Wikipedia talk:No original research/Archive 32

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

The present round of debates notwithstanding, I just made this one-word edit. this one-word edit Beyond this, understanding again that the debate(s) that is/are currently underway may change yet again, I recommend having at least another quick look at the third sentence in the subsection on "tertiary sources". (That sentence presently reads "Wikipedia itself is a tertiary source."). While I agree with this statement and have said so before, I should also point out that it is only a matter of time, IMO, until someone comes onto this page and points out that WP is often a secondary source too. I surely do not mean to "throw a monkey wrench into the works", but as with all encyclopedias, WP often, or at least occasionally, is a secondary source. Perhaps the most important distinction of all, if I understand the main policy mandate correctly, is that WP may, on the whole, summarize multiple primary and secondary sources, but according to policy is never to be used as a primary source. (For example, w.r.t. an eyewitness to a traffic accident, presently used as an example of a primary source, Wikipedia is under no circumstances permitted to first publish such an eyewitness report, and we are permitted only at the absolute extreme to rely on a previously published primary source of such an eyewitness report, within the limits described under the "primary source" part of WP:PSTS.) .. Kenosis (talk) 07:08, 18 December 2007 (UTC)

You mean this edit?

Dreadstar †

07:50, 18 December 2007 (UTC)

Yes. Thank you. ... Kenosis (talk) 08:01, 18 December 2007 (UTC) I've stricken it and corrected it according to Dreadstar's correction of the appropriate link. ... Kenosis (talk) 08:06, 18 December 2007 (UTC)

In articles about wikipedia, wikipedia is a primary source. The individual wikipedia articles about wikipedia may be considered both primary and secondary. --SmokeyJoe (talk) 11:37, 18 December 2007 (UTC)

I think I understand what you mean, but nonetheless, if articles about wikipedia are primary sources for any of the information contained within, they're technically in violation of

WP:NOR. Note that the article about Wikipedia, for example, has well over a hundred footnotes, most of which are citations to sources other than Wikipedia itself. ... Kenosis (talk

) 21:47, 18 December 2007 (UTC)

I meant to gently ridicule the seriousness and absolutism that is sometimes cropping up, including in your comments above. Generally, I agree with everything you say, except that maybe you push a little too far. The one word edit is absolutely fine, but with regard to your comments: "Wikipedia itself is a tertiary source." is basically true, but to insist that it is absolutely true, and enforceable, by writing it in black and white in core policy is going too far. It is like squeezing something so hard that force seepage out of the cracks. All this PSTS stuff is perspective dependent, and there are too many unique perspectives here, sometimes unstated. Wikipedia can easily be considered a secondary source or a primary source if you find the right context and perspective. A prohibition against using wikipedia as a primary source is absurd. We can of course, quite properly, use and summarise primary sources that are used and summarised by cited secondary sources. Perhaps a prohibition against “citing primary sources that have never been elsewhere cited” would be closer to what you really intend? --SmokeyJoe (talk) 00:14, 19 December 2007 (UTC)

Understood, as before, but understood somewhat better. Thanks. ... Kenosis (talk) 04:22, 19 December 2007 (UTC) ... The "bottom-line" rule is that WP is never to be used as a primary source. With all these intelligent people involved, I imagine that there eventually should be a widely accepted, completely defensible way of expressing it that'll be understandable to newly arrived and more experienced editors alike. But that's just my imagination at the moment. ... Kenosis (talk) 04:26, 19 December 2007 (UTC)

As before, “WP is never to be used as a primary source” is not strictly true. I am free to use wikipedia as a primary source in my paper on “The endlessness of consensus deliberations on contentious abstract issues when the number of participants exceeds a half-dozen or so”.

I think that what you mean to express is expressed perfectly well by “Wikipedia does not publish original research or original thought.” Unfortunately, the message is obscured by subsequent excessively detailed pitiful prose.

I’ve been thinking, that what we need, more than more detail on the abstract concept of “no original research”, is a page of specific examples of near-borderline things that are original research and other examples that are not. --SmokeyJoe (talk) 06:14, 19 December 2007 (UTC)

No, in the article

WP:NOR. The founding director's mandate included the rule that decisions be made by WP:Consensus, and that the core content poicies themselves were to be interpreted by consensus, but never superceded by consensus. Thus far, at least, I see no signs of a widespread mutiny concerning these mandates set in motion by the founder of WP (though of course any of us, or any group of us, is always free to start out own competitive wiki, as has already been done multiple times by a number of people). . ... Kenosis (talk

) 06:45, 19 December 2007 (UTC)

“WP is never to be used as a primary source” is a clumsy statement, and on analysis is false. I’ve given a counter example. It seems to have critical context or qualifiers that are omitted. Such statements are unacceptable on policy pages because they will be quoted out of context. This is important because what you write can be read in unintended ways. If the quote is historically or otherwise significant, please point me to the source precisely. Jimbo has said many things, but mostly it is not been ex cathedra. --SmokeyJoe (talk) 11:42, 19 December 2007 (UTC)

As I said, I think I understand what you're trying to say. Admittedly the language of the policy can be quoted out of context, almost no matter what the language used to express it. It's quite reasonable to debate where the transition is between original prose and original research. IMO, it's also reasonable, in principle at least, that the language is once again being debated by several participants. ... Kenosis (talk) 15:20, 19 December 2007 (UTC)

Someone argued above that primary sources and the issue of notability aren't directly connected to OR, but they are in fact central to it.

Here is an example of Jimbo removing original research — material based entirely on primary sources — from a BLP.

This was OR because it relied solely on primary sources. The sources were good (including trial transcripts), and the intepretation of them was, I believe, accurate. But no secondary source had seen fit to mention the subject. It was not regarded as relevant to this person's life by anyone other than the Wikipedia editors who added it.

It is for this reason that secondary sources are required in Wikipedia articles -- to answer the questions: "Who cares about this other than you?" and "Who is interpreting the primary sources like this, other than you?" If you can't answer those questions, you're probably engaged in original research. SlimVirgin ^{(talk)(contribs)} 01:26, 17 December 2007 (UTC)

All of these things do interrelate, but it may be that there are better and more logical ways to present the same information so that it causes less friction. --Lquilter (talk) 01:44, 17 December 2007 (UTC)

There's no reason for it to cause friction. The definitions are very simple; they're backed up by academic sources; the distinctions are understood and frequently used by good editors; they've been in place for three years; articles that lack secondary sources are routinely deleted by administrators and via AfD because they contain OR and aren't notable. As I said earlier, this policy has to reflect what good editors do, not only prescribe, and it currently does describe what good editors do. If you doubt that, please show me a good article that contains no secondary sources. SlimVirgin ^{(talk)(contribs)} 01:48, 17 December 2007 (UTC)

Every topic has its own definition of notability. For example, since there are so many college professors,

talk

) 01:50, 17 December 2007 (UTC)

I find a short edit summary unconvincing. It could be read to mean that any use of primary sources is original research. Or it could be read to mean that using legal primary resources in connection with biographies of living persons is original research, since it is difficult for the non-lawyer to determine which of many legal documents is the difinitive decision on a case. If you succeed in saying that it is ever original research to quote a primary source, I will say the definition of original research is so thoroughly screwed up that I will reject the entire policy and refuse to read it and do whatever seems right to me, and if you don't like it, you can block me.

Please note the example given of a revert by Jimbo Wales is about adding material to an article, not about the notability required to have an article in the first place. --Gerry Ashton (talk) 01:54, 17 December 2007 (UTC)

It was about the notability of that issue in the life of that person. Secondary sources hadn't written about it. Therefore, it was deemed not notable, an example of OR, because based entirely on primary sources. SlimVirgin ^{(talk)(contribs)} 02:21, 17 December 2007 (UTC)

I think the best rational for the removal would be undue weight, because there was no source saying the legal matters were on a par in importance with the rest of the article. (Since documents would have had to be reviewed to reach such a conclusion, such a source would necessarily have to be secondary in this case.) A second-best rationale would be that even if it satisfied policy, the material just wasn't interesting enough to retain in the article. Saying it is original research to decide to include material from primary sources unless a secondary source says it's important puts us on a slippery slope, because deciding what to put in and leave out of an article is essential to writing an encyclopedia; that's source-based research, not original research. --Gerry Ashton (talk) 02:49, 17 December 2007 (UTC)

I agree with this assessment — that undue weight would be the primary reason for removal of that material, not a blind anti-primary sources rule. Such a rule would be doing us a disservice in many topics; for example, there were several important aspects of Orion (mythology) that just were not covered in secondary sources, just because it's an obscure subject. What could possibly be a good reason for objecting to such (non-controversial) material?--Pharos (talk) 02:57, 17 December 2007 (UTC)

(ec twice)

Slim! What are you saying?!

1. The link is not in fact removing original research because it is based on primary sources. The edit comment states the removal was because it was original research. Thats it. Not a word about the source having been primary, which in fact it is, but has nothing to do with the OR violation.
2. (please ready this very carefully) "original research" is the process by which an editor takes a statement in a source, and misinterprets that statement in wikispace.
   This is what that editor in the link evidently did.
   NOR on the other hand is the policy that aims to prevent misinterpretation/misrepresentation of sources.
   It does not matter at all whether the source being misintepreted is primary or secondary or tertiary. Mis-representation can occur with any kind of source.
3. Notability is another policy and is neither influenced by OR/NOR, nor does it influence OR/NOR.

-- Fullstop (talk) 01:55, 17 December 2007 (UTC)

It is my interpretation that Jimbo considers the absence of secondary sources to constitute original research. Had’ve the content been using the sources for their commentary value about something else, rather than it being about the material in the sources themselves, then it wouldn’t have been original research. I disagree that here there was any difference between original research and the use of sources that were solely primary sources.

I also disagree with point 2. WP:NOR is not about misinterpretation/misrepresentation, it is about not creating content that hasn’t been published elsewhere even if the content can be sourced.

Finally, I disagree about Wikipedia:notability, as it definitely sits on the foundations of WP:NOR. WP:N says that you cannot write an article about a subject that hasn’t been elsewhere written about. WP:NOR says that you cannot write content that hasn’t been elsewhere written. --SmokeyJoe (talk) 02:47, 21 December 2007 (UTC)

It may seem that there is no reason for it to cause friction, but it clearly has caused friction. I'm just looking for the cause, and to me it seems that the cause is that there is a lack of clarity in the terms. Just reading over the talk pages for the past couple of weeks I've seen people clearly talking about two different things when they talk about "PS" or "SS". --Lquilter (talk) 02:14, 17 December 2007 (UTC)

There are people claiming there is a lack of clarity, but is there really? The terms seem quite simple for the most part. A primary source is produced by someone close to or involved in the situation. A secondary source is one step removed, someone uninvolved. It isn't rocket science.

Of course there will be unusual, borderline, complex examples, just as there are with anything -- with defining the word "reliable," or the word "published," for instance. But so what? Most people, most of the time, understand what these words mean and what they imply. Ditto with primary/secondary sources. SlimVirgin ^{(talk)(contribs)} 02:19, 17 December 2007 (UTC)

Yes, there seems to be a lot of original research on this page defining them. That's why I provided all those links above. Many of them define primary, secondary and tertiary sources for all fields.

talk

) 02:22, 17 December 2007 (UTC)

Indeed. Thank you for the links. Our policy is consistent with them, but there are people on this talk page who are just inventing their own definitions. SlimVirgin ^{(talk)(contribs)} 02:24, 17 December 2007 (UTC)

Indeed, the definitions seem simple enough to paraphrase in a sentence. But evidently they are not simple in application. People have fouled themselves up over peer reviewed research in the sciences & humanities. It may seem incredibly simple to some people, but this talk page & its arguments is replete with people who are confused about the terms & their application. Ergo, it is not simple. --Lquilter (talk) 03:11, 17 December 2007 (UTC)

Yeah, lets see who is inventing his/her own definition here...

• To reiterate Jimbo on lists that SlimVirgin herself used in the big fubar that became today's PSTS, with tags in green to note which current policy covers that (NB: Jimbo's comment is from the days before RS policy, i.e. when RS/NOR were not distinct)

"The basic concept [of 'original research'] is as follows: it can be quite difficult for us to make any valid judgment as to whether a particular thing is _true_ or not. It isn't appropriate for us to try to determine whether someone's novel theory of physics is valid, we aren't really equipped to do that. But what we _can_ do is check whether or not it actually has been published in reputable journals or by reputable publishers.RS So it's quite convenient to avoid judging the credibility of thingsNOR by simply sticking to thingsNOR that have been judged credible by people much better equipped to decide.RS

http://lists.wikimedia.org/pipermail/wikien-l/2004-December/017557.html

• Further...

An article that makes no new low-level claims, but nonethless synthesizes work in a non-standard way, is effectively original research that I think we ought not to publish.
I agree completely.

http://lists.wikimedia.org/pipermail/wikien-l/2004-December/017591.html

• To reiterate
  WP:NOR
  "in a nutshell":

*Wikipedia is not a publisher of original thought, nor a forum for promoting one's own point of view; all material must be verifiable
* Facts must be backed by citations to reliable sources that contain these facts
* Interpretations and syntheses must be attributed to reliable sources that make these interpretations and syntheses

Note also the complete lack of any reference to type of source, and inversely the (virtually) complete lack of reference to NOR/OR in the PSTS section.

• To reiterate
  WP:NOTE
  "in a nutshell":

A topic is presumed to be notable if it has received significant coverage in reliable secondary sources that are independent of the subject.

Note the complete lack of reference to OR/NOR. Note also the dependency of the phrase "inclusion criteria" in the first sentence of the article, and that "no original research" can never determine suitability for inclusion.

Sorry, Slim, its evidently time for you to revise that idea of yours that OR and NOTE are somehow interdependent.

-- Fullstop (talk) 02:55, 17 December 2007 (UTC)

Sorry, Fullstop, but you write above go against your own arguments.≈ jossi ≈ (talk) 06:22, 17 December 2007 (UTC)

please don't feel pressed to actually say anything substantial. It might break your streak. -- Fullstop (talk) 14:09, 17 December 2007 (UTC)

What is the point of saying anything when there is an inability to listen? ≈ jossi ≈ (talk) 15:59, 17 December 2007 (UTC)

The fact that this is not a back & forth b/w Jossi & Fullstop but an open conversation in which a lot of people are trying to listen or read or understand. --Lquilter (talk) 17:27, 17 December 2007 (UTC)

(unindent) SlimVirgin requested "if you doubt that, please show me a good article that contains no secondary sources". Well, that depends on what you define as good, and what you define as primary source. The article Mendenhall Order only (directly) uses U.S. government publications, the very same government that issued the order. Yes, it does wikilink to articles that contain secondary sources. Yes, secondary sources could be found if someone wanted to find them. Of course, one could argue that since the present government officials have little incentive to cover up any deficiencies of the officials that served more than 100 years ago, the fact that they are all officials in the the same government is a mere formality, and the government publications are really secondary sources in this case. Or, you could argue that it's a bad article. --Gerry Ashton (talk) 02:27, 17 December 2007 (UTC)

That article contains references to secondary sources e.g. Barbrow, Louis E. and Judson, Lewis V. (1976). Weights and measures standards of the United States: A brief history (NBS Special Publication 447). Washington D.C.

I'm asking for an example of a good article with no secondary sources. SlimVirgin ^{(talk)(contribs)} 02:29, 17 December 2007 (UTC)

Barbrow et al. is only a secondary source if you are willing to allow that it is indepencent of the organization that gave the order. The order was by the United States. Barbrow et al. was published by the United States. Formally, they are not independent, but in practical terms, they are, because the actual officials who gave the order were long dead by the time Barbrow et al. wrote their work. --Gerry Ashton (talk) 02:39, 17 December 2007 (UTC)

It sounds like you are confusing "independent source” with “secondary source”. A secondary source may be either dependent or independent. An independent source may be a primary source or a secondary source. If the title reflects the contents, the words “A brief history” mean that Barbrow et al. is, without ambiguity, a secondary source. Its independence is ambiguous. --SmokeyJoe (talk) 03:58, 17 December 2007 (UTC)

Some days, this policy includes language that classifies dependent sources as primary, whether they constitute a review of other sources or not. I see that today, a secondary source need not be independent. --Gerry Ashton (talk) 04:35, 17 December 2007 (UTC)

It doesn't matter if there are good articles with no secondary sources or not (eg TV episode articles with GAs). The lack of SS does not reflect on whether that article is guilty of OR or not, and inversely, articles completely with SS/TS can be just as guilty. -- Fullstop (talk) 03:42, 17 December 2007 (UTC)

• Hmm. If we agree for the sake of argument that, as SV says, the PS/SS/TS distinction is important for both
  WP:OR then that would seem to argue that it should be separate, with a reference to it on both pages. --Lquilter (talk
  ) 03:16, 17 December 2007 (UTC)
  • Articles on
    talk
    ) 03:26, 17 December 2007 (UTC)

A separate policy or guideline on PSTS would require that WP:N and WP:NOR both be revised whenever any significant change was made to PSTS. That's quite apart from whether it is practical to accept any internal definitons. --Gerry Ashton (talk) 03:28, 17 December 2007 (UTC)

If PSTS were a guideline then it, like WP:N, should conform to WP:NOR (a policy), not the other way around. It should be written such that it provides helpful guidance to carrying out the NOR policy. Conflicts between guidelines do need to be resolved, of course. --Lquilter (talk) 20:19, 17 December 2007 (UTC)

• "Using the outside definitions is the only way to resolve conflict". No. The only way we can develop uncontentious policy definitions of what we want "primary" and "secondary" sources to mean is to invent our own new terms for them. It'll take time, but you'll all see the sense of this eventually :)  —SMALLJIM  09:03, 17 December 2007 (UTC)

The problem with re-defining commonly used terms is the [law of primacy] which holds that people will have difficulty adjusting to the new definitions, especially when under stress (such as in a content conflict). This is not a desirable situation, and we should avoid it as much as possible. Dhaluza (talk) 12:43, 17 December 2007 (UTC)

That's an argument against doing it, certainly - probably the only significant one. But think of all the disagreements and misunderstandings (just on this page!) that would be avoided. My proposal above has the side effect of working towards such a beneficial change: by necessarily defining two types of primary source it introduces the idea of "type 1a" and "type 1b" primary sources. It's then only a small further step to associate secondary sources with "type 2" and everything is in place for a gradual shift to the new terms.  —SMALLJIM  13:17, 17 December 2007 (UTC)

Making this change isn't dependent on adopting my proposal, of course. Just slipping new terms into the existing PSTS text, and referring to them when appropriate, would have the same gradual effect.  —SMALLJIM  18:38, 17 December 2007 (UTC)

I still think PSTS causes more arguments than it solves. But, your proposal to use type-1a, type-1b, and type-2 sources could be a reasonable compromise to the terminology issue. It avoids redefining common terms, while also avoiding creating completely new terms or neologisms, which has also raised objections. I think it also highlights the arbitrary nature of the distinctions. Also by splitting close sources from factual sources, the type-1a and 1b designation can avoid this conflation as well. Dhaluza (talk) 04:17, 18 December 2007 (UTC)

I think that moving to type -1a, -1b, -2 is to move in the wrong direction, making things more complicated, requiring move wikipedia-specific definitions and explanations, and making wikipedia less accessible to newcomers. I disagree with Dhaluza’s word choice “arbitrary”. Source type distinctions may be ambiguous (to the novice), fuzzy, use- perspective- or context-dependent, maybe even irrelevant, but arbitrary they are not. Given a subject, a source, a context and knowledge of how the source is being used, and how the subject is treated by the source (information that should be at hand when editing), I expect that we will be able to agree whether the source is a primary source, a secondary source, or is being used as both at the same time. --SmokeyJoe (talk) 02:23, 21 December 2007 (UTC)

I don't agree that using new terms would make WP less accessible to newcomers. Yes, it might cause those who are already familiar with the primary/secondary source distinction to pause for thought—but that's no bad thing because bringing one's own preconception of what the terms mean (depending on which academic field one's experience is in) is another cause of confusion and dispute. Let's be honest though, most WP editors are not from an academic background and won't have come across the primary/secondary source distinction before. The current PSTS section is a cryptic rat's nest, and is likely to remain so without a shift away from those external definitions. I strongly believe that our policies should be written so that they are easily understood by all: I don't think anyone would disagree with that?  —SMALLJIM  14:33, 21 December 2007 (UTC)

Thanks for your comments, which naturally I agree with. Even leaving aside any possible benefits of a change like this, it's quite obvious to me that there are major problems with the current P & S source definitions in

PSTS. The definitions are a perfect example of "design by committee" and are most likely the hidden cause of many of the disagreements here. Maybe these flaws are an unavoidable consequence of a consensus-based system with its lack of leadership, but I'm surprised that no-one else seems able to step back and see them.  —SMALLJIM

  00:00, 21 December 2007 (UTC)

Yes, the WP:NOR PSTS definitions got out of hand, and are either the root cause or a contributing factor to all of the problems with PSTS. Abandon the policy space redefinitions, and don’t re-introduce them without a clear case for why the many reputable sourced definitions already provided are inadequate or hopelessly contradictory (which to my reading they are not). --SmokeyJoe (talk) 02:23, 21 December 2007 (UTC)

Those external definitions might not be contradictory as long as you can spare the time and brain power to analyse them to see the underlying similarities. But apparent contradictions abound - here's an example in the

WP:NOR

footnotes: we have secondary source defined as "a work that interprets or analyzes an historical event or phenomenon…" (Berkeley) or one which "analyzes and interprets primary sources", is a "second-hand account of an historical event" (BMCC). On the surface those are clearly contradictory.

Also, if we were relying on external PSTS definitions, we'd need to decide which one of the myriad nuances is applicable to any particular WP citation, and then defend that decision against those who would disagree with us. We shouldn't require WP editors to navigate such a rich and varied field of research. It is no basis for good policy.  —SMALLJIM  13:52, 21 December 2007 (UTC)

Actually, Gerry Ashton is correct. In the field of metrology, the use of Barbrow and Judson in Mendenhall Order is as a primary source. This is because the source is a primary document by the U.S. Government (United States Department of Commerce) concerning evidence from the U.S. Government that directly examines not just the history of U.S. weights, but the collected data in regards to units of weights and measures and SI units from the Federal Register of the United States. —Viriditas | Talk 14:10, 17 December 2007 (UTC)

While NIST is the publisher, the linked sources I've looked at are evidently provided by modern historians describing the original documents and source material. It seems odd to define them as primary in relation to the subject, which is the U.S. Coast and Geodetic Survey, published in the nineteenth century. No doubt that could be resolved by editors of the article. As for notability, a historian's account (whether published by the government or by an independent publisher) would be appropriate to determine that the subject is significant enough to meet

WP:NOT in determining if the subject merits an article, and if a section of the article was written from a primary source (such as the text of the original order itself) it would be likely to be appropriate to have a secondary source, such as a mention by a modern historian, to support the interpretation and show that the section is significant enough to be included in the article. Which brings us back to the start of this section, where Jimbo is rejecting as original research a section which is fully sourced from a primary source, but needs a secondary assessment of significance to justify its inclusion in a BLP. For many of us that's a useful way of clarifying the need for independent analysis when assessing and summarising a primary source. .. dave souza, talk

18:59, 17 December 2007 (UTC)

The subject of the Mendenhall Order article concerns the origins and history of the change from English to metric standards, a change that was implemented by the U.S. National Geodetic Survey in 1893. In 1901, the National Institute of Standards and Technology (NIST) took over the responsibility for weights and measures previously held by the Survey. The source, NBS Special Publication 447, was originally published by the U.S. Government Printing Office, containing an overview of history and data used, while Appendix 2-9 contain the original documents related to the Order. —Viriditas | Talk 23:25, 17 December 2007 (UTC)

The KISS principle states that design simplicity should be a key goal and unnecessary complexity avoided. Complexity should, very simply, be avoided: simplicity becomes a goal in itself.

