, sometimes correctly, sometimes confusingly, as secondary sources and primary sources, respectively. Jacob Haller 22:07, 31 August 2007 (UTC)

There's no one simple usage outside of Wikipedia, which is part of the problem. But for our purposes, can't we just focus on the terms "primary" and "secondary" and just say that a primary source is a source that is "primary"? In other words, an original source? A secondary source is a source that is "secondary". In other words, not primary. Of all the ways the terms are used outside of Wikipedia, that seems to be the core idea. The original source of an idea, conclusion, or set of data is the primary source, while any source of that same idea, conclusion, or set of data that isn't the primary source, is a secondary source.

We could also switch to neologisms, such as "reliable source" and "auxiliary source", but I don't see the point, since reliable source is already well defined at WP:RS, and the new term auxiliary source would just be something that isn't a reliable source. COGDEN 21:15, 31 August 2007 (UTC)

If you're concerned about the restriction of sources, that would worsen the situation. Many secondary sources, and even many tertiary sources (particularly signed articles), engage in original research. I don't believe this policy was intended to limit the citation of original research published by reliable third-party sources (in fact, WP:NPOV, WP:V and WP:RS would contradict such a position). The focus of this policy is prohibiting original research (and to an extant, just bad research) by editors of Wikipedia. Additionally, historical sources were a large part of the abuse. Any definition that allows the direct citation (without the support of secondary sources) of the Bible, Tacitus, Freud, Murray or other ancient and obsolete sources is severely problematic. Promoting discredited and obsolete theories is no less original research than promoting a novel one, and is especially troublesome since it contradicts modern reliable publications. Vassyana 18:50, 1 September 2007 (UTC)

If it's a discredited theory, original research is not an issue. We'd be talking in that case about WP:NPOV and WP:FRINGE. Original research, as Jimbo has described it, means original ideas that haven't (yet) been published in a reliable source. COGDEN 00:04, 3 September 2007 (UTC)

How about "sources of facts" and "sources of interpretations"? --Coppertwig 22:52, 31 August 2007 (UTC)

Should the NOR policy explicitly encourage or privilege the use of secondary sources?

I would say, not exactly. It should give advice on the appropriate use of primary vs secondary sources, as primary sources are much more likely to lead to inadvertent synthesis, but can be used without it being OR. SamBC(talk) 12:34, 31 August 2007 (UTC)

"I" think the policy should be modified very slightly , so that the word "should" is italicized (since the entire sentence is already in bold), to help distinguish the emphasis. I don't see the wording anywhere (perhaps I missed it?) that articles "must" rely on secondary sources, nor do I think it should. To me, making any more of distinction of endorsement would seem to contradict the definition. This policy should clearly describe "Original Research", and only that, without branching out into endorsements of one type of "source" over another, or possibly creating new definitions of a "source" that may differ from what it links to. OR can easily be created from all three types of sources, though it may be more likely to originate from primary sources. An article based solely on secondary and/or tertiary sources, unless strictly an entire copy and paste from those sources (is that plagiarism?) would be edited somewhat, correctly or incorrectly. During the editing process, condensing the article, or re-arranging the contents to allow the data from multiple sources to co_exist in a coherent flow, etc., the article can easily change the meaning or conclusion from the published source. I've seen this several times on various Military History articles.wbfergus 12:52, 31 August 2007 (UTC)

I suggest going in the opposite direction. At the top of a policy page it says, in part, "This page documents an official policy on the English Wikipedia. It has wide acceptance among editors and is considered a standard that all users should follow" (emphasis added.) So the word "should" is already there and applies to everything in the page. The second appearance of "should," in the sources section, is redundant. Rewrite the section so that the blanket "should" suffices. (Of course I actually advocate moving the entire sources section into "Wikipedia:Reliable Sources." I'm more making a point here. It's a good point.) --Minasbeede 20:52, 1 September 2007 (UTC)

This makes sense. Would it, in that case, make sense to pare down the policy to a definition of original research and why it's bad, possibly with examples, and link to a new guideline giving advice on how to avoid and/or spot original research? I know this isn't what's being talked about principally, but if it's a plausible solution to the current mess (for want of a better word) then it renders some other elements of discussion relatively moot. SamBC(talk) 12:59, 31 August 2007 (UTC)

Correct (I think). Endorsing one source type over another (even with a consistent definition of what the source type is), does not preclude anybody from still injecting OR into an article. Any source type can still be used to create OR. Just because somebody creates an entire article from ai

I hate getting linked all over just to try and get a simple explanation, but I think in this case, regarding the primary-wiki-source issue, see Primary source for the full-blown definition. It lists many examples that do get published (like official reports). wbfergus 14:47, 31 August 2007 (UTC)

• Yes. Secondary sources are necessary for writing encyclopedic articles. People read this policy for guidance as often as they do for dispute resolution and it should offer this guidance.--BirgitteSB 14:50, 31 August 2007 (UTC)

I don't entirely disagree with BirgitteSB, but I feel a further clarification may be needed. I think it should be Reliable secondary-wiki-sources, but those may not neccessarily be available to the current editor. At the same time, we (Wikipedia in general) should not prohibit that editor from making their edits either, but perhaps have a "guideline" that in these cases "good faith" editors flag that as something that needs to referenced to a reliable secondary-wiki-source. This would allow potentially valid information to be included, while having that flagged as something that needs to be verified 9similar to using the "cite" tag in an article, so somebody else looking the article over at a later date who may have access to such material can properly cite it. I also think that this should be tied in somehow with the "FA" or "A-Class" criteria, if it isn't already. All of this ties in very closely with how many articles are created on Wikipedia (refer to Wikipedia:About#Using Wikipedia as a research tool). There is (usually), and ongoing process of gathering data from all the various sources, and then several major edits and numerous minor edits; at nay one time proper sourcing, etc. may or may not be present in the article or even available to the current editor, while another potentially interested editor may have that "reliable source" information readily available, but doesn't know it needed or where to put it until after seeing the article.

Anyway, from what I see on on here already today, it looks like we are making progress and may readily reach concensus on these two issues. wbfergus 15:08, 31 August 2007 (UTC)

I find the distinction Reliable secondary-wiki-sources to be very confusing. ≈ jossi ≈ (talk) 15:14, 31 August 2007 (UTC)

Yeah, not the best choice of words, but the best I could think of, since there are reliable and unreliable secondary sources, and then there's evidently the "real-world" definition of a secondary source vs. the Wikipedia use of the term secondary source. Just to me, primary seems to connote that it should have a higher precedence than secondary. It's not until after you get all involved in the discussions and read numerous articles that the distinction is suddenly made over precedence. wbfergus 15:19, 31 August 2007 (UTC)

The precedence is inherently different on wikipedia compared to other pursuits, simply because the use of sources in wikipedia is different to that in a typical historical analysis, say. I think that if there's a separate guideline on avoiding original research, distinct from the policy defining and forbidding it, then it could be constructed in a way to make this easily understood. SamBC(talk) 17:13, 31 August 2007 (UTC)

See below. I wanted the terminology to be cleaned up first. It still needs to be done. --Minasbeede 15:21, 31 August 2007 (UTC)

I was just looking over the Military History Project's "source" definition, and while it (the concept) does encroach on another "Guideline", I think it may be a start for this question. It states:

Policy requires that articles reference only reliable sources; however, this is a minimal condition, rather than a final goal. With the exception of certain recent topics that have not yet become the subject of extensive secondary analysis, and for which a lower standard may be temporarily permitted, articles on military history should aim to be based primarily on published secondary works by reputable historians. The use of high-quality primary sources is also appropriate, but care should be taken to use them correctly, without straying into original research. Editors are encouraged to extensively survey the available literature—and, in particular, any available historiographic commentary—regarding an article's topic in order to identify every source considered to be authoritative or significant; these sources should, if possible, be directly consulted when writing the article.

I think this may be a good start to try and reach a concensus on, though of course the actual wording needs to be changed to accurately reflect the goal of this part of the "Policy". Anybody want to try their hand at rewording this as it pertains to the "Sources" section of the "Policy"? wbfergus 18:12, 31 August 2007 (UTC)

No or not without major exceptions. I noted earlier that the Communist Manifesto is a primary source on Marx's and Engel's beliefs, and a secondary source on their various opponents' beliefs. It is a trustworthy description of their own beliefs, and a rather untrustworthy description of their opponents' beliefs. I think many such works are reliable when used as primary sources and unreliable when used as secondary sources, not the other way around. Jacob Haller 20:32, 31 August 2007 (UTC)

That exception is already contained in Wikipedia:Verifiability (WP:SELFPUB). All that's needed is a brief acknowledgment of that other policy for coherency. Vassyana 20:50, 31 August 2007 (UTC)

No. The policy should not privilege either primary or secondary sources. The pros and cons of each are to complicated to be found in a policy. Maybe some distinctions can be introduced first in an essay, then possibly in a guideline, but there should not be any blanket preference or discouragement of either in this most important of policies. The archived discussions brought out plenty of instances where primary sources are far preferable to secondary sources. But the main reason I don't think there should be a preference is that we can't yet agree on what the difference is between a primary and a secondary source. It makes no sense to say that one or the other should be "rare" or "relied upon" when we don't agree on what either means. Also, any stated preference of one over the other should describe widespread Wikipedia practice, not lead it (see WP:Policies and guidelines). This policy cannot change Wikipedia practice. It is bound by that practice. COGDEN 21:02, 31 August 2007 (UTC)

The policy cannot distinguish between primary and secondary sources unless there is a consensus as to the difference. In my view, it's pretty simple: a primary source is something original, right from the horse's mouth, while a secondary source is a re-telling of a primary source. In another sense in common usage, a secondary source is a source that discusses a primary source; however, any new conclusions by the secondary source is also a primary source as to those new conclusions. On this basis, there is absolutely no reason to favor secondary sources, unless the primary source is unpublished, in which case it would not be a reliable source.

Moreover, since polices must describe widespread Wikipedia practice, not lead it (see WP:Policies and guidelines), the fact that primary sources (such as journal articles, interviews, and fictional works) are widespread, and even more likely to exist in featured articles than in non-featured articles, would seem to show that if any distinction is to be made between the two, primary sources should actually be favored. Personally, I don't think WP:OR should make a decree either way. COGDEN 20:49, 31 August 2007 (UTC)

Certainly, in the original uses of the terms (from historiography), drawing new conclusions from primary sources does not create a new primary source, it's a secondary source. The primary source is the "data" the secondary source is the "analysis" and "results", to draw parallels from scientific fields. SamBC(talk) 20:59, 31 August 2007 (UTC)

I don't think so. In historiography, if you are tracing the history of a historical idea, the primary sources are the original sources of that idea, regardless of whether or not the source is also a secondary source with regard to other ideas. In fact, by definition, all primary sources in historiography are also secondary sources. For example, in an article about feminism, you might cite the historiographic work Newton et al. (1983), Sex and Class in Women's History. This work cites many other secondary historical works, but cites them as primary sources, because what the authors are citing is not the history itself, but the original historical ideas found in the (otherwise secondary) historical works.

Just so you know I'm not just making this stuff up, here's a quote from a library [www.library.mcgill.ca/human/subguide/pdf/primsour.pdf website]:

"A primary source is a document in which an event is described by its witnesses or first recorders, for example: The Diaries of Louis Riel (McL FC3217.1 R54A3). It is the raw material or data which the historian must evaluate in studying the history of a period, event, or individual. The historian produces a secondary source based on primary sources, e.g. Prairie Fire: the 1885 North-West Rebellion by B.Beal and R.Macleod (McL FC3215 B42 1984).

"The concept is relative: any document may be used as a primary source. For example, if you were to study the historiography of the Riel Rebellion, the book by Beal and Macleod may be used as a primary source."

COGDEN 21:59, 31 August 2007 (UTC)

I understand what you're saying, and I agree that any document can be a primary source - not least in a biography of the author. However, that does not mean that referring to an analysis of the second world war in order to report (on wikipedia) on the overall course of the war or the postulated reasons for certain decisions does not mean that we are using the source as a primary source - it's being used as a secondary source. If we were using it in an article on "historical commentaries on the second world war" it would be a primary source. SamBC(talk) 03:37, 1 September 2007 (UTC)

In reference to descriptive vs. prescriptive, that is generally true, but in this case we also have a duty to the pillars and to consider the usefulness to the mission of wikipedia. Both of those points said, I don't entirely disagree with the conclusions you reach from your (in my opinion) incorrect reasoning. SamBC(talk) 20:59, 31 August 2007 (UTC)

I'll escalate. Do the words mean the same thing (exactly) in the NOR policy as they do in the Wikipedia:Reliable Sources guideline article?

Well, that's a complete bust. The words don't appear there. Nevermind. (Or, maybe, pay very close attention to the status of things.) —Preceding unsigned comment added by Minasbeede (talk • contribs) 21:04, 31 August 2007 (UTC)

I honestly don't think we have a problem knowing the difference between the types of sources, we just have a problem describing the difference. Who here really does not understand the distinction we trying to make between sources that reliably make "analytic, synthetic, interpretive, explanatory, or evaluative claims" and those that either make purely descriptive claims or cannot offer an reliable perspective on any such claims? Does anyone truly not understand that distinction?--BirgitteSB 21:23, 31 August 2007 (UTC)

I think you just hit the nail on the head. This is exactly the distinction we should be making in this policy. Vassyana 21:38, 31 August 2007 (UTC)

People claim to "know" the difference, but they nevertheless disagree. That people disagree on the difference is evident by how people categorize different types of material. Some people say peer-reviewed journal articles are primary sources because they are the primary source of new conclusions, others say they are secondary because they draw on raw data and make conclusions. Some people say that newspaper interviews are primary because they are the original source of what someone said, others say that they are secondary because they are drawn from a reporter's notes or a tape recording. I suggest, based on the above definition, that Birgitte would say that Darwin's Origin of Species was a secondary source because it contains "analytical, synthetic, interpretive, explanatory, or evaluative claims", whereas I (and others here) would say it's a primary source because it's the primary, original source of the theory of evolution. COGDEN 22:07, 31 August 2007 (UTC)

Please read my comment again and note I never wrote "primary" or "secondary". Do you really not understand the distinction be made there?--BirgitteSB 22:18, 31 August 2007 (UTC)

Sorry. I was referring to your definition immediately above. We were discussing primary vs. secondary, and I understood you were defining sources with "analytical, synthetic, interpretive, explanatory, or evaluative claims" as secondary sources. Whatever we call them, however, the fact they are analytical shouldn't be the defining factor. This type of distinction would exclude any factual information that isn't sufficiently "analyzed", which would still pretty much prohibit the creation of articles on current events, or the inclusion of most data and raw information in an article, even if it's not controversial and clearly accurate. COGDEN 00:25, 3 September 2007 (UTC)

No, no privileging of one type of source over another. It has been pointed out that many FA use a lot of primary sources. The key thing is that any "analytical, synthetic, interpretive, explanatory, or evaluative claims" come from sources, not from Wikipedians. --Coppertwig 22:57, 31 August 2007 (UTC)

Agree with Coppertwig. Get guidance from guidelines. If people come to the policy statement for guidance refer them to the guidelines from within the policy. That way there'd also be just one set of descriptions, not two. There could even be an examples page (that provides further guidance) of what meets the criteria and what doesn't, with an explanation of why or why not. These could be real cases, if that doesn't offend any of those involved. The examples could be slightly (or greatly) altered to avoid offense, if offense to someone is a problem. --Minasbeede 03:26, 1 September 2007 (UTC)

Above, Jaco Haller wrote, "I noted earlier that the Communist Manifesto is a primary source on Marx's and Engel's beliefs, and a secondary source on their various opponents' beliefs. It is a trustworthy description of their own beliefs, and a rather untrustworthy description of their opponents' beliefs." This is precisely why I think Wikipedia should encourage secondary sources. Wikipedia is about verifiability and not about truth because it assumes that all views are just that: views. One reason for favoring secondary sources is to help limit the chances of an editor putting hs or her own view in ... but another at least as important reason, which Haller brings up, is it helpe ensure that we can clearly identify the views oof others that we are including in the article. NPOV makes it clear, there is no truth, just multiple points of view. Some views may be more notable, but they are just views of the truth, not the truth itself. In some case the view is not ideological but professional - for practical reasons jurists and psychologists may have different views on insanity, or botanists and molecular geneticisms may have different views on what a species is. in other cases, the views may be ideological, so we would expect Republicans and Democrats to have different views of George W. And as Haller rithgly points out, Marx and Engels had their own view of other European socialists. We favor the secondary source because we are not claiming "this is the truth about 19th century socialism" but rather, "Here is Marx and Engels view ... and here is Y's view ... and here is Z's view..." I think this is a very important reason for the policy and for favoring secondary sources, and it shows again how much NPOV and NOR fit together. Slrubenstein | Talk 09:36, 1 September 2007 (UTC)

I don't see the verifiability problem with going to Marx (primary source) for Marx's views, Owen for Owen's views, Proudhon for Proudhon's views, etc. Jacob Haller 16:22, 1 September 2007 (UTC)

It depends. I think there may well be non-controversial times when ou could. But I know that there are many times when, at least in writing a Wikipedia article, you shouldn't. The problem is, Marx left over 25 volumes of writing, and there is a veritable cottage industry of different interpretations of what Marx really meant. The Bolshevicks and Menshiviks argues over it, then Trotsky and Stalin argued over it, then Kruscheve and mao argued over it - that is politics; in academia there are even a greater variety of "Marxisms." I do not think it would be at all controversial (or a violation of NOR) to write in an article "in x, Marx wrote '....'" - but as soon as one does so in order to make a more general point about what Marx believed or "Marxism" I think one is slipping into original research. I do think that there are a handful of principles and claims in Marx's writings that all so-called Marxists, as well as biographers and critics of Marx, would agree Marx "stood for" or "meant" - I could probably come up with six or seven points that any follower or critic of Marx would agree were things Marx meant. But that would still be my reconstruction of his though, based on making careful selections ... and it would not be hard at all to find a secondary source to support such a claim ... so we should back it up with a secondary source. Slrubenstein | Talk 16:46, 1 September 2007 (UTC)

Same answer, but clarification with an example. To go back to the 25+volumes of writing he have of Marx, how do we know what was speculation and what was bedrock for marx? How do we know the weight he himself gave? You might say "well, those things he had published in his life-time." but there are biographers and historians who believe that many people express themselves most honestly in work they did not have published including diaries and private letters. Moreover, many Marxists rely on unpublished works like the Grundrisse. And of Marx's published work, well, he wrote vor a variety of audiences, and over the course of a lifetime. How do we know the extent to which Marx was adapting his ideas to reach a particular audience, or just expressing himself with abandon? And how do we know the extent to which what Marx wrote at one time in his life he held to with the same conviction at another time? there are many scholars - including self-identified marxists and people who do not consider themselves "marxists," who argue that there is a profound difference between the "early Marx" and "later marx" - and also scholars who reject this view. Did you know about this? If so I think you would have understand exactly why I think one is doing original research when one uses quotes by Marx to make claims about "what Marx thought." If you didn't know this, well, this is just another reason I believe so strongly in the distinction between primary and secondary sources, and the importance of favoring the latter. Here is what it boils down to to me: if all you want to claim is "In a lecture Marx gave to the German Workingman's Club of Brussels in 1847, Marx said that 'The same commodity is offered for sale by various selers. Whoever sells the commodities of the same quality most cheaply, is sure to drive the other sellers from the field and to secure the greatest market for himself. The sellers therefore fight among themselves for the sale, for the market .... Thus there takes place a compeition among the sellers which forces down the price of the commodities ofered by them .... But there is also a competition among the buyers; this upon its side causes the price of the proffered commodities to rise,'" well, I would not object and would not worry about OR. If however you were to make synthetic or general claims about what Marx believed, I will want to know that you have read most of the available works of Marx (all 25+volumes) and have studied his biography and the social and historical context in which he wrote, and understood the various audiences to whom he wrote and spoke, and the influences upon him. Fact is, you would have to be a Marx scholar and know and understand Marx and his times well enough to have been published in a peer-reviewd journal, or a book published by an academic press, or a book or article that is notable among some group of people who know something about Marx - in short, I will have expected you to have published what we here call a secondary source. And if you have, we can quote and cite it! If you haven't - well let's look for one by someone else. So this is another reason why I think it important that Wikipedia favor secondary sources. Slrubenstein | Talk 18:49, 1 September 2007 (UTC)

The problem isn't finding different secondary sources for Marx, but for Owen, Proudhon, et al., let alone obscure figures. Generally, if commentaries, encyclopediae, etc. rely more on their critics' descriptions of their beliefs than their own descriptions of their beliefs, then we have a problem, but one which their own works, if available, can solve.