This is a follow-up to the “Life beyond PSTS -bloat reduction” section above, but focused upon the “Other options” section this time. This section is designed to grant a limited exempt from WP:NOR to certain article pages stating: “These pages may contain original research; that is, research for which there is no reference other than projects in the Wikipedia namespace. Original research that does not have Wikipedia as its object should, however, be avoided on these pages too.” I am purposing that we eliminate this section from WP:NOR for the following reasons.

1. It directly covers only two articles in all of Wikipedia: Wikipedia:Statistics Department which is tagged as inactive, and another page,Wikipedia:WikiProject Wikidemia, that while not tagged as inactive, might as well be. There may or may not be other pages that are covered but not named. We expect that editors will read the policy pages, and in return editors expect that what they read will be relevant to what they are doing. Policy that focuses on such a limited area is relevant to only the tiniest fraction of editors who will ever read it. We are asking editors to take the time to familiarize themselves with all three core policies (“editors should familiarize themselves with all three.”) If we say that and mean that, then it’s too much to ask to include policy that applies to only 2 pages.
2. In terms of reliability, anything within Wikipedia is automatically a Questionable Source.
   WP:SELFPUB
   already grants the same basic exemption, using different wording, to any questionable source including the pages covered here.
3. As Fullstop pointed out in a conversation we had: “The really simple reason why it does not belong in article policy is that its all about non-article thingies. A less obvious reason is that policy is WP:NOT a directory (either).” -3 of the 4 bullets act as a directory of non-Wikipedia entities that allow OR. Good point.

If anyone has a good reason for retaining this section, I am all for hearing it. Brimba (talk) 04:59, 20 December 2007 (UTC)

I am all in favor of taking exemptions out of this policy. Are there, and will there be a few articles should be exempt (ie where OR is appropiate and makes the article better)? Sure... And if consesus says a specific article needs to include OR, so be it. That is why we have WP:IAR. But I see no reason to highlight the exemption in the policy. Blueboar (talk) 14:33, 20 December 2007 (UTC)

Operating on the general principle that a policy should be as simple and understandable as possible (which will help consensus) I agree with the idea of removing this section from the policy. Exceptions can and should be hashed out in guidelines. These can be a little bit more flexible and responsive to change, etc. --Lquilter (talk) 17:19, 20 December 2007 (UTC)

I agree with the above. I say take it out. COGDEN 17:31, 20 December 2007 (UTC)

Despite my belief that an awareness of the primary source / secondary source distinction is very important, I agree: Take PSTS out. Consider starting again with a summary section, after the question of relevance has been sorted, and do not use policy to redefine incidental terms. --SmokeyJoe (talk) 20:40, 20 December 2007 (UTC)

Have you missed on the ongoing discussions, SmokeyJoe? PSTS is not going anywhere soon.... Brimba's proposal is about the removal of the "Other options" section. ≈ jossi ≈ (talk) 03:31, 21 December 2007 (UTC)

Sorry, clearly I was a bit off the rails. I have no issue with the removal of the "Other options" section. --SmokeyJoe (talk) 08:58, 21 December 2007 (UTC)

I agree with Brimba's proposal. SlimVirgin ^{(talk)(contribs)} 02:19, 21 December 2007 (UTC)

Good call, Brimba. ≈ jossi ≈ (talk) 03:29, 21 December 2007 (UTC)

Agreed. ... Kenosis (talk) 04:03, 21 December 2007 (UTC)

Excellent, Brimba.

Dreadstar †

09:44, 21 December 2007 (UTC)

Kenosis, regarding your recent edit, it's a minor point, but the thing about signed encyclopedia articles is that they're often not tertiary sources, but secondary ones. They often represent quite a distinct POV -- sometimes strongly representing it -- as opposed to the usual disinterested overview. I'm thinking particularly of signed articles in the Encyclopedia Britannica. It might be worth pointing that out. SlimVirgin ^{(talk)(contribs)} 19:05, 21 December 2007 (UTC)

Geez; I'm tryin' to help here. Firstly, why confuse the issue of what a tertiary source is? As a practical matter, encyclopedias are inherently tertiary sources, almost by definition. Secondly, Encyclopedia Britannica features articles signed by experts? Where? Please provide an example. In my experience only specialized encyclopedias provide articles signed by experts, which is part of the stock-in-trade of many specialized encyclopedias. General encyclopedias don't ordinarily do that. Third, If there's going to be wiggle-waggle about some being more reliable than others, then there should be a link to

WP:V#Reliable sources. Since it's place at the end, that pretty much covers the interaction between PSTS and RS for PSTS in general. Could it be written differently? Of course, as always. If it makes poor sense the way I edited it, change it back. But please leave out the confusion about Encyclopedia Britannica and signed articles and all that, and leave it with a simple referral to the proper place to discuss reliable sources, which is either WP:RS or the RS section in WP:V. ... Kenosis (talk

) 19:35, 21 December 2007 (UTC)

The EB has lots of articles by signed experts and those articles would be regarded as secondary sources. For example, "Why wasn't Auschwitz bombed?" by Michael Berenbaum. You'll need a subscription to read it. SlimVirgin ^{(talk)(contribs)} 01:49, 22 December 2007 (UTC)

Ah. This is on the Britannica website? Is the caveat on the project page about reliability of tertiary sources and the referral to

Evolution and the Roman Catholic Church, Women in the United States Senate, etc. (Again, I'm not trying to confuse things, but only to keep them in perspective and seek language that hopefully has a better chance to remain reasonably stable the next time around.) ... Kenosis (talk

) 03:22, 22 December 2007 (UTC)

Yes, fair point, keeping it simple is good. :-) SlimVirgin ^{(talk)(contribs)} 03:36, 22 December 2007 (UTC)

(1) Encyclopedias are by definition tertiary, but they may still have a viewpoint. Tertiary doesn't guarantee objectivity; it just means synthesizing previously published sources and arguments. Those arguments could still be selectively chosen, leaving out, for instance, some critiques and putting in others. (2) The classic example of signed encyclopedias is the 1911 ed. of Encyclopedia Britannica. Kropotkin wrote the article on anarchism, Freud wrote the article on psychiatry, and so on. These expert-authored articles are one reason why that edition is so acclaimed. (* Actually I'm not totally certain if those articles were "signed", per se, or if attribution is just widely known. Other encyclopedias do frequently include initials at the end of an entry and the initials correlate to an author page; so, signed. I'll try to generate a better example.) (3) All that said I agree with the general thrust of Kenosis' points which is that we should discuss reliability of particular sources on RS or V, and ultimately on the relevant article talk pages. --Lquilter (talk) 20:10, 21 December 2007 (UTC)

Understood-- yes, the articles with author's initials could be tracked to a noted author, I suppose. And yes, I would think some of these issues should be conveyed at

WP:V#Reliable sources too, perhaps with even greater emphasis in that policy area. ... Kenosis (talk

) 20:39, 21 December 2007 (UTC)

Just to make myself clear, lets say there is an article on Wikipedia about a high school I attend and there are a few things I know about the school, but can't be backed up online. Is this original research? 24.151.192.38 (talk) 02:33, 22 December 2007 (UTC)

If it cannot be backed up by published sources, online or printed, you will be violating Wikipedia:Verifiability. ≈ jossi ≈ (talk) 03:47, 22 December 2007 (UTC)

The discussion on the policy issues have become somewhat fragmented with the RfC and RfA comments and discussion going on now. I saw something that Fullstop posted at Wikipedia_talk:Requests_for_comment/COGDEN. He summarized the issues with the PSTS section as follows:

a) the relevance of a distinction between "primary" and "secondary" sources with respect to "original research";

b) the applicability of the terms "primary" and "secondary", to include the arbitrariness of such definitions, and the problems that such definitions cause;

c) the pitiful prose in that section.

I believe this is a very succinct and accurate summary of the questions that have been raised on this talk page that still remain unaddressed, and should be discussed further here. I also think these items are sufficient justification for applying the dispute tag to the PSTS section to flag the issues and encourage further discussion to reach a true consensus. Dhaluza ([[User talk: Dhaluza|talk]]) 14:36, 15 December 2007 (UTC)

So since there are no cogent arguments disputing this summary, can we reapply the {{

Disputedtag}} tag to the PSTS section now? Dhaluza (talk

) 17:33, 16 December 2007 (UTC)

OK, since no one has disputed this summary, I have reapplied the tag. Please discuss this here before reverting. Dhaluza (talk) 20:53, 22 December 2007 (UTC)

One problem that is still not being addressed in all of our drafts is the fact that the PSTS section does not clearly tie what it is saying to the concpet of "No Original Research". We can debate definitions till the cows come home... but it does us no good unless we clearly explain to the reader what all of this has to do with NOR.

From the discussion on this talk page, I know that several people feel it is central to the concept of NOR, and I have seen a few good attempts on this talk page to explain why... but the problem is that none of that is in the policy itself. At the moment, the reader has to "read between the lines" to see what the tie is. The policy hints at it with phrases like "because it is easy to misuse them"... but it never actually says it. We need a statement in the policy (not here on the talk page) that clearly ties the definitions to the concept.

One way to do this is to return the introductary statement: "Wikipedia is not a primary source for information". That was a statement that was directly tied the definitions that followed (the rest of the PSTS section) to the concept of NOR. If people have other ideas, however, I am more than willing to discuss them. Blueboar (talk) 16:42, 19 December 2007 (UTC)

See the attempts being made at Wikipedia:Evaluating_sources#Original_research_and_verifiability ≈ jossi ≈ (talk) 17:04, 19 December 2007 (UTC)

Copied here, for ease of access

All sources should be used in a way that does not give rise to new analysis, syntheses or original conclusions that are not

verifiable

by the intended audience, which may require technical knowledge and expertise.

Any interpretation of primary source material requires a reliable secondary source for that interpretation, and that to the extent that an article or particular part of an article relies on a primary source, that part of the article should:

* only make descriptive claims about the information found in the primary source, the accuracy and applicability of which is easily verifiable by any reasonable, educated person without specialist knowledge, and
*make no analytic, synthetic, interpretive, explanatory, or evaluative claims about the information found in the primary source.

If an interpretation is found within the primary source itself, such as a peer-reviewed journal article highlighting original scientific research, it may be cited directly without mediation by a secondary source if that interpretation is quoted or cited accurately without new commentary by the editor. This must be done very carefully due to the potential for abuse. To avoid any problems, always attribute the author of the interpretation in the body of the article.

Drawing conclusions not explicit in the reference cited is original research regardless of the type of source. It is important that references are cited in context and on topic.
Tertiary sources can be useful in providing context and avoiding original research in topics where there exist very large amounts of primary and/or secondary sources. "Common knowledge" claims may be cited to tertiary sources.

What "Wikipedia is not a primary source for information" intends to express is...
Encyclopedias are reference works. As works of reference, they only refer to information available elsewhere, and do not themselves include novel information.
-- Fullstop (talk) 17:25, 19 December 2007 (UTC)

I see two problems with the blockquoted text above. First, "Primary sources that have been published by a reliable source may be used in Wikipedia, but no original interpretations or conclusions may be drawn from primary sources." is odd because no original interpretations of any sources are permitted. Second, "If an interpretation is found within the primary source itself, such as a peer-reviewed journal article highlighting original scientific research," is odd because, for the purposes of NOR, only certain types of scientific research papers are considered primary sources. — Carl (

CBM · talk

) 17:41, 19 December 2007 (UTC)

Yes, this first issue Carl brings up has been the issue most complained about lately. The part of PSTS that introduces PSTS and the part that deals with primary sources already covers this because it says that it refers to the "thing being talked about". If that thing being talked about is a particular brand of analysis or interpretation found in a primary source for that particular interpretation, we'll either need to quote it or closely paraphrase it and attribute it to its primary or original source, or, if any further analysis or synthesis is involved beyond that found in the primary source, we will need to use a secondary or tertiary source to support it. Or at least that's the ideal that is set forth in the policy (though it take it as granted that the policy's "language-by-self-appointed-ad-hoc-committee" still has some weak points).

As to the idea, alluded to by both Carl and Fullstop and several other participants along the way, that PSTS is unnecessary, as well as to Jossi's observation that PSTS isn't clearly enough tied into NOR, I say "well, maybe so". I could argue, and in fact do argue right here, that WP:NOR can be reduced to the four words "stick to the sources". Fullstop's statement in bold font just above also captures the essence of the policy, though IMO it's only peripherally relevant what other encyclopedias do, because this is a specific type of encyclopedia that has managed to carve its own large niche in the world, in part by taking a unique approach to writing an encyclopedia. (Fulltop's observation, incidentally, is presently covered by the statements in the paragraph about tertiary sources that identifies WP as a tertiary source.) While the principle of WP:NOR can be reduced to a few words or a sentence such as Fulltop's, in practice WP:PSTS serves as a further barrier to original research than would be feasible by merely saying "stick to the sources, people." It arises out of actual experience with the kind of editing actions that some users will occasionally show a propensity to do. For example, WP:NPOV needed to be qualified with a section on

WP:PSTS to handle situations where people dig into the backwaters of documents and other obscure marginalia in support of an original synthesis. Do these things overlap to some extent? Sure they do. If we wish to take a minimalist position on editorial policy, NPOV shouldn't need a section on undue weight because they ought to know that NPOV doesn't mean you get to demand equal treatment for some marginal perspective that has no support in secondary or tertiary sources. And people shouldn't have to be told about "reliability" because it should be obvious that something someone said in a blog doesn't have equal weight with an idea that has been stated in many peer-reviewed journals and which has already been well treated in several encyclopedias. Similarly, people shouldn't need to be told that "no original research" involves being much more stringent with how primary sources are used, because everybody should know better. Right? ... Kenosis (talk

) 14:37, 20 December 2007 (UTC)

If the policy wants to prevent people from abusing sources, it needs to be careful not to also prevent others from using them correctly. This is my main concern in participating here. — Carl (

CBM · talk

) 15:11, 20 December 2007 (UTC)

Understood. ... Kenosis (talk) 04:31, 21 December 2007 (UTC)

• Comment: I think it would be a good idea if someone could clearly articulate on this talk page why the concepts of "primary", "secondary", and "tertiary" have any relevance to the concept of No Original Research. Sourced material is by definition not original, and badly sourced material is simply material sourced to an unreliable source. Why doesn't the concept of reliability cover everything that really needs to be said? I suggest an attempt to clearly refute the skepticism that's been expressed on this page. The quote above never directly explains, why the source categorization should be considered essential to NOR. It describes things in a way that seems to simply assume it. The quote abruptly transitions from referring to Wikipedia policy to a series of statements beginning with "Primary sources that have been published by a reliable source may be used in Wikipedia, but no original interpretations or conclusions may be drawn from primary sources..." This abrupt transition does not appear to articulate (let alone convince a reader of) any relationship, logical or otherwise, between the sentences before and the ones after. Best, --Shirahadasha (talk) 18:12, 19 December 2007 (UTC)

Indeed. As said above that text is an attempt to put all the different tidbits of material related to sources and source typing, as they appear in different policies and guideline, into one place for the purpose if seeing what works and what does not, which elements have agreement, and which ones do not. You can follow the discussion at Wikipedia talk:Evaluating sources. ≈ jossi ≈ (talk) 19:36, 19 December 2007 (UTC)

Blueboar, I don't think that the warning "...because it is easy to misuse them" does have more than a peripheral relevance to NOR. I've suggested above this replacement text: "Particular care should be taken with type 1b primary sources because they contain no independent check on the rationality or impartiality of the author…" To my mind that's the main danger in blindly relying on them rather than secondary sources from reputable publishers - it's more an issue of NPOV. The main stricture against WP editors performing OR themselves applies equally well to all sources, as Francis Schonken pointed out earlier[1].

What are the uses of the P/S distinction? 1. Notability is currently defined in terms of "reliable secondary sources". 2. Some say that primary sources should be used with care, but (as above) others disagree that the problem is specific to primary. 3.

WP:NOR currently states that Wikipedia articles should rely on reliable, published secondary sources - others dipute this. Does the primary/secondary distinction crop up anywhere else?  —SMALLJIM

  20:22, 19 December 2007 (UTC)

The policy doesn't clearly express the relationship because there is no clear relationship. The concept of "no original research" means at a minimum that wikipedia does not, itself, publish original research. Research that has never been published anywhere else does not belong on Wikipedia. That, I believe, we can all agree to.

Since syntheses of information is "research" we do not permit syntheses. I believe almost all experienced Wikipedian editors get this, too.

So that leads us to requiring cites and references for content in Wikipedia articles (

WP:V

). But clearly some cites and references are better than others. Almost all of us would agree that a pseudonymous posting on a newsgroup is a the bottom end of any reliability scale. But beyond that there is significant dissent. Defining a reliability scale that is broadly applicable is not so easy, and how to establish what's at the top is not so easy. We say in WP:NOR that "peer-reviewed research" at professional journalism is at the top. However, the PSTS material, which was inserted I imagine to try to provide some way of thinking about this, leads some people to put some kinds of peer-reviewed research (science publications) in a lower-tier of reliability and usability; this is a direct internal inconsistency. Some people take this to an extreme to suggest that research published elsewhere cannot even serve as a reference here at Wikipedia; they take "no original research" to mean, apparently, (a) no original research published here at wikipedia; and (b) no original research mentioned here at Wikipedia. This is in my view a serious confusion and troublingly and non-functionally overbroad interpretation of NOR, and I believe it stems in part from a rigid and discipline-insensitive application of PSTS.

Finally, it is my impression that some people have a fairly broad idea of what "NOR" means -- they think one cannot cite to, or must use with caution, original research that has been published elsewhere. I think that is wrong-headed in the extreme

The PSTS material also fails to fully address significant reliability issues; for instance, the distinction between an editorial in a newspaper and journalism in that newspaper. Efforts to address this latter problem with

WP:RS

have apparently not met with success. I'm unclear what's been happening with the former problem, but from my recent engagement with it it's a mess.

--Lquilter (talk) 22:32, 19 December 2007 (UTC)

A mess indeed... Any fresh ideas on how to clean up the mess would be very appreciated. ≈ jossi ≈ (talk) 22:39, 19 December 2007 (UTC)

I'm trying my best and I think a lot of other editors are too. It feels "fresh" to me but I'm pretty sure it's a retread. I'm rarely given to quoting the Bible but I'm pretty sure that Ecclesiastes 1:9 applies to anything I can produce, sorry to say. Disheartening. Why can't everyone just agree with me? --Lquilter (talk) 22:50, 19 December 2007 (UTC)

yeah, its a retread, or rather, a wall that has already collapsed being given a new coat of paint.

Incidentally, why are you trying to qualify "reliability" here anyway? After all "no original research" is not there to to "address significant reliability issues." That what the eponymous WP:RS is there for. Whether they succeed or not is another matter, and not one we need to care about here, right? -- Fullstop (talk) 23:24, 19 December 2007 (UTC)

If you want to get really depressed, get a sample list by clicking "Random article" half a dozen times and consider how the last n months discussions here relate to what's actually out there in article space. Here's one I prepared earlier:

Brown Stadium, Nižná, Tvrdošín District, Havengore, Rick Cerone, Christ – The Album, Llanbradach. (I rejected one tiny stub.)  —SMALLJIM

  23:37, 19 December 2007 (UTC)

I think part of the problem is that the section has taken on a life of its own. If you go through the history, it started when the phrase: "Wikipeida is not a primary source" was added. Soon afterwards, a (very short) definition was added to explain what a "primary source" was... followed by a definition of "secondary source" to show what Wikipedia should be. As time went on, the definitions became longer and longer. What started as a short sentence in the lead, soon became a section in its own right. Somewhere along the line, however, the phrase that started it all got lost. Someone cut the line "Wikipedia is not a primary source". From that moment on, we lost the entire reason why we were explaining the terms primary and secondary in the context of this policy.

However, the text remained useful as a policy statement in other ways. It was handy to have what amounted to a "primary sources are bad (not forbidden, but strongly cautioned against)" statement in a core policy... even if it did not directly relate to the policy itself. So, bit by bit, the definition wording was shifted from talking about what Wikipedia should not be, to talking about what types of sources we should use. Unfortunately, we now are at the state where the section is useful, but is so divorced from its original intent that people are wondering why it is there in the first place and no longer think it belongs in the policy.

Thus we are faced with several options... do we 1) break off a useful, but unrelated, policy statement into a new policy or guideline? 2) rework the discussion of primary and secondary sources so that it is referring to Wikipedia itself and not the sources we are using within wikipedia (ie return to the original intent of including the definitions)? 3) a combination of both?... or 4) none of the above. Personally, I would opt for number 3. I think PSTS should be policy... I just don't see it as being part of this policy. However, If we return to the concept of "Wikipedia should not be a primary source" then I can see a reworded section to explain what that means. Blueboar (talk) 02:16, 20 December 2007 (UTC)

1. Yes, you have correctly identified how the cruft came to pass. What took you so long? :)
2. "Wikipedia is not a primary source" means "Wikipedia hasn't got anything that hasn't been said before."
3. What Wikipedia is "not" is thoroughly defined at WP:NOT.
   Including what is actually meant
   by "Wikipedia is not a primary source."
4. Re: the options: If proponents of the belief were right about the section being useful, then the section should easily be able to become a policy in its own right. Unfortunately, the proponents of the section are not sufficiently sure of their own footing. So, to stay on the "safe" side, they avoid exposing it to the test, and so simply insist that it stay snuggly in NOR.

-- Fullstop (talk) 03:28, 20 December 2007 (UTC)

Yes Blueboar, PSTS has taken on a life of its own and has grown (like a cancer?) overly large. I think: Let’s abandon the definitions in favour of mainspace definitions until new definitions receive consensus support over the mainspace definitions. (Probably won’t happen - the multitude of reputable definitions already sourced all seem to say much the same thing and I’m yet to see someone highlight differences.) The attempts to tie down definitions of these imprecise concepts has been detrimental to the question of relevance to WP:NOR and to the quality of the prose.

This would leave a much shortened section that hopefully will better focus on the question of relevance.

As for “Wikipedia should not be a primary source": It should be abandoned. It is a clumsy sentence, unclear in meaning, and well-capable of misinterpretation. It is a very poor re-statement of the succinct core of the policy which is “Wikipedia does not publish original research or original thought.” --SmokeyJoe (talk) 03:32, 20 December 2007 (UTC)

ooooh! I like that. “Wikipedia does not publish original research or original thought.” Yes!