For Marx, much of the later debate among Marxists focused on addition to his theories, e.g. adapting them to Russian conditions, and "Marx believed this" or "Marx believed that" were not final arguments. So disagreement among Marxists didn't always mean disagreement about Marx.

IMO, the ideal "x believed y" reference would have both (1) a primary source by x stating his/her views about y, and (2) a secondary source about x, though I'm personally more concerned with finding (1) than (2). If these contradict each other, I think that a full discussion should outweigh a short comment. Jacob Haller 19:28, 1 September 2007 (UTC)

I don't see it. If any personal analysis is made of what Marx said that's OR and it should be ejected. If someone quotes anything Marx said, with no interpretation, then it stands, probably.

If the quote is intended to show how some (properly sourced) interpretation, etc. of what Marx said is wrong then in most cases (and I can't think of what the exceptions should be) it should be ejected (more properly, treated just as such cases are routinely treated.) Why should anyone care if it's historic-research-primary or Wikipedia-primary or historic-research-secondary or Wikipedia-research-secondary or anything else? If it's OR, bye-bye to it, or change how it's used, or replace the other material with the new material, as appropriate - and I see no way anyone can, a priori, declare what the disposition should be on any basis, not just on the basis of what type of source was cited (here I assume that the citation is valid else it's a trivial case that needn't be discussed.) I can see and appreciate that this sort of situation may most frequently occur with material regarded (somewhere/somehow) as "primary" but the issue isn't whether it is Wikipedia-primary or Wikipedia-secondary, (or ...) the issue is that it's OR and the disposition should be removal (with the understanding that some other action might be appropriate so using the word "removal" is more of an example than an expression of a policy nature.)

I get it. Some people quote the Bible, etc. in a pugnacious way. That has to be dealt with. Changing the policy does nothing that helps. If the policy is made so restrictive that the policy alone is reason for removal (the apparent goal of those who want source-typing in the policy) then the policy is going to be over-restrictive and will, with a 100% guarantee, be misused elsewhere, harming Wikipedia.

Nowhere do I say that such material (abuses arising from misuse of the Bible and other "primary" sources) belongs. I'm only saying that pumping up a policy with restrictions is a bad solution, if it's a solution at all. I lean toward it's being no solution. --Minasbeede 20:29, 1 September 2007 (UTC)

There are already phenomena for which Wikipedia itself is a primary source, such as the Bogdanov affair, and since the page histories record edits as they occur, they are thus a reliable source of which editor accounts were used to write what things. The problem with OR or usage of primary sources might be biases or the creation of hoaxes. In some of those cases, the policy is justified. In some other cases, however, in which a phenomenon develops that numerous bloggers, message board users, etc., notice before the general or mass media does, OR might be the best way of describing the phenomenon, at least as the bloggers or message board users see it. I would suggest the following: if something is quite controversial and argued over by numerous people, the controversy should be discussed comprehensively from all sides; if something such as the cancellation of a TV show provokes an enormous campaign with a rapid accumulation of signatures that suggests numerous contributors, Wikipedia should take note of that as well. Try to give the benefit of the doubt when you can. — Rickyrab | Talk 14:51, 7 September 2007 (UTC)

Regroup

Let's review. As I understand what I've read here, Wikipedia policies should be descriptive, not proscriptive. Yet so far nobody has been able to accurately describe or define the primary/secondary distinctions (that are made in a special Wiki way and aren't the same as occurs outside Wiki), which looks to me like the policy section that is so very important can't even be enunciated. Surely if it can't be said it doesn't describe anything.

So let me ask for a few examples of articles where this distinction is so important: if it can't be described maybe it can be illustrated. I looked at the Wiki article on "Evolution" last night and found that, to the level I checked, it looked like it was indeed a long listing of facts with nary a word of explanation, analysis, or an any other of the things that Wiki editors are forbidden to do unless they're quoting someone else. I think the Wiki "Evolution" article is a very fine article but apparently (unless I missed something, which is possible) it cannot serve to at all illustrate why the emphasis on the superior nature of secondary sources is so necessary, as is asserted here by some. I'd think that in the over 100 sources cited there are both primary and secondary sources but one might suspect that if all the content is simple facts that all of the content could and did appear first in primary sources. So, what are a few articles that illustrate why the emphasis on secondary (as the term is used in Wikipedia) sources should be given special emphasis? --Minasbeede 11:26, 1 September 2007 (UTC)

This is not a debatable topic. Since you cannot identify the "analytic, synthetic, interpretive, explanatory, or evaluative claims" in Evolution, I must conclude you are not capable of helping us resolve this discussion. Please set aside and let those of us with the basic understanding of the issue work it out. --BirgitteSB 12:03, 1 September 2007 (UTC)

Nonsense. All topics are debatable. — Rickyrab | Talk 14:53, 7 September 2007 (UTC)

Who asked for a debate? I said that the Evolution article seems to be a fine article yet it also seems to be a serial recitation of facts, something that some assert here is undesirable. Not only does it seem to be a serial recitation of facts I can see why that particular article needs to be that and must strenuously avoid "analytic, synthetic, interpretive, explanatory, or evaluative claims" - so probably it is indeed a bad article to look to for what is claimed. Still, the Evolution article seems to me to be an article that illustrates that the claim made here by some (that you have to use secondary sources in order to include analytical material) isn't universal. So I asked for examples of articles that do show what is claimed.

The title "regroup" is what I came up with. the real motive was to split the section so editing would be easier. Any title is fine - and if a different title is inserted this can be removed. --Minasbeede 14:09, 1 September 2007 (UTC)

Are journal articles primary and secondary sources?

I've been asked to contribute on this topic and don't think that I can't really add any new arguments, as they have been all repeated many times. Instead, I think it would be useful to look at scientific papers - one thing I know well - to illustrate that it is impossible to classify some sources as either primary or secondary.

Take this 2004 paper as an example A trypanothione-dependent glyoxalase I with a prokaryotic ancestry in Leishmania major.

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

In the first paragraph of the Discussion on the other hand, I discuss previously-published data on Methylglyoxal synthase in the light of my results and bring these data together and interpret them. Here, if a Wikipedia article cited the paper on the possible problems caused by methylglyoxal synthase, it would be using the paper as a secondary source.

Scientific papers are neither primary nor secondary sources, they contain elements of both. Tim Vickers 01:28, 3 September 2007 (UTC)

I believe that, in practice, many sources are thus - it's one of the problems of using the distinction at all. I think one thing that we've started to tread into that will be a fruitful avenue for discussion would actually be to forget the distinction for a while and talk about why we ever wanted to distinguish, and see if there's other ways of acheiving the goal. SamBC(talk) 09:55, 3 September 2007 (UTC)

I think the idea was brought in so as to have a good hammer to beat the kind of people who find some out of context quotes from, say, Marx, and then quote them out of context and put their own bizarre interpretations on them and make them into some hugely disproportionate part of the article. Clearly, our policy needs to make it clear that that kind of nonsense is unacceptable. john k 14:03, 3 September 2007 (UTC)

If I may extend the metaphor, it makes sense to do that in a way that isn't using a sledgehammer to crack a nut, even if it's a particularly annoying nut. SamBC(talk) 14:23, 3 September 2007 (UTC)

Why not just say that all material must be cited, with facts backed by citations to reliable sources that contain these facts, and interpretations and syntheses attributed to reliable sources that make these interpretations and syntheses? Tim Vickers 17:24, 3 September 2007 (UTC)

A journal article could also be considered a tertiary source if it collected related primary and secondary source material from other articles. But if you quoted it as a source about itself, then it would be used as a primary source. This is all very interesting in some sense, but I don't see what it has to do with the WP NOR policy. The NOR policy is really about keeping WP from becoming a primary source. So whether a source is written or used as primary, secondary, or tertiary is not material, it is only when the material is used to fashion a novel interpretation that is relevant here. Dhaluza 17:59, 3 September 2007 (UTC)

I agree we're getting away from NOR policy. However, I think how you classify a citation to a scientific article depends on what you're citing the article for. If you are citing the original research, conclusions, or analysis found in the scientific journal, you are citing it as a primary source: it's the original source of the new idea. If you are citing the journal article solely for its references to previously-published information, you are citing it as a secondary source. If you are citing an author for some novel synthesis of previously-published sources or data, then you are citing to a primary reference for the new synthesis; however, such a reference would obviously be secondary as well, to the extent it is just repeating prior sources. Really, the thing that ought to be classified is the citation, not the source, because any source can cited as either primary or secondary. COGDEN 02:30, 5 September 2007 (UTC)

I would tend to disagree with the statement: "If you are citing the journal article solely for its references to previously-published information, you are citing it as a secondary source." If you are simply using am otherwise secondary (or tertiary) source simply as a witness to a fact, perhaps because you cannot access the original primary source, then you are still citing it as a primary source, but not an original source. A bare statement of fact would remain primary source throughout its chain of custody. If you also included commentary or contextual information from the source, then that use would be citing a secondary source. So, for example, if you were to write simply: "Archimedes calculated the value of π lay between 3 + 1/7 and 3 + 10/71" citing a math textbook or encyclopedia because you don't have access to Archimedes' primary source documents, you would only be using the cited reference as a witness to the fact, and it would still be a primary source, even though it was published more than 2000 years after the fact became known. If you went on to cite the source saying that his method of calculation was a major advancement in mathematics at the time, that would be a secondary source citation. I think the definition we have for primary sources does not make the distinction between primary and original sources clear. Dhaluza 11:06, 5 September 2007 (UTC)

Above, Tim write, "Why not just say that all material must be cited, with facts backed by citations to reliable sources that contain these facts, and interpretations and syntheses attributed to reliable sources that make these interpretations and syntheses?" Every bolicy starts with a box explaining the policy "in a nutshell." I'd have no objection to this being the nutshell, or part of it. Slrubenstein | Talk 15:01, 5 September 2007 (UTC)

Well, it is only material that is controversial or likely to be challenged that needs to be cited. We don't require citations for common knowledge because that is obviously not OR (to most people, at least). So such a broad statement could be mistakenly used for wholesale deletion of unreferenced content, which would not be supported by current consensus. Dhaluza 10:29, 6 September 2007 (UTC)

To a limited extent, that's already in WP:V#Burden of evidence, and is not limited to controversial material: "Any edit lacking a source may be removed, but editors may object if you remove material without giving them a chance to provide references. [...] Do not leave unsourced information in articles for too long, or at all in the case of information about living people." Jakew 10:38, 6 September 2007 (UTC)

Why has this become so complicated?

At bottom "No original research" means that Wikipedia should not be written as a primary source. This policy used to state this explicitly as: "Research that creates primary sources is not allowed." until this significant rewrite during the time that the controversial depreciation of primary sources was being integrated into the policy.

I think that we should go back to basics and focus on the problem--editors generating new not-previously-published primary source material in WP articles, rather than trying to limit the re-use of existing published primary source material. The arguments about whether basing an article on primary or secondary source material creates laudable or deplorable articles is not relevant to this policy. We are only concerned here with whether the material is new or novel, not whether it is good or bad. Trying to stretch this policy to cover these types of subjective value judgments has only caused confusion and contention over what should be a very simple concept. Dhaluza 17:59, 3 September 2007 (UTC)

I've come to agree over the past few weeks. Although I think there's merit in creating some guideline for when different types of sources are preferable to other types of sources, this is a complicated issue, and we need to build up a consensus from scratch, since nothing we've been discussing has emerged with any sort of consensus yet. We can just not discuss the issue here at all, and leave that for a future guideline such as Wikipedia:Classification of sources. COGDEN 02:19, 5 September 2007 (UTC)

The discussion of sources vs. article quality clearly does not belong here, because it is not directly related to NOR. And since that is what is at the bottom of the edit warring, those who want to passionately pursue these arguments should take them to another forum as you suggest. And cutting this policy down to the basics and building up consensus from there is what we need to do now to repair the collateral damage this sideshow has caused. Dhaluza 02:52, 5 September 2007 (UTC)

I think there should be a page that could be entitled something like "Wikipedia:Wikipedia policy for Dummies" (or possibly something more original) that gives a clear and consice version of the main policies and guidelines that avoids what is known as instruction creep. MrMurph101 03:13, 5 September 2007 (UTC)

WP:A? That conveys the main issues of WP:NOR and WP:V clearly and concisely, in a way that's much more useful for beginners and Dummies.... dave souza, talk 11:28, 5 September 2007 (UTC)

Answering the main question from the section title, this has become so complicated because a simple policy about NOR, over time, was "expanded (wrongly in my opinion) to include things that are not about OR/NOR but instead trying to re-define existing definitions from other subject areas, such as "Sources" or what "Synthesis" is. wbfergus 13:01, 5 September 2007 (UTC)

That seems unfair - I doubt anyone was trying to (re)define those for the sake of it. The tone seems to ascribe all sorts of ulterior motives to the additions. SamBC(talk) 13:32, 5 September 2007 (UTC)

Sorry about that, I didn't mean it that way. To me, it seems as if over time, as issues cropped up, people added a bit here, a bit there, and pretty soon this policy was no longer strictly about NOR, but "encroached" into other subject areas that had no real bearing in a policy about NOR. Those issues, while very real and needing explanation, should have been handled on different (and linked) pages. It's just an inadvertent consequence of trying to accomodate people's questions without proper forethought, in my opinion. wbfergus 14:10, 5 September 2007 (UTC)

I agree that at best, this is a case of WP:CREEP, but more likely this is really a case of WP:COATRACK where people have taken a simple policy and obscured it by hanging extra things on it that were not supported by a broad consensus (whether well intentioned or not). And the only way to address this is to take the coats off by "brutal reduction" as that essay suggests to get back to the essence of NOR. Dhaluza 10:19, 6 September 2007 (UTC)

We do not need a policy for dummies. What we need is patient editors who are willing to explain things to newbies - that is one of the things that makes this a collaborative project. Slrubenstein | Talk 14:59, 5 September 2007 (UTC)

Are examples original research?

Many articles on law, science and other complicated stuff have examples to help the reader understand. The examples are not referenced. Are they original research? (For example, "Concepts" section of mens rea). --Kaypoh 06:26, 5 September 2007 (UTC)

The examples you mention are not OR, just unsourced statements. However, giving an example that is not clearly exemplary of the thing in question might well constitute OR. If I were to say "Musical genius has occurred in a wide variety of cultures over time, for example Bach, Beethoven and Britney Spears" that would be OR, because without a citation to support the rather strange assertion that Britney is a musical genius, I would be advancing a point of view that is not proven. --Slashme 07:06, 5 September 2007 (UTC)

Well, I don't know whether they are unsourced but true, or someone's analysis. It's common for people to include examples in articles based on their own analysis rather than from a citable authority (c.f. the huge lists of examples attached to fourth wall which it took four AFDs to finally delete). In general we want to use examples which out sources cite as examples. Mangoe 14:12, 5 September 2007 (UTC)

Some people might not consider Bach or Beethoven to be musical geniuses. So it violates NPOV. — Rickyrab | Talk 14:56, 7 September 2007 (UTC)

Should that read "outside sources"? --Minasbeede 14:19, 5 September 2007 (UTC)

There is a difference between examples, which can easily lead into OR or NPOV issues, and a recition of facts from a primary source, like the transcript of a speech would be. Kaypoh's "examples" aren't really examples, they are facts from a primary source, which if needed, should be included verbatim (if allowed) in Wikisource, and appropriately linked in the article. wbfergus 14:25, 5 September 2007 (UTC)

I think the Concepts section of mens rea was not a good example of this type of "example used to explain" (don't know what to call it). Below are a few examples of what I really mean.

See this from Fallacy of many questions:

An example of this is the question "Are you still beating your wife?" Whether the respondent answers yes or no, he will admit to having a wife, and having beaten her at some time in the past. Thus, these facts are presupposed by the question, and if it has not been agreed upon by the speakers before, the question is improper, and the fallacy of many questions has been committed.

And this from Intention (criminal):

For example, suppose that A, a jealous wife, discovers that her husband is having a sexual affair with B. Wishing only to drive B away from the neighbourhood, she goes to B's house one night, pours petrol on and sets fire to the front door. B dies in the resulting fire. A is shocked and horrified. It did not occur to her that B might be physically in danger and there was no conscious plan in her mind to injure B when the fire began. But when A's behaviour is analysed, B's death must be intentional. If A had genuinely wished to avoid any possibility of injury to B, she would not have started the fire. Or, if verbally warning B to leave was not an option, she should have waited until B was seen to leave the house before starting the fire. As it was, she waited until night when it was more likely that B would be at home and there would be fewer people around to raise the alarm.

And this from Mistake (criminal law):

For example, if a defendant goes into a supermarket and places eight items in a basket which is presented to the cashier for payment in the usual way. Both honestly believe that all eight items have been scanned, and the defendant pays the sum shown on the bill. A store detective, however, notices that a mistake was made by the cashier so that only seven items were actually priced. This detective arrests the defendant after leaving the store. Since the defendant honestly believes that he has become the owner of goods in a sale transaction, he cannot form the mens rea for theft (which is usually dishonesty) when he physically removes them from the store. Accordingly, he should be acquitted.

And this from Self-defense (theory):

For example, a person who unknowingly chances upon two actors practising a fight would be able to defend his restraint of the one that appeared to be the aggressor.

--Kaypoh 14:48, 5 September 2007 (UTC)

For the examples you just posted above, I don't see why those can't be included in the article if they can be found in what is (currently) called a secondary source and properly referenced. If they can be referenced, then they are not original research. If they can't be referenced, then they may be classified as original research since you wouldn't be able to prove that they were not your "ideas", correct or incorrect as they may be. wbfergus 14:54, 5 September 2007 (UTC)

The examples are original research. Track the edit history - I would bet that these passages were added to the articles before 2003 when this policy took full effect. It should not be hard to find sources for the examples, and one way to improve Wikipedia articles is to find stuff just like this, and look for soruces! Slrubenstein | Talk 14:58, 5 September 2007 (UTC)

If the examples are obvious and straightforward, noncontroversial extensions of clearly-verifiable ideas, then I don't think it should count as original research so long as a person without special expertise would be able to verify that the example merely illustrates the ideas that have already been set forth and attributed. In that case, you are not actually putting forward original research; you are just re-stating what has already been said and attributed in a more user-friendly way, by an example. Other encyclopedias use examples, and it's not often that you find suitable examples in primary or secondary sources. Tertiary sources like textbooks, maybe, and if such examples can be found, all the better. But if nothing "original" is actually put forward by the example, and specialized knowledge isn't required for the reader to verify the correctness of the example, I don't see how it's original research. COGDEN 19:11, 5 September 2007 (UTC)

Considering these comments, I would add only this: sometimes what is original research is a judgement call and that judgement is best left to those editors who have been working on an article. As COGDEN suggests, eidts that are uncontroversial sledom if ever require editors to turn to policies for guidance; the more the controversy, the more salient the policies. Slrubenstein | Talk 21:46, 5 September 2007 (UTC)

I think the key is whether the examples are illustrative or interpretive. If they simply explain what has already been set forth, and there is no new thought, then this is no different than paraphrasing a source to avoid copyvio. But if the examples involve interpretation or new ideas, then that would be OR without a RS to attribute them to. Dhaluza 10:25, 6 September 2007 (UTC)

Is merging data tables from diverse sources OR?

I have looked at articles that had a couple of funky tables and then looked at the referenced sources as well as others, and wondered if constructing a single table that merged the data from the various sources would constitute OR.

In the case of oil wells, there is a list of them with data pulled from multiple sources which seems not be OR, while a I have a feeling that creating a similar kind of compilaton for Palestinian census data would be seen as OR. I've concluded for the time being that I will fill in existing tables with info I find, but I won't create any because I now see it as minefield - one's structuring of data is a POV.

But perhaps this should be handled with a policy more like that for images and other original artwork? The objective is to present the diverse facts and analysis in some coherent fashion, right? Mulp 01:53, 6 September 2007 (UTC)

As long as it's somebody else's analysis, and is backed up with proper references for the analysis. Regarding the table, on the surface it appears that it would not be OR, if for example, you created one large table from a source of oil wells in Oklahoma and then another source of oil wells in Texas. There wouldn't be any overlap of oil well data in that case, so it would not be up to you to decide how to handle the cases where there was overlap. It might also be contentious if the data from the two sources were in different formats, and you performed some conversion of the data so the table could be uniform, like if one was reported in meters and the other in feet. Suddenly, even though it's a simple conversion, you changed the original data, so some could claim that as OR. Personally I wouldn't, but I think quite a few others would, at least enough to cause a major discussion on that page and here. That's just my opinion though. In that particular case, it may be "better" to contact the source of the information and ask if they could provide it in the same format as the other, that way nobody can claim that you modified the original data. If it's USGS data, I happen to know some of the folks over there, and it probably wouldn't be that big of a deal to accomodate that kind of request. I do it all the time with many of the data requests I get (I'm in a different group). wbfergus 10:20, 6 September 2007 (UTC)

Dictionary

Is it OR to say that X term, that some people describe as a neologism, is not found in the OED or [insert name here] other specific dictionary? Obviously the OED is a primary source in this case and the factual accuracy is easily verified. 136.152.153.36 22:21, 6 September 2007 (UTC)

I think the neologism rule has gotten way out of hand. Sometimes an article needs just the right adjective to modify a noun. If the adjective-noun combination doesn't crop up in any sources, or "worse" yet, does turn up in primary sources, some editors will object that it is a neologism.

Anyway, the OED is a dictionary, and dictionaries lead to analytical OR. Jacob Haller 23:27, 6 September 2007 (UTC)

I'd say it's not original research, because whether or not it exists in the OED is something anyone can verify just by looking at the book, without making a novel interpretive leap. I don't understand what Jacob Haller means by "dictionaries lead to analytical OR". Certainly dictionaries, especially one as academic as OED, may contain original research, but citing them is not original research unless you inject novel interpretations of OED into the mix. COGDEN 23:34, 6 September 2007 (UTC)

Dictionaries offer tidy definitions, even when historical usage does not. Tidy definitions encourage people to draw analytical conclusions, e.g. Benjamin Tucker supported individual ownership and free markets, "therefore" (analytical OR) he was not a socialist and/or he was pro-capitalist (he was involved in the socialist movement of the time, and his economic ideas are basically adaptations of Proudhon's; anyway there have been a variety of definitions of socialism).

There are some recurring analytical OR edit wars which are just wrong:

• "Libertarianism refers to opposition to the state and opposition to collective ownership."
• "Socialism refers to a form of collective ownership which requires the state."
• "Therefore, libertarian socialism is an oxymoron."

And there are some where I agree with the analysis, but not its inclusion:

• "Anarchism refers to opposition to the state and opposition to hierarchies."
• "Capitalism refers to a form of economic hierarchy which depends on the state."
• "Therefore anarcho-capitalism is an oxymoron."

It's reasonable to note that some groups currently called L. or most groups ever called A. oppose S. and C. respectively, but these arguments should draw from history not from dictionaries. Jacob Haller 23:59, 6 September 2007 (UTC)

Another editor redefined left-libertarianism based on a dictionary. In this case, the dictionary is just wrong, and would define most left-libertarians out of left-libertarianism entirely. However, the dictionary was a tertiary source, while the descriptions by left-libertarian organizations, individuals, etc. were primary sources, and were deleted. Jacob Haller 00:03, 7 September 2007 (UTC)

Where do you get that anarchism refers to "opposition to hierarchies"?? A dictionary maybe? If we go by your reasoning, then anything that calls itself a form of anarchism is a form of anarchism, including anarcho-capitalism, and that therefore we should adjust the definition of anarchism to include anarcho-capitalism. That would require removing from the definition of anarchism that it is opposed to hierarchies. Yes? —Preceding unsigned comment added by Sorojourner (talk • contribs) 01:11, 7 September 2007

I get the definition from discussion with other anarchists, and find it in some written sources as well. I think it's fair play to challenge the credential of "anarcho-capitalism" on anarchism or anarchism and capitalism, with proper sources, but not in the intro to anarcho-capitalism.

I just got sick of hit-and-run OR "clarifications" to libertarian socialism and the same standard which limits these on LS also limits them on AC. It's Wikipedia's role to describe ideologies, and their positions, and the controversies they are involved in, not to tell readers whether the ideologies are self-consistent or not. Jacob Haller 17:40, 7 September 2007 (UTC)

This seems to have gotten off-track with respect to addressing the issue I was interested in. There is an article that says "X is a term that...". Some verifiable academics say this term is a neologism. Some Wikipedians want to mention that the word "X" was published much earlier in books A, B, etc. Other Wikipedians want to mention that "X" is a not included in the OED or any other dictionary. Both of these statements are based on easily verified facts discovered by Wikipedians, and potentially they both add useful context to whether or not X is a new term (or perhaps a rarely used old term that has taken on a new meaning). So, consistent with the OR policies regarding primary sources, is one allowed to mention both, neither, just one, etc.? 136.152.153.15 21:25, 7 September 2007 (UTC)

Proposal for removing the coats from the coatrack

The arguments for and against the use of specific types of sources in specific situations need to be discussed more fully, and in the context of WP:OR, WP:NPOV, and WP:Notability. I think we can work toward a potential guideline for this. In the meantime, however, to get a consensus, original research policy that we can all live with, why don't we get back to basics. We can avoid the controversy for now by paring back to the fundamental ideas we all agree with, which I think are represented in the following text, which I would propose as a near-term replacement for the "Primary, secondary, and tertiary sources" section:

===Use of sources===

Research that consists of collecting and organizing material from existing sources within the provisions of this and other content policies is encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia. However, care should be taken not to "go beyond" the sources or use them in novel ways.

When citing a source in support of a proposition in a Wikipedia article, the proposition should (1) be verifiable to the source by any reasonable, educated person without specialist knowledge, and (2) make no novel analytic, synthetic, interpretive, explanatory, or evaluative claims based on the source.

Even if people want there to eventually be more than this in the policy, maybe this is an acceptable step that we can build on. COGDEN 23:59, 6 September 2007 (UTC)

How about offering comments or suggestions, or even making some edits to the two variations I put together in #Notice of draft revision above? It's an attempt to address this specific situation, of taking the "parts" that don't pertain to NOR and getting them into their own article, so this policy can be less contentious. There's also another variation in the works that I've seen as well, but I'll let the other person "announce" it once they're satisfied with it's content. wbfergus 10:07, 7 September 2007 (UTC)

While I think the alternative proposals are useful as illustrative examples, we should be able to edit this policy here to preserve the edit history. I've taken a shot at diffusing one of the controversial items already. Dhaluza 11:24, 7 September 2007 (UTC)

What you've done is good. Preserving the edit history is good. I think the classification of sources language needs to be moved elsewhere. I suspect you left them in to avoid an edit war. I think that was a good idea and action (or lack of action.) It looks like there is a visible consensus for moving them elsewhere. Can that be assumed and discussion turn to the issue of where to move them and how to reword that section, if that is needed? --Minasbeede 13:38, 7 September 2007 (UTC)

Are immediately verifiable statements OR?

I would like to give an instance that possibly hones the edge of the point being made by the "example" discussion above.

I recently edited an article of a person X where I wanted to say that "X is often referred to on the web as Prof. X". Later I show that his official biography says he was not a prof. This I felt was important to the article, since many readers may come to the article with a view that he is a Prof. An encyclopedia is being written against the backdrop of this common perception, after all.

Now in stating that "X is often referred to on the web as Prof. X", I have no "source" that makes this full claim. However one can cite hundreds of webpages that say he is a prof. It is a claim that is immediately verifiable, and backed up by these sources, which are completely "reliable" for this purpose - ie. they are clearly referring to him as Prof. X.

We are not introducing a new theory / original idea / term/ argument - it is just the immediately verifiable statement that "X is often referred to on the web as Prof. X". My question is - should it constitute original research?

I am referring to the article on P. N. Oak, where I have made this claim recently. If there is a consensus here that this is OR, I will remove the statement. mukerjee (talk) 09:05, 7 September 2007 (UTC)

I think the problem comes with the word "often" - it's OR to assess the frequency of such references in any way. You can say that he is referred to as such with a single citation, although more is better. To say that such references occur often without it being OR, some reliable source must have made the assessment. SamBC(talk) 18:07, 7 September 2007 (UTC)

135,000

David Irving has repeatedly pushed the number 135,000 as the number killed in the bombing of Dresden in World War II. This number like the man is now throughly discredited. See

• Introduction Note the dates on many of these estimates particularly given this letter by Irving to the The Times on 7 July 1966.
  • The 'Final Report' of 15 March 1945. (See note 9 and following that discuss the letter to the times linked above).
  • The real TB 47.
  • Further bending of reliable sources
  • Further misuse of figures: refugees, burials, and excavations.
• general conclusion Not one of [Irving's] books, speeches or articles, not one paragraph, not one sentence in any of them, can be taken on trust as an accurate representation of its historical subject. All of them are completely worthless as history, because Irving cannot be trusted anywhere, in any of them, to give a reliable account of what he is talking or writing about. -- Evans

The trouble is that the number still appears in tertiary sources like encarta Dresden. This is not a problem at the moment because of the reliance on secondary sources. But how would one handle this if the new wording for sources is used? See Talk:Bombing of Dresden in World War II#Population statistics and use of napalm.

In a second case take an example from a primary source like Protocol I that says "The Parties to the conflict shall take all feasible measures in order that children who have not attained the age of fifteen years do not take a direct part in hostilities and, in particular, they shall refrain from recruiting them into their armed forces." Unless one is an expert on international law the nuance of the words "feasible measures" will be lost unless one had read the ICRC commentary on this that makes clear, this is not a complete ban on the use of children in conflict. The ICRC had suggested that the Parties to the conflict should "take all necessary measures", which became in the final text, "take all feasible measures" which is not a total prohibition on their doing so because feasible should be understood as meaning "capable of being done, accomplished or carried out, possible or practicable". See Military use of children#International humanitarian law for sources on this.

It seems to me at first glance that doing away with primary, secondary and tertiary sources may open up a can of worms and I would like others who are considering these radical changes to comment on how the new proposed policy will deal with these types of issues. --Philip Baird Shearer 20:43, 20 September 2007 (UTC)

For the first example, it's a case of editorial judegement (etc etc). In the second case, anyone quoting that protocol and saying that this utterly prohibited child soldiers would still be committing OR, because they would be interpreting it. SamBC(talk) 20:57, 20 September 2007 (UTC)

Editorial judgement is a problem, is Encata not as reliable source as Götz Bergander in this case? Under the current rules the answer is no. But under the new rules who says? Just look at the section on the bombing of Dresden given above to see the potential problems this opens up. --Philip Baird Shearer 21:04, 20 September 2007 (UTC)

(ec)SamBC's response is correct. Additionally, the primary/secondary/tertiary distinction would not affect the 135 000+ claim, as it is widely repeated in secondary sources. A brief survey of available sources seems to indicate that most recent scholarship has revised the number down to the 25 000 to 35 000, largely based on recent German estimates and the death rate for similar bombings. However, even books published in the last five years still repeat the 135 000+ claim. It seems more a case of NPOV and using editorial discretion to judge the most reliable sources, than one of original research or source typing. Vassyana 21:13, 20 September 2007 (UTC)

"even books published in the last five years still repeat the 135,000" which book since the Lipstadt libel suit has published 135,000? --Philip Baird Shearer 21:32, 20 September 2007 (UTC)

(ec)Well, if you want to know about reliability of sources, check WP:RS. That defines what's reliable. This policy is just about what is or isn't OR. SamBC(talk) 21:19, 20 September 2007 (UTC)

Without 1/2/3 the RS is undermined because Encarta is usually considered to be a reliable tertiary source, but under the new rules that will be gone from the supporting policy and so undermining RS. --Philip Baird Shearer 21:32, 20 September 2007 (UTC)

As to the second point about OR yes under the current rules this is clear (because it is a primary source), but under the new proposed rules, how do you tell the difference between a summary of a source and interpretation of a source given that there is no specific difference between a primary and a tertiary source? --Philip Baird Shearer 21:12, 20 September 2007 (UTC)

It would still be clear under the proposed replacement. Since making such an assumption would be providing interpretation and/or analysis not present in the source, it would be original research. Source typing does not affect this one way or the other. Vassyana 21:14, 20 September 2007 (UTC)

If Wkipedia is not to be reduced to a series of quotes how does one tell the difference between a summary of a source and interpretation of a source if the distinction between primary and tertiary sources is lost. Currently if one takes a paragraph from a tertiary source and summarises it, it is acceptable. But if one takes a paragraph from a primary source one may only quote it. Under the new rules it seems that one would be able to summarise it, and given the highly specific nature of some primary sources -- like the one I have given above -- who but an expert is to judge if it is a true summary or an interpretation? I think that this change may lead to all sorts of edit wars. But this is the first time I have read this proposal and I am open to persuasion that it is a step forwards. --Philip Baird Shearer 21:44, 20 September 2007 (UTC)

In short, some sources that seem at first glance like they ought to be reliable disagree. That problem will always be with us. Editors will just have to form a consensus that certain superficially reliable sources are actually not so. I doubt that following a general rule that articles published in peer-reviewed journals are always better than encyclopedias will preserve us from this problem; while encyclopedias are less up-to-date and often written by non-experts, peer-reviewed journals publish articles that by definition have not been presented to all the scholars in a field until after journal publication, so are not thoroughly reliable until sufficient time has passed for rebuttal articles, or at least letters to the editor, to be published. (Perhaps the problem with journals is somewhat reduced through electronic preprints.) --Gerry Ashton 21:17, 20 September 2007 (UTC)

Philip... not allowing the 130,000 number simply because the current policy contains a caution(not a ban) against using primary sources is a gross misuse of the clear intent of the policy (even in its current form)... An editor adding the number and citing to encarta (or even to Irving) is NOT adding their original thinking to the article in any way shape or form. They are appropriately repeating information they assume is from a reliable source. If the source is incorrect, that is an issue for V or RS ... but it certainly should not be an NOR issue. Blueboar 21:59, 20 September 2007 (UTC)

Perhapse it should be moved to WP:V (It was issues like this why I was not against WP:ATT in principle, just the current implementation). I have yet to be persuaded that this section should not be in one of the core content policies. --Philip Baird Shearer 22:49, 20 September 2007 (UTC)

RS is a mess, so I'm not sure anything useful belongs there; but this is a reliability issue. PBS is making the case that Irving and those who use his figure are unreliable sources on Dresden. This is a perfectly reasonable argument to make, and source criticism is a part of editorial judgment. I would note, however, that, unless some WPian is using the actual surviving records from Dresden directly, primary/secondary won't help: Irving is secondary. Septentrionalis PMAnderson 22:14, 20 September 2007 (UTC)

Not quite. Irving is an unreliable source so we can strike him. But Encarta is under the current definition a tertiary so it is not given the same weight as a secondary source "Unsigned articles may be less reliable, but they may be used so long as the encyclopedia is a high quality one." AFAICT we are loosing that level of granularity and so who is to say that a well researched well cited secondary source carries more weight than encarta if these changes take place? I have chosen two contentious examples because AFAICT in the real world of every day editing of controversial articles the current wording gives a lot more guidance to people who may not be familiar with working with sources than the proposed new wording. I am worried that such a change is being viewed through the restraints that the current wording gives and without such guidance (and constraints) it we will be opening up Pandora's box.--Philip Baird Shearer 22:43, 20 September 2007 (UTC)

Philip Baird Shearer, First, there is no consensus for what the wikipedia meaning of the phrase "primary source" is. Second, there is no consensus that any given secondary source is a more reliable source than any given tertiary source (and for good reason, I might add). Finally, this is a reliable sources issue and not an issue that is related to someone making a claim that is not supported by the cite if one is given, or making a claim that is believed to be not supported by any reliable published source if no cite is provided, which is the definition of OR and the proper subject of this policy. WAS 4.250 23:29, 20 September 2007 (UTC)

I thought that there was a consensus, and it is expressed in the policies of Wikipedia specifically the definitions given in the section "Primary, secondary, and tertiary sources". If not then what do the sentence "When consensus is referred to in Wikipedia discussion, it always means 'within the framework of established policy and practice'. Even a majority of a limited group of editors will almost never outweigh community consensus on a wider scale, as documented within policies." in WP:CON mean? --Philip Baird Shearer 23:43, 20 September 2007 (UTC)

Further (to address you second point) within that consensus it state "Some tertiary sources are more reliable than others, and within any given tertiary source, some articles may be more reliable than others. For example, articles signed by experts in Encyclopaedia Britannica and encyclopedias of similar quality can be regarded as reliable secondary sources instead of tertiary ones. Unsigned articles may be less reliable, but they may be used so long as the encyclopedia is a high quality one." which would appear to me to endorse the idea that unsigned articles in a otherwise reliable tertiary source may be less reliable than a reliable secondary source. --Philip Baird Shearer 23:50, 20 September 2007 (UTC)

Please take some time to review the archives for the past month. They clearly illustrate the lack of consensus for PSTS within this policy. Beyond that, your concerns about reliable sources may or may not be valid, but regardless this is not the place to address such concerns. WP:RS handles reliable sourcing and WP:NPOV handles appropriate weight. This policy addresses original research, which is a separate issue from the one you are raising. Vassyana 01:19, 21 September 2007 (UTC)

WP:RS is a guideline, not a policy. I think this this needs to be in a policy document. It is quite easy to argue that it is part of NOR because using primary sources and interpreting them is original research. If this section is removed from the policy documents there will be people who argue that they are not interpreting international treaties because they are merely summarising their content, this is prohibited currently and this change will have a very big impact on the way primary, secondary and sources tertiary sources are used in articles. This will have a direct effect on the way that the data sources used to extract information for Wikipedia articles are researched. --Philip Baird Shearer 11:35, 21 September 2007 (UTC)

Let me give you another example to try to illustrate my concerns. The wording of the Genocide treaty says Genocide is "... any of the following acts committed with intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such: (a) Killing members of the group; ..." There are whole articles and a quite a few court cases that analyse each clause of that sentence. It is not the same as reading a similar sentence in a secondary source. What does "intent to destroy" mean? What does "in part mean"? For example is intent to destroy the group the same as the intent to kill members of the group? If one person is killed does that qualify as part of a group or does it take more than one? If more than one person how may makes a part? Also what if the perpetrators kill all those in a group within their reach, but if that is only a small part of the total group (because most of the group are not within the reach of the perpetrators) does that effect the meaning of "in part"? If a Wikipedia editor tries to summarises that sentence on Genocide one is almost guaranteed to get it wrong and mislead the reader. This is very different from summarising a tertiary source. --Philip Baird Shearer 11:35, 21 September 2007 (UTC)

Philip, the question of whether there is consensus to change the policy is what this discussion is all about. Consensus can change. We are in the process of determining if there is a new consensus. So far, there seems to be a fairly clear majority in favor of making a change... and while that does not mean the same thing as having consensus, it does indicate that something is wrong with the policy as it stands. We are not rushing into this... one thing I learned from the ATT debates is you have to give a LOT of notice and time for comment when you make a major change to a policy. You have to convice a lot of people who will be warry of changing anything that the change is needed and that the proposed wording is good. We are in the process of doing that convincing. As to WP:CON... This seems to pop up whenever two large groups of editors disagree about the wording of a policy page... It is always a bit of a problem to know what "within the framework of established policy and practice" means when it is exactly that "policy and practice" that is the subject of debate... One side always argues that you need show consensus to change... and the other argues that you need to show consensus to keep. I reject both arguments ... you simply need to determine what the consensus IS. Blueboar 01:51, 21 September 2007 (UTC)