Now,

where have I seen that before? :) -- Fullstop (talk

) 05:02, 20 December 2007 (UTC)

Blueboar's restatement was really helpful (wouldn't it be nice to have a /history page that explains this sort of thing, with links to key changes & major disputes?) and SmokeyJoe's comment is really on point. It's always better to use the same language rather than introducing a second set of language that needs interpretations and introduces the possibility of ambiguity. --Lquilter (talk) 17:23, 20 December 2007 (UTC)

• Comment: So far the comments in this section have all been by skeptics of the necessity of a PSTS section in the NOR policy. Skeptics have argued that it is an independent policy which (whatever its independent merits) was incorporated into NOR for historical/convenience reasons. I would like to hear what proponents have to say. What are the reasons for having a PSTS section in the NOR policy? What makes it essential to a workable NOR policy? Best, --Shirahadasha (talk) 06:09, 20 December 2007 (UTC)

Note: Given a decision to focus on a separate proposed WP:Evaluating sources policy/guideline describing PSTS in detail, perhaps the fate of a PSTS section belongs with that policy/guideline proposal rather than here. Best --Shirahadasha (talk) 14:12, 20 December 2007 (UTC)

Hi Shirahadasha, to answer your question, the issue of primary-source use is crucial to NOR. It's often by misusing primary source material that NOR violations occur. For example, if I were writing your biography, I'd look at secondary sources (newspapers, books), and I'd find you were (say) a writer, PhD, worked for several charities, were once arrested for not paying a speeding ticket. These are all issues that secondary sources have seen fit to mention in relation to your life, so my research is good in that I'm using reliable sources and the details are notable, because written about by third parties.

However, suppose I dislike you. I want to go a step further than New York Times. I start searching through birth and death records, divorce courts, family court, criminal court. I discover you've been divorced twice, and that your second husband hired a psychiatrist to advise him about you. I find out you spent three weeks in jail once. I look up your birth certificate and discover you lied to the New York Times about your date of birth.

Even though all the material is documented, and has been published by reliable sources, they are primary sources, and therefore it is not allowed into Wikipedia, because it is being used contentiously. Any use of primary-source material on Wikipedia may be used only to describe, not to analyse or synthesize, and if anyone challenges it as contentious, a secondary source must be found. That's why that section is in the policy. SlimVirgin ^{(talk)(contribs)} 02:27, 21 December 2007 (UTC)

Slim, you've made two statements:

1. "Any use of primary-source material on Wikipedia may be used only to describe, not to analyse or synthesize."
2. "Even though all the material is documented, and has been published by reliable sources, they are primary sources, and therefore it is not allowed into Wikipedia, because it is being used contentiously."

Now...

• Say I have all those records you described, then #1 allows me to post them. Further, WP is not censored, "contentious"-ness is not an issue unless WP:BLP, which is *not* however what NOR can be invoked for.
• WP:BLP already notes "must be written conservatively, with regard for the subject's privacy." The results of dumpster-diving won't pass muster anyway.
• What happened to "codification of general practice"?
• Further, if I were taking all that trouble, I'd sell/give away the research, and additionally distill the result of the press onto Wikipedia. Nationwide coverage + secondary sources for WP + dogfood for the anti-WP whackos.

Not a good rationale/example you have there. There are also better ways to deal with such a scenario. -- Fullstop (talk) 03:21, 21 December 2007 (UTC)

• BLP example which is just as cogent but is clearly "purely" OR-specific? Thanks, --Shirahadasha (talk
  ) 03:40, 21 December 2007 (UTC)

I'd have to agree that a better example is needed. As written most of what you say above could probably be deleted merely through violations of

WP:EVALUATE

, where we try to tie all of the various and miscellaneous 'pieces' together, but I could be missing something as well.

I'd also have to wonder about your last paragraph. I don't see those sentiments expressed anywhere in this (NOR) policy. I just did a quick search for 'content' on the policy page (so I could get any variation of 'contentious') and that term is not there, so therefore your statements about those primary sources not being allowed appear misdirected in regards to this policy, as those are 'almost' what

verifiability. The verifiablity standard is met if those are in fact 'true' records, so the only 'thing' prohibiting the source would be how the source is used, which is where NOR 'may' come in. I just don't clearly see the tie to 'source-typing' in the example. wbfergus ^Talk

13:45, 21 December 2007 (UTC)

Let's take

WP:BLP out of the equation... suppose the article was about someone who lived in the 1800s, and not someone living today ... does Slim's example constitue a NOR violation? I would say that it would depend on how you included the information in the article... are you simply stating facts (That the subject of the article had been divorced twice; that the subject's second husband hired a psychiatrist to advise him about her; that she spent three weeks in jail once, etc.) or are you going beyond the simple facts stated in the sources to make a conclusion, analysis or synthesis? Remember, Primary sources can be used in articles... we just need to take care in using them. Blueboar (talk

) 15:03, 21 December 2007 (UTC)

I'm not sure we even have to bring BLP into it. Slim's example -- which is a great example of something we should prohibit -- is straightforward synthesis, and should be prohibited as "original research that has never been previously published". It should be so prohibited whether it's a living person or an ancient historical artifact, because the synthesis has never been previously evaluated by any third party, and so wikipedia has no way of knowing whether it's accurate or notable. The individual cites allow each fact to be verified but the gestalt is still unverified. So we don't really know that the facts are tied together accurately. We need to prohibit it but the question is what is the best way.

Slim's example also points out an insufficiency in using PSTS classification to prohibit this kind of synthesis. The same kind of synthesis, with all the same kinds of problems, could be done with secondary sources. Imagine an editor sees an article in which "Tom Jones" describes using multiple pseudonyms, "John Thomas", "Tom Johnson", and "Tim Jones", in a variety of track and field competitions. Now the editor goes to various newspapers and sees that people with those names in various track & field competitions. The editor goes to Wikipedia and in the "John Thomas" article puts all those facts together to assert that the competitions were won all by the same person, John Thomas. We might be able to knock this article out on any number of grounds, but we should also be able to knock it out on

NOR

as an unpublished synthesis. In my view it is vital that we be able to use NOR, because we have no idea if this argument is accurate or not -- although the bare facts are cited, we have no idea if the people behind them are actually the same person or just people who happen to have the same name as the pseudonyms chosen by Tom Jones. Even if there's no living person (BLP) and no controversy, we still need to be able to get this out of here. We need NOR to take care of it.

The same would hold true for use of any sources in any "PSTS" class: We need a source to support, not just the individual facts, but the argument; the synthesis. Here we would need some published source that basically establishes the synthesis itself. Someone else has to have evaluated the argument. So I would propose that we strengthen the synthesis part.

--Lquilter (talk) 15:49, 21 December 2007 (UTC)

So, by removing BLP from the equation (say substituting Thomas Edison instead of Shirahadasha), all of the statements about contentious material get thrown out the window (at least as far as NOR is concerned). That is not expressed anywhere in NOR, especially within PSTS. So, as Blueboar pointed out, it is solely a metter of how the material is used, not whether it can (or should) be used. Using your example, that seems to clearly be a case where the 'best way' to handle it is on the article's talk page. If that doesn't conclude the issue, then it could be brought here or maybe even an RfC, but policies shouldn't try to account for every possible permutation that could occur, just maybe 98% of them. Your's is a better example of synthesis I think than the policy has, but that's a different issue for another discussion. wbfergus ^Talk 16:28, 21 December 2007 (UTC)

It would seem that the issues presented so far could be described as

WP:SYNTH issues -- problems with how sources can be used and with combining sources to make a synthetic argument -- rather than problems with the sources themselves. Why can't we cover this by simple requirements like "don't go beyond what the sources say" and "don't synthesize sources into an argument not covered in any source" with a paragraph or so explaining these requirements? Why is a source typology needed? WP:Evaluating sources is currently very long and detailed in its discussion of what kinds of sources can be used in what field, and seems to be getting longer. Best, --Shirahadasha (talk

) 17:33, 21 December 2007 (UTC)

I don't see where your "synthesis" is coming from, Lquilter. You seem to be assuming a lot about SV's example that simply isn't there - she said nothing about the WP editor weaving those primary sources into a synthesis. Based on what you say I assume you wouldn't argue against the addition of just one of those primary-sourced facts to the article - there can be no synthesis in that case, can there? What if two primary-sourced facts are stated in different sections of the article, under "Marriage" and "Death", say - where's the synthesis there? On the other hand my reading of SV's analysis of her example is that she would object to just one of those facts being added because it is "being used contentiously", i.e. the WP editor is trying to sully the subject's name. (I think there's a non-sequitur there, but that's a different matter.)  —SMALLJIM  17:46, 21 December 2007 (UTC)

I got "synthesis" from two places. (1) First, from SV's description itself. She says such documents can only be used in Wikipedia "may be used only to describe, not to analyse or synthesize". (2) Second I got it because the behavior described is exactly what research & synthesis is, by definition. Going to a courthouse to dig up records is research. Putting them together in a form that has not previously existed is synthesis. Posting them on Wikipedia is publication. Since it's the first publication of information (a synthesis), it's not permitted.

As I read SV's example, she is also noting that the behavior is used "contentiously" and she's implying that the information is personal or private. That raises BLP issues and so it's a good example of how these things all inter-relate. But, in my view, WP:OR should be able to take care of unsourced, unpublished syntheses that constitute OR, whether or not they present privacy or BLP or contentious issues. I don't want someone presenting these arguments about the best way to cultivate apple trees, or any other non-controversial topic. So I'd like the WP:OR policy to address this issue as purely and economically as possible in a way that is as broadly applicable to different subject-matters. OR is a problem regardless of its BLP issues. It does indeed present a particular problem with BLP issues, which should be addressed clearly on the BLP policy. How to assess issues of relevance, notability, defamation, and privacy concerns of particular pieces of information (birth, divorce, illness, children, location, etc.) should all be handled at BLP, in my view. I don't think this policy (NOR) has to carry the weight of every other policy goal. If a particular fact is challenged on a BLP article we should really be looking at the underlying reasons why that fact is subject to challenge -- if the underlying concern is privacy or defamation, then that should be addressed directly. In SV's example the concern is something like privacy & defamation as a result of a stitched together (synthesized) picture of the person that "goes further" than the NYT article. WP:OR should provide clear guidance to get rid of the synthesis, without reliance on the BLP concerns; BLP should provide clear guidance to get rid of any inappropriately private or defamatory comments, regardless of whether they are OR or not. --Lquilter (talk) 20:20, 21 December 2007 (UTC)

Going to a courthouse to dig up records is research. No, that is 'source based research', which is allowed, not 'original research' (putting forth new ideas basically). We are allowed to conduct research of various materials (sources) which may be pertinent to an article. But, after finding the material, then care must be taken to use them appropriately, which is the area you hypothosize SV may have been going with her example. The artcile itself just have easily said "On February 29th, 1913, Thomas Edison spent three weeks in jail.", and then cite that to the source (say Baltimore Police Deptmartment, record #XYX). No conclusions, interpretations, analysis, just a simple statement of the facts as stated in the source. That is allowed. I don't see anything in any policy that explictly states that a court record is not allowed, as long as it still meets the reliability and verifiability requirements. Whether it went through a publishing house or a web site is inmaterial. The fact that it is a public record alone makes it a published source in this case. There are many unpublished primary sources listed in the table on

WP:EVALUATE. Court records generally are primary sources, but once they become public records, they are for all intents and purposes, published. wbfergus ^Talk

20:43, 21 December 2007 (UTC)

Perhaps the following misunderstanding could be cleared up: The example above discussed the use of primary sources "contentiously". My understand of "contentious" use is use to advance a position in an argument or dispute. Thus, "contentious" use would seem to suggest that some sort of

WP:SYNTH policy either covers or perhaps be clarified to cover. Am I wrong in this? Could someone explain or give an example of "contentious" use that does not involve the construction of an implied argument and does not imply a synthesis? Are primary sources considered a problem in the absence of "contentious" use and if so, would it be possible to provide examples? In the example, many things seem addressable without having to classify sources. Claiming a person lied to the New York Times because some other source disagrees (regardless of what kind of source), would seem to be not merely a synthesis but an unjustified one and this would seem clear under very simple statements of the policy. The fact that different sources have variant facts does not mean anyone has lied. Best, --Shirahadasha (talk

) 00:41, 23 December 2007 (UTC)

I don't know where to post this, but I am wondering if this is original research in the form of synthesis. Thanks. Ra2007 (talk) 00:32, 22 December 2007 (UTC)

Most definitely. Other Wikipedia articles should not be cited as references. Also, the individual facts should be cited, not pasted together to support a broader conclusion. Vassyana (talk) 00:37, 22 December 2007 (UTC)

The main problem is that the author is trying to demonstrate the validity of a conclusion by citing sources to support the conclusion, rather than citing a source that actually made the conclusion. This would be fine in a research context, but not on Wikipedia. — Carl (

CBM · talk

) 00:43, 22 December 2007 (UTC)

Yes. As Carl/Vassyana note, the OR occurred by using sources to "support" the editor's own conclusion.

This is the most prevalent kind of OR, and while acceptable in secondary sources, it is not acceptable in a reference work like Wikipedia.

As Carl notes, what the editor should have done is cite a source that actually made the conclusion. -- Fullstop (talk) 16:53, 22 December 2007 (UTC)

This was added:

It is important not to add Original Research to Wikipedia, as that would make Wikipedia itself a primary source for that information.

I don't think this is clear in any way, and I can't see how to fix it. Adding synthesis would make WP a secondary soruce, not a primary source. Adding previously unpublished eyewitness accounts of things would make WP a primary source, but that's primarily a matter for WP:V rather than WP:OR. The policy as it stands does not use the convention that any first publication of a theory is a primary source; only eyewitness accounts, experimental data, etc. are classified as primary. — Carl (

Seems clear and simple to me. This was the original intent of what became WP:PSTS. The only thing that might make it clearer is to use the original formulation from

WP:FRINGE to say that WP should not be the primary witness, rather than primary source, and avoid the whole PSTS issue. Dhaluza (talk

) 18:09, 22 December 2007 (UTC)

I completely agree with "Wikipedia should not be the first place in which facts or theories are published." The problem here is just finding a wording that works in that particular sections, while keeping the policy cohesive. — Carl (

CBM · talk

) 18:46, 22 December 2007 (UTC)

Yeah... I had trouble with my addition myself. As I said in the edit summary, it was a poor attempt to do the right thing. I feel that it is important to return the PSTS section back to its roots - explaining what the term "primary source" means in the context of the statement "Wikipedia should not be a primary source"... but, unfortunately the section's wording has shifted from that original context enough that it is hard to add that context back. Not sure how to fix this. I was hoping someone would come up with a better solution in subsequent edits. Blueboar (talk) 19:22, 22 December 2007 (UTC)

And related to that, I have attempted to set up a basic history of this policy with links to major revisions. There's been a lot of churn on this since around April 2006, but the goal is to get important major revisions & language marked -- to serve as a sort of index to the page history. That should hopefully (a) help editors new to this discussion get up-to-speed; and (b) help us locate significant changes if we want to go backward. Ideally it will also include pointers to significant discussions in the talk archives. I started including editor names in a few places thinking that would be helpful so that people could explain what they meant, but it might be better to not have that sort of thing -- the document is not intended to be used as any sort of "let's blame this editor for taking us here", but simply a way to annotate & access major revisions. (I was inspired by Blueboar's summary which was helpful to me, so I added that in; if there are other concise prose descriptions of history of the policy they should get added too.) Cheers, Lquilter (talk) 20:08, 22 December 2007 (UTC)

Well done, Lquilter. -- Fullstop (talk) 20:31, 22 December 2007 (UTC)

Blueboar, 'Wikipedia should not be a primary source' does not make it necessary to "[explain] what the term 'primary source' means in the context of the statement."

1. we have a WP article on 'primary source'. Linkage should suffice.

2. policies need to be in plain language; jargon and ambivalent terminology should be avoided.

3. Even this one solitary use of 'primary source' does not justify the need for a PSTS section!

-- Fullstop (talk) 20:31, 22 December 2007 (UTC)

As this is a generic principle, and the meaning should be obvious from the context, a precise definition of primary source is not needed. But using the alternative "primary witness" further refines the meaning of source to say that WP is not vouching for the content. Dhaluza (talk) 20:43, 22 December 2007 (UTC)

IMO, the advantages notwithstanding, "primary witness" is still too non-plain. -- Fullstop (talk) 21:50, 22 December 2007 (UTC)

(outdent)

Carl, why were you unhappy with your own "Wikipedia should not be the first place in which facts or theories are published"? After all, that pretty much mirrors whats expressed at WP:NOT#Content. -- Fullstop (talk) 21:50, 22 December 2007 (UTC)

I would also support ""Wikipedia should not be the first place in which facts or theories are published." "Primary source" is a jargon term, while this sentence is in ordinary English. A sentence in plain English is always better than a sentence in jargon. Best, --Shirahadasha (talk) 02:31, 23 December 2007 (UTC)

Fullstop: I am happy with that quote. I am not happy with an unqualified that being the first place of publication would make WP a "primary source". I edited the policy earlier to rephrase Blueboar's quote in a way I thought was more cohesive with the rest of the section. — Carl (

CBM · talk

) 03:24, 23 December 2007 (UTC)

Carl, I have no problems with what you wrote... exept that it does not reflect the reason for my addition ... which was to return the clear tie between the section and the basic concept of NOR... that Wikipedia should not be a primary source. Blueboar (talk) 03:39, 23 December 2007 (UTC)

When you say "Wikipedia should not be a primary source" I think you mean Wikipedia should not be the first place of publication for a fact or theory. I agree with that. But I don't use the term "primary source" in that way, and more importantly NOR doesn't use it that way. The definition in NOR is essentially the historiographic definition; (roughly) eyewitness accounts qualify as primary sources, and analysis as a secondary source. So, the way NOR is written, adding new analysis still doesn't make WP a primary source, because of the way primary source is defined. I don't think it's cohesive to put text about creating new analysis in a section of the policy that is about eyewitness accounts. Perhaps you could rephrase it to not use the term 'primary source' and put it in a different section? — Carl (

CBM · talk

) 03:50, 23 December 2007 (UTC)

And there is the problem in a nutshell... The definition used in the policy has shifted. Until about a year ago, NOR did define 'primary source' in that way. The PSTS section started off by using the term 'primary source' in the context of Wikipedia not being the first place of publication. It was that context that directly tied the definitions of PSTS to the rest of NOR. I think it is importatnt to re-establish that context and re-tie PSTS back into NOR. Without using the term "primary source" to mean "first publication", or at least include "first publication" within its meaning, PSTS no longer relates directly to NOR. Blueboar (talk) 15:40, 23 December 2007 (UTC)

This is a proposal to add a new section to WP:NOR covering citing sources. I started writing a detailed rational for moving some material on citing sources from the guideline to policy. Then I figured that if it does not look logical at first glance, anything that I could write would probably not change many minds.

I have tried to give a simplified but accurate layout to citing sources -quick and easy to understand and does not require a novice to make any technical decisions. By far the most common method of citing sources on Wikipedia is the footnote, so I have omitted any mention of Harvard referencing. Anyone inclined to use Harvard referencing would still be allowed to under the phrase “other methods”. People comfortable with Harvard referencing will appreciate being given the choice. Far more importantly those who are clueless will not be forced to make a choice.

Most of our policies ask people to use sound judgment. Asking people to cite their sources is different; it is asking them to take specific actions. Anything we can do to streamline and simplify that process is good, as long as it does not degrade the final outcome in any significant way. Again "as simple as possible, but no simpler”.

How to cite sources

Citing sources and avoiding original research are inextricably linked: to demonstrate that you are not presenting original research, you must cite reliable sources that provide information directly related to the topic of the article, and that directly support the information as it is presented. Any reader should be able to verify that material added to Wikipedia has already been published by a reliable source. Material that is challenged or likely to be challenged, and all quotations, should be accompanied by a clear and precise citation, normally written as a footnote; other methods, including a direct description of the source in the article text, are also acceptable.

Articles can be supported with references in two ways: the provision of general references—books or other sources that support a significant amount of the material in the article—and inline citations, that is, references within the text, which provide source information for specific statements. Inline citations are needed for statements that are challenged or likely to be challenged, including contentious material about

All citation techniques require detailed full references to be provided for each source used. References must contain enough information for other editors to identify the specific published work you used. When citing books and articles, provide page numbers; page numbers must be included in a citation that accompanies a specific quotation from, or a paraphrase or reference to, a specific passage of a book or article.

Full references for books typically include: the name of the author, the title of the book or article, the date of publication, and page numbers. For journal articles, include volume number, issue number and page numbers. References for newspaper articles typically include the title of the article in quotes, the byline (author's name), the name of the newspaper in italics, date of publication, page number(s), and the date you retrieved it if it is online.

Various templates can be used to help format full references more consistently. Templates exist for specific formats, such as {{cite journal}}, {{cite book}}, {{cite web}} and {{cite news}}. There is also a generic {{Citation}} template.

Don't let complicated citation structures prevent you from adding a source: if you don't know how to format the citation, provide as much information as you can, and others may fix it for you.

Brimba (talk) 04:12, 23 December 2007 (UTC)

IMO, on the whole it's an excellent piece of editing advice. I would support a link-to from WP:NOR to whatever location this is placed in, even if it's on an

essay page. ... Kenosis (talk

) 04:51, 23 December 2007 (UTC)

Ahh, that’s the thing, I am proposing it as a logical part of NOR itself. Brimba (talk) 05:00, 23 December 2007 (UTC)

It seems to me (however well written) to really just be about

WP:V. I don't see what it has to do with OR, aside from, obviously, that OR is inherently improperly sourced. Someguy1221 (talk

) 05:01, 23 December 2007 (UTC)

I thought about suggesting it as part of V. However, V is about the quality of a source, while NOR encompasses the need for a source. In truth there is considerable crossover between the two; but it seems better in NOR to me. Brimba (talk) 05:11, 23 December 2007 (UTC)

It is definitely the case IMO, that WP:NOR is in a very important sense an extension of WP:V, or at minimum more closely intertwined than either of the two are with WP:NPOV. This is, I feel sure, why a significant number of editors proposed that

WP:A be a consolidation of V and NOR. That particular effort was rejected by the community after being given the status of policy that superceded V and NOR. I happen to believe that this rejection was in very large part due to an expectation of continuity in policy among those many users who had found the current tripartite method a workable and acceptable one upon which those users had come to depend. ... Kenosis (talk

) 05:18, 23 December 2007 (UTC)

The way I see it, given that the purpose of WP:OR is to describe what OR is, how to spot it, and how to avoid it, there is no need on this page to offer detailed descriptions of proper methods of citing sources, which could be handled by a link to...You know what, I'm going to leave that the way it is and admit I can't think of a perfect target.

WP:CITE is about the accepted way of presenting those sources. I think your text above gets some points across that maybe don't perfectly fit in any of those three, but it may do better to be split up. The first paragraph could probably find a home here, but the rest of the text just seems to be too tangentially related. It seems to just be a summary of some nitty gritty from WP:CITE.Then again, it couldn't hurt to summarize WP:CITE on this policy, but I think you go into unecessary detail for that. Someguy1221 (talk

) 05:32, 23 December 2007 (UTC)

Cite needs a complete rewrite atm. Having said that, I would not see the use in offering anything here short of what I am purposing. It gets too simple to be of much usefulnss. The current in-line linkage would do everything a short summry paragraph would do, without the added wordage. I’ll see if I can reword it and cut it back some more though. Brimba (talk) 05:44, 23 December 2007 (UTC)

Version 2
{{main}}
Citing sources and avoiding original research are inextricably linked: to demonstrate that you are not presenting original research, you must cite reliable sources that provide information directly related to the topic of the article, and that directly support the information as it is presented. Any reader should be able to verify that material added to Wikipedia has already been published by a reliable source. Material that is challenged or likely to be challenged, and all quotations, should be accompanied by a clear and precise citation, normally written as a footnote; other methods, including a direct description of the source in the article text, are also acceptable.