Philip, welcome to our discussion! I would ask you to please take a step back for a moment, and please review the examples you provided and the arguments you are making about them. Next, please think long, hard, and honestly about how they apply to "original research". Quite simply, they don't. What they boil down to are variations/interpretations of "Source" issues, as has been previously stated. I would also like you to take a look at the "Sources" section in this current policy (not this proposal). Next, take a look at WP:UNDUE and the section immediately following it, WP:NPOV#A_vital_component:_good_research. Next, open another browser window and take a look at WP:V#Sources. Again, open another window and take a look at WP:BLP#Sources. Now the real fun begins. Looking at just these sections, imagine looking at them through the eyes and mindset of a brand new editor to Wikipedia. While ultimately they may eventually wind up stating the same general 'thing', they all do so in a non-standardized way, making many new editors' eyes glaze over. Further complicating the issuse, try comprehending those 4 'sections' along with WP:RS, Primary source, Secondary source, and Tertiary source. All of these can probably be condensed into an article (preferably a guideline) on the types of sources with examples of how different 'things' can be classified different ways, depending on other mitigating factors, like conclusions, synthesis, etc, and then further linked to the appropriate "type of source", like primary, etc. In short, all of this "Source" stuff is very convoluted, especially given that various disciplines use the same terms to mean different things. It doesn't achieve anything leaving it in the policies, it just muddies the water and detracts from what the policy is actually about. The whole "source issue" can be greatly simplified by moving them out of the 4 main policies to one standardized (though probably expanded) article, with only a brief description in each policy about how the source issue pertains soley to that policy, "for further definitions and examples, see Types of Sources" (whatever it gets called). wbfergus ^Talk 11:09, 21 September 2007 (UTC)

Blueboar, I understand the arguments that both Vassyana are presenting, and as I said above it may be that the current section should be moved into WP:V, but if it is removed and replaced with the suggested text then it will also alter the reading of other sections of the content policy and guidelines. For example I think the section WP:V#Sources would need a rewrite, because quite obviously a treaty is a more reliable text on that treaty than any "peer-reviewed journals and books published in university presses" as are court transcripts and UN Security Council resolutions, all of which are quoted in articles other than those about themselves. I also think that WP:RS#What is a reliable source? is affected in a similar way. So before any change is made I would like an analysis made of how such a change would effect the reading of the whole content policies and guidelines.--Philip Baird Shearer 11:35, 21 September 2007 (UTC)

I'm not sure how you got here, but the points you just address are why the proposed policy change was announced at the WP:VP. We need input from others on how this 'proposed' change may affect the other policies as well. I view this a step towards eliminating much confusion amongst the various policies. wbfergus ^Talk 12:03, 21 September 2007 (UTC)

This discussion about changing the emphasis on secondary and tertiary sources rather than primary sources is somewhat misplaced and misguided, IMO. Primary sources that report different empirical or historical calculations than widely reported, as was mentioned in the example at the beginning of this thread (whether 135,000 is = or ≠ to the number killed in Dresden), fall under WP:NPOV#Undue_weight and WP:RS, which work in tension with one another. Primary soucces are already stated in WP:NOR to be validly used to report plain facts without interpretation or synthesizing of ideas that may arguably follow from the plain facts. Moreover, WP:NOR, in its emphasis on secondary and tertiary sources for interpretive and synthetic issues, works in tension and in tandem with WP:VER and WP:RS, along with WP:NPOV or course. These tensions are natural and intentional, and the founding Wikimedia Board would appear to have clearly understood this when the three basic editorial policies were first put forward. WP:Consensus, of course, is the method by which these tensions are balanced or resolved in the process of writing the articles. The push to alter this balance by substantially rewriting one or more of the three most basic editorial policies neglects that the solutions to the dilemma already exist within present WP policy. The dog should move the tail here, not the other way around. ... Kenosis 13:54, 21 September 2007 (UTC)

But which dog? This is an attempt to bring some cohesion to the policies, that if it's there, is very difficult to properly decipher, especially to newcomers. wbfergus ^Talk 14:29, 21 September 2007 (UTC)

Let's be at least halfway real about this, please. This recent push here is an attempt to make major changes over minor editorial problems and exceptions to the general trend on the wiki (the tail trying to wag the dog), and/or for one group to try to get their way of doing things at a few articles implemented wikiwide (also the tail trying to wag the dog). An equally informative tome on one project page is equally difficult to decipher as are three separate pages expressing three interactive principles. This was tried recently in the WP:A fiasco, raised to the status of policy and fairly promptly demoted after a wiki-wide "wtf?". The cohesion already exists in interaction of the three basic editorial policies, with consensus as a procedure for mediating among them. Those three basic policies operate both cooperatively and in tension with one another, and they do it extremely well -- so well, in fact, that the very existence of WP is not only a function of modern processing power and the wiki method, but also of the balance and tension between the three basic policies of WP:NPOV, WP:VER and WP:NOR. ... Kenosis 15:14, 21 September 2007 (UTC)

There is no general trend on the wiki to disallow primary sources in general - it's been pointed out during this process that a large number of featured articles make heavy use of primary sources. If you look back to the earlier stages of this process, you will also see that this isn't a homogeneous group who turned up to get things changed to something they (as a whole) thought it should be. Opinions varied massively at the start of the process. Your "halfway real" description seems to mischaracterise both wikipedia practice (the use of primary sources) and the nature and makeup of the group of people who've taken part in this process. SamBC(talk) 15:43, 21 September 2007 (UTC)

There is only one use of the word "disallow" on this page prior to my quoting Sambc's usage right here (so now there are two, but I didn't say this or anything close to it). Please do not put words in my mouth or on my "pen", so to speak. The present version of the relevant passage in WP:NOR states: "An article or section of an article that relies on a primary source should (1) only make descriptive claims, the accuracy of which is easily verifiable by any reasonable, educated person without specialist knowledge, and (2) make no analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on primary sources should be careful to comply with both conditions." This reasonably expresses the concept. ... Kenosis 16:25, 21 September 2007 (UTC)

As to my "let's be at least halfway real..." comment, this refers to what appears to be the present notion that this is just a minor proposal that can be implemented without further scrutiny because several WP users have worked on it and believe it is nearing consensus.My review of the situation indicates that this is in fact intertwined with an effort to significantly change the policy, up to and including rewriting the rules on "original synthesis" and other major changes. As to the "process" being referred to by Sambc, that process has been a discussion among several WP users, indeed properly regarded as a preliminary discussion, advocating one or more significant changes in WP policy w.r.t. WP:NOR. The "tail attempting to wag the dog" here is my reference to the exceptional cases driving attempts to change the basic, fundamental balances of WP:NPOV, WP:VER and WP:NOR. The use of anecdotal examples of users having problems in interpreting and resolving issues in individual articles, such as, e.g. the Dresden bombing example in this section, are properly regarded as a {{sofixit}} issue at the individual articles, not valid reasons to change the policy across the wiki. ... Kenosis 16:47, 21 September 2007 (UTC)

Material close to the subject..

If we're writing an article about, say, Encarta, we have to treat what it or its authors and publishers say about it as primary material, and aim to base interpretations or opinions on that material on third party secondary sources. That's widely understood from the current policy, and should not be lost in this push to avoid "source typing". ... dave souza, talk 21:25, 20 September 2007 (UTC)

That's a concern for NPOV, verifiability and reliable sourcing. However, it seems to have very little to do with original research. Are there potential concerns about biased, unreliable and/or inaccurate information in sources close to the subject? Absolutely. However, that's not a question of original research. Vassyana 21:41, 20 September 2007 (UTC)

But it is closely tied to this proposed distinction between "Statements of fact" and "Statements of interpretation, analysis or conclusion", so it's logical to point it out in this context. Since I've got to cook and eat something, could you please draw on your expertise to point out where this is explicitly dealt with in these policies? Thanks, .. dave souza, talk 21:55, 20 September 2007 (UTC)

I have a dinner I must attend, but I will look over those pages when I return. However, I was not intending to point to any particular section, but rather the principles themselves. Remember, the rules aren't intended to cover every possibility. They should present the principle as clearly as possible, not constrain it to the exact language used. Vassyana 22:22, 20 September 2007 (UTC)

I'm post dinner, so not a reliable source. I would just like to say that I am really impressed how people have engaged with the substance of the concerns on source typing, and everyone should be congratulated for engaging with good faith. Spenny 22:47, 20 September 2007 (UTC)

Dave... When we use the terms "statement of fact" and "statement of interpretation/etc". we are talking about what is written in Wikipedia... not what is in the source. As far as NOR goes, any statements about what the authors or publishers of Encarta say about Encarta end up falling into one of these two categories depending on how they are written: if an article states something like: "According to the Publishers of Encarta, 'Encarta is great'" (cited to where the publisher says this) then we are dealing with a statement of fact. No NOR violation.

Now, if the article just stated something like: "Encarta is great" we also have a statement of fact, but not one for which the publisher of encarta is a reliable source. It isn't an NOR violation, but it should be challengable on other grounds.

Then, if the article stated something like: "Because it's publisher thinks encarta is great, we should all rely on it" again citing the publisher, we no longer have a statement of fact, but one of conclusion... the statement is not matched by the source... instead we need a reliable source that makes the connection between the publisher's opinion and encarta actually beign being reliable. Blueboar 23:30, 20 September 2007 (UTC)

WP:SELFPUB already addresses this issue, does is not? ≈ jossi ≈ (talk) 23:43, 20 September 2007 (UTC)

Selfpub deals with limited circumstances where an unreliable source can be used. Whether Encarta's the ideal example or not, this is an issue where a source is reliable but rather close to the subject, and so hinges on the rather undefined question of "third party".

Am in general agreement with Blueboar's argument, but this tends to hinge on what other policies and guidelines call "third party sources" without spelling out what that means – by accident or design this policy goes further in warning of the danger of using a source close to the subject for anything other than plain fact (including the fact that the source expresses an opinion) and the preference of finding a secondary / third party source doing the analysis or summary, rather than selecting the summary ourselves. If the detail of this point is moved elsewhere it's still appropriate to have a summary-style outline here, directly related to the sourcing of facts and interpretations. .. dave souza, talk 23:58, 20 September 2007 (UTC) fix typo dave souza, talk 00:15, 21 September 2007 (UTC)

Considering that our proposed change does not use "third party", it isn't a factor... what I meant was a source not directly connected to the topic. Say: External source if you wish. Blueboar 01:30, 21 September 2007 (UTC)

As to the term third party, I have always been unhappy to see that in Wikipedia policies and guidelines. It derives from contract law, where many contracts have two parties; anyone who didn't sign the contract is a third party. The extention to contracts with more than two parties is fairly obvious, but when extended to the world of publications, it becomes a rough metaphor with flexible interpretation. An example of why the term does not work well is that official announcements, laws, orders, and the like from governments can be quite useful as sources, but they are first party publications. --Gerry Ashton 00:10, 21 September 2007 (UTC)

Which is why it's valuable that this policy is at least clearer about it, and though improvements are possible, we shouldn't be ditching parts of policy without being sure that there's an equally clear replacement readily accessible. Bedtime now, ... dave souza, talk 00:15, 21 September 2007 (UTC)

But why this policy? What does it have to do with original research? I think sourcing, particularly in relation to what should be considered reliable or taking in considerations of bias, is much better addressed at WP:V, WP:RS and WP:NPOV. Many of the editors here have a point that this policy is about original research and considerations that go beyond that should be addressed in the most relevant place. Vassyana 01:15, 21 September 2007 (UTC)

Material close to the subject can be more biased and less reliable than other material; however, the opposite could just as easily be true, too. A journalist recording an interview is more likely to capture the interviewee's words and intent than a scholar reading that interview 100 years later. A modern Jehovah's Witness reading the book of Hebrews is less likely to accurately describe early Christian practice than Paul was, writing in the 1st Century. It really depends on the source and what it is used for, and overgeneralizing can easily cause problems here. In many cases, facts and conclusions recorded by someone who was close to the facts is better, particularly if the issue is controversial. In other cases, facts and conclusions drawn by a modern scholar might be better, particularly if there is a modern consensus. COGDEN 00:36, 21 September 2007 (UTC)

Which is why the policy alerts the editor and urges caution. .. dave souza, talk 00:53, 21 September 2007 (UTC)

None of this is in the realm of Original Research... You are debating the reliability of sources, not whether they are being used to support original research. Any source can be used to back some sort of statement (if nothing else, it can be used to back a direct quote from that source). The type of source being cited does not determine if it's use constitutes OR... HOW it is used determines that. Thus, this policy needs to discuss the how (ie the type of statement the source is being used as a citation for) and not the what (ie the source itself). Blueboar 02:06, 21 September 2007 (UTC)

WP:NOR is one of WP's core content policies and has stood in essentially the same form for a very, very long time, and benefits from broad acceptance. It is uncontroversial and broadly accepted. I can tell you for a fact that any attempts to alter it's original meaning and formulation, such as diminshing the distinction between primary and secondary sources, are going to fail regardless of any agree between the handful of editors participating in the discussion here. FeloniousMonk 14:31, 21 September 2007 (UTC)

Thing is, that the original meaning and formulation didn't distinguish between primary and secondary sources. During this ongoing and long-winded review, the history of the page was looked at and the circumstances of that addition examined. SamBC(talk) 14:38, 21 September 2007 (UTC)

Wrong. The original formulation makes the distinction. By original formulation obviously we do not mean the very first stup of a policy - that is a meaningless criteria in a collaborative environment. Several people collaborated on crafting the original policy and by the time it took its first stable form - within a month or so of its first being forwarded as a policy - it made the distinction between wources and has been that way for four years, the vast, vast vast majority of its existence. Slrubenstein | Talk 14:47, 21 September 2007 (UTC)

Well, a quick review of the history indicates that it appeared about two years ago. 4 years ago, the policy was, I believe just being introduced (or hadn't been yet). In the early versions including it, the definition was included in order to say that OR creates primary or secondary sources, which is actually an interesting idea, although it has trouble holding water in terms of consistency. I'm continuing to review the history now, but so far it seems to bear out what other people have said earlier in this review (now somewhere in the archives). In any case, saying it's been there for 4 years is quite simply incorrect. SamBC(talk) 15:08, 21 September 2007 (UTC)

Further to this, I find the following extract from a January '06 version interesting:

In most cases, Wikipedia articles include material on the basis of verifiability, not truth. That is, we report what other reliable secondary sources have published, whether or not we regard the material as accurate. In order to avoid doing original research, and in order to help improve the quality of Wikipedia articles, it is essential that any primary-source material, as well as any generalization, analysis, synthesis, interpretation, or evaluation of information or data has been published by a third-party reputable publication (that is, not self-published) that is available to readers either from a website (other than Wikipedia) or through a public library. It is very important to cite sources appropriately, so that readers can find your source and can satisfy themselves that Wikipedia has used the source correctly.

I believe that this was included largely as commentary on the rules that preceeded it, which included a note that it was acceptable for articles to be based entirely on primary sources in appropriate situations (ie, the article has no analytic etc content). I think this is the kind of thing that the current situation grew from. SamBC(talk) 15:15, 21 September 2007 (UTC)

The same version contains a marvellous list of what's actually excluded further down the page, which appears to be the "meat" of that version of the policy. I include it hear because I think we can learn from it:

"An edit counts as original research if it proposes ideas or arguments. That is:

• it introduces a theory or method of solution; or
• it introduces original ideas; or
• it defines new terms; or
• it provides new definitions of pre-existing terms; or
• it introduces an argument, without citing a reputable source, which purports to refute or support another idea, theory, argument, or position; or
• it introduces or uses neologisms, without attributing the neologism to a reputable source; or
• it introduces a synthesis of established facts in a way that builds a particular case favored by the editor, without attributing the synthesis to a reputable source."

This seems to have been the actual body of the "rules" of the policy at the time. The sourcing information appears to have been by way of explanation. SamBC(talk) 15:20, 21 September 2007 (UTC)

The fact remains is that this policy and WP:RS has made the distinction between sources far, far longer than recent attempts to weaken and dillute it. Any revision of the NOR policy that fails to make a distinction between the quality of sources based on whether they are primary or secondary will never gain sufficient community support from long term, credible contributors to see the light of day. FeloniousMonk 15:18, 21 September 2007 (UTC)

That seems a remarkably incivil way to refer to a weeks-long review and consensus-building effort, announced to the wider community through the village pump on multiple occaisions, that has included as participants long term "credible" contributors. Especially when this discussion has been happening on-and-off for even longer. Are you aware that your statement could be read as suggesting that those who have participated in this process are not "credible contributors"? SamBC(talk) 15:23, 21 September 2007 (UTC)

Over the course of the past month over two dozen editors have taken part in the discussion over the "Sources" section of this policy. It is clear there is a lack of consensus for the primary/secondary/tertiary distinction that has been present in policy over the past year. Some of the comments that a NOR policy lacking this feature will never succeed in gaining consensus are misguided and troubling. Quite the opposite is true, or rather, a NOR policy containing these distinctions has already been shown to lack support. Policy must be back by strong consensus. I originally started off in strong support, and making a strong defense, of the primary/secondary distinction. However, the lack of consensus for such a section in this policy is clear. Knee-jerk opposition to any proposal that omits an approach clearly lacking consensus borders on disruptive. Please take the time to review the archives for the past month, which demonstrate the lack of support for the standing section and illustrate the evolving nature of this discussion. Reviewing the same archives, as well as those the months previous, will clearly demonstrate that the PSTS distinctions were the cause of repeated confusion and conflict. Vassyana 16:25, 21 September 2007 (UTC)

Not quite... lets go back and see how things developed in the early days...

• In Jan '05 the words Primary source are in the policy... but they refer to not making Wikipedia a primary source... it does not talk about source materials. In other words, it talks about not adding original material to wikipedia.
• By March '05 we see the secondary materials mentioned, and some definition and distinction between primary and secondary... but look at the context: "Original research is research that produces primary sources or secondary sources." We are still talking about Wikipedia and not cited sources.
• By August '05 the terms are broken out into their own section... but again the context is Wikipedia and not what is being cited: "Within Wikipedia research that creates primary sources is not allowed".

It was not until sometime after that we started to caution about primary sources... but as late as mid '06 we still had language that focused things on not creating a primary source. this was a key concept that somewhere along the line got lost in the shuffle... we began to focus on what the sources were, instead of how the sources were being used. In answer to Monk's concerns... I don't see this as an attempt to alter the policies "original meaning and formulation"... I see it as an attempt to restore the policy's original meaning and formulation. Blueboar 15:24, 21 September 2007 (UTC)

It is true that the policy of WP:NOR evolved to differentiate between primary and secondary/tertiary sources, based upon practical experiences with users attempting to create original syntheses in WP. WP is not an alternative publisher for new ideas. Rather, to restate the very basic, most obvious nature of the project, it is an encyclopedia. An attempt today to remove this important differentiation of primary and secondary/tertiary sources would be to turn WP over again to the temptations of some users to use it as a publishing vehicle for their own "original research", or original syntheses. People who wish to do this sort of thing will need to first get published, then we'all can consider drawing on what they've published, being cautious about WP:COI in such cases, of course. ... Kenosis 15:33, 21 September 2007 (UTC)

Okay, right, does the current proposal open some sort of loophole? Is this issue of synthesis not addressed by the current proposal? If it isn't, help make it include it, but there's a lot of confidence that it isn't necessary to refer to the 'type' of the source to do so. Or did I misunderstand you? SamBC(talk) 15:38, 21 September 2007 (UTC)

This "weeks-long review and consensus-building effort" that Sambc referred to above obviously took some work among more than one participant. But please understand one thing here. When people get involved in the wiki, they do not ordinarily keep track of every single discussion on every single policy and guideline talk page. Rather, ordinarily, we expect some continuity of policy, guideline and practice. The work of the last three weeks is laudable, but it's work by a few users with an idea, one that hadn't drawn wider attention until now, when it's beginning to draw wider attention. The "current proposal" is only one piece, AFAICT, of an ongoing effort over the past two months to rewrite the rules here. AFAICT, we're not even sure which arguments are referring to what proposal, and upon what justifications. This limited participation that led to this perception of a developing consensus, like it or not, will need to be subject to greater scrutiny by many more users to be seriously considered. Incidentally, where's the thread at the Village Pump? ... Kenosis 15:56, 21 September 2007 (UTC)

If this were an effort to create new policy, I would understand the concern, but nobody is trying to change policy here, nor could we. This is about fixing a not-quite-correct description of common Wikipedia practice that has been in the policy page for a while, and seemed for a while to be actual "Wikipedia policy", but turns out not to be so in its current form. True Wikipedia policy is not what's written on the policy page. It's what the consensus is, and sometimes that doesn't exactly match the words written to describe and document it here. I think everybody here can agree, given a particular concrete example, which statements are original research and which are not. That as-yet-unexpressed instinct we all have is the true policy. Nobody's trying to change policy. We're trying to express it in a way that we can all agree with. COGDEN 00:56, 22 September 2007 (UTC)

Except for the part about there being an effort to "rewrite the rules"... I agree with everything you just said. I certainly do not want to "rewrite the rules"... I wholy support them. I just think the rules got off kilter in this one section and we need to re-focus them. As for the thread... see: Wikipedia:Village pump (policy)#Proposed change at WP:NOR Blueboar 16:06, 21 September 2007 (UTC)