Full references for books typically include: the name of the author, the title of the book or article, the date of publication, and page numbers. For journal articles, include volume number, issue number and page numbers. References for newspaper articles typically include the title of the article in quotes, the byline (author's name), the name of the newspaper in italics, date of publication, page number(s), and the date you retrieved it if it is online.

Various templates can be used to help format full references more consistently. Templates exist for specific formats, such as {{cite journal}}, {{cite book}}, {{cite web}} and {{cite news}}. There is also a generic {{Citation}} template.

Brimba (talk) 05:55, 23 December 2007 (UTC)

Your last comment convinced me; I like both versions. I think you should alter the "full referneces" paragraph to note "For journal articles, also include..." Let's not be giving the impression people shouldn't provide the journal name and article title. (And if you replace "greek" with "chinese" and the second to last sentence with "eat whatever the waiter puts in front of you," I've totally been there.) Someguy1221 (talk) 06:07, 23 December 2007 (UTC)

Yea, that was fun to write. ;0

I can certainly make that change. I like the first version better, but I suppose the second would do as well. Remember, I have been staring at this for a while, so my opinion is a little slanted.

Making things simple and easy is key to getting people to add sources. Right now it is not simple, nor easy. Or if it is, that’s because you have figured it out in the past, I am speaking of someone who has never added a source before. Brimba (talk) 06:21, 23 December 2007 (UTC)

Along these lines, I think it could would also be nice to insert a new "section 1" to WP:CITE, something along the lines of "If you are a total newbie, feel free to just do <insert simple instructions here>; an experienced Wikipedian will inevitably come along and fix it." Someguy1221 (talk) 06:30, 23 December 2007 (UTC)

I have a problem with this... It is a prime example policy creap. While in itself the proposed addition is a very good policy statement, it only tangentially relates to the concept of "No original research". It belongs somewhere... but not here. Blueboar (talk) 15:52, 23 December 2007 (UTC)

“It belongs somewhere... but not here.” That’s the problem.

Citing a source has gone from a nicety to a requirement in certain cases. However, we have no policy on what exactly that means.

We do not need yet another policy page; we have enough of those already.

Accurately citing a source is fundamental to both NOR and V. Stating how to do that should be clearly laid out in one of those two, and yet its just tangent enough to both for someone to make a good argument that it does not belong “here”.

CITE is a guideline (reference), and functions as such, meaning its all things to all people. And as currently written it contains not just the kitchen sink, but the kitchen sink from next door as well. That’s not going to change, because it’s a guideline.

The biggest stumbling block we have on sourcing is we have nothing clear and concise written on how to do it; anyplace. If someone new adds content that is then fact tagged, we point them to CITE. Even if they have the information in front of them, they have to be motivated enough to sort that page out. Unintentionally, CITE acts as a divider between those in the “know” and those who are clueless. Even if the page was rewritten it would include too many details and too many choices because its function, to cover all the various details. It should exist for those who have questions on certain points, not as the first place we send people wishing to cite the source they found.

The only other location that makes sense is WP:ATT. That’s the most natural location of all.

I was going to add some more, but have to go off-line.

Brimba (talk) 19:22, 23 December 2007 (UTC)

Throughout Wikipedia are plot summaries (of books, films, etc.). In few cases are any sources other than the underlying thing be summarized cited. That is, the wiki entries don't cite to some other published plot summary, but instead proceed to summarize the plot based on an editor's reading / viewing / etc. of the thing being summarized. How is this not original research? If it is, should there be an exception to the general policy stated in the policy? —Preceding unsigned comment added by 24.91.117.236 (talk) 06:32, 23 December 2007 (UTC)

Care must be taken, but plots can be summarized from the primary source as long as they don't introduce analysis, synthetic, speculation, or the like. As there are generally many eyes that watch such plot descriptions, any points that may be considered contentious should be cited to a source other than the primary source. (for example. "In the movie, Mr. Smith motive is not explained, but the producer in an interview in Blah Magazine stated that..." is appropriate; "Mr. Smith's unstated motive is to ..." is not). --MASEM 06:38, 23 December 2007 (UTC)

(ec)This is covered in

it's all about the spirit, not the letter. As long as the editor doesn't analyze the plot, merely reiterates it, it's probably not a violation of OR. Someguy1221 (talk

) 06:42, 23 December 2007 (UTC)

So as not to repeat myself, and to not run the risk of incorrectly paraphrasing the statements of others, I would recommend taking a look over this recent discussion on the matter of fiction in Wikipedia:

Wikipedia talk:No original research/archive30#Fiction in Wikipedia. Vassyana (talk

) 06:52, 23 December 2007 (UTC)

In writing any summary, decisions need to be made as to which elements should be included and which can be skipped over. If the flaw involves something that is crucial to the plot, then it should probably be mentioned in the article (as a simple statement of fact... no "analysis" or editorial comments included)... In most cases, a flaw such as a change in hair color would not count as being crucial to the plot, and so mention of the flaw can be safely left out of the summary (as it would be trivial) ... if hair color (or whatever the flaw is) is crucial to the plot, then it probalby could be mentioned. Of course, chances are that if the flaw is crucial to the plot, some critic will have noticed the flaw and mentioned it in a published review. In which case you can cite the review. Blueboar (talk) 03:33, 24 December 2007 (UTC)

Wikipedia is

WP:NOT

an indiscriminate collection of information. So, drop the change from blond to brunette unless it contributes to the story. And no, NPOV and NOR never conflict; they are (at least nominally) completely distinct. NPOV is for the relative representation of sources appropriate to their weight, NOR is to ensure that editors don't make a source say something it doesn't.

Having "only plot elements that are consistent" stay in and "plot elements that are puzzling or straightforwardly goofy" removed is part of good article writing. Like I said, Wikipedia is not an indiscriminate collection of information. -- Fullstop (talk) 04:27, 24 December 2007 (UTC)

As Faullstop says, it depends on the importance of the material. Some continuing fictions are characterized by deliberate frequent and incompatible changes; sometime in fiction it will mark a major development of change in the series; sometime it will be an inconsistency so major as to excite published comment, but often it's the essence of fancruft. These article need to be a medium between being too sketchy to be meaningfull, and so detailed as to lose perspective. My personal guide is that the article has to make sense to people who do not know much or anything about the fiction in advance. It was the irrational overdetail of many pot articles which has led to the present over-reaction. DGG (talk) 04:35, 24 December 2007 (UTC)

OK, here's another suggestion for you all to ignore :-) I was going to work it up into an essay, but this seems like an apposite moment to present it, incomplete and poorly worded though it is.

This is a proposal for a complete rewrite of the primary/secondary distinction in terms of the people that originate the sources, rather than relying on features of the sources themselves, as the present policy does.

• Updated the Definitions and Usage sections below.  —SMALLJIM  19:37, 17 December 2007 (UTC)
• Revised the definitions and examples again, per discussion below.  —SMALLJIM  10:01, 25 December 2007 (UTC)

I'd be grateful if someone would test it against a few cases to see if it runs, limps, or falls over straight away.

Primary and secondary sources

Rules for this subsection:

• Always take the last option that fits without forcing (1a > 1b > 2)
• A topic is a Wikipedia article or a part of it.
• "someone" and "other people" must be identifiable non-fictional people.
• The singular "someone" includes the plural "some people".

Definitions

1. A primary source for a topic is either:

• a. an observed thing that is relevant to the topic, or

*b. a source in which someone describes their own thoughts about the topic or describes things that they have observed about it.

• b. a source in which someone describes their own thoughts about the topic or makes statements about its properties.

2. A secondary source is a source in which someone discusses at some length someone else's thoughts or observations about the topic.

2. A secondary source discusses at some length other people's thoughts about the topic and the statements they made about its properties.

Using sources

• Wikipedia editors must always take care only to describe what is included in the source and never to introduce their own thoughts (original research) into any topic. The aim is that any reasonable, educated person should be able to confirm that the information contained in the topic agrees with the cited sources, with nothing added.
• Particular care should be taken with type 1b primary sources because they contain no independent check on the rationality or impartiality of the author. Because of this concern,
  reliable
  , published secondary sources are preferred.
• For clarification: there are two types of original research (OR) to beware of. Firstly Wikipedia editors must not add their own OR to articles, whether it is something they have thought of themselves or through making assumptions about what the sources say. Remember that Wikipedia is not a primary source and must not contain any OR. Secondly, although it is not forbidden to report on OR by other people (this is usually of most concern in type 1b sources), it is always important to consider
  notability
  .

Examples

Example one. The Gundestrup cauldron is a (type 1a) primary source because it is an "observed thing" obviously relevant to its own topic. Now, Fred Smith has written up his own thoughts about the cauldron in a book The Gundestrup Cauldron - My Way. That book is also a primary source (type 1b), because it describes Smith's thoughts about the cauldron, but doesn't mention anyone else's thoughts or observations about it. However, Jane Jones' well-referenced book All About the Gundestrup Cauldron is a secondary source for the cauldron because in it she discusses other people's thoughts about it and the statements they made about its properties.

Example two. Many sources are both primary and secondary depending on the context. For instance Jim Black's book A History of London is a secondary source for London's history, because in it he discusses other people's thoughts and statements about the city. However, the book is also a (type 1a) primary source on Jim Black because it is an observed thing (a book) relevant to Jim Black. Jim Black's autobiography is a (type 1b) primary source on Jim Black because in it he describes his own thoughts about himself.

Example three. A photograph of a car is a type 1a primary source for that car. It might be a type 1b source, if it's photographed in a way that could be said to express the photographer's thoughts about it (that's probably forcing it a bit). It certainly isn't a secondary source, though.

Comments on the paradigm shift

That's it (E&OE). It probably needs some further rules about the meaning of "thoughts" and "discussing" (discussing certainly includes interpretation, synthesis etc).  —SMALLJIM  19:13, 16 December 2007 (UTC)

It seems to me that the biggest problem with PS/SS/TS is how it is used and what it is used for. The definitional problems suggest, to me, that it is not always helpful and in fact may be more frequently unhelpful and confusing. I do appreciate fresh attempts to cut through this mess, though! --Lquilter (talk) 21:11, 16 December 2007 (UTC)

Yes. This discussion caught my attention when I was checking to see if I could use a photograph of someone playing a musical instrument to cite the claim that they were a player of that instrument. I'd consider that using the photo as a secondary source. It would only be a primary source if the article were about that specific photo. The primary/secondary distinction as made in composition classes everywhere hinges on the relationship of the source to the subject. The primary source is the subject. Everything else is secondary but only in the context of a given piece of writing. For most wikipedia articles there will be no citable primary source. The primary source for an article about carrots would be a carrot. A photograph of a carrot is a secondary source by virtue of the fact that it's not a carrot. It would only be a primary source in an article about that particular instance of a photograph of a carrot. Anything else is a deviation from standard usage and quite possibly original research. --Vlvtelvis (talk) 00:30, 17 December 2007 (UTC)

I disagree that a photograph of a carrot is a secondary source. Maybe, perhaps, if it is particularly artistic, thus "saying" something, or shows a peculiar view of a carrot, or shows a use of a carrot, or an extreme of a carrot, though in most of these imagined cases, a caption would be needed, and the caption would be secondary. However, if it is simpy a carrot, I cannot see how you can see it a secondary. What transformation of iformation is there? Can you provide an example of someone stating that a mere photograph (uncaptioned) is a secondary source? --SmokeyJoe (talk) 02:16, 21 December 2007 (UTC)

Hi Vlvtelvis. I'm not sure how much you actually appreciate about source typing in Wikipedia, so I apologise if this is a naive response. You seem to think that there can be only one primary source for a topic. That is not at all the case. It is also important to distinguish between the subject of a topic and the sources that are offered in support of it. In no case (whether under

WP:PSTS or my suggestion above) can a photograph of someone playing a musical instrument ever be more than a primary source. Hope this helps.  —SMALLJIM

  18:49, 17 December 2007 (UTC)

Thanks for the comment. I've just added a first attempt at a "Usage" section to the above, written from the same "people" point of view. I think it explains the intention of PSTS far more concisely than the existing text does.  —SMALLJIM  23:05, 16 December 2007 (UTC)

This model would work much better for religion and philosophy articles than the current PST model. The PST model assumes a monotonic regress from the source: primary=close, secondary=observations of, teriery=observations of secondary, etc. With religious thought (and sometimes literature and philosophy and art as well) the relationship is more cyclical with layers of interpretation based on interpretation that eventually are reintegrated into direct reading of the original primary source. The advantage of this reframing is that it does not assume a monotonic regress from the source.

RavPapa above suggested that religion articles offer a good litmus test for OR problems so I'd like to see how this works with the bible commentary example I gave above:

• bible manuscript - primary source (1a)
• bible edition - (1b) - somebody/some committee's opinion of the "real" or "best" text based on observing manuscripts, possibly (2) when an annotation discusses variant readings and cites sources for those - as is sometimes done in critical editions
• midrash - 1a for an article on midrash, 1b for an article on bible text
• rashi - 1a for an article on rashi, 1b on bible text (Rashi's opinion of the bible), 1b on midrash (Rashi's opinion of midrash)
• academic article on rashi - 1b to the extent author draws his/her own conclusions directly from Rashi's commentary and is writing about Rashi, 2 if the author is writing on the bible and using Rashi's interpretation of a biblical passage, also 2 if the author does a lit review of other works on rashi's commentary
• Plaut commentary on bible (based on critical scholarship) - 1a if someone is writing an article on modern biblical commentaries, 1b if Plaut is drawing conclusions about the meaning of a passage directly from its textual content, 2 if Plaut is summarizing what other commentators have said

The definitions above seem to work well for material like biblical commentary. I'd personally prefer this alternate approach. Egfrank (talk) 13:05, 17 December 2007 (UTC)

Thank you for the comments. Your worked examples are from a field that I know nothing about, but the important thing from my point of view at this stage is that the model is easy to understand and gives sensible results. I know that in the current environment here my proposal doesn't have Shadrack's hope in a fiery furnace of being adopted, but you've given me some confidence that it will be worth writing up as an essay intended to give a different perspective on the issue. Anyone else care to contribute to something a little more productive than most of what's going on here at the moment?  —SMALLJIM  18:45, 17 December 2007 (UTC)

Smalljim, I think it would be great if everyone wrote up their own "essay" (proposed guideline?) about this, say, for the next week (through 12/24); then we all discuss the relative merits of the various proposals. Repeat as necessary until we have consensus or a small set of clearly defined options to take to the larger forum. Maybe if some people break free of the model with creative thinking some beautiful solution will arise. --Lquilter (talk) 20:16, 17 December 2007 (UTC)

The proposal is hard to interpret because 'observe' can mean different things. If it means to relate eyewitness testimony, then that type of source is certainly primary. If it refers to interpretation, synthesis, or logical deduction, then the sources called '1b' above are what are ordinarily considered secondary sources - they provide analytic and synthetic claims about primary sources. I see no reason why we should attempt to redefine such sources as a type of primary source. They are exactly the type of source we should be using on Wikipedia, and by removing them from the things we consider secondary sources, we essentially redefine secondary sources to mean tertiary sources. — Carl (

CBM · talk

) 04:28, 18 December 2007 (UTC)

Thanks, Carl. Apologies for the late reply. You're right that observe was a poor choice of word. I certainly intended it only in the sense of reporting the properties of and not in the sense of making observations about e.g. "that looks like a fast car": which falls under the thoughts about part. I shall do some rewording.

Regarding your second point, the real problem is that there's a huge grey area in the existing system because the definitions of primary and secondary are not mutually exclusive. I think my proposal removes this. Consider Fred Smith's book in Example one above - under

WP:PSTS as it is written today, you could argue that it is a secondary source because it is a "secondhand account[s], at least one step removed from an event" (I suppose the "event" is the making of the cauldron?), whereas someone else could use the examples of primary sources to argue that it must be primary because it is "published experimental results by the person(s) actually involved in the research" or an "original philosophical work" or an "artistic and fictional work". My model clearly puts Fred Smith's book as a type 1b primary source, which is the status that we'd actually want it to have, I think. _{I've retitled this section so I don't miss any more comments posted here.}  —SMALLJIM

  15:59, 19 December 2007 (UTC)

The main problem with the current article is that the primary/secondary/tertiary distinction is primarily a scientific/historical paradigm - in the humanities and certain parts of the social sciences it is hardly discussed mainly because it breaks down and leads to self contradictions. Instead one is trained to focus on particular analytical flaws. Another editor has suggested that needs to be the focus (see #Why are people upset with this policy?)

Above I gave an example from biblical hermaneutics (see #PST in religion articles), but the same problem also exists in psychology. Take Carl Jung's Memories, Dreams, Reflections. This book is at one and the same time a

• "primary" source (1a) in the theory of psychology both by virtue of it being the topic of numerous articles by Jung and by virtue of the fact that Jung includes observations based on his clinical experience.
• "primary" source (1b) in the critique of Freudian theory
• "secondary" source (2) in psychology, anthropology, religion, and mythology - Jung explicitly and implicitly uses the work of many other scholars in those fields.

Frankly though if I had to use the current classification scheme on the article page I wouldn't know how to classify this book at all (or for that matter any of the post Freudians). Nor would any other editor. The result would be argument after argument on the talk pages. A policy that creates arguments rather than giving people guidelines for resolving them likely causes more damage than help. Best, Egfrank (talk)

It should not be necessary to explicitly determine whether the type of source in order to determine whether it is used correctly. The main question would likely be whether to explicitly attribute certain claims to Jung or to state them as accepted fact. That would come down to a general sense among editors of whether the claims are generally accepted by contemporary scholars, not to the type of source that the book represents. — Carl (

CBM · talk

) 14:47, 18 December 2007 (UTC)

In the humanities, something is good OR if you are the first to think it and are in coherent dialog with what others have already said about the topic. It is bad OR if you developed it without bothering to even check if others have already thought it or misread or distort what others have said. Similarly, something is a good source if the author developed it in dialog with others and a bad source if the author wrote in isolation of others. On rare occasions an author writing in isolation can become a good source, but only if his or her own lack of dialog with others is compensated for by a critical literature that develops around his or her inventive theory. So the categorization of 1a/1b/2 is very much to the point for assessing source quality and OR in the humanities. Egfrank (talk) 12:03, 19 December 2007 (UTC)

I believe that original research would help wikipedia in some very limited cases. There are some articles which have such poor or hard to find coverage that original research may be neccessary to increase the overall quality of the encyclopedia. Now granted this is a rare case and should not be used to tarnish articles such as those on celebrities or to add made up and crazy numbers, but there is a limited use for Original research I believe. It should be more of a 'warning' and placed under the microscope rather than disallowed entirely and purged. It should be enforced based on the nature of the article and the nature of the addition.

Once again I will state that this is a rare case and should only be used where non-original research is very difficult or impossible to find due to a lack of coverage and such usage should be placed under a microscope. Klichka (talk) 19:52, 19 December 2007 (UTC)

Of course. See

WP:PPP. There is no need to explicitly allow for rare exceptions in the policy, and doing so might lead to more harm than good. Experienced editors know when to ignore rules. Contributors with good stuff at their fingertips are urged to be bold. --SmokeyJoe (talk

) 20:27, 19 December 2007 (UTC)

I think we all understand that there are situations where a bit of original research is unavoidable and forgivable in moderation. For instance, I made this edit about a year ago, just a few hours after personally travelling the highway in question. This type of information, while relevant to an article that delves into the status of construction activity on a highway that's currently undergoing expansion, isn't really individually notable as a news story, so media coverage was nonexistent and I was technically engaging in original research by basing the edit on personal experience. But this edit wasn't advancing a theory or an alternative interpretation of anything — it was just a minor status update on a project whose bigger-picture aspects are well-sourced, and which was physically verifiable by anyone travelling the same stretch of highway.

It's important to remember that

WP:NOR, as written, applies principally to the advancement of original interpretations, opinions and theories. It doesn't entirely prohibit the sparing inclusion of objective and straightforward facts that can be externally verified through alternative means. Though at the same time, if a topic is so obscure and uncovered that original research is required to even assert its existence, then it probably shouldn't have an article at all. Bearcat (talk

) 10:39, 20 December 2007 (UTC)

Bearcat is right; there's no argument. If the fact changed significantly alters an argument or point ("the highway is poorly constructed and that's what led to its collapse") then personal observation doesn't work. But if it's not a linchpin fact then it's simple fact-checking. --Lquilter (talk) 15:52, 21 December 2007 (UTC)

I look at it from the point of view of someone trying to verify the information. The only physical displacement required, if any, should be to a well-stocked library, not the site of the article's subject. I don't see anything wrong with making the edit you describe, but I don't think you'd be in a strong position to defend it if it were challenged (or simply removed). Ultimately, editors are the checks on the system, the text of the policy is not; and the unclear implication here is that NOR policy should make more room for exceptions.

WP:IAR covers that. WP is not a news site, and is ideally not a repository of unverifiable information. Blackworm (talk

) 02:47, 26 December 2007 (UTC)

There is a discussion going on at Wikipedia:Reliable sources/Noticeboard#Wikinews: Please post definite answer. (essentially about the reliability of quoting and citing to Wikinews interviews) that touches on WP:NOR issues.... some people here may want to participate. Blueboar (talk) 13:17, 25 December 2007 (UTC)

My contribution to the PSTS debate. Bark here or at Wikipedia talk:Wikipedia is a tertiary source. --Francis Schonken (talk) 18:20, 18 December 2007 (UTC)

I have reverted the change of the policy to this proposal because the proposal takes the approach that article that lack primary sources are poor articles. This is wrong. For subject matter that has been around for a hundred years or more, the primary sources may be inaccessible, in a foreign language, or in topics with heavy math content, use outmoded and confusing notation. --Gerry Ashton (talk) 20:30, 18 December 2007 (UTC)

Which part of the policy says or implies that, Gerry? SlimVirgin ^{(talk)(contribs)} 20:39, 18 December 2007 (UTC)

This version states:

Wikipedia illustrates such summaries and descriptions with material that is as close as possible to the primary source(s) on the described topic.

A Wikipedia article without illustrations is a poor article

I took those statements, in combination, to indicate that articles without primary sources (when available) are poor articles. I disagree; I think the use or non-use of primary sources is a matter of good writing. I note, for example, that university professors teaching introductory physics courses never ask their students to read

Philosophiae Naturalis Principia Mathematica. I also noticed that my electromagnetics professor never asked me to read A treatise on electricity and magnetism. If university professors can decide it isn't always necessary to use primary sources, why can't Wikipedia editors do the same? --Gerry Ashton (talk

) 21:02, 18 December 2007 (UTC)

I get a 404 error on your first link, Gerry.

Dreadstar †

21:09, 18 December 2007 (UTC)

They work for me, both in Internet Explorer and Firefox. The first is to a Wikipedia article about Isaac Newton's Principia. The second is to volume 1 of James Clerk Maxwell's seminal work. --Gerry Ashton (talk) 21:25, 18 December 2007 (UTC)

Oh, sorry, it's the very first one: "This version states:". Does that one work for you?