The first stub for the policy was December 21 2003. After a few edits no one worked on it until February 2004. Reddi, the third person to work on the stub, first introduced the term "primary source" in the twenty-first edit to the policy, in an edit summary, on February 13 2004. Late that same day, a few edits later, he introduced the distinction between primary and secondary sources. I think that one of the first twenty edits to the policy still falls under "original" and February 2004 was more than three and a half years ago - and thousands of edits ago. And we are talking about the 23rd or 24th edit. Really. Slrubenstein | Talk 16:16, 21 September 2007 (UTC)

But as people have pointed out, the distinction then was not being used to refer to the sources people use. It was saying "original research creates" primary (and later secondary) sources. The introduction of identifying OR in any way relating to the PST nature of the sources used is vastly more recent, as has been pointed out. SamBC(talk) 16:23, 21 September 2007 (UTC)

I second the concerns raised by FeloniousMonk and others. This discussion is interesting but do not think that a change of emphasis or dilution of one of the more long-standing formulations of this policy can happen at this stage. ≈ jossi ≈ (talk) 16:28, 21 September 2007 (UTC)

Jossi, you know I strongly favour a distinction between primary and secondary sources. However, it's fairly clear there is a lack of policy-level consensus for such a section, at least in this policy. I've dropped the issue after extensive discussion because of that continued lack of support after vigorous debate and discussions. Vassyana 16:40, 21 September 2007 (UTC)

In reply to Sambc - as others have pointe dout, the policy developed based on editors' experience. After the original expansion of the stub in february 2004 the policy was relatively stable until it was overhauled in February and March 2005. In other words, with a year's worth of experience with conflicts and ArbCom ruling we were able to explain the idea more clearly and define terms more precisely. There is nothing unusual about this process. I have no doubts that every change made in february 2005 were fully in the spirit of the february 2004 version. And after march or April 2005 the policy has been fairly stable - that is two and a half years in something very close to its current version, and well over three and a half years - and I emphasize: the vast majority of the policy's existence - in which the distinction between primary and secondary sources was central to the policy. You can argue all you want but you can't change these facts. Slrubenstein | Talk 16:33, 21 September 2007 (UTC)

I would refer you to the brief history given by Blueboar earlier in this section. In late August 2005, the policy did indeed mention and define primary and secondary sources, although I disagree that they were central. It specified that there are circumstances in which an article may be entirely based on primary sources. In that version, the distinction between source types simply does not refer to the sources of an article and is not used to prohibit, or even discourage, one type of source in preference for another. As Blueboar stated, even into last year the language relating to sources focussed on not creating a primary source. SamBC(talk) 16:41, 21 September 2007 (UTC)

That was a very subtle shift, and I think that's probably why I didn't notice the change until about half a year after the fact, having been so conditioned under the original formulation from 2004 to fall 2006. I don't think it really matters, though, how the shift came about, because what's important is to create a consensus description of Wikipedia practice now. Some editors are insisting that if the policy statement doesn't include their positive language, they will not concede. This is not how it works. There certainly has to be consensus to include a new statement of policy, but consensus is not negated because someone refuses to consent unless their controversial statement is included. If the policy says anything, it has to reflect true Wikipedia policy. If something is too controversial, it stays out. COGDEN 00:56, 22 September 2007 (UTC)

Summary of concerns on source typing

• The definitions of primary and secondary sources are unique to Wikipedia and this causes confusion.
• Sources in the context of policy are not always primary or secondary but may display elements of both.
• OR refers to the introduction of concepts and as such the classification of the source is not at the granularity of the presentation of the concept.
• Taking into account the above, source typing does not necessarily allow an automatic assessment of original research.
• Specifically, qualification of acceptable use of primary source material (which could be the ideal source for the accurate portrayal of facts) is clouded by a presumption of a secondary source being a de facto better source of a fact or concept or analysis, whereas it is the editing, not the source that is contaminated.
• Source typing is used in Wikipedia to discredit valid information presented correctly purely on the basis of it being a primary source without consideration of its actual credibility and the relationship to the edit in question.
• The potential for source typing to be a primary segmentation over the quality of source allows for more dubious journalistic sources to be granted a higher credibility than sound but primary sources.
• Valid secondary sources in the general sense have been discredited as primary sources as they present a novel interpretation or analysis.
• Source typing is potentially a technical debate, beyond the scope and interest of many typical WikiEditors who may not be academic.
• Source typing as a methodology of determining suitable sources does not transfer well to articles outside the academic sphere such as pop culture, which may still be notable and encyclopedic to many.
• The quality of a source is not something for NOR to be concerned about: that is in the remit of other policy. This is therefore policy creep.
• The underlying principles of presentation of information that exists elsewhere in sources do not require source typing, so by avoiding this concept, we can present policy in a less contentious way and gain wider credibility.

Just to be clear, the above is not a neutral summary, but the case against. Spenny 10:20, 21 September 2007 (UTC)

I would sum this up with:

• It is the use of sources, not their nature, that creates Original Research.

-- Dhaluza 11:08, 21 September 2007 (UTC)

Some counterarguments as to why categorizing sources is important -

• Primary versus secondary sources are a fundamental distinction in research, see, e.g., here. We're free to adopt a standard definition or adjust it for our purposes.
• If a source has elements of both, we can acknowledge that but it doesn't change the distinction
• There are OR concerns with respect to primary sources
• There is a significant problem on Wikipedia with editors who do not understand sourcing drawing their own conclusions from primary sources in a way that makes the material unreliable, and unverifiable even if true.
• Drawing a distinction between primary and secondary sources is the best way to educate these users about what is wrong with the way they are using primary sources.
• There is also a problem with people using tertiary sources (either wikilinks, or links to other encyclopedias) that increases likelihood of errors, and also makes verifiability difficult.
• If sourcing is not mainly an OR issue then fine, move it to some other policy or guideline page like WP:RS, but if it's important don't don't delete it without putting it somewhere.
• Indeed, a source can be more or less reliable whether it is primary, secondary, or tertiary. We can say so. Wikipedia editors are perfectly capable of understanding that quality issues overlay category distinctions.
• This is not hard to understand. Educating editors on the subject is a worthwhile task because it leads to better articles. Understanding sources is also good for people's analytic understanding of the world beyond Wikipedia.
• Source categories are transferable and especially important outside of academia. For example, people need to learn that youtube hit counts, sales data, interviews, source documents, song lyrics, etc., can be improperly used as primary sources and that (for instance) notability is established by a secondary source saying something is important, not from the editor's evaluation of the importance of the material.

- Wikidemo 12:06, 21 September 2007 (UTC)

I totally disagree with Spenny's characterizations of the policy and the comments in support. In actuallity:

• The definitions of primary and secondary sources given in the policy are derived from the common definitions used in academia and journalism and are not a source of genuine confusion. This "confusion" upon review is most often contrived in order to side-step conforming to this policy and WP:NPOV.
• Sources are almost always easily characterized as primary or secondary. There are almost no instances when a source displays elements of both.
• OR refers to the introduction of concepts that are not verifiable through reliable pubslished sources, and the classification of the source is necessary. The "granularity of the presentation of the concept" is not.
• Source typing is both necessary and sufficient for assessing whether a concept/view is original research.
• As every journalist knows a secondary source is indeed de facto a better source than a primary source for determining whether a concept or view is original research. "Contamination" in this sense is a non sequitur: Multiple secondary sources remove any such doubts.

The NOR policy is not going to so significantly change that editors will not be required to consider sources based on their relationship to the view or concept. That's simply never going to fly. FeloniousMonk 14:49, 21 September 2007 (UTC)

Monk... read the proposed language... editors are still required to consider sources based on their relationship to the view or concept. All we do is shift the emphisis from the source being used, to the statement that it is being used for. I have to ask... are you perhaps resisting any change simply because it is change, or are you thinking about why the change is being proposed and what the proposal actually says? Blueboar 15:30, 21 September 2007 (UTC)

There have been multiple proposals put on the table recently, which, taken together, amount to a radical restructuring of WP policy. Which "proposed language" is being referred to here? ... Kenosis 15:41, 21 September 2007 (UTC)

The most recent version (as of these sudden and rather amorphouse oppositions) is, I believe, at #Draft as of 9/20/07 (minor update) above. SamBC(talk) 15:44, 21 September 2007 (UTC)

"[T]hese sudden and rather amorphouse oppositions" are a function of receiving information that this discussion has been proposing significant changes to WP policy. Surely the implication is not intended to be that upon receiving notice of these proposed changes others should bow to the developing consensus of several users because they've already discussed it among themselves? ... Kenosis 16:14, 21 September 2007 (UTC)

Most current proposed language below... Blueboar 15:54, 21 September 2007 (UTC)

Thank you. ... Kenosis 16:14, 21 September 2007 (UTC)

<undent> Thanks for listing the concerns, Spenny. Right at the outset the first assertion really has to be substantiated:

1. The definitions of primary and secondary sources are unique to Wikipedia and this causes confusion.

> If evidence for this is well presented in the archives, it needs to be linked from here so that newcomers don't spend days hunting for it. There are issues with the other points which I'll try to come back on shortly, but time doesn't permit just now. .. dave souza, talk 18:08, 21 September 2007 (UTC)

Tertiary sources

Stepping back from the specific issues:

1. Is there agreement that we do not want to produce WP by collating Encarta with Britannica and expressing the result in our own language? (Many articles would be improved by this, but that's another problem.)
2. Is there agreement that one reason not to do this is that encyclopedias are no better than their sources, and so we should consult the sources?
   • PBS has come up with a particularly glaring example of this, but it's not alone.
   • On the other hand, secondary writing is, on average, better than an uncritical reading of primary sources.
3. Should this be policy?
   • And if so where, if not here? Septentrionalis PMAnderson 16:16, 21 September 2007 (UTC)

This seems to be mainly a case of reliable sourcing. Cheers! Vassyana 16:33, 21 September 2007 (UTC)

That's not an adequate answer. It begs the question of whether it should be policy; and if not, why not? Septentrionalis PMAnderson 16:38, 21 September 2007 (UTC)

I just don't think this is the place to discuss it. Our concern here is original research. If it's a question of NPOV or reliable sources, it should be addressed at those places. SamBC says it well below. Vassyana 16:43, 21 September 2007 (UTC)

Well, in response to number 1, then I expect most would agree about that. Ditto point two itself. The subpoints warrant more considerations. For point 3, that's the meat. Whether a source is suitable just isn't a matter of original research. The originality of an edit isn't affected by what source is used to write it. There must simply be a source for the information or analysis/conclusion. If someone has a reliable source as to, say, the population of a county (say, census data), then the inclusion of that in the article on the county article isn't original research, and calling it that is misleading. The policies on verifiability and point of view discuss what sources are appropriate and when, and do so (pretty) well. If they don't currently caution against what you describe, and NOR does, then that caution should probably be moved into the more appropriate policy, or into the guideline on reliable sources. SamBC(talk) 16:33, 21 September 2007 (UTC)

Agree to 1... Mostly agree to 2 (I have quibbles about the specific sub-points)... 3 is the hard part. I see this mostly as a reliability issue (the determining the reliability of Encarta and other such encyclopedias) not a NOR issue. Citing another encyclopedia is poor research (its not original research). I am not sure if we can have a policy statement about conducting poor research, so I am not sure where it should be mentioned at policy level ... but I certainly would mention it at a revised WP:RS. Blueboar 16:45, 21 September 2007 (UTC)

Agree to #1 and #2. An important corollary to #2 is that for purposes of verifiability and further research, the user will have to consult the sources even if we don't; using a tertiary source puts the reader at least one more link farther away from verifying and studying the underlying material. Re. #3, I don't see the difference betwen making it policy or guideline. Either way it's a rule to follow. It really is a matter of RS, not NOR directly - they're two pieces of the same puzzle but this stuff is in that adjacent piece. A proposal to simply remove the source criteria won't fly, but a proposal to move it to WP:RS might be fine. WP:RS is about as hard and firm as a guideline gets. In the process we should bolster RS to say that the reliability of sources is a combination of various factors, one of the weightier issues being whether they're secondary. We should make explicit that when there is a choice of sources that stand for the same thing, or more importantly when they contradict each other, the strongest source should be used. Weaker sources can still be used for further confirmation or detail, but may be removed if they add nothing. Wikidemo 17:02, 21 September 2007 (UTC)

Wikidemo's uncertainty about positions is an unfortunate result of the recent policy creep. Policies used to be aims, which we agreed on; guidelines to be rough and ready discussions of how to achieve them. Not collating the other encyclopedias sounds like an aim, and so a policy. I should add that WP:RS is a particularly bad choice of guideline; its status and existence are warmly disputed, although not by me. Septentrionalis PMAnderson 17:40, 21 September 2007 (UTC)

"We do not want to produce WP by collating Encarta with Britannica and expressing the result in our own language" is a strawman. Of course we prefer articles that are based on source research in secondary and primary sources over rewording an encyclopedia. But since this is a policy page, do we want to forbid it? Isn't it better to have an article that rewords another encyclopedia than no article at all? If an editor encounters an article that is an unsourced POV pile of manure, but he/she has neither the time nor the background to do good research, to rewrite it based on encyclopedias, rather than leave it as it is? The situation reminds me of Charles Babbage, who lost his financial backing because he was such a perfectionist that he seldom finished anything. --Gerry Ashton 17:50, 21 September 2007 (UTC)

Personaly, I agree with all three. However, historically this policy has wavered back and forth concerning tertiary sources - sometimes it has addressed the issue, sometimes not. I think that there is enough community support to say that we should strive to draw on the same kinds of sources as other encyclopedias, rather than other encyclopedias themsevles ... but I think the support is for a preference and not a prescription. In my own proposal, I suggest that consulting other encyclopedias is a useful part of research, especially when they provide bibliographies, but should not be used as sources unless signed (non-anonymous authorship) i.e. can be pegged to an identifiable view. I think for the sake of consensus the policy should express these values but as preferences not as prescriptions, unless we want to open up a big debate on the issue. The fact is, LOTS of wikipedia editors rely on other encyclopedias. I happenb to think this is a very bad idea for all sorts of reasons but the fact remains, they do it, and if they were invited in on this discussion I bet they would express real resistence. I agree with Jossi and FeloniousMonk and others that there is a longstanding consensus (or as close to one as anyone will ever get at Wikipedia) for the distinction between kinds of sources. But there just isn't a strong enough redord concerning opposition or strict limitation to use of other encyclopedias. It IS a discussion we ought to have, but (1) maybe not here and (2) if here, definitely not now. One step at a time. Slrubenstein | Talk 17:53, 21 September 2007 (UTC)

Slrubenstein makes very good points... but fails to explain how the classification of sources relates to the concpet of "No Original Research". Perhaps someone who was involved in the developement of the PSTS section in this policy could explain this to me?

Folks, this isn't about "getting rid" of PSTS ... it is about the fact that PSTS does not belong in this policy. Source typing just isn't an NOR issue. ANY source can be misused to support OR... the type does not make it OR... the misuse does. Blueboar 18:43, 21 September 2007 (UTC)

Fair enough on the issue of tertiary sources. A complete replacement of the section on Primary, secondary and tertiary souces with the proposal developed here in the last two or three weeks would be a radical reformulation of the policy without effective community feedback. More plausible, it appears to me, would be a revisiting of the issue of other tertiary sources (WP itself being reasonably termed a "tertiary source"). I should caution, however, that tertiary sources are valuable sources in many areas of the wiki, due to the numerous specialized encyclopedias covering various topic areas, e.g. philosophy, linguistics, mechanical principles, etc., etc. Go look in a good library sometime-- they're just filled with them, valuable specialized summary resources on just about every topic area you can think of. The notion that tertiary sources are limted to just Encarta, Britannnica and the like, is simply a mistaken view of the range of tertiary sources that are widely drawn upon in writing articles throughout the wiki. ... Kenosis 18:49, 21 September 2007 (UTC)

But what has this to do with "No Original research"? Nothing. This 'subject' can, and should be, included as either a "guideline" or possibly as a watered down policy all on it's own. Again, the type od source material used does not constitute "original research". It's how the source material is used (especially after inclusion in Wikipedia), that may lead to "original research". The type of source material has nothing to do with it. Any kind of source material can be twisted, mis-interpreted, or just flat-out "changed" by an editor to insert "original research". wbfergus ^Talk 18:58, 21 September 2007 (UTC)

What it has to do with WP:NOR has to do with describing the range of syntheses that are already available, and how they fit, in general, into the concept of using published material in writing articles devoid of original syntheses, a.k.a. original research. The descriptions assist in clarifying the flow of information from primary sources through summary resources, in order to allow users a better perspective on how to avoid original research. This information is valuable to WP users because WP is quite different from "research papers", where we're commonly expected to draw on sources and come up with an original synthesis. The distinction is extremely important in placing in perspective the range of sources available to users. As such, it does cross over into WP:V to some extent, but not to a very great extent. In this project page, it is set in the context of avoiding original research while still making editorial decisions about how to write an article. ... Kenosis 19:12, 21 September 2007 (UTC)

Blueboar notes, "but fails to explain how the classification of sources relates to the concpet of "No Original Research"." And asks for an explanation. Wikipedia has an NPOV and a V policy which incouraged editors to cite sources as away of includin diverse views in articles in an unbiased way. Several editors became aware that some people were citing sources to introduce material in articles in biased ways. Mere compliance with V was not enough. And biased edits often seemed to comply with NPOV because they clearly identified the point of view of the cited material. How were people seeming to comply with NPOV and V, while violating them, at least in spirit? By selective quoting, by duxtapposing different quotes in order to make the editor's point, and by using primary sources to argue (the editor's argument) that a primary source was wrong. So it became evident that it was not enough to insist that editors had to use reliable sources. We needed to emphasize that editors' views shouldn't go into articles, and we needed to show how one could seem to comply with V and CS and still be forwarding one's own arguments, which meant we need to show some dangers in how editors could use sources, and encourage other ways to use sources. Slrubenstein | Talk 22:24, 21 September 2007 (UTC)

Thank you, Slrubenstein... I agree that selective quoting, Juxtapposing different quotes in order to make a point, and misusing sources are all NOR... but I don't see how classification by type gets into it. Can't the missbehaving editor do all those things with any type of source? And I think our proposed wording address these problems...

I would say that selective quoting is making a statement of fact and not citing something that clearly demonstrates that fact, because the fact is taken out of context... this is covered in our proposal. Juxtapposing different quotes to make a point is a form of sythesis... and we are not removing the synt section in this proposal. Making an argumentative statement without citing to a reliable source that makes that same argument is covered in the proposal. The types of sources used seem irrelevant to the NOR violation. Blueboar 23:52, 21 September 2007 (UTC)

Proposed language as of 9/21/07

(To replace the section entitled: "Primary, secondary and tertiary sources"...)