Dreadstar †

21:33, 18 December 2007 (UTC)

I don't think the current version says that anymore, Gerry. SlimVirgin (talk)(contribs) 21:31, 18 December 2007 (UTC)

Francis, do you even know what you're reverting here? The policy has been restored by several people and copy edited by Brimba. You're reverting to a version that I'm not aware has much, if any, support, and the writing in it really does need to be fixed. SlimVirgin ^{(talk)(contribs)} 21:30, 18 December 2007 (UTC)

I have no idea what you're talking about. You reverted to a version of the article that seems to have no or little support, and which needed a copy edit. Please don't do that. If there is one point in it that you object to, please say which on talk. SlimVirgin ^{(talk)(contribs)} 21:34, 18 December 2007 (UTC)

Honestly, I have no idea what you're talking about. I only saw that your recent edits to the WP:NOR page gathered no consensus, see #Discussed and undiscussed edits (that is, the section above this section), so I reverted them. Take it from there.

Well, if you had applied less editing to the WP:NOR page ([2]), you might maybe have understood what I was talking about. So, are you interested in what I was talking about, or do you simply say "I have no idea what you're talking about" because you don't want to hear? --Francis Schonken (talk) 21:51, 18 December 2007 (UTC)

Francis, you've written a nice essay which appears to have no bearing on avoiding original research, and tried substituting it for the section here which clearly requires editors to only use primary sources for facts, and secondary sources for interpretations of these facts. Your logic needs a great deal of explanation, and be assured that it does not have consensus. .. dave souza, talk 22:26, 18 December 2007 (UTC)

Dave souza misunderstands the policy; editors may quote interpretations of facts if the interpretation is contained in the primary source. There isn't an OR problem unless the editor is doing the interpreting. --Gerry Ashton (talk) 22:33, 18 December 2007 (UTC)

Gerry Ashton doesn't seem to realise that editors can state facts, including the fact that interpretation is included in the primary source, but must look to a secondary source for interpretation or evaluation of such facts. Unless, of course, we change the policy, rather than clarifying the wording of the policy. .. dave souza, talk 22:53, 18 December 2007 (UTC)

Dave souza and Gerry Ashton are talking past each other and probably have different internal models of what a "primary source" is. --Lquilter (talk) 22:46, 19 December 2007 (UTC)

Anyway, inspired by dave's suggestions I highlighted some parts of the essay that are hard "no original research" requirements. [3] --Francis Schonken (talk) 22:48, 18 December 2007 (UTC)

<undent> Thanks, that helps. The crucial paragraph is Wikipedia is a tertiary source : Wikipedia summarizes descriptions, interpretations and analyses that can be found in secondary sources, and/or bases such summaries on tertiary sources. Wikipedia illustrates such summaries and descriptions with material that is as close as possible to the primary source(s) on the described topic.
The "that can be" leaves the way open for them being found in other, dare we say primary, sources. The current careful balance of requirements is needed: a first suggestion would be to change that to "should be found in secondary sources, or high quality tertiary sources" which deals with the point that tertiary sources are often lacking in necessary depth. The terms "illustrates" and "illustrations" seem rather confusing to me, suggesting pictures rather than text. The phrase "material that is as close as possible to the primary source(s) on the described topic" doesn't really equate to the current requirements when using primary sources. .. dave souza, talk 23:08, 18 December 2007 (UTC)

Don't think it was all that complicated to improve the text on that point, replaced "can be" by "are" [4] --Francis Schonken (talk) 18:06, 19 December 2007 (UTC)

Actually, we dare not say. The more essays/policies/live usage I read in the phrases "primary sources" and "secondary sources" appear, the more convinced I am that there is only one word for the use of these phrases. And that word is 'evil.'

There is no other word for all the grief those two phrases have caused.

Frchrissakes, is there no way to express a "crucial" paragraph in plain English, using unambiguous words and without wiki-speak jargon? Sheesh.

-- Fullstop (talk) 23:36, 18 December 2007 (UTC)

I'm not too convinced that the benefits outweigh the added complication of having to explain the P/S/T sources distinction. Of course, for those who are acquainted with these distinctions it is an easy paradigm, that helps in enhancing quality. I don't think Wikipedia can ignore these concepts, but maybe treating them at policy level is contraproductive. Anyhow, a policy that first goes through the trouble of explaining the P/S/T sources distinctions, and then ends up saying something like (1) one shouldn't make interpretive claims about primary sources; (2) one shouldn't make interpretive claims about secondary sources; (3) one shouldn't make interpretive claims about tertiary sources; and (4) one shouldn't make interpretive claims about any sources, whether they're primary, secondary or tertiary, and that's the reason why we explained the difference between these types of sources... makes no sense - alas that's more or less the state of the current PSTS section. But that doesn't mean I think ignorance should rule the world. --Francis Schonken (talk) 18:06, 19 December 2007 (UTC)

The "tertiary source" part is even more tricky. We haven't been discussing it as much as the P/S distinction, but there is no really clear separation in he academic world or in common usage. For example, an advanced textbook with citations partially to the primary literature can be regard as either. DGG (talk) 01:32, 19 December 2007 (UTC)

Tertiary sources aren't all that much written within academic disciplines (e.g. encyclopedias are written by "encyclopedists" - not representatives of some sort of an academic discipline), so I don't worry too much that some academic disciplines give little attention to tertiary sources. But to raise the awareness of wikipedians (which are a type of encyclopedists) for the "medium" they're writing in is not by definition bad imho. --Francis Schonken (talk) 18:06, 19 December 2007 (UTC)

Right, the distinction isn't at all hard-and-fast. Same as with WP:NPOV. ... Kenosis (talk) 04:07, 19 December 2007 (UTC)

Agree --Francis Schonken (talk) 18:06, 19 December 2007 (UTC)

Proceeded with the replacement of the updated

Gerry, I suppose your comment below was appended to the wrong section: the block quote you refer to is in the next section --Francis Schonken (talk) 20:41, 20 December 2007 (UTC)

In the block quote above, "any interpretation of primary source material requires a reliable secondary source for that interpretation" is longer than necessary. It suffices to say that any interpretation of source material requires a reliable source for the interpretation. The source containing the interpretation would be secondary (according to the current definition in NOR) because it is interpreting a another source, but readers of the policy don't need to know that. It would be better to phrase it as "Wikipedia editors must not interpret source material, and should draw interpretations of source material from other reliable sources." Using active voice avoids ambiguity about which author or editor is doing the interpreting. --Gerry Ashton (talk) 20:31, 19 December 2007 (UTC)

Commenting only on the two versions GA presents here, the latter phrasing is an improvement in my opinion. --Lquilter (talk) 22:44, 19 December 2007 (UTC)

I've always found the wording of this section, and the examples given, somewhat confusing and/or unclear. When trying to explain what sort of edits constitute original synthesis, I generally use the following examples:

The first kind

John says A. Bob says A implies B. Therefore, John and Bob say B. - This example constitutes original research by advancing a position unsupported by either John or Bob; John never mentions B, and Bob does not claim A is true.

The second kind

John says A. Bob says not-A. Therefore, John is wrong. - Bob never claims that John is wrong, so this extension of his argument constitutes original research.

The third kind

John says A. Bob says B. Steven says John is wrong. Therefore, Bob is right. - Claiming one assertion to be wrong is not an endorsement of an alternative assertion, and thus this extension of Steven's claim constitutes original research.

I think adding something along these lines would greatly clarify the meaning of "original synthesis." I'm presenting it here to get thoughts and opinions (and because I can't actually add it myselfWhoops, mistook the color of the protection template. Well, I still don't like editing policies myself). Thank you. Someguy1221 (talk) 01:04, 22 December 2007 (UTC)

Comments

Also, very often it occurs in a "John says A, Bob says B, therefore C" fashion. ["C" is the result of "A"+"B", is the essential assertion.] From my experience, this is the most common "formula" of original synthesis. Vassyana (talk) 01:15, 22 December 2007 (UTC)

OR is the process of using sources to state a conclusion not immediately evident in those sources. On Wikipedia, synthesis is the term used to describe the process of "adding two plus two together" and then stating the result as if it were a source's.
For example, a book on horses might say horses are quadrupeds, and that horses are mammals. The synthesis would then occur when a Wikipedia article cites that horse book to conclude something else, like a) quadrupeds are mammals, b) mammals are quadrupeds, c) horses are like cows because in a book on cows those were also said to be quadruped mammals.
All three cases are synthesis; the truthfulness of the conclusions are irrelevant. -- Fullstop (talk) 17:35, 22 December 2007 (UTC)

• The second example seems to me to be more a violation of
  WP:NPOV - two different sources of equal weight (in the example) and a definitive assertion of which one is correct. --Lquilter (talk
  ) 18:41, 22 December 2007 (UTC)

huh? NPOV has nothing to do with it. NPOV is the fair representation of all sources that say different things about the same subject. Not applicable here. -- Fullstop (talk) 20:39, 22 December 2007 (UTC)

"John says A. Bob says not-A. Therefore, John is wrong." is taking the side of John in Wikipedia. NPOV explanation says "None of the views should be given undue weight or asserted as being judged as "the truth"". "Therefore, John is wrong" is an assertion that Bob's statement is the truth; hence, a straightforward violation of NPOV. One could characterize this as "OR" but it's a bit of a stretch; what's the research? It's not a synthesis, just a preference of one side over the other. If I'm missing something, please clarify. --Lquilter (talk) 21:33, 22 December 2007 (UTC)

oh, that (Someguy1221's) example. The indenting led me to suppose the "examples" you were referring to were those I had just previously provided. But that second example doesn't have enough material to call it NPOV either; after all, the editor is providing both sources. It is however still OR because the "is wrong" is a conclusion that the editor is making himself: "saying not-A" isn't the same thing as "A is false." -- Fullstop (talk) 21:44, 22 December 2007 (UTC)

Sorry about the confusion; I didn't want to interject in someguy's thread. I think we're going to have to agree to disagree on this one, though; describing "is wrong" as a (research) conclusion seems less intuitive than describing it as an opinion. Luckily I don't think it really matters -- we can putter about with it on our talk pages if we care to. --Lquilter (talk) 21:59, 22 December 2007 (UTC)

(Outdent) I think an effort to clarify and better explain the

Here's another example of synthesis which to me is the same idea as the plagiarism one, but easier to understand. Suppose source A says, "Any bird can easily escape from a leopard by flying away." and source B says, "An ostrich is definitely a bird." and the Wikipedia article says, "Any bird can easily escape from a leopard, and an ostrich is a bird." That's OR synthesis. It's not stating that an ostrich can easily escape from a leopard. It's implying it. But it's still not allowed by the NOR policy. --Coppertwig (talk) 23:03, 25 December 2007 (UTC)

It's original synthesis by juxtaposition of facts. It's usually part of subtle attempt "the first kind." It's also usually very easy to deal with as one of the facts is almost invariably irrelevent to the sujbect of the article. Noticing completely irrelevent facts tends to be a good way of spotting attempts at leading the reader to an original conclusion. Someguy1221 (talk) 23:52, 25 December 2007 (UTC)

I agree that an implied synthesis is not only possible and often happens. For example, if source X asserts A, source Y says B implies C, and source Z says C implies not-A. Putting sources Y and Z together to contradict source A is an implied synthesis even if "therefore not-A" is not explicitly stated. However, there's a grey area. As long as one isn't constructing an argument, it's OK to string statements from different sources together, and every Wikipedia article does this. Sometimes it's not necessarily clear whether an implied argument is being constructed or not. Best, --Shirahadasha (talk) 06:08, 26 December 2007 (UTC)

The current version of

Dreadstar †

18:44, 26 December 2007 (UTC)

• secondary sources: in liturature an article or essay that analyzes or interprets one passage in a book or connects two or more passage to describe a theme within the book would be considered a secondary source. Within a religious tradition, there are many, many such sources. However, these interpretative sources over time become the subject of yet further interpretation and integrative work. This happens internal to the tradition and, in the last 200 hundred years, it has also happened within academia. To further complicate the issue, liberal streams of both Judaism and Christianity reincorporate the academic tradition building upon it yet another layer of interpretation upon interpretation. Example:
  • bible (primary source)
  • midrash - interpretive material from the talmudic period (approx 200-700CE)
  • rashi - medieval commentator using midrash to read the bible. secondary source? - most Jews would say yes. Functionally Rashi is used this way and many religious readers are not always aware of his relience on midrash. Others view him this way because they feel that midrash isn't really interpretation - it is filling in the blanks and is as much God's word as what was written in the text itself. Still others (mostly academic) observe that rashi used midrash but artfully selected material to reflect his own interpretative beliefs. Tertiery source? - many non-orthodox and academics would say yes to this as well)
  • modern academia - many studies and disputes about what rashi really means (secondary source on rashi? tertiery source on bible?)
  • reform/conservative Judaism bible commentaries - interpretation of a biblical passage using rashi seen through the eyes of one or more academic sources. Secondary source? arguably since this is a direct commentary on the bible passage. Teriery soruce? arguably because the interpretation generally reflects a survey of all prior commentary and academic analysis and reads medieval commentators such as rashi through the lens of integrative and analytical academic studies of the same.

So what here is a secondary or teriery source? Sorry to muddy the waters further, but I'd like to see the religion articles improved and a clear policy on sourcing them would help immensely. Egfrank (talk) 06:40, 16 December 2007 (UTC)

My own sense here is that primary sources can be used to illustrate a point made by an integrative or analytic source but that any claim that is made directly from a primary religious text source is OR. In some cases even a claim based on an integrative source may also be OR, if there is a history of disputes about how to interpret that integrative source. Egfrank (talk) 06:48, 16 December 2007 (UTC)

I would contend that Rashi is an original source, insofar as he is presenting his original insights into the Midrash and the Bible. I would even contend that Onkulus (the Aramaic translation of the Hebrew biblical text) is an original source whenever its translation elucidates the meaning of a passage. In other words, religious dialectic is the mapping of a text to a moral or religious problem, and any effort to do that is original. Even if you are merely repeating what some previous interpreter has said, you are in some way adding your own understanding through language and context.

Writing about religion boldly clarifies the central paradox of this whole discussion, which is - there cannot be encyclopedic writing which is not original. This is the Wittgenstein paradox: an encyclopedia article is creating a map of existing knowledge. But such a map necessarily requires a key - a way to map the map back to the original subject. And that key is always original.

Which is why I believe that this whole discussion is doomed to failure. In practice, we all know what we want not to be in the encyclopedia: people publishing their own experiments of mice running around in mazes, or people writing articles proving that they are the latest true prophet. Our efforts to reduce that intuitive understanding to a set of formal definitions is bound, for strictly logical reasons, to fail. --Ravpapa (talk) 09:00, 16 December 2007 (UTC)

I would Rashi is a primary source, as a historical source of religious writings. To make Judeo-Christian parallels, the writings of the Apostolic Fathers and the Hadith are also primary sources. To make respective parallels with more distance of time, the Church Fathers and historical commentaries on ijma are also primary sources. These writings are not only primary sources of the historical interpretation of Judeo-Christian religions, but indeed primary sources regarding the religions themselves. This is especially true since the writings mentioned are central material in the modern orthodoxy of those faiths. Vassyana (talk) 12:59, 16 December 2007 (UTC)

Rashi has been very thoroughly peer-reviewed and found almost always a notable authority within the field he represents and cited in almost all contemporary works in the field, and this makes a big difference. The commentary in the very recent and well-known Schottenstein Edition of the Talmud, for example, is based mostly on Rashi. Given this thorough peer review, why should we treat Rashi differently from the contents of other kinds of peer-reviewed research? If we disallowed all works in which authors provide "original insights", what sources would be left? Many fields have abstruse jargon and disagreements about what people meant. Best, --Shirahadasha (talk) 00:57, 23 December 2007 (UTC)

I find myself in thorough agreement with

viewpoint neutrality, Wikipedia necessarily made reliability relative as well. Best, --Shirahadasha (talk

) 01:03, 23 December 2007 (UTC)

I honestly have no idea how my comments relate to NPOV in such a manner. Could you please elaborate? Vassyana (talk) 21:41, 27 December 2007 (UTC)

I see serious NOR (and NPOV) problems here: http://en.wikipedia.org/wiki/Talk:Adamic_language but two editors keep reverting my deletions, ignoring my argument, and even deleted my RfC. I´d appreciate it if people who care about this policy would check out this case and tell me if I am off-base. Thanks, Slrubenstein | Talk 22:27, 26 December 2007 (UTC)

I noticed you called for an RfC categorized solely in Language and Linguistics. It might be useful to add an RfC in the sociology and religion category since the article subject appears to represent an intersection of religious and linguistic concepts and viewpoints. Best, --Shirahadasha (talk) 22:32, 26 December 2007 (UTC)

Unpublished synthesis of published material

I am proposing that we add a new lead paragraph to

Material can often be put together in a way that constitutes original research even if its individual elements have been published by reliable sources. Synthesizing material occurs when an editor tries to demonstrate the validity of his or her own conclusions by citing sources that when put together serve to advance the editor's position. If the sources cited do not reach the same conclusion, or if the sources cited are not directly related to the topic of the article, then the editor is engaged in original research. Summarizing source material without changing its meaning is not synthesis - it is good editing. Best practice is to write Wikipedia articles by taking claims made by different reliable sources about a subject and putting those claims in our own words on an article page with each claim attributed to the source that makes the claim.

This would be followed by the current lead paragraph:

Editors often make the mistake of thinking that if A is published by a reliable source, and B is published by a reliable source, then A and B can be joined together in an article to advance position C. However, this would be an example of a new synthesis of published material serving to advance a position, and as such it would constitute original research. (Mr. Wales disapproves of synthesized historical theories and states: "Some who completely understand why Wikipedia ought not create novel theories of physics by citing the results of experiments and so on and synthesizing them into something new, may fail to see how the same thing applies to history." (Wales, Jimmy. "Original research", December 6, 2004)) "A and B, therefore C" is acceptable only if a reliable source has published this argument in relation to the topic of the article.

Brimba (talk) 06:18, 28 December 2007 (UTC)

I'd support that. FYI, in an early version of WP:ATT, we had:

"Material can often be put together in a way that constitutes original research even if its constituent parts have been published by reliable sources. If you have reliable sources for the edits you want to make, be careful that you're not analysing the material in a way that produces a new idea or argument of your own. Just because A is published by a reliable source, and B is published by a reliable source, doesn't necessarily mean that A and B can be joined in order to advance position C."

SlimVirgin ^{(talk)(contribs)} 06:26, 28 December 2007 (UTC)

I have no objection to this basic approach. It's quite consistent with the "spirit" of WP:NOR and also is consistent with a primary-secondary-source analysis where editors need to debate whether something is original research as differentiated from simply expressing previously published concepts in one's own words. ... Kenosis (talk) 05:36, 29 December 2007 (UTC)

I like it, it's very clear and helps introduce what we're trying to say in WP:SYNTH. Once again, Brimba..nice work! Do we need to specify that it is the editor's position or just leave it as advancing a position, making it somewhat more neutral?

Dreadstar †

05:44, 29 December 2007 (UTC)

I agree. This is a very good proposed addition. It provides a solid clarity to what is being discussed in the section. I also agree with Dreadstar's suggestion for the more "neutral" wording. Vassyana (talk) 05:57, 29 December 2007 (UTC)

The only caveat I'd want to offer is to suggest that the language should perhaps not include such phrases as "editors often make the mistake of... [fill in your own perception of 'mistakes often made']". Possibly some (please pardon this word in this context) "synthesis" of Brimba's proposed language and the earlier proposed language from the

WP:A experiment might be useful in expressing the principle. ... Kenosis (talk

) 06:01, 29 December 2007 (UTC)

Go Brimba! Go! ≈ jossi ≈ (talk) 16:12, 29 December 2007 (UTC)

Added in the paragraph with the “the editor's” changed to just “a”. Got to run. (still open to improvements.) Thanks, Brimba (talk) 16:47, 29 December 2007 (UTC)

Certain aspects of this new lead bother me. First the phrase "an editor tries to demonstrate the validity of his or her own conclusions", because this implies that original research is something done by a Wikipedia editor; in other parts of the policy, any unpublished original thought is original research, whether it was the editor who did the research, or it is something the editor heard about through unpublished channels.

Another problem is the phrase "with each claim attributed to the source that makes the claim". This is only necessary for disputed claims; for material where virtually all the sources agree, it suffices to list the sources in the "References" section. --Gerry Ashton (talk) 17:25, 29 December 2007 (UTC)

Regarding point #1, perhaps we could add the word "often", making it: "Synthesizing material often occurs when an editor tries to demonstrate the validity of his or her own conclusions.. From my own experience, editors trying to validate a pov they agree with or believe to be true almost always do this very thing. This section is specifically about using published sources in a way that creates new research; unpublished original thought is covered by other elements of the policy.

Dreadstar †

17:38, 29 December 2007 (UTC)

I also support this approach. I believe the language as proposed helps explain the policy more clearly than current wording and would support it as is. I also agree further improvements in details are possible. --Shirahadasha (talk) 03:15, 30 December 2007 (UTC)

Concerning Gerry Ashton’s point #1; if someone is combining what are clearly reliably sourced statements to bolster some claim that they heard but that did not originate with themselves, then it would not technically be their own claim. I can see your point; however, the editor is still the one deciding that this particular A combined with that particular B equals the C that they heard on the street. They are still the ones concluding that the sources are appropriate to the claims made, not the guy on the street who first thought it up. Really I think if the claim is not supported by the sources, and the claim did not originate with a Wikipedia editor, then its not OR, and the problem with sourcing falls back on V, not NOR.

Point #2, yes a slight rewording is in order. Everything must be verifiable, but not every claim must be directly attributed, only those that are challenged. The current wording is not as clear on that point as it should be. Good call.

Concerning Dreadstar’s idea of adding the word “often”, I think I have negated the need for that in my answer above.

Anyway that’s my thoughts on the subject. Brimba (talk) 05:29, 30 December 2007 (UTC)

I just noticed the suggestion got changed to: "Synthesizing material occurs when an editor tries to demonstrate the validity of his or her own conclusions by citing sources that when put together serve to advance a position."

That's not really correct. All material will tend to advance a position. What matters is that it shouldn't advance the editor's position i.e. a position not in any of the sources. SlimVirgin ^{(talk)(contribs)} 06:27, 30 December 2007 (UTC)

Actually it's probably okay to leave it. I just think that in a few months time someone is going to arrive objecting, but perhaps we should wait until they do. :-) SlimVirgin ^{(talk)(contribs)} 06:29, 30 December 2007 (UTC)

I have changed with each claim attributed to the source that makes the claim to with each claim being attributable to the sources. Hope that fixes the problem. Brimba (talk) 06:41, 30 December 2007 (UTC)

It's good, Brimba. Thanks for writing it. SlimVirgin ^{(talk)(contribs)} 06:50, 30 December 2007 (UTC)

Much better, Brimba. Thanks!

Dreadstar †

06:53, 30 December 2007 (UTC)

Just notifying about a parallel discussion: Wikipedia talk:Verifiability#Primary sources --Francis Schonken (talk) 17:47, 1 January 2008 (UTC)

Crum375 and SlimVirgin have made major undiscussed changes to this policy.[5][6] While they have not raised their significant changes on this talk page,[7] they demand that others discuss their changes to build consensus and deign to declare consensus outside of discussion.[8][9][10][11] The only discussion of these edits was someone expressing concern over these "bold" changes (see

If these major undiscussed changes are not substantiated and discussed, I will undo them.

On the other hand, the replacement for secondary sources was discussed (see

If these well-discussed changes are not reasonably and substantively opposed, I will reinstate them.

Thoughts? Comments? Vassyana (talk) 14:25, 18 December 2007 (UTC)

Regarding the first diff. I'm really not thrilled with either of the first sets of red highlights, though I think your's is slightly better. The second set of red highlights however, yours is clearly better and reflects fairly accurately what at least 20 different 'participants' on this page have expressed numerous times over the last 4 months.