Citing the Right Source Materials (alternative suggested section title: Sticking to the Sources)

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

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

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

Comments on 9/21 draft

above is the most recent as of today. Blueboar 15:55, 21 September 2007 (UTC)

Thanks. I would argue that the wording there is not only dense, but it dilutes substantially the current formulation of this policy. ≈ jossi ≈ (talk) 16:30, 21 September 2007 (UTC)

What problems or loopholes do you see? Vassyana 16:44, 21 September 2007 (UTC)

The wording is quite convoluted, and it needs wider community input. After that WP:ATT disaster, I am unwilling to contemplate wide-ranging changes to policy without such. ≈ jossi ≈ (talk) 17:19, 21 September 2007 (UTC)

Jossi, how exactly does it dilute it? I'm not saying that you're not accurate, just identifying the specific problem is what's needed. SamBC(talk) 17:53, 21 September 2007 (UTC)

• Summarily reject the proposal. The proposed replacement above for Primary, secondary and tertiary sources should be summarily rejected, at least until it is thoroughly scrutinized by the wider community for an extended period of time. It represents a complete change to the ongoing efforts of numerous editors and long discussion over the years, in favor of the work of a focus group of three or four users at this talk page. ... Kenosis 16:59, 21 September 2007 (UTC)

It is being proposed exactly so it can be thoroughly scrutinized by the wider community. Obviously those who have worked on it hope it will gain larger community consensus... but if you don't propose it, you can't find out if that larger consensus exists. And while it does represent a significant change to the ongoing efforts of editors... it does not represent a significant change to the INTENT of this policy... in fact it clarifies that intent. Kenosis mentions, in a thread above, his concern that this is the "tail wagging the dog"... This to me is actually a good discription for why I decided this change was needed. The tail of source typing has been wagging the dog of NOR. We need to get back to NOR and away from source typing. I can see source typing being useful... and if we can figure out where to put it, I would not be opposed to having it in some other policy or guideline. It simply isn't right for the NOR policy. Blueboar 17:16, 21 September 2007 (UTC)

Summarily rejecting something until it is scrutinized is silly. No-one's asking for a final decision now, or at any specific time. We're discussing ideas along a general proposal, and the proposal needs work, no-one is likely to disagree with that. I think there's some misconception that, if these new objections hadn't been changed, the page would've been changed by now, or imminently (imminent in the sense of hours or 1-2 days or something). SamBC(talk) 17:52, 21 September 2007 (UTC)

The comments I read above show no indication whatsoever of preparing to be scrutinized by the wider community. Rather, it appears that several users are advocating implementation of the proposed change based on the very limited participation on this talk page thus far, claiming either consensus or near consensus at present. I note that it was just mentioned at the Village Pump two days ago. ... Kenosis 01:03, 22 September 2007 (UTC)

Kenosis, we're no longer really discussing the removal of the Primary/Secondary/Tertiary formulation. We don't need wider community consensus to know there's no consensus for that. It's pretty obvious at this point. We're just trying to do something positive here. The alternative is just to have a blank section, and nobody wants that. I really haven't seen any real criticism of the proposal above. All I hear is defenses of the already-proven-controversial PST formulation. That ship has already sailed. COGDEN 01:10, 22 September 2007 (UTC)

For my own part, I am relieved to hear that, because the change in that explanation of sources from "primary and secondary sources", which had been in place for at least three years, to "primary, secondary and tertiary sources", implemented last October, solved many conceptual problems around the wiki. The proposal to remove the entire section was, to say the least, a bit much. Thank you for so noting as you did, COGDEN.. ... Kenosis 02:19, 22 September 2007 (UTC)

Kenosis, I think you've misunderstood Cogden, what was meant was, I believe, that we had stopped debating the removal of the PSTS section (from NOR, not from all policy) simply because discussions had made it clear there was no consensus for it being there, now. Sure, there was consensus before, but consensus can change. I disagree with Cogden somewhat, because there seems to be no consensus in either direction, which is especially clear thanks to the new participants in this discussion. However, there being no consensus in either direction would imply that there needs to be a discussion to establish how to satisfy the concerns raised by removing PSTS (from NOR) without having to keep what has now become a very controversial section. SamBC(talk) 02:34, 22 September 2007 (UTC)

That said, it has become less objectionable to many since it was changed to no longer give clear and general preference to secondary sources (over primary ones). However, IIRC it was changes around that, among others, that led to protection for the page. SamBC(talk) 02:34, 22 September 2007 (UTC)

Where are you going to propose this, so it can be scrutinized by the community? Or perhaps, if sufficient opposition is expressed, it will not even be proposed?--Filll 17:13, 21 September 2007 (UTC)

Notice already posted at the Village Pump (policy) page... and I posted notice at V ... Please let me know if there are other venues where I should post a notice. I definitely would like to get broader community imput and consensus. Blueboar 17:22, 21 September 2007 (UTC)

Does the proposal include putting the primary/secondary/tertiary source section somewhere else like WP:RS? If so then it's more of a housekeeping function, moving the sourcing stuff to where it belongs and leaving in this policy only the part that directly relates to original research. This work has already been noticed in multiple policy pages and notice boards. When people settle on a final version then perhaps it's time to remind people again on other pages and give them another chance to review and comment.Wikidemo 17:19, 21 September 2007 (UTC)

I think people generally envisage that happening; exactly where is still a matter of debate, I think, and I doubt anyone would "jump the gun" and lose the guidance totally. Thing is, the different care to be taken with different types of source (in the PSTS sense) isn't a matter of policy, it's a matter of guidance. If the policy is to include guidance, it would make sense for that guidance not to be intermingled with the actual rules. SamBC(talk) 17:59, 21 September 2007 (UTC)

I think it is too early. The intention is commendable, but the wording needs a lot of work as not not lose the distinctions. ≈ jossi ≈ (talk) 17:20, 21 September 2007 (UTC)

To Wikidemo, Re: moving the PSTS section... I have no objection to moving it somewhere. The sole reason why I am making this proposal is that it does not belong here. It isn't a NOR issue. It probably works best at RS conceptually, but as that page is in such a shambles, I am not sure that is realistic... and I know some people want it at policy level. Could it work in V?

To Jossi... what distinctions are you referring to? Blueboar 17:28, 21 September 2007 (UTC)

Reject NOR by definition is a proscriptive policy. While we should be clear about what is allowed and what is not allowed, the emphasis is on what is not allowed - e.g. original synthesis. The Roosevelt example is unhelpful because, even if it is permissable, what readers of this policy (which says "no") are going to need most clarity on and explanation of is what they should not do. We have "Cite sources" to encourage people to cite sources. This policy is about the different ways sources (which in general we have to use) can be misused and discussion of sources shoudl be with this purpose in mind. Slrubenstein | Talk 17:37, 21 September 2007 (UTC)

No... the policy is NOT about the different ways sources can be misused... it is about including original research in Wikipedia articles... and it says "No" to that. I agree that sources can be misused in a way that fits OR ... but the type of source is not the problem when that happens, the misuse is the problem. Any type of source can be misused, be it primary, secondary or tertiary. There are also appropriate uses for all three types of sources. Again, where it comes to the concept of No Original Research, the type of source is not the problem... the misuse of the source is the problem. Blueboar 18:52, 21 September 2007 (UTC)

Comment there was widespread disagreement about what the current written policy, particularly the source-typing sections, mandates, but there was more agreement about what it aims for; there was also disagreement about how it is used in practice, which may vary from page to page. The proposed text is supposed to describe the current policy (in its aims and in practice) more clearly than the current text does. The old source-typing can be improved on another page, to remove discrepancies between Wikipedia descriptions, and to address differences between Wikipedia and various scholarly fields. Can anyone point to one place where the old text discouraged a practice, which was generally rejected, and the proposed one encourages it, or vice-versa? Jacob Haller 18:37, 21 September 2007 (UTC)

Okay, either I missed something in the wording above or I need new glasses. Somehow I missed that this (today's) version of a "proposed" substitute for the current "Sources | Primary, secondary, and tertiary sources" section was being put to a vote. It solely appears to me that it is simply todays version for people to comment on or suggest additional additions or other changes.

In regards to the various comments from some new names I see up above, let me try to clarify a couple things again. For at least a month now, there has been much discussion primarily focused on the "Sources" section of this policy. Many of us (actually all of us) who have been in engaged in the 'limited' discussions on this agree that we do not want to water down the policy in way, shape, or form. But, many of us have also come to agree through many 'spirited debates', that including the "Primary, secondary, and tertiary sources" section confuses the policy to a point that it currently provides more confusion than guidance, especially when viewed in conjunction with the other policies and their rewording of the same principles. We (the many editors participating in this discussion over the last month or so), are not trying to back-door anything. We thought that it was simply best if we could get close to something that we could agree on, and then tell others what we are working on, so that maybe reaching some kind of concensus would have been simplified by many arguments already having been discussed and argued back and forth for a while. That is why we finally posted this 'proposal' over at Wikipedia:Village pump (policy)#Proposed change at WP:NOR. If we wanted to backdoor this, we definately wouldn't have advertised our intentions or asked for further input from others.

So again, I will ask you to think about why a policy (in this case one about 'No Original Research') needs to also include extraneous definitions and examples of what the various types of source material are? What does that have to do with "No Original Research"? We can clearly state that original research is definately not allowed without worrying about how to classify different kinds of source material. If not, then I think we should all go back to school for a few more years.

Including the "Primary, secondary, and tertiary sources" section in the policy is like having an article about Apples, and then having a sub-section devoted to disambiguations, so road-apples and horse-apples aren't included as well. It may be related, but nothing at all to do with the main subject of the article.

Instead of just saying Disagree, try offering some constructive criticism on how to make it better.wbfergus ^Talk 18:52, 21 September 2007 (UTC)

I think we have two questions. Both came from the source-typing debate, but each can be dealt with individually:

1. Should source-typing be in NOR or someplace else?
   1. If we remove source-typing from NOR, are there other concerns its removal may raise in NOR?
2. Should the proposed section be added to NOR?

There were several attempts to (1) fix source-typing within NOR (to resolve different definitions) and (2) address various concerns regarding appropriate uses of primary sources, appropriate uses of secondary sources, etc. within NOR. I think we eventually agreed that RS could resolve the occasional issues with the sources themselves, and that source-typing didn't help with the more common issues with their misuse. One proposal referred to "primary material" and "secondary material" within each source, then "interpretative material" for secondary, and eventually that led to this. Jacob Haller 19:07, 21 September 2007 (UTC)

• Summarily reject the proposal Oh yeah, that's just plain silly, right? Except that it doesn't mean that one didn't look at the proposal; it means that one looked at it and chucked in the trash as being, well, plain silly and quite stupid and not worthy of a long response. &#0149;Jim62sch&#0149; 23:48, 21 September 2007 (UTC)
  • Well, it's only polite to give people constructive feedback. To describe something that a number of editors have worked on seriously as "plain silly and quite stupid" can't really fail to be disparaging of those editors. Why is there this view that we're all being troublemakers? What happened to WP:AGF? SamBC(talk) 00:24, 22 September 2007 (UTC)

• Summarily reject the proposal as Not An Improvement. I fail to see, in all this debris, what is perceived as "broken" that is being supposedly "fixed" by the proposed changes. NOR by definition is a proscriptive policy; it defines what is not acceptable. Muddying the waters with non-germane ruminations about other things is inappropriate to this policy. KillerChihuahua^?!? 23:57, 21 September 2007 (UTC)
  • If that's the case, then how does the current section on sources not muddy the waters? It spends a lot of time on things other than describing what's not acceptable. On the flip side, it's not exactly unusual to define something, in part, by that it is not, as well as by what it is. It generally aids understanding. See the current section on synthesis, which follows a similar format to the proposal. SamBC(talk) 00:22, 22 September 2007 (UTC)
    • Are you speaking of WP:SYN? Please link rather than say "current section" - I don't know whehter you're talking about on this talk page where there is a discussion about the WP:SYN section, or the actual WP:SYN on teh policy page. Specificity helps, thanks much. KillerChihuahua^?!? 00:28, 22 September 2007 (UTC)
      • Sorry, yes, I mean what's currently in the policy - I've gotten out of the habit of linking universally in this discussion because it was becoming too redundant (and by this discussion, I mean going back further than the last couple of days). In general, if I say "current" I mean what's on the project page now, whereas "current proposal" means what's currently proposed. It seemed clear to me, so I apologise. SamBC(talk) 02:10, 22 September 2007 (UTC)
        • Thanks, I'd rather be a bit anal about ensuring we're talking about the same thing, than risk talking at corss purposes and confusing one another. Your comparison of Synth and the proposed changes to PSTS omits the rather important distinction that whereas Synth describes and identifies what Synth is, and why it is OR - hence one definition of a Bad Thing - PSTS has the task of describing three things, none of which are Bad Things, bur rather are different levels or types of Good Things - sources - and states clearly NOT having them is a Bad Thing. In other words, the comparison is inept. It isn't even apples to oranges; its more like poisoned mushrooms to haystacks, where the goal is to feed a horse. KillerChihuahua^?!? 10:15, 22 September 2007 (UTC)

Does anybody actually have any specific criticism of the proposal? Because "it's not an improvement" is not a criticism. The question is, is it an improvement over saying absolutely nothing on the subject? Because that's our alternative. If we can't find consensus on this topic, we're going to have an empty policy that helps nobody. I'd encourage people to take a good look at the proposal and see if they can't improve it. It's the closest thing we've had yet to a consensus proposal. COGDEN 01:14, 22 September 2007 (UTC)

The section on Primary, secondary and tertiary sources is a vital section that puts WP:NOR into perspective for users around the wiki. The proposal to replace the section or move it elsewhere is a major change in this policy page that I find unacceptable, at least without something that represents a clear improvement on it. Since well over three years ago, a distinction had been made between primary and secondary sources and what uses constituted original research or original synthesis. A significant number of WP users remained confused about the distinction, e.g., "well, if WP is not supposed to synthesise material directly from primary sources (except for purely factual matters), and if we're supposed to use secondary sources, then how is WP a secondary source?" The extra note about tertiary sources was inserted into WP:NOR nearly a year ago, and suddenly it made better sense to most other participants and much of the confusion diminished substantially. The description of encyclopedic matter as primarily tertiary helped to resolve this confusion right on the policy project page. Since then, it's tended to serve us well in resolving confusion about what is original synthesis, or original research, as distinguished from original wording and editorial decisionmaking.

...... Admittedly there remain residual bits of confusion. One user mentioned Encarta and Britannica, which are also "tertiary sources" according to the delineation presented in WP:NOR. I pointed out somewhere in these rapid-fire multiple threads today there is a wealth of more specialized tertiary sources such as specialized encyclopedias, to which we can also add many textbooks as tertiary sources. Indeed one could readily also argue that there are numerous sources that are quaternary, quinary, senary, septenary, octonary, etc. etc. Categorizing sources as primary, secondary and tertiary has, nonetheless, served quite well in clarifying the basic types of sources from which material in WP is drawn. If there is a proposal to better clarify the distinction or explain that the delineations between them are not always sharply drawn, or whatever, that would, as I said in the section below, be quite welcome and potentially helpful. But the proposal to replace it with the language at the top of this talk section is, IMO, completely unacceptable on the justifications given thus far. I believe I used the words "summarily reject" to covey my summary opinion about the proposal. ... Kenosis 02:47, 22 September 2007 (UTC)

Attention newcomers to this article

In an effort to minimize confusion regarding how this current discussion of replacing the "Sources" subsection on PSTS came about, please refer to the Archive for the last month or two. It's an awful lot of reading I know, but it aslo way to much to even begin to try to re-link back into here. This is not a "let's get this done today" proposal. We have been discussing this for a while and will continue to discuss even longer.

Take your time to read the Archives and get caught up on what you think the pertinent issues are before posting comments, especially derogatory or accusatory comments. Probably the very earliest this "proposed change" would or could (or even should) take effect would around the end of October. We need plenty of time to assimilate everybody's constructive feedback, comments, and suggestions. Thanks you. wbfergus ^Talk 19:16, 21 September 2007 (UTC)

Having spent quite a bit of time trying to get up to speed on these arguments by reading the archives, please be assured that a coherent case needs to be put together and "read the archives" doesn't present your arguments properly. .. dave souza, talk 20:10, 21 September 2007 (UTC)

Okay, a new day, starting to get my caffeine and nicotine levels back to normal, and I had about 6 hours of sleep, so let's see if I can coherently try to state what we are attempting here. This perticular subject of what we are attempting has been scattered all over this page the last few days, so maybe this section title will draw a few of the newer people to this discussion here.

First, we are not, in any way, shape or form, attempting to weaken the NOR policy. We are attempting to make the policy itself more coherent and concise, by having this policy only deal with NOR related issues.

Second, this mainly came about because of numerous edit-wars on many different kinds of articles where one of the parties stated something was "OR". Many of these also related directly to the interpretation of primary, secondary or tertiary sources that were based solely on this policy's brief definition of them, not on the more detailed explanations (and examples) offered on their respective pages at Primary source, Secondary source, and Tertiary source.

Third, "Sources", in one form or another, appear of multiple policies and/or guidlines. For an example, open each of these in a new window and compare:

• WP:PSTS (from this policy)
• WP:UNDUE and the section immediately following it, WP:NPOV#A_vital_component:_good_research
• WP:V#Sources
• WP:BLP#Sources

These all attempt to describe the same issue, but in different ways. Especially in regards to source-typing, why should each policy even attempt to re-define these, when they are easily linked to the aforemention articles created solely for the purpose of defining the various types of sources?

Over the last month that I have been here, there have been numerous examples of how OR can be inserted into an article, either intentionally or unintentionally, regardless of the type of source used. As a matter of fact, as several other threads on this particular page (not even in the archives yet) have shown, secondary source material can be mis-applied just as easily (if not even more so) to create original research than just quoting the raw data from a primary source can. With secondary material, one could easily cherry-pick various sentences from the source(s) to build an argument promoting a POV that is not even contained within the source. This is usually much more difficult to achieve with data from primary sources, unless the editor tries to add their own conclusion or interpretation, which can easily be challenged to cite an appropriate source to verify the claim.

Trying define a type of source (some sources can also be multiple kinds of sources, depending on the context of what is being cited to it), and general blanket statements can easily be either misconstrued or otherwise misapplied by different editors at different times. The type of source has nothing at all to do with whether or not OR is being inserted into an article.

In short, many of us (around 20-30 different editors in the last month), though still not in complete agreement on wording or implementation yet, are starting to lean more towards this policy (and maybe the others) being more like an insurance policy. Everything is not defined just in the one policy. Much of what is in your insurance policy actually comes in whole or in part, from either another policy or other 'document'. The entire policy is the result of multiple "definitions" taken from various documents, and there is no overlap of those definitions when all the pieces are put together. With the current state of the policies, there is much overlap and redefinition of the "sources" subject over and over and over, with variations in those definitions. To alleviate confusion, we would like to remove this confusion from this policy and place it in a more logical place, that (hopefully) the other policies could link to similarly. We are not proposing to remove WP:PSTS, just move it out to a better place, so that PSTS would still be part of the policy, just not defined within this policy. Similar in wording to how this policy and the others, state "Since NPOV, V, and NOR complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all three". Instead,, probably something along the lines of "Since NPOV, V, NOR, BLP and 'source typing' complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all, as a good Wikipedia article will adhere to all".

I probably worded that last sentence poorly, but I think you should be able to get the idea of what we are trying to accomplish. If the "source typing" was moved into it's own article, having the common definition that the others can all share makes this concept easier to comprehend. I think more thought and effort needs to go into what and where this "source typing" article should be, like a guideline isn't strictly enforcable, but do we really need a policy to define the types of sources?

Thanks for reading this, hopefully this may alleviate some concerns about this being conducted behind closed doors as an attempt to weaken current policies. wbfergus ^Talk 11:02, 22 September 2007 (UTC)

should we make a proposed change essay page?

Would it make sense to create a "proposal" page... explaining the rational for the change, the proposed language, and the various options on implimentation (such as a) moving PSTS to some other policy or guideline, or b) creating a new guideline out of PSTS)? I know something like that is really meant for entire page revisions or new proposals.... and it is rare for policy change discussions, but it might help people FIND what is being discussed, and get a sense of the basic issues (pro and con) without having to sort through all the chatter or the archives. If people like this idea, I would suggest such a page say prominently that that comments continue to be made here, on this page. (My concern would be that someone would come along later and the say... "gee, I didn't even know the issue was being discussed"... my suggestion is so that both the proposal and it's discussion are prominent. Blueboar 19:39, 21 September 2007 (UTC)

I think it would probably be very helpful for newcomers, to quickly find out what is going on, why, etc., and clearly see that has not been a "behind closed doors" discussion by a few people to slip in some wording to water down the policy. wbfergus ^Talk 19:44, 21 September 2007 (UTC)

Thanks, as stated above it's really necessary for the case for each aspect of this change to be presented in a clear and accessible way with appropriate references to allow informed discussion. ..dave souza, talk 20:13, 21 September 2007 (UTC)

That would probably help the situation to move forwards. You seem very capable with these presentation matters, Blueboar, would you like to do the honours? SamBC(talk) 20:30, 21 September 2007 (UTC)

Hardly any point, really. Like others have pointed out, the idea is fatally flawed and will "never fly". Can't say I'd ever support it, myself. The proposed changes I've seen are a stereotypically Bad Thing. Odd nature 21:13, 21 September 2007 (UTC)

I would not stop anyone to work on improvements to policy pages. If Blueboar and others want to start a sandbox page to display their proposals for improvements, why not? ≈ jossi ≈ (talk) 21:22, 21 September 2007 (UTC)

Would you care to do us the courtesy of explaining how it is "fatally flawed" or "a stereotypically Bad Thing"? The people who've been working on and towards this aren't all newbies. SamBC(talk) 21:31, 21 September 2007 (UTC)

Yes, that's a rather dismissive comment, and is not at all helpful. Would you please explain why we should not just dismiss it instead? Dhaluza 22:48, 21 September 2007 (UTC)

Agreed that a "I refuse to even read this" or "I wont' consider this" or "it will never work" is not a bona fide expression of support or opposition and pretty much counts as no opinion. An "I don't like the wording" without a statement of what is wrong with the wording is not much better. However, do take heed because it may well be an accurate assessment of what other people do in fact think. I'm still a little lost as to exactly what's being proposed. I don't care to participate in the drafting and debate buf this is merely an effort to put the primary/secondary/tertiary source in its proper place I'm wholeheartedly behind that in principle and I would probably support a proposal if a specific, firm one is made. If it's a proposal for gutting that section without finding it a good home somewhere else, I still need to be convinced and from what Jossi and others seem to be saying that proposal would likely be DOA.Wikidemo 22:56, 21 September 2007 (UTC)

BTW, if WP:RS is a bad home for it merely due to quality concerns then we should clean up WP:RS. If it's that WP:RS is only a guideline then we should make it a policy in cleaned-up form or merge it with WP:V or something. Or put the sourcing policy in WP:V. Wikidemo 22:56, 21 September 2007 (UTC)

(NOTE: I wrote this the following comment was made, so some overlap). Thanks Wikidemo for the constructive comment and feedback. You've peeped in here now and again so you have some idea about what's going on, even if you don't know all of the minor details of each discussion. Others are recently arrived to the discussion and basically have no clue, so if you are confused, the newer memeber of the discussion must totally be lost, as I think we've seen from other posts. I just went back through the edit history for this page to see when I first started, and it was Aug. 23, with around 2,000 comments/edits since then. That's an awful lot of 'stuff' to read through to try and get a feel for where this subject just when I first joined this discussion, and it looked like this was going on before I joined.

I think that what is being proposed is that the primary/secondary/tertiary section only be replaced here with something more NOR related, but that the information in that section needs to be moved into another article/page/guideline/policy, which could then be easily referenced from all of the others using similar language. I really don't think that anybody has found a good 'home' for it yet though. Is it more appropriate in WP:V, WP:RS or maybe a new 'page' is beyond me. I think if we can agree on where this part of the policy would be more appropriate and work with the folks actively discussing that 'page' (policy or guideline), then it will be easier for others to see that we are not 'gutting the policy', as that guidance in whatever form, will still be applied, but instead proposing to make this policy easier to understand. There, I think I said all that correctly, if not, I'm sure others will correct me. wbfergus ^Talk 23:26, 21 September 2007 (UTC)

OK... I will write up a proposal sub/page this weekend ... I would welcome input from those opposed to the proposal as well as those for it.

As for what the proposal is ... today some new ideas were put up for consideration which I think have some merrit, but need further discussion. Here is my suggestion....

• Proposed: That the PSTS section of WP:NOR be replaced with the "Citing the Right Source Materials" section contained in the 9/21 draft (above). The PSTS section will be moved to either another existing guideline or policy page, or to its own unique guideline/policy page (exactly which page to be determined by community consensus prior to the replacement and move.)

Are we agreed that this is the proposal? (note, I am not asking if you agree with the proposal... I am just making sure that this is the proposal.) Blueboar 23:14, 21 September 2007 (UTC)

I think we can break it into two complementary proposals:

1. to add "Citing the right source materials"
2. to move PSTS to another suitable page (V, RS, Classification of Sources, etc.) and

(However, CRSM should cover any gaps which PSTS' move would have opened, so the two parts go together). Jacob Haller 23:25, 21 September 2007 (UTC)

I would suggest a cautious approach. I do not think that any of these wide-ranging changes will have any traction. Work on improving rather than changing may be a more fruitful approach. ≈ jossi ≈ (talk) 00:00, 22 September 2007 (UTC)

Jossi, there was a lot of effort to improve it, and nothing met with anything approaching consensus even within a small group. The first thing that consensus was apparent for was that the current PSTS stuff has lead to confusion and difficulties, which was shortly followed (as I recall) by a general acceptance that this distinction between sources isn't actual vital to prohibiting, identifying, or helping people to avoid OR. I have a feeling that the recent post to VP wasn't the only one about this during the whole discussion either, so it's not as if the discussion has been behind closed doors. In essence, though, the point of this "change" is improvement. SamBC(talk) 00:09, 22 September 2007 (UTC)

"Right" sources? That's hovering dangerously near "truth" - "right" is inherently open to interpretation, and is subjective. Primary, Secondary, and Tertiary are clearly defined and not open to abuse of interpretation. I do not see how reducing clarity and confusing the issue and inviting arguments by using subjective phrasing is in any way anything but a recipe for disaster. KillerChihuahua^?!? 00:02, 22 September 2007 (UTC)

If I might interpret that into a constructive suggestion, what about "most appropriate" rather than "right"? Several of us have seen abuse of interpretation of the PSTS stuff, although few would disagree that the distinction between primary, secondary, and tertiary sources is useful in editing wikipedia. SamBC(talk) 00:11, 22 September 2007 (UTC)

Bear with me here, this is going to take a bit. I don't see the point in merely changing the title to something which is not clearly an improved title. While "Citing the most appropriate source materials" avoids the word "right", if I understand you correctly, you wish to have some sort of explanatory content there, and shuttle the PSTS information off somewhere else, is that correct? KillerChihuahua^?!? 00:22, 22 September 2007 (UTC)

That over-summarises a little... the intent is to provide more useful information as to sources being appropriate to content in order to demonstrate that content isn't OR, which isn't directly related to whether the source is primary, secondary, or even the slightly-less-well-defined (in policy and in general) tertiary. PSTS isn't directly and especially relevant to OR, although it is relevant, and it's also relevant to V and RS, to my mind. The definitions (and possibly guidance) in the new location could then be referenced from any policy or guideline which needs to do so. I personally imagine that this may be (briefly) referenced from the new section in OR, stating that secondary sources are typically good sources for analytical (etc) content, while primary sources are typically most useful for factual statements in content. Having the whole definition in the policy for the purpose of that would seem excessive. SamBC(talk) 00:30, 22 September 2007 (UTC)

Apologies for "over-summarizing" the language and ideas here have become so convoluted and complex I'd prefer to start with simple and work up to details, if you see what I mean. If I follow what you're saying, its a qualified Yes. I can see the point of clarifying the PSTS section, but I don't see that adding another layer, by first discussing "sources being appropriate to content in order to demonstrate that content isn't OR" (a statement I find confusing, actually, and would appreciate if you would attempt to rephrase, as your meaning is not clear to me), is going to help with anything at all. While I agree that PSTS is relevant to V and RS (one of the reasons for the up then down ATT), I don't see that adding another layer of complexity does anything much but offer a way to back-burner this rather important distinction, and possibly open the door to yet more creative interpretations by edit-warriors, as rather than a simple PSTS they will have a discussion about appropriate sources which gives two opportunities for muddiness rather than one. KillerChihuahua^?!? 01:25, 22 September 2007 (UTC)

Sorry about the complexity, I think those of us closely involved are sometimes unaware of the collective thoughts that have been built up and a situation-specific "shorthand" may have developed. Hopefully with new participation this can be unravelled, which should be of benefit even to the people who've been at this for weeks. Basically, knowing the PSTS status (in the general, classical sense) of a source simply doesn't give as much help in determining the appropriateness of the source as the wording in the current policy seems to indicate to many people. Analytic statements don't require a secondary source, and factual statements can come from almost any source, but the primary source for the fact is quite often "better" (more reliable, neutral, etc). I don't believe that that aspect can be eliminated, although tight WP-specific definition of PSTS would help. However, if we're going to have tight WP-specific definitions, then it becomes important to ensure that they are WP-wide as well as WP-specific, and that the same definition is used on all policies that refer to them. In that case, it makes sense for it to be on a seperate page from the policies in question, although there may be other things worth putting on that page (this was the impetus behind the original creation of Wikipedia:Classification of sources, which has since become confused and a tad stale). With or without this seperation, however, it has been felt that NOR would benefit from guidance on selection of appropriate sources to demonstrate that content isn't OR that isn't based squarely on PSTS (for the reasons discussed earlier in this paragraph).

Apologies for my writing being less than clear - I'm not at my best at the moment, due to getting over an illness IRL (you don't want to know what, but it does mean that I'm on a pretty high dose of codeine). Also, the preceding paragraph is a combination of my summary of the discussion here over the last few weeks, and my own thoughts. I've tried to indicate what's what, but have found it hard. SamBC(talk) 02:22, 22 September 2007 (UTC)

Sorry to hear about the illness, and hope you get over it soon. It seems to me that there's great significance in the statement you made: "Analytic statements don't require a secondary source, and factual statements can come from almost any source, but the primary source for the fact is quite often "better" (more reliable, neutral, etc)." The second part I can agree with to some extent, though often the primary source needs specialist interpretation which is why there's the caution about it at present. However, I find the assertion that analytic statements don't require a secondary source disturbing. It's my understanding that WP:V requires us to rely on reliable, third-party published sources, and the NOR policy at present gives an exemption in that primary (or first party) sources can be used for statements of fact that require no analysis, in the context of analysis by secondary (or third party) sources. This issue bears directly on NOR, and should be made clear as part of the policy. .. dave souza, talk 09:05, 22 September 2007 (UTC)

You make the assumption that all primary sources are first-party, which seems to be a common misunderstanding (because of the use of ordinality in the terms, perhaps). A source can be primary and not first-party quite easily, because the term "first-party" means that the the subject of the source is the same as the "author" (or source) of the source. For example, in Quaker history, there are a number of first-hand (but not first-party) accounts regarding Quaker shops, from customers who found their fixed prices terribly confusing. These sources are not first-party as they are not from the shopkeepers themselves, or any investor or so on, but they are first-hand, which is a better term to relate to "primary". In those sources (which I'm afraid I don't have references for, it's not something I've studied formally) the non-Quaker customers make a number of analytical/conclusive statements. Of course, quoting those (especially the ones along the lines of "these Quakers are crazy!") would most likely require phrasing that puts an opinion or conclusion or analysis as a statement of fact ("customers found that this indicated that Quakers were crazy"), which is of course one way to address the issue. However, a common interpretation is that peer-reviewed journal articles are often primary, and they are certainly sources of analytical/conlcusive statements. A lot of problems were found with the idea of a tight wikipedia-specific definition of PSTS, and even more with that being on this page when it's useful to other policies and not core to the OR issue. SamBC(talk) 11:23, 22 September 2007 (UTC)

For the record, different interpretations of PSTS fed several edit wars on other pages, with arguments that any secondary source trumps any primary source, that we can't use direct statements in an author's own works as evidence for the author's opinions, but we can use his critics' works, that we are not allowed to use primary sources for any purpose, etc. Moreover, there was no agreement on the meaning of "primary source" and "secondary source." Many ancient histories are written from older, often lost, histories. In particular, scientific papers were debated, and generally considered especially appropriate primary sources. The scientific paper issue resulted in suggestions like "wikiprimary" and "wikisecondary" which led into the effort to define NOR without using PSTS. Jacob Haller 01:58, 22 September 2007 (UTC)

I am well aware of the edit wars, as I imagine are most editors of more than a few weeks duration. This is an argument for clarifying the P, S, and T definitions and statements, not for shuttling it off to another page, and certainly not for inventing neologisms. I'm not seeing how this relates directly to the current discussion. KillerChihuahua^?!? 02:02, 22 September 2007 (UTC)

Some personal history on this issue... I got involved in this discussion (about a week ago now) because I too assumed that this was about weekening PSTS to allow what I thought were inappropriate sources. I argued long and hard for leaving PSTS alone. However, in the course of my arguing, I came to two realizations... 1) the definitions in PSTS are based those of historians - other disciplines use the terms differently. This has led to a lot of controversy and argument. 2) talking about types of sources in the NOR policy was a problem. The type of source does not automatically lead to Original Research. There are appropriate NOR uses for any type of source (direct quoting for exmple... citing the source for a quote is not OR), and there were OR misuses for any type of source (taking the quote out of context, or using it to back an argument). And for me, that was a key moment... It wasn't the type of source that led to OR, it was the misuse... OR that involves sources is when the statement in the article was not accurately reflected by the source that was cited. Blueboar 02:31, 22 September 2007 (UTC)

Thanks for the background Blueboar, we can only hope that will help people to share your, well, for want of a better term, epiphany. This does illustrate a practical problem any proposal is likely to face - it will be difficult and stressful to wait for/induce the same epiphanic realisation in other editors who have the same (entirely understandable) knee-jerk reaction. To clarify, I don't mean knee-jerk judgementally. It's a natural and almost always useful reflex in anatomy, and the same is generally true in the metaphorical use. Just sometimes it gets in the way, and this case would seem to be one of them. SamBC(talk) 02:38, 22 September 2007 (UTC)

Yes, Blueboar, I think you have it right. I don't think I have seen any of the serious contributors to this discussion arguing that we should lessen the standards for NOR, in fact I think we are trying to strengthen them by focusing on the root cause and addressing that. What we need to show is that the use of PSTS actually weakens NOR by being an unnecessary distraction. We need to stop using the terms primary, secondary and tertiary altogether, and describe exactly what we mean (without resorting to neologisms). Dhaluza 09:43, 22 September 2007 (UTC)

Just a point that needs being made: although some edit conflicts will be caused by editors who realy want to do original research, and resist our policies (and they need to learn, or go away), and other conflicts will be caused because the policy is unclear (and needs to be clarified) I think that there are also many conflicts that are unavoidable and in a way desirable. What I mean 9by this third caegory of conflicts) is conflicts among editors who understand the policy quite well and have reasonable difference over how to apply it to the particular case. I do not think that when there is such a conflict, it is because an editor is acting in bad faith, nor because the policy is flawed or poorly written. Rather, I think that policies like NPOV and NOR are not simple policies that are easy to follow. The demand us to act and write in ways we would not in perhaps any other context. People in good faith will have different interpretations, and will debate them. I simply do not see this as a bad thing. I get the feeling that some people here want to avoid edit conflicts at all costs, even if that means getting rid of a policy or radically revising it. I agree that there are many unnecessary edit conflicts and we all agree those are bad. But I believe some edit conflicts are good - I have been involved in amny myself, where the end result is both parties to the conflict end up with a much better understanding of one another and the topic (and the policy) and the edits to the article are much better. In these conflicts, policies express core values and provide guidelines, but it is unreasonable to ask them to provide solutions to all our problems. We need to do that, on the talk pages of articles, and it really can involve heated arguments among editors acting in good faith. That is what makes this a wikipedia - anyone can edit, strangers collaborate, articles are the proeucts of complex collaborations among editors with diverse views. I hope all of you have had the experience I have had - of being involved in a tense argument with another editor that eventually gets resolved on our talk pages and leads to a much better understanding of things. Such conflicts do not require changing policies (or blocking users), the just require patience and a recognition that people in good faith will disagree, including disagree over how to apply a policy to an article, but if they are willing to keep at it, they can work through to the solutions. Policies provide the princples, but it for editors, on talk pages, to do the hard work of finding solutions to complex problems. Slrubenstein | Talk 09:51, 22 September 2007 (UTC)

KillerChihuahua said, "Primary, Secondary, and Tertiary are clearly defined and not open to abuse of interpretation. " I disagree. There is a large class of very useful sources: ordinary peer-reviewed scientific articles, for which there is no consensus here as to whether these are generally to be defined as primary sources, secondary, or a combination of the two. This lack of clarity/consensus in the definition is one of the fundamental problems in the current dispute. I see many other problems with the various proposed definitions, too: for example, I pointed out that laboratory notes might contain interpretations. I think the complexities in categorizing the various sources mean that it isn't feasible to define them in a policy, but it should be done in a guideline where there's more flexibility. It just isn't feasible to define them clearly in a way that covers every case. --Coppertwig 16:09, 22 September 2007 (UTC)

Proposed addendum

WP:V requires us to rely on reliable, third-party published sources, and the NOR policy at present gives an exemption in that primary (or first party) sources can be used for statements of fact that require no analysis, in the context of analysis by secondary (or third party) sources. It seems to me that this distinction is being lost in changing the emphasis to what's really an expansion of "However, care should be taken not to "go beyond" the sources or use them in novel ways." If we accept the analysis that the primary/secondary/tertiary analysis is so broken that we have to make the radical change proposed (and evidence of this breakdown is needed for those of us who can't find it in the archives), we have to ensure that any relevant requirements under the present policy are fully covered by the new policy, and also have to provide clarification for editors used to the current terminology. This applies both to experienced editors who haven't re-read the policy for a while, and also to new editors advised in discussion that use of a source goes against WP:NOR in the old terminology. Here's a suggested addendum to the current proposal:

Care must be taken to accurately represent the overall sense of the source, and not introduce our own interpretation in summarising or selectively quoting from the source. Statements of fact can draw on documents or people very close to the situation being written about, often described as primary source material. Statements of interpretation, analysis or conclusion must draw on third party material, often called secondary sources, which draw on primary sources to make generalizations or interpretive, analytical, or synthetic claims. For example, we can state the fact that when Charles Darwin was finishing proofs of The Origin of Species he wrote to a friend "So much for my abominable volume, which has cost me so much labour that I almost hate it." However, we can't use this source from the author to state the conclusion that "the book is abominable or almost hateful", but for an assessment must find a reliable third party source giving the opinion of a recognised current expert in the field.

Obviously the example given[1] is the first that came to hand, and no doubt a better one can be found. There may be other points from the current formulation that need to be covered. Aspects of source categorisation not closely related to NOR could be covered in WP:RS, or in the proposed new guideline. If the "third party" definition needs clarification, that has to be reflected across all policies and guidelines that use the term. .. dave souza, talk 10:00, 22 September 2007 (UTC)

That is a very acceptable draft proposal Dave. However, based upon my understanding of discussions here over the last month, might I suggest that a couple minor changes be made. To help clarify the interaction of the various policies, add in parentheses which applicable policy or guideline the different parts are covered by, as in "Care must be taken to accurately represent the overall sense of the source (WP:NPOV and WP:UNDUE),..." etc. Maybe with an example that shows how all of the applicable content policies and their associated guidelines all act together, others can more readily see the cohesion between the policies. wbfergus ^Talk 11:13, 22 September 2007 (UTC)

Perhaps an example using the Dan Rather episode might be applicable (or maybe a completely fictious example?), as it also would have a tie in WP:BLP and probably WP:RS as well. The more policy points and guidelines that can be tied together in one example would highlight how with a good article, all of the policies and their guidelines are tied together, and for this particular policy, OR is avoided. wbfergus ^Talk 13:07, 22 September 2007 (UTC)

The only problem with this reasoning is that, as I've explained above, primary and first-party are not synonyms, not is there an implication of first-party if a source is primary. I assume that the confusion comes from the use of ordinals in the terms. Primary is more first-hand than first-party, as first-party means that a source is writing/talking/whatevering about itself. SamBC(talk) 11:36, 22 September 2007 (UTC)

There is no agreement or consensus on the meaning or application of the phrase "primary source"; but there is understanding and agreement that claims must be supportable by cites to reliable published sources that make the same claim and that the claims of notability we used to judge inclusion into Wikipedia must not be from biased sources (we specify not being biased by reason of being neither from the "first party" or the "second party" with regard to the claim, but the point is being unbiased even if this is the only objective criteria we've managed to specify to date). WAS 4.250 12:50, 22 September 2007 (UTC)

WAS 4.250, do I correctly understand your assertion to be that there is not general consensus about what's meant by the difference between primary, secondary and tertiary sources, or more specifically about what's meant by a primary source as differentiated from derivative sources and interpretive sources? ... Kenosis 15:23, 22 September 2007 (UTC)

The term third party should be avoided because entities that strive to be unbiased can still be a first or second party. Some examples are every decision of the supreme court of any country where another branch of that government was a party to the case. Another would be the 9/11 Commission Report because it reported, in part, on actions of the U.S. federal government, and was also set up by the same government. --Gerry Ashton 14:40, 22 September 2007 (UTC)

The whole "third party" issue is, indeed, another potential mess, but it's a good general rule. It's quite possible that it would be best to add a not to all policy pages mentioning third party in that way, to indicate that certain sorts of non-third-party sources are generally as good as third-party ones (such as courts, and respected government statistics agencies). It also seems to lead to the odd misunderstanding regarding first-party and primary, which is in many ways the same sort of confusion that led to the whole debate in the first place. SamBC(talk) 15:36, 22 September 2007 (UTC)

You both mention examples that provide good reason why third-party sources are preferred and primary sources where restricted with a "take caution" disclaimer. The meaning, impact and context of significant government review reports, like the Commission, are strongly debated by political scientists, historians and sociologists. Similarly, the meaning, impact and context of significant court decisions are strong debated by political scientists, historians and legal scholars. If experts in those fields engage in a vigorous debate, we should not be citing those "primary sources" to support claims. Conversely, if the experts in those fields are in general agreement, then it should not be difficult to cite a textbook, review literature or another tertiary source that notes the universal interpretation, or strong scholarly consensus. Outside of complementing the points of third-party references, citing such large and dense primary sources amounts to little more than cherry-picking at the whims of the contributor. Again, the examples are good examples for this, since both the Commission and court decisions in the United States must be taken in the context of case law, which includes existing conventions, past court decisions and past applications of law. This complex context cannot be accounted for without engaging in original research or referring to a third-party source. Just some thoughts. Vassyana 16:28, 22 September 2007 (UTC)

Vassyana, your cautions on the use of my examples are cautions that apply to using primary sources, not to using third-party sources. Being a third party means having no physical saftey, financial, or egotistical connection to the topic, and not being an employee, officer, or member of organizations connected to the topic. Being a secondary source means all the data and concepts in the source came from other sources, and are not new. These are different concepts. --Gerry Ashton 00:45, 23 September 2007 (UTC)

In reply to dave souza re evidence of a breakdown: I think there is/was no stable version of the policy, at least not relatively recently, and as I see it there is a three-way conflict among three groups. There is consensus that ordinary peer-reviewed scientific articles are usable as sources and are not to be restricted as "rare", "used sparingly," etc., but one of the groups feels strongly that some of the previous wordings of the policy imply deprecation of such sources and that the policy should not do that. Another group feels strongly that editors must be warned that "primary sources" (or what they mean by primary sources, which could be worded differently) must be used particularly carefully. A third group strongly believes that the phrase "primary sources" must include ordinary peer-reviewed scientific articles, because that's how it's defined elsewhere or because "primary" means "first" and such articles present new results, and this third group seems to include new people who show up from time to time and make definitive statements on this talk page that scientific articles are primary sources, causing alarm among the first group. No wording which satisfied all three groups was found until the proposal that many of us developed over the past week or so under Blueboar's leadership. Another way out might be for the policy to mention scientific articles specifically, for example to state that they are not generally included in the definition of primary source here. However, there is some difficulty in getting consensus for such a definition since I think some people believe that obviously only the raw data is primary and the peer-reviewed article is secondary, while others believe that obviously the peer-reviewed article is primary; others admit that a peer-reviewed article can be seen as containing both primary and secondary material.