Regarding the second diff, I think Kenosis' version is more clear on the highligted sections, as a few of the other discussions above seem to be addressing as well (interpret that as others are also having problems with the revert). wbfergus ^Talk 14:40, 18 December 2007 (UTC)

Either way, the notions that something has to be preserved because "it has 'consensus'" or because "it has been that way since X" have both been demonstrated to be false. -- Fullstop (talk) 16:27, 18 December 2007 (UTC)

Has anyone thought of making a sandbox copy of the policy page and doing all the editing on the copy? Agreeing a few simple ground rules would make this a much less contentious way of working. The rules might be:

• agree that all edits are to be made to the sandbox page until consensus is reached that a change can be copied to the live page.
• Once a change is made to the live page there is no reverting of it. All reversions and further edits must be done on the sandbox page until consensus is reached again.
• If an "outsider" edited or reverted the live page, his edit would stay temporarily, but he would be very strongly recommended to "join the party" and quickly reach consensus on the fate of his edit. No reasonable editor could refuse such a request.
• All discussion to be on the policy talk page, not the sandbox one.
• 3RR applies, like everywhere else.

A prominent html comment in the policy page would help ensure compliance. Hope this helps,  —SMALLJIM  16:31, 18 December 2007 (UTC)

(e/c) I've said this many times: I wish everybody would leave that mainspace alone until there is consensus. (What if all of us were this incautious about editing or reverting the page?) But before there can be consensus, we have to determine what consensus looks like so we'll know when we have it - and there have been strong disagreements expressed here about what consensus even means. But I do think that the argument that editing main space is justified as long as it's accompanied by volumes of repetitive discussion on the talk page is a bit unrealistic. Engaging in discussion is not the golden ticket which gives you the right to make mainspace policy changes--community consensus justifies policy changes. And there is no consensus evident on this page. If it does exist, it isn't evident to me. All the intervening

mainspace edits, and their reverts, are just provoking hostilities. Professor marginalia (talk

) 16:43, 18 December 2007 (UTC)

(ec)I would generally agree except to say that community advertisement and discussion are the best means we have to reach and measure consensus (and they are widely touted as the best routes to building and identifying consensus). The other common measures, such as common good practice or community will, are matters identified through community discussion. In the absence of discussion, it's nothing more than a cacophony of opposing editors each claiming their "consensus" is the "true" consensus. Additionally, discussion is the means that we employ to identify when community consensus has changed. Of course, this discussion does not always take place on policy pages, as some of those changes are noted in "live" practice. For example, there have been occasions when policy has changed due to very active and broad noticeboard discussions or similarly broad and active discussions in response to ArbCom cases and actions. Regardless, discussion and community exposure are the fundamental elements of consensus building and identifying consensus. Vassyana (talk) 17:00, 18 December 2007 (UTC)

Junk those edits. Per comments by SmallJim, PM and for formal reasons (violations of WP:POLICY "Editors should be careful that any changes they make to a policy page reflects consensus") -- Fullstop (talk) 16:58, 18 December 2007 (UTC)

The silence, both in the justification of the undiscussed edits and in opposition to the secondary sources replacement, is deafening. Vassyana (talk) 17:59, 19 December 2007 (UTC)

Yes, the silence speaks volumes. With no voice of dissent, we quite obviously have consensus. Take a quick straw poll (only yea/nay, no comments) and we're done. -- Fullstop (talk) 01:39, 20 December 2007 (UTC)

No, it does not. SlimVirgin and Crum do not seem very active recently, I would suggest to wait a few days. ≈ jossi ≈ (talk) 01:48, 20 December 2007 (UTC)

SlimVirgin was plenty active enough to make multiple comments to this page ... just none putting her undiscussed changes up for discussion and none addressing this issue. I will say that a straw poll isn't going to be helpful. Some justification for unjustified edits and explanation of opposition would be very helpful though. Vassyana (talk) 03:23, 20 December 2007 (UTC)

Oppose. I oppose whatever it is Vassyana and Fullstop think they have consensus for. This section is gibberish. I have no idea what change you are proposing. --Gerry Ashton (talk) 02:29, 20 December 2007 (UTC)

If you have no idea what's being discussed, I'd recommend you look over the post that opens this section. Vassyana (talk) 03:20, 20 December 2007 (UTC)

I looked at it again. It's just too much work to try and figure out what it is you propose. --Gerry Ashton (talk) 19:16, 20 December 2007 (UTC)

There's not a whole lot to figure out. There were undiscussed edits (outside of a single objection). If they're not discussed, they should be undone. There were discussed edits (some discussion linked) reverted without discussion. Unless the objections are substantiated, the discussed edits should be reinstated. I really don't understand what there is to "figure out" about that. Vassyana (talk) 14:37, 21 December 2007 (UTC)

As I understand it, on the first two links show the two undiscussed changes, merely showing the diffs. All the other links are for various places on the talk page highlighting the lack of discussion for those changes, while advocating that changes be discussed. Fairly simple, only a couple of mouse clicks to see that. It's also fairly simple to see above that it is proposed to re-revert the undiscussed changes to the original form. If you can find where those changes were discussed, so they shouldn't be re-reverted, then please kindly provide them and this should then be a dead issue, if there was agreement to implement them. wbfergus ^Talk 15:02, 21 December 2007 (UTC)

Nine days without a substantiation of the undiscussed edits or the reversion of the discussed edits. I will wait another 24 hours to allow for a full ten days, but barring any substantive comments and discussion, I will revert the undiscussed edits and reinstate the discussed edits. Vassyana (talk) 21:38, 27 December 2007 (UTC)

Five additional days have passed. Since the undiscussed edits have not been justified and the removal of discussed edits has not been substantiated, I will be reverting the undiscussed changes and reinstating the discussed changes. Vassyana (talk) 06:19, 2 January 2008 (UTC)

Yes this looks better now. I have further trimmed some of the mealey-mouthed language to make it more understandable (without changing the intent). Dhaluza (talk) 11:09, 2 January 2008 (UTC)

Excellent rewording! Thanks. Vassyana (talk) 13:06, 2 January 2008 (UTC)

Alrighty, what are issues regarding these changes to nutshell?

Even if the old nutshell was accurate, it’s trying to say too much. Some people will never read anything more then the nutshell; in such cases stating things in simple clear terms is better then trying to give a more rounded view (good intentions lead simply to wordiness). Of course that is just my opinion, but its well grounded in the concept of usability. Besides if the language is clear, they might be more likely to take the time and read the lead, which is where the broad overview should be. Brimba (talk) 06:17, 30 December 2007 (UTC)

As is clear by now, people grant a lot of authority to the nutshell, which ought to be more stable than the general text of the article. The older nutshell is superior because it makes clear the relationship with other content policies and is clearer about what the policy requires. The other lines of the nutshell are not incorrect. Brimba's point seems to be that NOR only applies when something is challenged. In practice, that may be true - policies only come into play when someone accuses another of not following a policy. So what? our policies should make clear what our ideal is. But if this is Brimba's only problem, it is easy to change attributed to attributable. This policy has been under assault by a variety of people including those who claimed to support the polciy but took issue with the language. When this nutshell was edited, people on all sides, people who were adamant about including the PST distinction and those opposed, people who felt the language of the policy was fine and those who didn't, all agreed on this version. Why reject it, wholesale? At least, let's have more discussion rather than one person vetoing an agreement reached by many diffeerent editors. Slrubenstein | Talk 19:21, 30 December 2007 (UTC)

I've trimmed the first bullet, as it was redundant with the two points below, and the

WP:V policy. I think this is better than the previous formulation above because it does not redundantly repeat the title either. Dhaluza (talk

) 21:59, 30 December 2007 (UTC)

Let me start off by stating the very obvious: The statement “Facts must be backed by citations” is NOT POLICY; not on this page nor on any other, nor has it been in the past.

Your statement that “The older nutshell is superior because it… is clearer about what the policy requires.” fails. The problem is that the statement “Facts must be backed by citations” does not represent anything NOR requires. So why is it there? When I pointed this out, you response was “So what? our policies should make clear what our ideal is.” If it’s not policy, and its NOT, then your statement is a misstatement.

“When this nutshell was edited, people on all sides, people who were adamant about including the PST distinction and those opposed, people who felt the language of the policy was fine and those who didn't, all agreed on this version.” Where is this discussion? I could not locate it in the archives. Please feel free to point me in right direction; I would like to see it. “all agreed on this version” of the nutshell? That’s quite a strong consensus; I would hate to go against it.

“agreement reached by many diffeerent editors.” Again, where’s the discussion on the nutshell?

Policy is formulated on the page, not within the nutshell. Doing so is a case of the tale wagging the dog. The nutshells job is to explain the policy in a concise manner. And it is not to “makes clear the relationship with other content policies”. It is to explain the policy of the page; end of story, there’s no secondary purpose to a nutshell.

“the nutshell…ought to be more stable than the general text of the article.” I might point out that your version only existed for 96 days. It in turn replaced a version that had been stable for over 7 months, and most of the wording of that version had existed for far longer. The only explanation surrounding any part of the change was this statement “I have been trying to edit the intro section for clarity and brevity - I also changed the "nutshell" based on a comment by Tim.” Not sure who Tim is, or what the comment was.

The nutshell you replaced was:

This page in a nutshell: Wikipedia is reliable sources without further analysis. Content should not be synthesized to advance a position.

I am not sure what you found problematic about that wording? Or why it was changed? I would not object to restoring this version (meaning the one that existed for 7 months) if you have no objections to doing that.

Brimba (talk) 05:15, 31 December 2007 (UTC)

I agree, the original version is better than the current one:

This page in a nutshell: Wikipedia is reliable sources that contain these facts Interpretations and syntheses must be attributable to reliable sources that make these interpretations and syntheses

I was simply using the exiting version as a starting point, and not looking at prior history. The old version does not have the overlap with NPOV in the first bullet, and retains the important point in the third that synthesis to advance a position is the problem (not synthesis to make a coherent encyclopedia article from multiple sources). I have no objection to restoring the previous version. Dhaluza (talk) 02:58, 3 January 2008 (UTC)

Gerry, read the reference -- a secondary source is one step removed. "One step removed" in English means one step, or two, or 50, or 5,000. "Secondhand" in English means secondhand, third hand, fourth hand, 50,000th hand. SlimVirgin ^{(talk)(contribs)} 19:46, 2 January 2008 (UTC)

I reverted a change to the policy that indicated that secondary sources are secondhand accounts. The first hand is the participant or eyewitness. The second hand is the publisher. Most eyewitness or participant accounts are secondhand by the time a Wikipedia editor sees it, even though the account is a primary source. --Gerry Ashton (talk) 19:47, 2 January 2008 (UTC)

Gerry, I'm sorry, but you're just making this up as you go along. The eyewitness is the first hand and the publisher the second? :-( SlimVirgin ^{(talk)(contribs)} 19:51, 2 January 2008 (UTC)

I disagree. From my point of view as a Wikipedia editor, firsthand information is information about things I witnessed or participated in (i.e. I've operated a fire engine and can therefore spot certain bogus information in the Fire engine article. Secondhand information is accounts communicated directly to me by a witness or participant (my uncle told me that in the Western Theater in World War II, some days the soldiers were ordered not to take prisoners). Everything I read that was not written by a participant or witness is thirdhand information. --Gerry Ashton (talk) 19:53, 2 January 2008 (UTC)

You are seeing things only from your own point of view i.e. as if you were the primary source. Your uncle's information is firsthand (primary source) because he was there.

Look Gerry, the point is that words in this policy have to mean what they mean everywhere else in the English-speaking world. The definition of the terms has to be the definition used by universities and professional researchers and publishers. The policy should be written clearly and succinctly -- should be well written, not confusing, not tortured, not ill-considered, not implying things we didn't mean to imply, not language gone on holiday. I'm sorry if I'd being short, but I'm still recovering from an attempt to copy edit Wikipedia:Evaluating sources, the confusion of which some people are trying to import into this policy. SlimVirgin ^{(talk)(contribs)} 20:10, 2 January 2008 (UTC)

There is no consensus for the following, and it has never been in the policy before to the best of my knowledge (it's also false, if anyone cares, that secondary sources are ipso facto more likely to be biased than primary sources):

"Secondary sources ... can be biased by the views of their author(s) and/or publisher, so caution is required to preserve the neutral point of view."

SlimVirgin ^{(talk)(contribs)} 19:50, 2 January 2008 (UTC)

It has been in the policy until you removed it at least twice (see diffs posted above); so the best of your knowledge is not that good. It really strains your credibility when you repeatedly make demonstably false assertions. This is a wiki, and anyone can check the historical facts with a little digging. I suggest you do some checking before relying on your imperfect memory to back your historical arguments. Your assertion that this statement says secondary sources are more likely to be biased is also not supported by the quote you are citing, so ironically you are engaging in a form of OR. Dhaluza (talk) 13:20, 3 January 2008 (UTC)

Please say which diffs rather than vaguely referring me to something posted above, and quit the digs -- I've really had enough of it. If I did remove it, it may have been because it was added without consensus, so you'd need to do more than provide a diff showing removal. The very existence of the sentence does imply that secondary sources are more likely to be biased, because no mention of bias is made in relation to primary sources, so why mention it here? But the more important point is that the issue of source bias is separate from the issue of OR, and this is the NOR policy. SlimVirgin ^{(talk)(contribs)} 16:10, 3 January 2008 (UTC)

Let me add that the only possible firsthand information about Wikipedia or any other publication is the physical attributes of the publication, such as "Wikipedia is responding fast today" or "My American Heritage Dictionary weighs 6 pounds". For practical purposes, the only firsthand sources that could be used by a Wikipedia editor are sources written by the editor him/herself. --Gerry Ashton (talk) 20:03, 2 January 2008 (UTC)

You have misunderstood what a primary source is, in that case. SlimVirgin ^{(talk)(contribs)} 20:10, 2 January 2008 (UTC)

Gerry, perhaps we could write the policy from your perspective by having firsty sources, secondary, thirdy, fourthy, and so on. Would that help, do you think? SlimVirgin ^{(talk)(contribs)} 20:13, 2 January 2008 (UTC)

As usual, this policy does not use terminology consistent with the viewpoint of the reader of the policy, a Wikipedia editor. Primary, secondary, and tertiary are not directly related to firsthand, secondhand, and thirdhand. Primary, secondary, and tertiary are properties of sources; a secondary source will remain a secondary source until it becomes so old it becomes a historical source. Firsthand, secondhand, and thirdhand describe how a piece of information came to be known by a particular person. This policy should either avoid the word secondhand altogether, or be written from the point of view of a Wikipedia editor who has aquired a piece of information and is thinking of puting it in an article. --Gerry Ashton (talk) 20:20, 2 January 2008 (UTC)

"[A] secondary source will remain a secondary source until it becomes so old it becomes a historical source." Could you explain the difference between a secondary source and an "historical source"? SlimVirgin ^{(talk)(contribs)} 20:46, 2 January 2008 (UTC)

Some argue that sources that would have qualified as secondary sources at the time they were written (

Historia ecclesiastica gentis Anglorum for instance) should now be treated as primary sources because they are outdated, and while they may not have seemed close to the events related when originally published, they seem close from a modern perspective. I don't take a position about whether this is a sound definition of secondary source or not. --Gerry Ashton (talk

) 21:04, 2 January 2008 (UTC)

Yes, that's correct. A secondary source may become a primary source with age. You confused me by calling it an "historical source." SlimVirgin (talk)(contribs) 22:44, 2 January 2008 (UTC)

I have removed the word secondhand from the project page. Note that the source cited for that passage, http://www.lib.berkeley.edu/instruct/guides/primarysources.html#secondary does not use the word secondhand.

1. there is no "more" or "less" of anything that can be attributed to any characterization of "primary" or "secondary." There is *no relationship* whatsoever between source-type and bias (or reliability or ease/likelihood of misuse etc). Typing has no influence on the ability to misuse/misrepresent a source.
2. A dog is a dog, and regardless of whether it is a poodle or a great dane, any dog can be maltreated. Defining (and arguing about the definition of) the type of dog is not an effective way to inhibit maltreatment. It does not matter whether 'poodle' is a "simple and useful concept."
3. Similarly, the insistence that PSTS is the only (?) way to prevent misuse of sources is nauseatingly un-constructive. The distinction of primary/secondary is not the magic bullet.

-- Fullstop (talk) 20:15, 2 January 2008 (UTC)

Exactly. In 4 months of bickering back and forth, I haven't seen one single (real) example presented of how a primary source, when used correctly, injects OR into an article more than the same conditions based upon a secondary source. Once it is published in a reliable and verifiable source it is no longer original, regardless of the source type. So what if the source type is primary? If it is quoted, cited and used correctly how is that more prone to OR than something from a secondary source that used incorrectly, and how does that make the statements based upon the secondary source inherently more reliable? wbfergus ^Talk 20:47, 2 January 2008 (UTC)

It has been explained a thousand times. You can't have been reading very carefully, with respect. SlimVirgin ^{(talk)(contribs)} 22:42, 2 January 2008 (UTC)

Slim, could you please point me to one or two of these explanations? I seem to have missed them. Blueboar (talk) 01:50, 3 January 2008 (UTC)

I must have missed them somehow also. I've been browsing through the archives and can't seem to find any concrete examples. All I've been able to find are just some generalized statements that it happens, but no examples. Any help in pointing out these examples would be appreciated. wbfergus ^Talk 17:15, 3 January 2008 (UTC)

It looks like secondhand, "1a: received from or through an intermediary;" covers not only " secondary, thirdy, fourth, etc" sources but any content that is "not written by a participant or witness"..thirdhand is still seocondhand information - it appears that thirdhand is pretty specific, but secondhand is more general. If I buy a secondhand car or other merchandise, I don't take it for granted that it only had one previous owner. So, generally speaking, it looks like anything that is not firsthand is secondhand.. And, um, no revert warring please...

I propose to work out some wording that addresses the issues of PS and SS as it pertains to their use in controversial, disputed, and/or contentious subjects. ≈ jossi ≈ (talk) 16

10, 3 January 2008 (UTC)

I look forward to Jossi's wording... however, I am concerned that it will be yet another addition that has no relation to the concept of "NO ORIGINAL RESEARCH"... Jossi, please make sure that anything proposed directly relates to this concept. If your intention is to say that OR is acceptable if it is non-contentious, I very much disagree. To my mind OR is bad no matter how contentious, and no matter what kind of source it cites to. Blueboar (talk) 17:04, 3 January 2008 (UTC)

I agree that we need to keep the page on-topic. Any discussion of bias, for example, belongs on NPOV. SlimVirgin ^{(talk)(contribs)} 17:19, 3 January 2008 (UTC)

I have no objection to having special provisions for "controversial, disputed, and/or contentious" subjects, but not articles. The point is that the dispute should be reflected in the sources, not just the article talk page and edit history. An internal Wikipedia dispute is a separate problem. We can discuss whether this belongs in this policy, another existing policy, or a new stand-alone policy based on the content. If it is directly related to the fundamental definition of original research, then it belongs here. But we should not expand the concept of OR just so we can attach important provisions to this policy as a rider. Dhaluza (talk) 01:09, 4 January 2008 (UTC)

Where does the idea come from that this edit was discussed and agreed? Bits of it are clearly not correct e.g. the implication that secondary sources are more likely to be biased than primary sources. If anything, it's the other way around, because the hope is that secondary sources will have looked at multiple primary and other secondary sources. That's why WP articles ought to rely on them, and it's why they're needed when edits based on primary sources are challenged. SlimVirgin ^{(talk)(contribs)} 11:56, 2 January 2008 (UTC)

(moved from SV talk) Your reversion was undone.[12] Undiscussed changes were reverted. Discussed changes were implemented. You were given ample time to comment and join in the discussion. Your reversion of discussed changes and undiscussed changes must be subject to discussion and consensus, just as you advise others. Please see Wikipedia talk:No original research#Discussed and undiscussed edits. Vassyana (talk) 12:22, 2 January 2008 (UTC)

(ec)I left a message on your talk page. For the benefit of others coming here, please see the very first section of this talk page. Vassyana (talk) 12:26, 2 January 2008 (UTC)

Vassyana, there were objections from good editors e.g. "I oppose whatever it is Vassyana and Fullstop think they have consensus for. This section is gibberish. I have no idea what change you are proposing." --Gerry Ashton (talk) 02:29, 20 December 2007 (UTC)

There were real problems with some of the changes, and no clear benefits. For example, this is just a minor point, but why would you want to remove that Wikipedia is a tertiary source? It's that kind of analogy that helps people to understand the distinctions. SlimVirgin ^{(talk)(contribs)} 12:28, 2 January 2008 (UTC)

Please elaborate on your issues with the changes. The changes you made were reverted because they were undiscussed except for someone raising a concern about them. I do not believe anyone would have an issue with the Wikipedia = tertiary statement (though it is a bit superfluous), so I will restore that. Vassyana (talk) 12:43, 2 January 2008 (UTC)

Gerry's objection was that he didn't understand what was being discussed (myself and another editor provided clarification). Gerry's previous contribution to the discussion was to strongly reject another editor's suggested revisions to the proposed replacement. (I wanted to make sure I checked the archives to verify my memory before responding about that.) Vassyana (talk) 12:52, 2 January 2008 (UTC)

It's also a question of clarity. The current version starts with:

• "Secondary sources are secondhand accounts, at least one step removed from an event."

You want to replace that with:

• "Secondary sources are references that draw on research and other references to make interpretive, analytical, or synthesized claims, or create a general overview."

Which version is more succinct, and which do you think most editors will understand more easily? SlimVirgin ^{(talk)(contribs)} 12:33, 2 January 2008 (UTC)

I believe it is a question of which is more accurate, essentially. Also, many editors felt the replacement was an improvement over the previous section. Vassyana (talk) 12:43, 2 January 2008 (UTC)

But many thought it wasn't. And your version is not more accurate. What does this mean, for example -- "However, they [secondary sources] can be biased by the views of their author(s) and/or publisher, so caution is required to preserve the

undue weight

and ensure the information cited is used in context."

Why would secondary sources be particularly biased? Why should care be taken with them in particular to avoid undue weight and ensure material is cited in context? SlimVirgin ^{(talk)(contribs)} 12:51, 2 January 2008 (UTC)

Very few thought it wasn't and even fewer commented at all to that effect. I do not necessarily agree with those sentences myself. They address a repeatedly stated concern from numerous editors over the past several months of discussion here. The language itself was worked out as a compromise between wide-ranging opinions voiced here during that time. It was revised according to the feedback received. Why is the advice a problem? How is it inaccurate? Vassyana (talk) 12:59, 2 January 2008 (UTC)

Because secondary sources don't have to be watched in particular for bias. SlimVirgin ^{(talk)(contribs)} 13:09, 2 January 2008 (UTC)

(ec)What about the many editors who have expressed concern? What about the widely expressed concern about biased sources? I should additionally note that most concerns I've seen people express about biased sources have been in relation to secondary sources. Again, I don't necessarily agree with the statements myself (I personally tend to think primary sources are more likely to be abused for POV), but they are reflective of the comments made here by numerous editors. Though I don't think the statements are necessary, I don't understand how the statements are problematic. I don't think it's inaccurate or harmful to tell editors to watch for bias. It's reflective of a broader view held by more people than my own. I hope that all makes sense. Vassyana (talk) 13:19, 2 January 2008 (UTC)

Please don't keep posting on my talk page. The discussion needs to take place here. SlimVirgin ^{(talk)(contribs)} 12:36, 2 January 2008 (UTC)

It was a message directly addressed to you (and therefore posted to your talk page, as is appropriate), not a general comment for discussion here. Vassyana (talk) 12:43, 2 January 2008 (UTC)

(moved from SV talk) Please do not restore that post to WT:NOR. It was not intended to be posted as part of the general discussion and I've retracted half of its content with an apology. Thank you for your understanding. Vassyana (talk) 13:01, 2 January 2008 (UTC)

Vassyana, you keep posting to my talk page, but when I move the posts here, you remove them. Everyone needs to see all the posts about this, not just me. Please keep the discussion in one place. SlimVirgin ^{(talk)(contribs)} 13:07, 2 January 2008 (UTC)

The discussion at the top of this page shows that the changes were agreed by Fullstop and Vassyana. That's not consensus by any stretch of the imagination. SlimVirgin ^{(talk)(contribs)} 13:10, 2 January 2008 (UTC)

Please read the section again and follow the links, as that is a gross misunderstanding/misrepresentation. The replacement section was discussed and revised by more than two people.