The wording you propose, dave souza, seems geared towards history or social sciences, not hard science. Typically, a peer-reviewed scientific article contains data and also contains statements of interpretation and conclusions about the data. It's not clear what "third party" means in this context, since the article is not about the authors of the article but is about some object of scientific study. Requiring a "third party" without explaining what this means for scientific articles could possibly be interpreted as meaning you can't use scientific articles because there's no "third party". Whatever the wording is, it has to be considered how it will affect the use of scientific articles. --Coppertwig 15:43, 22 September 2007 (UTC)

The policy has never been stable? Untrue. When the policy was first established, there was work on it between December 2003 and February 2004 which achieved a stable version (which included a distinction between primary and secondary sources). In February and March of 2005, several people (including myself) revised the policy - the revisions evolved via consensus and the result was a version that was stable until September 2006, when three or four elements were questioned by a number of editors. We worked through each one until consensus was reached - and the changes all had to do with wording, not the substance of the policy - and the policy has been stable since then. So i see no breakdwon of consensus let alone absense of consensus. I see a stable policy that was revised largely for style, not for substance, twice since it was initially developed. Both times revisions were collaborative and achieved via consensus, and both resulted in highly stable texts. And I repeat: the revisions had largely to do with wording not substance. The policy as such has been accepted by virtually the entire community since its conception. As to Coppertwig's analysis, I would say that virtually all Wikipedians accept the first two points, and I think it is relatively easy to add that many peer-reviewed journal articles, while being secondary sources, contain primary source material (and Coppertwig is wrong to suggest this distinguishes natural science articles from other fields; it is as true of articles in the humanities and social sciences as it is of the life and physical sciences) and that it is important not to take such primary source material out of context. It is not a major problem and to suggest is is is a red-herring. Slrubenstein | Talk 16:57, 22 September 2007 (UTC)

I agree with what you say, and especially like the example, but I disagree with how you say it. Rather than characterizing sources as primary or secondary, which is unnecessarily confusing, why not just call them what they are. Primary sources are factual sources, and secondary are interpretive or analytical as you said. Just stop right there. The sooner we jettison this primary/secondary baggage, the sooner we can make this policy about NOR and nothing else. Dhaluza 15:56, 22 September 2007 (UTC)

Here we have a problem... I think a lot of us will agree with the statement: "primary sources are factual sources" (at least in the way PSTS uses the term "primary"), but I don't think the coralary statement "factual sources are primary sources" is true. Factual statements can come from any type of source. Blueboar 16:45, 22 September 2007 (UTC)

In reply to Slrubenstein: I didn't say it had never been stable. Not a major problem? Great. I think there is generally pretty good agreement about what the policy should mean. It's pretty much just the wording we've been having trouble with. I think adding something to the policy that defines how peer-reviewed articles fit in might very well solve the problem. --Coppertwig 00:31, 23 September 2007 (UTC)

Wait -- "not a major problem"? Maybe there is no major problem here and maybe the page should not be protected? --Coppertwig 01:51, 23 September 2007 (UTC)

Errors in "reliable sources"

Back in another discussion taken from here, someone implied that primary/secondary source conflicts were "rare". I have now had to treat yet another case of a mistake in a secondary source where appeal had to be made to a primary source. This particularly seems to be a issue with BLPs of people who are notable enough for media sources, but who lack conventional biographies. These media sources make lots of mistakes; fact-checking (or at least cross-checking) is obligatory when questions arise.

As I commented over in the discussion of WP:RS, our policies don't treat this issue well. The problem in WP:RS is that the issue of source errors isn't addressed at all, so that being correct isn't a criterion for being reliable. The problem here is that the constant hammering on avoiding primary sources is ratifying the notion that secondary source errors are to be preferred over primary sources.

One point I consider non-negotiable: when a primary source is quoted in the article, the primary source should be cited if possible. In this case, a secondary source should be never, ever, be preferred over a primary source. Likewise, when a secondary source misrepresents a primary source, whether by accident or design, the secondary source statement must not appear in Wikipedia unless it is necessary to explicitly deny its accuracy. These are the absolute standards of non-fiction writing, and we cannot credibly contravene them. Whatever other rules we establish, these must not be contravened. The problem I've consistently seen in the discussion here is that the discussion of sources is so abstract that it doesn't consider the ways that secondary sources use primary sources. Some of these uses it is not our purpose to argue with, but some of them, we have to review. Mangoe 14:46, 22 September 2007 (UTC)

Whether primary, secondary, or tertiary, a source can still be wrong or unreliable. We shouldn't reprint obvious errors. When it's useful to provide a quotation, cite the primary source for the quotation if you've actually seen it and checked it. If not, cite the secondary source for what it reports about the primary. That's done in academia all the time. Don't doctor up a fake primary source out of information gleaned from the secondary, as if you had actually read the primary when you didn't. That would only mask any error. It's okay if an editor excludes a secondary source as an obvious misstatement or misquote of a primary source. But keeping in mind that Wikipedia is about verifiability, not truth, an editor shouldn't do his/her own original analysis of primary sources to confirm whether the secondary source is a good synthesis. Wikidemo 15:22, 22 September 2007 (UTC)

This issue of verifiability and reliable sources is one thing. The issue of how those sources are used in creating an article is another. Of course they interact. All three of the basic editorial policies interact -- WP:NPOV, WP:VER and WP:NOR. WP:NOR requires dealing further with the issue of what kinds of syntheses are generally available among the sources. The issue of whether they're "right" or "wrong", as has been explicitly stated in WP's policy growth and development from the very beginning of WP, is not the issue. Odd as it may seem at first blush, the original mandate from the founding Board (read that:"from Jimbo Wales") that the standard is "verifiability, not truth", still controls in Wikipedia. What's required are sources that show a published weight of opinion about a particular assertion of fact or conclusion that can be drawn from published information. But that's more for WP:VER and its offshoot WP:RS. For WP:NOR, the issue centers around relying on published sources and not formulating new syntheses or drawing new conclusions that are not already available in the published literature. When the published literature conflicts about the basic facts or conclusions, WP:NPOV#Undue_weight controls the method by which the material is to be presented. As I said earlier here and as others have stated too (not necessarily in this discussion), the "big three" basic editorial policies work in tandem and in tension with one another. That interaction and tension among the "big three" is mediated by consensus process at the local article level. For the purposes of this policy page, WP:NOR, it is useful and appropriate to provide a framework or basic breakdown of the range of sources available, which the primary, secondary, tertiary delineation does quite effectively. ... Kenosis 15:51, 22 September 2007 (UTC)

I think that any serious thinker on an issue who wants to present it properly has to do more research than is suggested. Currently, it seems to be suggested that we have to take secondary sources at face value: indeed it has been asserted to me that attempting to correct those sources, even when referring to what must be their source material, is original research - we would typically have the synthesis argument brought up. The point is that this concept does not transfer well outside of the domain of the most reliable sources, where we are thinking of proper reliable sources, such as high quality scientific journals. As soon as we allow ourselves to consider the less reliable sources, say such as the Times, we should be evaluating the evidence put to us.

For example, in the Times today, they have published a story on biofuels being worse for the environment than normal fuels. They quote some statistics. I could base an entry on biofuels asserting that opinion quoting that story from the Times. However, the Times story does give its source quite specifically. Its source is an article in Atmospheric Chemistry and Physics, noted as an article placed for open review. So we start to see that although the Times paper presents the evidence quite firmly, (and one of the points it makes is that one shouldn't be to hasty in jumping to conclusions - yet the article appears to have taken the point as proven). And is a scientific paper a primary source in this case, or a secondary source?

According to NOR, we apparently could dive in on the Times story and use it. We could check other newspapers, but we then need to consider whether they themselves are reliable. Often these sources are from news agencies and the end result is checking the same source, adapted for the readership of the paper. We find that the report, and we can see it has been misrepresented (and can quote it, not just argue this), surely it is correct to do this. (We could more readily argue that the information should not be in due to the quality of the source).

A couple of questions. Is that scientific report a primary source for the story, or is it a secondary source as it is a scientific analysis? What about for the specific claim that biofuels contribute more to global warming than normal fuels? Does that change anything as to how we look at the sources? Let's assume this has been added to an article, how would you deal with it with the policy as a whole? Spenny 15:26, 22 September 2007 (UTC)

Academic papers are often a mixture of primary and secondary. We can all point to errors in secondary sources. Any newspaper is guilty of factual errors and bad analysis, even the Economist, Financial Times, New York Times, etc. Even the best-researched books get things wrong here and there. But what's the alternative? Supplanting sources with individual editors' analysis of what is true opens the door to even worse, systematic errors. A newspaper story about a piece of technology or a scientific theory is only a medium-good source. Any subject worth its salt has been written up in books, reports, journals, etc., which are more reliable sources than newspaper articles. A weak / stubby article may just have a few sources picked up from a google search, in which case it's still important to use the best available reliable sources. Nothing wrong with those articles, they're just the first step towards a better article. As everyone knows you can't and shouldn't take Wikipedia articles as the truth about a subject, just an introduction and starting point for making your own mind up and finding more information. Wikipedia is a collaborative effort. As the article matures, its sourcing gets better and better as people make refinements.Wikidemo 15:47, 22 September 2007 (UTC)

Proposed wording as of 2007-Sept-22

(To replace the section entitled: "Primary, secondary and tertiary sources"... which is to be moved elsewhere)

Citing Appropriate Source Materials (alternative suggested section title: Sticking to the Sources)

Within Wikipedia articles we will find statements of fact and statements of interpretation, analysis or conclusion. It is important to cite appropriate sources to back those different types of statements. Statements of fact should be cited to reliable sources that contain that fact. Statements of interpretation, analysis or conclusion should be cited to reliable sources that contain the same interpretation, analysis or conclusion, often called secondary sources (see WP:PSTS). Collections of facts which tend to lead the reader to a certain interpretation should be supported not only by a reliable source for each fact, but also by a reliable source that contains the same interpretation.

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

When there are a number of reliable sources that interpret a particular piece of material, we need to be especially careful not to insert or imply our own interpretation of the original material.

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

Comments

I tried to address the concerns that had been expressed and to incorporate some of the wording from dave souza's proposal. Further discussion and proposed changes are welcome. --Coppertwig 15:19, 22 September 2007 (UTC)

How about combining the titles and dropping the "right" (vs. wrong) issue (and using wiki-caps) as follows: Citing appropriate source materials, and sticking to those sources -- Dhaluza 16:02, 22 September 2007 (UTC)

"Appropriate" might be a better word than "right". I have changed it. One other comment... facts can come from all sorts of sources, not just pimary ones. so I took out the "usually called primary sources" clause. While statements of analysis, interpretation etc. can be found in primary or tertiary sources, they are usually found in secondary ones, so I left that in. And given that one of the points of this proposal is that PSTS will not "disappear" but simply be moved elsewhere... I think it appropriate to point the reader to PSTS when mentioning secondary sources. Blueboar 16:12, 22 September 2007 (UTC)

• "Sticking to the sources" is, IMO, a reasonable proposal for the title of an additional section or subsection for possible inclusion in WP:NOR upon receiving scrutiny, discussion and feedback from the broader community. via the Village Pump, and perhaps other relevant fora (e.g. a mention of this discussion in WT:V and WT:NPOV), in addition to this talk page. The additional participation in the last day or two is a step in this direction. IMO, "Citing the right source materials" should be dropped from consideration, for reasons given by several including KillerChihuahua above. The explicit content of the proposal needs further work. the concrete example is not adequately broad to convey the range of situations encountered on the wiki, IMO, and I'm resistant to providing a bunch of examples like the Roosevelt example on a policy project page. Perhaps a new proposed guideline might be considered, e.g. Sample applications of WP:NOR or whatever. I don't know. IMO, the first three paragraphs of the proposal are reasonable proposed additions to the policy page, either leading up to the primary, secondary, tertiary distinction, or immedately following it, or in a separate section a bit farther down. Please consider how the proposed language integrates with the more-or-less existing language in this policy page. ... Kenosis 16:12, 22 September 2007 (UTC)

I see two problems with KillerChihuahua's comments: 1) he talks about adding the proposed language as "an additional level of complexity" instead of it substituting one level of complexity for another. I would agree with him if we were adding the proposed section on top of PSTS... ie keeping the PSTS section in NOR... but by moving PSTS and replacing it with the proposed language, we actually simplify the complexity. 2) His comments don't address the main problem with PSTS... the fact that PSTS is talking about things that are not really in the realm of NOR... The type of source being used does not determine whether there is a NOR violation. OR can happen with any type of source. PSTS is an important and valuable statement... but it is misplaced in WP:NOR. It needs to be moved elsewhere. Blueboar 16:26, 22 September 2007 (UTC)

Yes, your position and that of several other users on the section Primary, secondary and tertiary sources (referred to by some on this page as PSTS) has been made clear. Many other users disagree. I disagree in no uncertain terms. Because of the mass of talk that has been brought to bear on this page on multiple issues, I will simply reproduce the relevant comments I submitted on the same issue several sections above.... Kenosis 16:52, 22 September 2007 (UTC)

..... The section on Primary, secondary and tertiary sources is a vital section that puts WP:NOR into perspective for users around the wiki. The proposal to replace the section or move it elsewhere is a major change in this policy page that I find unacceptable, at least without something that represents a clear improvement on it. Since well over three years ago, a distinction had been made between primary and secondary sources and what uses constituted original research or original synthesis. A significant number of WP users remained confused about the distinction, e.g., "well, if WP is not supposed to synthesise material directly from primary sources (except for purely factual matters), and if we're supposed to use secondary sources, then how is WP a secondary source?" The extra note about tertiary sources was inserted into WP:NOR nearly a year ago, and suddenly it made better sense to most other participants and much of the confusion diminished substantially. The description of encyclopedic matter as primarily tertiary helped to resolve this confusion right on the policy project page. Since then, it's tended to serve us well in resolving confusion about what is original synthesis, or original research, as distinguished from original wording and editorial decisionmaking.

...... Admittedly there remain residual bits of confusion. One user mentioned Encarta and Britannica, which are also "tertiary sources" according to the delineation presented in WP:NOR. I pointed out somewhere in these rapid-fire multiple threads today there is a wealth of more specialized tertiary sources such as specialized encyclopedias, to which we can also add many textbooks as tertiary sources. Indeed one could readily also argue that there are numerous sources that are quaternary, quinary, senary, septenary, octonary, etc. etc. Categorizing sources as primary, secondary and tertiary has, nonetheless, served quite well in clarifying the basic types of sources from which material in WP is drawn. If there is a proposal to better clarify the distinction or explain that the delineations between them are not always sharply drawn, or whatever, that would, as I said in the section below, be quite welcome and potentially helpful. But the proposal to replace it with the language at the top of this talk section is, IMO, completely unacceptable on the justifications given thus far. I believe I used the words "summarily reject" to covey my summary opinion about the proposal. ... Kenosis 16:52, 22 September 2007 (UTC)

• I don't see any need for this, or for any alteration to the policy for that matter. We're only going to be altering core policies when there's any pressing need that overides the need for a stable policy, and I've yet to see evidence of such a need anywhere on the project but in claims made here, and anyone other than those making such claims. FeloniousMonk 16:45, 22 September 2007 (UTC)

FeloniousMonk, first congratulations on what looks like your approximately 3 year anniversary here on Wikipedia. Secondly, your opinions are valuable as an Administrator. However, I think it must be apparent that so many people have been arguing this back and forth for so long, with such passion, that there must be a "problem" somewhere along the line, or people wouldn't have been coming here to voice their opinions or to seek clarification in the first place. I (and many others) do see what we are proposing as neither weakening nor destabilizing to any of the core policies. Instead, we see it as a needed cleenup measure to minimize why so many people have been coming here to discuss this in the first place. If we can agree on moving PSTS somewhere else, and create that 'location' with whatever other changes may be required (additional examples, further defintions, whatever), and then replace the current PSTS section with this proposal, how do you see that as a destabilizing event in Wikipedia policies? Perhaps we can address that to alleviate your concerns. Thank you. wbfergus ^Talk 16:55, 22 September 2007 (UTC)

Whether WP:PSTS is critical to editors understanding of WP policies is not relevant here, it clearly is not critical to understanding WP:NOR because we can say what NOR really means without using PSTS as a crutch. Frankly I think it is not well defined, not widely applicable, and therefore widely misunderstood. Also the idea that removing it changes core policy would require changing the core policy on WP:CONSENSUS. It's obviously contentious, and not necessary, so it should not remain. We can define NOR without creating defined terms, which is a form of Neologism. Let's just use plain English. Dhaluza 20:45, 22 September 2007 (UTC)

<undent> Plain English or not, this now refers to secondary sources without making it clear that sources close to the subject, which for most of us is plainly and simply expressed by the term "primary sources", can be used for facts but not for interpretation etc. This is still less clear than the existing policy. The Darwin example I showed above was designed to bring out the point that a fact about an openly attributed opinion can be stated, but that opinion from, eh, a primary source can't be used as analysis, conclusion or whatever. That's a point which still seems to be lacking, and some thought is needed as to whether this proposal would have other unintended consequences. .. dave souza, talk 23:34, 22 September 2007 (UTC)

Would those who are against the proposal to state the NOR policy without PSTS please suggest alternatives? For example, would you try to find a previous version of the policy for which broad acceptance can be built up? We need to move forward somehow, to get out of this deadlock and get the page unprotected.

In reply to dave souza: when you say "sources close to the subject", do you mean an individual human subject? How would this wording apply when the topic is a mathematical theorem, a property of an astronomical object, or something of that nature, rather than a human subject? --Coppertwig 23:57, 22 September 2007 (UTC)

For that matter -- who objects to the current version of the policy, and on what grounds? --Coppertwig 02:50, 23 September 2007 (UTC)