Wikipedia_talk:No_original_research/archive27#Taking_the_dive_.28PSTS_proposal.29

)

Could you please say who actively supported your changes? SlimVirgin ^{(talk)(contribs)} 13:37, 2 January 2008 (UTC)

As a quick sampling: Blueboar, CBM (Carl), COGDEN, Dreadstar, Dhaluza, Fullstop, Jossi, Kenosis, Phil Sandifer, Professor marginalia, SmallJim, SmokeyJoe, Was 4.250 and wbfergus. That's a fairly diverse crowd representative of a wide variety of opinions regarding PSTS. Vassyana (talk) 14:00, 2 January 2008 (UTC)

Blueboar seems to be objecting below; Dreadstar has several times reverted your changes; and I seem to recall (can't be sure) that Jossi has too. Can you show me where all these people agreed with you? SlimVirgin ^{(talk)(contribs)} 14:23, 2 January 2008 (UTC)

Blueboar.[13] Jossi.[14] Jossi's only concern noted and addressed.[15][16] Dreadstar.[17] Also, Dreadstar reverted the changes once, not several times as you claim.[18] Please note that the changes he reverted were the same as those he supported. It was due to an expressed objection by Kenosis,[19] who supported the insertion of the secondary sources section.[20] Kenosis' only concern noted and addressed.[21] Vassyana (talk) 15:02, 2 January 2008 (UTC)

(outdent)Going up near the top of this section, while Vassyana'a version isn't as concise, I think it makes it more understandable to the new or even average user. I also strongly agrre with the additional warning about secondary sources. There are literally thousands of them floating around with an extreme bias. Take practically book on politics, and quite a few books on history. The books published by 'noted historian' Bruce Cumings are another example. His books are extremely biased, if he can claim that North Korean gulags aren't really all that bad, since families are kept together and they get to catch and eat 'small animals' (rats) to enhance their normally meager diet. And this from somebody who is generally praised (usually by communism supporters or anti-America contingents) as a 'noted historian'. Unfortunately though, even if he is a professor at the University of Chicago, his books aren't peer-reviewed through the University, but are usually published through outside publishers (W. W. Norton, New Press, etc.) Hope this helps clarify some things. wbfergus ^Talk 13:49, 2 January 2008 (UTC)

Hi, it's been a busy time of year, and this isn't the only page I edit. In my opinion SV's preferred versions is clearer, and the point about using secondary sources should appear in that section as well as having the brief mention under primary sources. However, I'm willing to try to consider each clause on its merits. Concerns about biased sources and NPOV apply to all sources – one common type of "primary source" is the self-published or other source closely linked to the subject, which will commonly be biased and must be set in the context of secondary source assessment and analysis. If the caveat is required, it should be placed at or near the end of the section, and apply to all kinds of sources. .. dave souza, talk 14:00, 2 January 2008 (UTC)

Once again, it seems that people are forgetting what this Policy page is about... the issue of whether a source is biased has nothing to do with NOR, nor with whether it is primary, secondary, or even tertiary. All three kinds of sources can be biased. As long as the article is accurately reflecting what the source says, however, it isn't OR to discuss what a biased source says. Now, I agree that biased sources are not the best sources to rely on for statements of fact. However, they can be good sources for statements of opinion, and way to deal with a biased source is through attribution. This is more an issue of WP:NPOV and especially

WP:Undue weight

.

Can we please keep our focus on the concept of "No OR", and resist the temptation to let other issues sidetrack this Policy page? Blueboar (talk) 14:16, 2 January 2008 (UTC)

Exactly. Any source can be biased, and arguably almost all sources are biased in some form or another. That in itself has nothing to do with OR. SlimVirgin ^{(talk)(contribs)} 14:25, 2 January 2008 (UTC)

Which why we have had at least 4 months of arguing about PSTS and the current work underway at WP:Evaluating sources to finally get it out of here and somewhere more appropriate. But, in a way the above is related to the NOR policy (as it's currently written anyway). Primary sources can pretty much only be cited verbatim. It's kind of hard to do that and push a OR point. However, with secondary or tertiary sources, the editor is given the freedom to re-word it, etc. In those cases, especially with a source that can be biased, or even challenged as biased (and somehow still clamored as being 'authoritive'), care must be taken to only present material in the same context as originally presented. wbfergus ^Talk 14:31, 2 January 2008 (UTC)

To get what out of here? If you mean the issue of bias, it was Vassyana who just added it. The issue of primary/secondary has nothing to do with bias, but with original research, so it's directly relevant to this policy. SlimVirgin ^{(talk)(contribs)} 14:40, 2 January 2008 (UTC)

To get the whole 'sourcing issue' out of here. It's been argued almost the entire 4 months that primary sources are more prone to misuse than any other source, and that secondary sources are much more accurate and reliable than primary sources. It's good to see though that you agree that any type of source can be biased and inaccurate. But quoting verbatim data from a primary source is a bit more difficult to use for OR, where using a biased secondary source, which doesn't have to be used verbatim, can therefore be twisted into OR more easily. But again, if something is published, it's no longer 'original'. wbfergus ^Talk 14:46, 2 January 2008 (UTC)

I guess that I should also say that it's good to see (from your comment on the policy edit page), that silence does not neccessarily equal concensus. That has been a sore sticking point to many of the 'discussions here over the last 4 months as well, with those who support keeping things as they are (or were) stating vehemently that since nobody openly disagreed with parts of the policy in the past, their silence amounted to tacit approval. It's refreshing to see you say otherwise. Maybe we can finally put those old arguments to rest now. wbfergus ^Talk 14:50, 2 January 2008 (UTC)

Misuse of primary sources

<undent> It's easy to misuse primary sources to produce original research, at its extreme it's called quote mining, and it's hard to deny that's what the primary source says – which is why we need good secondary source analysis or summary to avoid that as much as possible .. dave souza, talk 15:46, 2 January 2008 (UTC)

a) There is no evidence whatsoever that "primary sources" are easier to "produce" original research with.

b) quote mining (extreme example here) is just as "easy" to make with primary sources as it is with secondary (and tertiary) sources.

c) The *only* policy that quote mining can violate is NPOV, i.e use of a quote farm (not necessarily "primary" sources) to overwhelm the opposition.

d) "it's hard to deny that's what the primary source says" is no more/less true than "it's hard to deny that's what the secondary source says."

e) if the editor is simply reiterating the sources, then he/she is not in violation of NOR, regardless of whether the sources being used are primary or not.

f) the primary/secondary distinction has no relationship to the distinction between unreliable/reliable sources.

g) the simplest way to inhibit contentious use of sources is to require these to be substantiated by other sources that adhere to scientific method. For ease of understanding, this may be described by a policy that gives any unsubstantiated opinion/analysis/hypothesis/description etc zero weight.

-- Fullstop (talk) 18:18, 2 January 2008 (UTC)

Thanks for a complex example where there certainly seems to be original research supported by clear misrepresentation of sources. Hope that's getting sorted out to your satisfaction. My own take is that the need for secondary sources for analysis is a useful reminder to myself to look for a reliable outside opinion, even if the direct source looks good, and this came up with the Australopithecus afarensis example I cited a long time ago, where information from a scientific paper was usefully set in context by a good newspaper article about the findings. A subtle one that's still in progress appeared here,[22] where Darwin's words quoted accurately were set in a context that implied that they exemplified "academic racism was pushed by white supremacist Caucasians". The words seem as though they support that, but the context they come from needs knowledgeable assessment.[23] It's a case where an honest reading of the primary source can lead to misunderstandings, and a good secondary source is needed to avoid unintentional original research. I agree with your conclusion, but if we start demanding secondary sources for secondary sources we get into the sort of infinite loop that was the subject of debate earlier. .. dave souza, talk 19:54, 2 January 2008 (UTC)

1. Well, it was only on a fly-by that I initially came in contact with that so-called "article." Its junk, but thankfully also bereft of any import whatsoever. ;)
2. Yes, misreading Darwin (naively or otherwise) can superficially be described to be an example of "easier" misuse of a primary source. In reality however, the cause of the mangling does not lie with the type of source but with the editor's ignorance of Darwin's time, language and social mores. A secondary source that was contemporary to Darwin would be just as "easy" to misrepresent.
3. I'm not expecting secondary sources for secondary sources (note that I did not use the words "primary" and "secondary" at all).
   What I am suggesting is that any contended^* statement should not cite the source that "said it first", but should instead cite a source that refers (not necessarily positively) to the source that made the original statement.
   Such a clause...
   a) would restrict primary sources to only where they can do no harm.
   b) it would effectively establish 'notability' at the statement level.
   c) is actually about original research ("said it first"),
   d) would reinforce the concept of reliability,
   e) would provide a source that potentially lists alternative positions, so supporting NPOV.
   Thus, the purpose behind psts can be achieved without either dicking around with the ps/ss terminology or with their distinction or groping around for substitute terminology.
   The fallacy that "A causes B" is the same thing as "A is a prerequisite for B" is what the insistence of the ps/ss distinction is based on. But even if my suggestion isn't by itself considered worthy of merit, what it demonstrates is that one can arrive at 'B' even without 'A'. There is more than one way to skin that particular cat.

-- Fullstop (talk) 22:06, 2 January 2008 (UTC)

Personal attacks

Vassyana, I have had enough of your insults. You're previously me a bald-faced liar; you revert my edits calling them disruption; and now because I ask who agreed with your changes, you say I'm making wild accusations. [24] Quit it, please. SlimVirgin ^{(talk)(contribs)} 15:35, 2 January 2008 (UTC)

I self-reverted those comments immediately after making them, as clearly shown by that link. It's utterly inappropriate for you to dredge that up when it was immediately self-reverted before any comments were made. Since it was raised, those comments were an admittedly kneejerk response to your blatant implication that I was being dishonest. Vassyana (talk) 15:49, 2 January 2008 (UTC)

I'm not going to let this kind of nonsense go unanswered. Addressing your other two claims of NPA violation: The first is most certainly not a personal attack, but rather an observation of an obvious falsehood (at worst, poorly worded).[25] The second is not a personal attack, and quite notably it was in response to actions of the sort that you have repeatedly and forcefully told others is disruption. Vassyana (talk) 15:57, 2 January 2008 (UTC)

I am telling you again to quit the personal attacks on me. I won't take them from you anymore and every time you do it, I'm going to call you on it. SlimVirgin ^{(talk)(contribs)} 17:42, 2 January 2008 (UTC)

Vassyana, SlimVirgin: I respect you both enormously and find the above exchange quite disturbing. Could you make an effort and stay focused on the discussion at hand instead of dragging this debate in a personal direction? ≈ jossi ≈ (talk) 16:04, 2 January 2008 (UTC)

To address SV's observation that "The discussion at the top of this page shows that the changes were agreed by Fullstop and Vassyana."

That "discussion" is not about a proposition, but about SV's undiscussed changes to the policy page. There were no "changes" that were "agreed" to by Vassyana or myself, but rather the general agreement that no one should be writing undiscussed changes of any kind into policy.

To suggest that that talk section is about anything else is disingenuous and/or symptomatic of an inability to read, exemplified for instance in Gerry's "Oppose" for "I have no idea what change you are proposing" (itself a rather strange response to Vassyana's request for "Thoughts? Comments?").

Either way (and really all I had to say in that section), SV's unilateral modifications to policy have - once and for all - buried the specious supposition that PSTS is cast in stone (i.e. that things need to remain they are just because they have been a certain way for N years).

-- Fullstop (talk) 17:26, 2 January 2008 (UTC)

This policy has never said that secondary sources are more likely to be biased. That is a new thing added by Vassyana without consensus, and even if it had consensus, we shouldn't add it, because it's just plain wrong. Also, the policy needs to be written clearly so that people understand it, and so we look as though we understand it ourselves. The latest edits don't give that impression, just as Wikipedia:Evaluating sources doesn't, and for the same reasons.

SV, I must take you to task for making such a demonstrably false assertion. Here are two diffs [26] and [27] where you previously removed similar language yourself. So for you to say this "is a new thing added by Vassyana without consensus" is simply untrue, unless you mean that if you alone disagree there cannot be consensus. You have also made numerous changes to this policy without consensus, so that is not a moral high ground you have any legitimate claim to.

Your assertion that this says that "secondary sources are more likely to be biased" is also untrue--it simply says that secondary sources can be biased by their authors and publishers, which is completely correct, and an important consideration in using sources wisely, which is what this section is purportedly dealing with. The idea that primary sources are biased and easily misused, but secondary sources are somehow immune to bias or misuse is what's incorrect. Not to mention your further assertion that we should ignore consensus and follow your individual divine wisdom instead. Dhaluza (talk) 02:32, 3 January 2008 (UTC)

It really does seem to me that the people causing a problem over the primary/secondary thing simply don't understand it, because these are very simple and useful concepts, and they can and should be written about simply and with clarity. There's no need for tortured language and inaccurate descriptions. SlimVirgin ^{(talk)(contribs)} 17:42, 2 January 2008 (UTC)

SV... I disagree with your contention that "the people causing a problem over the primary/secondary thing simply don't understand it." But, for the sake of argument, let's say you are correct in that contention. If so, doesn't that tell us that a better explanation is needed? It's not as if there are only one or two people who have issues with the section, after all. I'm all for simplicity and clarity ... that is one of the reasons I don't think PSTS belongs in this particular policy... it isn't simple and clear. Blueboar (talk) 18:21, 2 January 2008 (UTC)

The problem is that they won't accept a clearer explanation. They want to make the language as tortured as possible, then they complain they don't understand it. How much clearer could this be? "A primary source is one close to and with direct knowledge of an issue. In contrast, secondary sources are at least one step removed, and provide secondhand accounts or a general overview." Where's the difficulty? SlimVirgin ^{(talk)(contribs)} 18:48, 2 January 2008 (UTC)

Discussion needed

Well, I hate am happy to say it, it looks like more discussion is needed. As Jossi does, I have enormous respect for Vassyana and SlimVirgin both. Now, I may be wrong, but it appears that Slim reverted a change that contained a new element she disagrees with and doesn't make sense to her - (secondary sources more likely to be biased and tortured language and inaccurate descriptions). Even though I've agreed with some (or all) of the proposed changes, I'm going to revert that out again so it can be discussed here. And I do admit that Slim's comments have given me pause...so I'd really like to see further discussion - from everyone...

I know this is frustrating, but the changes she disputes were discussed and implemented during a short time that coincided with a lot of other things, and with the amount of traffic this policy has seen recently, I think it's only fair to allow further discussion before implementing. There was a call for more time to be given for Slim and Crum to comment, and I believe an editor eariler had expressed confusion over the complexity of changes. Let's give it a bit more time, shall we?

And please, let's try to focus on the editorial content of the policy and not engage in fingerpointing or further comments about the other contributors. Let's start the New Year off right!

Dreadstar †

18:51, 2 January 2008 (UTC)

Thanks, it's clarified the two options being considered and it will be worthwhile going over it clause by clause with the aim of clarity and effectiveness, will try to find time to contribute. Perhaps a good idea to archive the personal part of this section? Happy new year! ... dave souza, talk 19:31, 2 January 2008 (UTC)

I will not do so myself, but I would request that the two significant undiscussed changes be reverted as well.[28][29] If a well-discussed version needs more discussion, undiscussed significant changes certainly need some discussion of their own. Vassyana (talk) 00:12, 3 January 2008 (UTC)

These principles have been discussed on this page going back years. Your reverting and talk page edits are very unhelpful, Vassyana. You've been editing this page only since July. In that time, you've become the 3rd highest poster (compared to the top two who've been editing it since July 2006 and March 2005), with 493 edits in five and a half months, which works out at roughly 22 posts to this page every single week. Most of your edits to the policy have been to try to undermine the use of secondary sources, or to revert anyone who disagrees with the changes. You're a relatively inexperienced editor with 1,400 mainspace edits, yet your attitude seems to be that you are right and everyone who disagrees is wrong, or isn't really disagreeing; and everyone who tells you that the material you're objecting to has been in the policy in one form or another for three years is also wrong, and if they produce diffs and remember editing the policy themselves back then, they're hallucinating. And when I complain about you calling me a bald-faced liar and for posting in edit summaries that my edits are disruptive, it's how dare I bring it up, and I'm being disruptive for mentioning it. The whole thing is too strange. SlimVirgin ^{(talk)(contribs)} 00:48, 3 January 2008 (UTC)

I politely request that you reconsider/retract your comments. Vassyana (talk) 01:15, 3 January 2008 (UTC)

Look, I'm not trying to stir up trouble, even if it looks that way, but it's increasingly hard to see what you're trying to achieve here, and also becoming hard to assume good faith, given the personal attacks and the sheer number of posts. The most significant differences between the two versions below are these:

• You left out of your version that "All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than original analysis of the primary-source material by Wikipedia editors." But this has been in the policy, in some form, for three years. Yet you somehow manage to twist my restoration of it as my addition of it, and you start complaining that I'm just editing whatever I fancy into the policy, while preventing you from doing the same. This is a distortion, I hope inadvertent.

• The second difference is that you have added: "Secondary sources are valued for analysis, broader context, and a more independent view. However, they can be biased by the views of their author(s) and/or publisher, so caution is required ..." This has never been in the policy. First, it's wrong, and indeed ridiculous, to imply that secondary sources are more likely to be biased than primary sources, when (if anything) the opposite is true. But more importantly, as several people have told you (even, I believe, Fullstop, who is on your "side"), the issue of bias is completely unrelated to the concept of NOR. And yet again, my removal of it to the long-standing version that makes no mention of bias is disruptive, while your very recent addition of it without consensus is fine.

This is a policy. It has to be stable, yet since July, it has been protected seven times. I've not yet looked to see how many of these protects were triggered by edit warring over your proposals. The need for stability doesn't mean it can't be edited, but if copy edits are made or the ideas are tightened, it has to be done in a way that doesn't change the fundamental concepts within the policy. That requires some understanding of the policy (letter and spirit) and some knowledge of its history and intent, and experience of how it's used and misused. But the people on this page who have that experience and knowledge are precisely the ones you want to pay no attention to, whereas if you would only collaborate with them, instead of constantly resisting, it would be a win-win situation. SlimVirgin ^{(talk)(contribs)} 01:33, 3 January 2008 (UTC)

There is no policy that says a policy has to be stable. But there is a policy that says all policies must represent community consensus. Stability is not an alternative to consensus, stability comes from consensus. Let's put things in proper perspective. Dhaluza (talk) 00:51, 4 January 2008 (UTC)

And look -- you constantly berate others for not discussing issues you decide we need to discuss, and if we don't want to, you take silence for assent, and go and change the policy. I therefore posted below the two versions side by side, and stood ready to discuss them with you. And what happens? No more mention of those issues from you (that I have seen). Just a request that someone else revert for you. SlimVirgin ^{(talk)(contribs)} 01:37, 3 January 2008 (UTC)

• Excuse me if I object to being misrepresented.

I didn't say "the issue of bias is completely unrelated to the concept of NOR."

I said "there is *no relationship* whatsoever between source-type and bias (or reliability or ease/likelihood of misuse etc)."

Which in plain English means "your distinction between primary and secondary is gratuitous."

• Moreover, I prefer not to be considered on any "side" other than Wikipedia's. And anyone with even a marginally intact sense of priority will have long realized that -- to Wikipedia -- the PSTS gunk is one of the most destructive things around; your "need" to edit, the grief it causes, the other edit warring, the reams of talk page round-about, two fork pages, and so on and so forth. And this is just whats happening here.
• Now, if that ludicrous ps/ss distinction had some purpose, that would be one thing. But it doesn't.

-- Fullstop (talk) 02:46, 3 January 2008 (UTC)

I would politely request that you reconsider/redact your comments. Vassyana (talk) 02:53, 3 January 2008 (UTC)

Done. -- Fullstop (talk) 03:10, 3 January 2008 (UTC)

Let me try to be both as blunt and as polite as possible here.

My high volume of talk page edits has a lot to do with working on proposals and efforts towards compromise and consensus. Contrary to your belief that my edits here have been unhelpful, a wide swath of editors have repeatedly praised my efforts towards that goal, even while disagreeing.

I strongly object to your insulting and inaccurate portrayal of my behavior. I strongly disagree with your characterization of some of my comments as personal attacks, and really wish you would drop the issue instead of waving it like a flag. I tried walking away from that particular discussion, but you have pushed the issue multiple times after Jossi (quite reasonably) implored both of us to move away from that.

Regarding your confusion and difficulty in assuming good faith towards me, I feel similarly towards you when you persist in using misrepresentations and outright falsehoods, and apparently hold yourself to a quite different standard than you exhort to others.

There is no distortion or misunderstanding. The policy page changed. I was not responsible for those changes. Saying it was X way at Y time in the past is not a valid excuse for circumventing discussion and consensus. Consensus can change, as you well know, and undiscussed restorations of some past version that completely ignores the discussions and changes that have occurred since is simply wrong.

I note you disagree with the bias statements, and as I have said (repeatedly), I have far more concern about primary sources than secondary sources. It arose as part of the compromise language in response to months of discussion and concerns raised during that time.

All I've asked is that opposition to discussed edits be substantiated and that substantial changes to policy be discussed. I've also suggested that editors who refuse to join in the discussion or do no more than !vote be discounted when considering consensus. This is nothing more than you have asked and suggested in multitudinous policy discussions. Regarding the side-by-side comparison, I find it saddening that my lack of immediate response within hours is used as a criticism. Vassyana (talk) 02:12, 3 January 2008 (UTC)

I respect everyone's judgment and experience here more than my own, which is why I've been content to sit on the sidelines with the confidence that the best Wikipedians are working on it. I like both versions a lot, but SlimVirgin's a little more so. They are nearly identical too operationally in how they instruct people to write then encyclopedia, even though it may not seem so if you've been looking at it up close for a long time. Without getting into the specifics, I think SV's sticks to the point more closely so as to define PSTS in a clear, concise way. Vassyana's includes some admonitions and editorial commentary (such as keeping NPOV and avoiding UNDUE) that, though important, are covered elsewhere. These are a little distracting because they tend to imply that those admonitions are somehow part of the definition of PSTS. My main hope is that once people do agree on a final version we can leave this alone for a while, and establish the kind of united front of consensus needed to defend the policy against those who would call the policy disputed or radically change the way we think of sourcing and NOR. Wikidemo (talk) 01:41, 3 January 2008 (UTC)

I agree that sticking to the point and avoiding overlap with with other policies are laudable attributes; but I disagree that on balance the version attributed to SV is better because it includes the contentious assertions as fact that: 1) primary sources are easily misused and 2) that Wikipedia articles should rely on reliable, published secondary sources. As to the first point, there has been no proof that primary sources are easier to misuse than secondary. That may be a widely held belief, but to assert an asshole estimate as fact in a policy is not good policy. As to item 2, secondary sources are valued for providing context where there are a range of opinions or a large volume of data that must be mined. In fact heavy areas, where there is little room for interpretation, we actually prefer primary sources for facts. For example, the salient characteristics of an aircraft are not the subject of any meaningful debate. They are established facts, and that is what an aircraft article is based on. Sure, we value secondary sources for background and additional insight or context to the extent they are available, but they are not essential to creating and encyclopedic entry on an aircraft. The other obvious example of this is populated place articles created from census data. Again, secondary sources make the articles better, but the thousands of bot created articles were based on a single primary source. There are probably other examples as well. So basing an article on secondary sources is valuable, but not essential in all cases. It has also been suggested that for very mature topics, basing an article on tertiary sources may even be preferable. So we need to jettison this excess baggage. Dhaluza (talk) 03:36, 3 January 2008 (UTC)

What is missing from your comment, Dhaluza, is addressing contentious, disputed, and/or controversial subjects and facts. In these type of subjects, the use of secondary sources is imperative, and the use of primary sources, discouraged. On non-contentious subjects, or those about which there are no disputes (such as the wingspan of a Gulfstream IV, or the number of inhabitants of Guam, the use of secondary sources is, I agree, ancillary. So, I would not call this excess bagagge", on the contrary. ≈ jossi ≈ (talk) 03:53, 3 January 2008 (UTC)

Dhaluza is at least admitting that the version I'm restoring contains 1) primary sources are easily misused and 2) that Wikipedia articles should rely on reliable, published secondary sources; the points Vassyana seeks to remove or dilute.

Dhaluza, you talked somewhere else about my "imperfect memory," but I have been writing and editing this policy for three years, and it's only a few hundred words long. I've therefore become very familiar with the text, and the nuances it can throw up, as well as the ways in which it's used and misused by editors. I can tell you that the caution regarding primary sources has been in it for three years, as has the recommendation in favor of secondary sources, howsoever worded. There is now a concerted effort to undermine those points, which is fair enough -- people are allowed to propose policy changes -- but what's not fair enough is the attempt to do it but pretend it's not being done; and the claims that those defending the policy are in fact trying to change it; and that if people don't take part in a discussion, they are ipso facto placing themselves outside consensus and may safely be ignored, and then attacked if they revert. That is the atmosphere that Vassyana wants to maintain here, and it's not on. SlimVirgin ^{(talk)(contribs)} 16:46, 3 January 2008 (UTC)

Once again, this is demonstrably incorrect, and I have corrected you on this point before. I have posted this diff several times to this discussion. It shows the definition of primary source was first added a little less than three years ago--the "recommendation in favor of secondary sources" did not come until later, so it was not here for three years as you assert. Regardless of how long it was here, how does the argument that something was in the policy for X years carry any weight? It has been presented frequently, but is the best argument that can be offered?

Furthermore, you mischaracterize the debate. I don't think there is any one seriously trying to undermine the basic principle of getting interpretation and analysis from sources, not from editors. There is disagreement over how to express this, and what other baggage it carries into this policy. You further mischaracterize Vassyana's objection to your failure to discuss changes on the talk page during an extended comment period, while immediately criticizing them once posted in an a reversion edit summary. He is not suggesting that your input can be ignored, but he is objecting to how you express it. Dhaluza (talk) 00:51, 4 January 2008 (UTC)

One thing to keep in mind is that of the 2+ million articles, only a small fraction are contentious. They may take up a lot of editor's time, but we still should not base policy solely on this. The problem is with the absolute statements which are not absolutely correct. Since they are narrowly focused, they have the potential to do more harm than good if widely applied. While these absolute black/white policy statements may seem helpful in beating editors over the head, that is not how things are supposed to work in a consensus based community. So for a broad based policy, this is excess baggage. Perhaps we need a separate policy for addressing contentious articles, so we don't clutter up the core policies with things intended to address special cases. Dhaluza (talk) 06:06, 3 January 2008 (UTC)

Slim, once again, the changes are not reflective of my personal opinion. I am personally strongly supportive of secondary sourcing and adverse to primary sourcing. I have simply sought to work out a compromise between the very diverse views of the matter.

If someone does not participate in the consensus building process, they are not part of the consensus building process. It's pretty much common sense. One editor disagreeing does not give them the right to revert discussed changes nor does it give them the right to revert to a previous version without discussion or consensus. Anyone who disagrees with the standing wording of a policy is free to post to the talk page and join in the discussion. Simply "voting" means nothing. Reverting discussed changes without discussion is disruptive. I'm not asserting anything out of the ordinary in this, but rather just reiterating the fundamentals of a consensus-based wiki process. Your implications otherwise are confused and/or rooted in acute misunderstanding, assuming good faith. Vassyana (talk) 17:55, 3 January 2008 (UTC)

No, you're not getting away with that. You are not going to blackmail people into discussing everything you think needs to be discussed, and if they won't, you embark on revert sprees of your changes until the article needs to be protected. If people don't respond to you, it may mean they don't agree with you. If they revert you, it's confirmation that they don't. Their revert is their comment to you.

(outdent) SlimVirgin, it's hardly blackmail. It's exactly what you have exhorted to people on multiple occasions, so please pack away that false righteous indignation. Similarly, your claim that reversions are comments is contrary to both policy and what you have repeatedly told people in the past. This double-standard treatment and outright distortion is simply revolting and ridiculous. Vassyana (talk) 08:39, 4 January 2008 (UTC)

This is a core content policy, not an article, and the presumption is in favor of stability, not change. From now on, unless I actively agree with one of your changes, please take my silence to mean that I disagree, and my revert to mean that I disagree strongly. SlimVirgin ^{(talk)(contribs)} 18:48, 3 January 2008 (UTC)

Well, looks like Wikipedia is closed now, per Slim's comments above. Her silence will mean she disagree's with anything anybody says, and she is therefore to revert freely without discussion to show that she strongly disagrees. The only way anybody will ever know if she agrees with anything is if she actually writes a comment stating so. So, in all essence, this is now Slim's Wikipedia to do with as only she sees fit. Any changes made by anybody else can be reverted at her whim, no discussion or explanation neccessary. wbfergus ^Talk 19:28, 3 January 2008 (UTC)

No, the point is that there are people who feel that unless everyone instantly responds to their edits, it means their edits are automatically "approved by consensus". This is a false premise, especially so for core policies. Not everyone is inclined to respond to every wild claim and engage in discussion. As SlimVirgin noted above, the presumption for core policies is on stability, and any non-trivial change requires consensus. Consensus is achieved by getting the editors who have historically contributed to a given policy to indicate their approval in an active manner. If such an editor reverts, it is a clear sign of no consensus. The true

owners of the core policies are all of us — we need stability, and we trust the established editors that contributed the most to the policies to maintain that stability for us. Crum375 (talk

) 19:49, 3 January 2008 (UTC)

There was no demand for "instant response". Instead, there were months of discussion, including policy RfCs and many advertisements of discussion at the policy village pump. Additionally, you and Slim were given opportunity, and even directly encouraged, to justify your edits and reversions. So, please do not engage in such distortions. Reversions are not a replacement for discussion. If you object, you have to explain yourself on the talk page just the same as any other editor. You are not exempt from participating in consensus-building if you wish to have your voice heard in the determination of consensus or non-consensus. Vassyana (talk) 08:53, 4 January 2008 (UTC)

Crum375, just what exactly is an "established editor"? I've seen SlimVirgin use the same phrase in similar situations. So, again, what exactly is an "established editor"? Cla68 (talk) 13:06, 4 January 2008 (UTC)

There is no such presumption in the WP:Consensus policy, nor does it support your statement above. What policy are you following? Dhaluza (talk) 00:51, 4 January 2008 (UTC)

Per WP:Consensus:

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.

Our core policies are the foundation of the entire project — it is essential that their stability be maintained, otherwise, like all buildings with a shaky foundation, everything else will crumble. This means that to make fundamental changes to core policies a wide consensus is required, not a majority of a few editors on a talk page. Crum375 (talk) 01:04, 4 January 2008 (UTC)

I agree that to make fundamental changes to core policies a wide consensus is required, not a majority of a few editors on a talk page, but that is a two-way, not a one-way street. It is essential that policies represent consensus, from which stability will come. It is not essential that we maintain stability in lieu of consensus. Dhaluza (talk) 02:30, 4 January 2008 (UTC)

By default, any existing Wikipedia policy represents consensus, until proven otherwise and changed. Crum375 (talk) 02:33, 4 January 2008 (UTC)

I strongly disagree with that assertion. Typically a policy description should be edited over time with ample discussion and input to form a broad and stable description of consensus. But this is not always the case. The simple fact that policy page was edited to make some point, particularly when there was only limited input and discussion on that point, does not make it by default represent consensus. You are fundamentally misapplying the principle of "silence implies consent". Consensus is what it is, and the written description of it is always imperfect, and sometimes just plain wrong. Dhaluza (talk) 10:20, 4 January 2008 (UTC)

We especially can't have someone who is posting around 22 comments to this talk page every week for six months dictate to users who don't want to respond to them that we're thereby placing ourselves outside consensus. Several good editors who helped to maintain this page for a long time have been driven away because of the absurd amount of discussion, and the expectation that even the tiniest copy edit be discussed, and investigated, held up to the light, and twisted this way and twisted that, lest the change of a comma to a semi-colon imply that, ZOMG, primary sources are totally banned!! Enough already. SlimVirgin ^{(talk)(contribs)} 20:02, 3 January 2008 (UTC)

Further polemics. My high volume of talk page edits has a lot to do with working on proposals and efforts towards compromise and consensus. Contrary to your belief that my edits here have been unhelpful, a wide swath of editors have repeatedly praised my efforts towards that goal, even while disagreeing. I beg you to drop this antagonism and distortion. Vassyana (talk) 08:39, 4 January 2008 (UTC)

Random break

Okay, I suppose that you must mean things like this [30] must be discussed and approved then, since they were pretty massive changes, far from just afew minor copyedits. And as you've stated before, the silence of the community evidently doesn't amount to consensus for those changes, it just took some of us a while to notice the changes, and since then many have disagreed, so there therefore is no consensus for them. Regarding Crum's statement above about "instantly responds", it also seems like that is how Vassyana was being held to regarding the side-by-side versions, yet when he requested comments over a week long period, that wasn't time enough.

I don't mean to nit-pick or attack you. You seem to have the respect of several other editors that I respect, yet there seems to be a double-standard at play here as well. I could easily go back through the last 4 months of archives, and pick out statements by supporters of your position that when applied to the other posistion, don't carry any weight, like they (the supporters of your position), clearly stated numerous times, that silence did in fact equal agreement, so therefore consensus was met. Now, when Vassyana makes an edit after being offered for dsicussion for a week or longer, changes made when their were suggestions or objections, etc., suddenly the silence no longer means consensus, but tacit disapproval? There's many other similar cases of what seems to be double-standards at play here as well, that's just one example off the top of my head.

Can we just agree to stop all of the nit-picking, etc., lay down some new 'ground rules' for here on out, and everybody agree to play nice and treat each other with the same respect and courtesy they expect from others, and that everybody plays by the same set of rules? That doesn't seem to much to ask. wbfergus ^Talk 20:28, 3 January 2008 (UTC)

That would be nice, wbfergus, but it would require people to inform themselves about the content of this policy, and stop claiming that things they have personally not seen before are new. I know it's a nuisance to have to search through diffs, but it's important not to accuse editors of adding new material if you're not willing to do it (I don't mean you personally).

The other problem is that some people seem to see changes in meaning where there aren't any, and don't see substantive change when it's obvious. But yes, in principle, if we can overcome these issues, I'm very willing to help set up ground rules if it will stop the nitpicking. SlimVirgin ^{(talk)(contribs)} 22:11, 3 January 2008 (UTC)

"Contended" statements

(partial/edited cross-post from elsewhere)

1. Yes, misreading Darwin (naively or otherwise) can superficially be described to be an example of "easier" misuse of a primary source. In reality however, the cause of the misuse does not lie with source-type but with the editor's ignorance of Darwin's time, language and social mores. A secondary source that was contemporary to Darwin would be just as "easy" to misrepresent.
2. What policy should say is...

Any contended^* statement should not cite the source that "said it first", but should instead cite a source that refers (not necessarily positively) to the source that made the original statement.

Such a clause...

a) would restrict primary sources to only where they can do no harm.

b) would effectively establish 'notability' at the statement level.

c) is actually somewhat related to 'original research' ("said it first").

d) would reinforce the concept of reliability,

e) would provide a source that potentially lists alternative positions, so supporting NPOV.

(end of adapted cross-post)

The above-mentioned a—c are the crux of what the whole PSTS shebang is purportedly there for.

• Yes, there is no ps/ss distinction needed to express what they want to say, and no need to use the words "primary" or "secondary" either.
• Yes, the reason why they can't explain it is because they themselves have only the vaguest idea of what it is that they were groping for. In their minds keeping cruft out has vaguely something to do with ps/ss which has vaguely something to do with NOR. Not wrong per-se, just all very muddled. Of course, what they meant by "primary" all along is that what comes first.
• Yes, they will probably write my two-liner off too. But my point is that one can arrive at 'B' even without 'A'. There is more than one way to skin that particular cat.

And yes, I've know this for some time, but until today I didn't imagine that there were actually people who supposed that they were more discerning than anyone else. Fine, call me naive. -- Fullstop (talk) 06:35, 3 January 2008 (UTC)

When facts are contested by other reliable sources (not just by WP editors) then we should defer to other reliable sources to characterize and contextualize the debate, rather than allow editors to add their own views or analysis. Note that it is not necessary to use source typing terms like secondary or seccond-hand to address this--plain language works just fine. Dhaluza (talk) 13:03, 3 January 2008 (UTC)

Regarding Fullstop's point above that the policy should say "Any contended* statement should not cite the source that "said it first", but should instead cite a source that refers (not necessarily positively) to the source that made the original statement."

That's more or less what it does imply, Fullstop, though it would be wrong to state it as strongly as you have (your formulation would reduce primary source use even more, and there's clearly no consensus to do that). SlimVirgin ^{(talk)(contribs)} 17:23, 3 January 2008 (UTC)

Well, the restriction applies only when a statement/attribution is contested (cf asterisk next to the word "contended"). This is conform with both Dhaluza notes above as well as what Jossi notes above what I said.

It is easy enough to determine whether the idea has "no consensus" or not, or whether its just a question of phrasing.

But the real hurdle is whether you are personally prepared to get over the superfluous ps/ss stuff or not.

Reaching consensus on the phrasing of the two-liner is (comparatively) trivial once people...

a) are aware that it is not necessary to use source-typing to deliver the message and

b) are prepared to get rid of what is redundant.

c) agree to work on the phrasing of the idea of the two-liner in full awareness that it makes the PSTS section redundant.

The ball is in your court Slim, the PSTS cruft is your baby (now turned monster).

-- Fullstop (talk) 20:02, 3 January 2008 (UTC)

(outdent) No, the PS/SS/TS classification of sources is an extremely useful way to explain our sourcing and content policies. It helps editors understand and qualify sources, and sources are the essence of this site. This is a widely accepted classification, not one invented by Wikipedians. If you understand the distinction between PS/SS/TS, odds are you understand sources, and will know how to use (or not use) them. Crum375 (talk) 20:17, 3 January 2008 (UTC)

1. >>"classification of sources is an extremely useful way to explain our sourcing and content policies"
   Explaining "useful ways" is not the purpose or objective of THIS (or any other) policy. Besides, "useful way" is by definition GUIDELINE.
2. >>"This is a widely accepted classification, not one invented by Wikipedians"
   Incorrect on three counts:
   a) The subject of this policy is "original research" by Wikipedians.
   b) The meaning of those terms is not universal, leave alone a "widely accepted classification."
   c) it is (arbitrarily) chosen by Wikipedians with a meaning and purpose specific to Wikipedia. It applies nowhere else.
3. >> "If you understand the distinction between PS/SS/TS, odds are you understand sources"
   I'm terribly sorry for you if your idea of "sources" is formed by the understanding of the distinction between a fountain, a hydrant, and a faucet. St. Augustine had already figured this out 1700 years ago. One doesn't have to know what a "poodle" is in order to know what a "dog" is, nor does one comprehend what a "dog" is by understanding the distinction between a chihuahua and a great dane.
   Even a four year old child knows what a dog is, even it doesn't know (or need to know) what breed. Its also not necessary to distinguish breeds in order to establish that dogs can bite.

PSTS is not there to gratuitously classify sources. PSTS is there to make a point. And that point does not require ps/ss distinction to make it.

-- Fullstop (talk) 21:43, 3 January 2008 (UTC)

See also my reply below, timestamped "21:03, 3 January 2008 (UTC)". NOR is here to explain to editors not to add their own research. PSTS is an extremely useful framework to explain that, and evaluating sources in general. Policies are written to be understood, so they can be properly used. Crum375 (talk) 21:50, 3 January 2008 (UTC)

I just found that it was Fullstop who originally wrote this: "In the humanities, materials that are potential objects of academic research but do not themselves constitute academic research are considered 'primary sources'." Could you say what it means, please, Fullstop? SlimVirgin ^{(talk)(contribs)} 22:05, 3 January 2008 (UTC)

Fullstop, please read this. It is in the policy:

"Appropriate sourcing is a complicated issue, and these are general rules. The decision as to whether primary or secondary sources are more suitable on any given occasion is a matter of common sense and good editorial judgment, and should be discussed on article talk pages."

We can't pin things down as much as you want to. Sometimes even when edits are contested, using primary sources will be appropriate. What the policy says is that it's easier to misuse primary sources, and articles should therefore be based on secondary sources. It doesn't mean primary sources can never be used. You're all reading things into the policy that aren't there, then demanding they be removed, then writing new versions in poor English that don't say what you want them to say, and do say what you don't want them to say, and demanding UN Security Council resolutions before a copy edit's allowed that might clarify things.

The bottom line is that every educated native English speaker understands what primary/secondary means, even if they'd disagree about the application in borderline cases, and the distinction is helpful and used all the time on talk pages. SlimVirgin ^{(talk)(contribs)} 20:36, 3 January 2008 (UTC)

yes, sometimes when edits are contested, a primary source will be appropriate. That is however not the issue. The issue is that the position being described by the primary source should (not 'must') be acknowledged by a secondary source. If for no other reason than to establish that the reading is correct. Which would then also be conform with what NOR policy is all about.

ok, you say the pathetic prose of the PSTS section is the root of all troubles. I'm inclined to disagree because I believe that the pitiful prose is a natural consequence of trying to define a chimera. However, I'll give you the benefit of the doubt, and see what you come up with.

I once wrote a complete PSTS in 4 sentences, but which apparently wasn't conform with the then "prevailing" idea of what p/s was and the insistence that ps/ss/ts could only be adequately "summarized" with an essay. Your ideas as more in line with mine, so perhaps you're going in a sustainable direction. I've given up trying to track changes on the policy page itself, so please give me a heads-up some other way. Thanks. -- Fullstop (talk) 22:03, 3 January 2008 (UTC)

I believe you wrote this description of a primary source: "In the humanities, materials that are potential objects of academic research but do not themselves constitute academic research are considered 'primary sources'." Could you say what it means? SlimVirgin ^{(talk)(contribs)} 22:13, 3 January 2008 (UTC)

That was not one of the sentences I was referring to.

Ignore it. This is an inversion of what a secondary source is, which I inverted to (sort of) line up with what the "in the sciences" part was then saying (thats all gone now).

1. A primary source is a novel^* intimation^{(abstraction, concept, conclusion, conviction, determination, design, doctrine, experiment, estimate, exposition, explanation, hypothesis, idea, impression, interpretation, judgment, notion, perception, rationale, scheme, solution, story, suggestion, supposition, theory, thought, view)} that has (to anyone's knowledge) never been made/scrutinized before. Additionally, any statement that does not reflect scientific method is a primary source.
2. A secondary source is a methodical exploration^{(acknowledgment, analysis, examination)} of a primary source.

In both cases, "source" is relative to the target statement that draws on that source. Of course, someone is going to argue about those definitions, but thats not what you asked for. :)

This still isn't what I was referring to when I said "I once wrote a complete PSTS in 4 sentences". I can't find it, its buried somewhere in the archives.

BTW: As far as WP need care about contestable determinations, those that have never been scrutinized/reiterated by any other source "do not exist." This is an implication of the due-weight clause in WP:NPOV since editors may not assign weight themselves (which would again be OR) but must have a source to do it for them.

-- Fullstop (talk) 23:53, 3 January 2008 (UTC)

My position here is intermediate. I believe source typing is an effort to address some important underlying policy needs, such as ensuring we do not rely on subject/viewpoint proponents to establish the notability or significance of their own subject/viewpoint even though we can rely on them for other facts. I believe the typology can be useful. However, I believe that source typing is sometimes imprecise and/or unnecessary and hence is sometimes "overkill" for addressing the actual policy requirements, and moreover that its appropriate use will vary somewhat from field to field. For these reasons, I believe that PSTS should appear in a guideline, not policy, and suggest that the proposed

WP:NOR attempt instead to articulate, clearly and directly, the actual underlying policy needs which led to PSTS being proposed. Best, --Shirahadasha (talk

) 20:49, 3 January 2008 (UTC)

I think understanding PSTS is essential for understanding sources in general, and for NOR specifically. Many editors simply fail to understand why, for example, adding a minor comment along with a source is considered OR. Once they fully grasp the distinction between PS and SS, they realize (for example) that they took a PS, and by adding their minor comment it, Wikipedia effectively became a secondary source, hence they added OR. These subtle points are much harder to make without the underlying framework of PSTS. Yes, PSTS does not always have sharply defined boundary lines, but the concept is always there, and it is extremely useful, and widely used, as a general guideline in assessing sources and pointing out OR. Crum375 (talk) 21:03, 3 January 2008 (UTC)

1. If the subtle point is supposed to be what PSTS is clarifying, then it has botched the job. That is why there is so much grief about PSTS itself.
2. insisting that only a ps/ss distinction can make the point is sheer stubborness. There is not only one way and no other.
3. the point that the misrepresentation of a PS can make Wikipedia a secondary source is painfully redundant. Misrepresentation of any kind of source is what NOR is supposed to be preventing.
4. Please read what I said at the top of this sub-section. Your point about misrepresentation of "primary sources" is addressed there. It does not require a distinction of ps/ss to make that point. The point can be made even without the essay on distinction.

-- Fullstop (talk) 22:21, 3 January 2008 (UTC)

PSTS has not "botched the job". When I joined WP and read the policies, it was very useful in helping me understand the spirit behind the rules, so "it worked for me". PSTS is the conventionally and widely accepted way to classify sources. Strictly relying on sources is our job here. NOR is all about relying on the proper sources, and not adding to them or changing what they say. If you have a better way to explain the NOR concept, please propose it and obtain consensus. My own inclination to understanding any rule is to understand the underlying concepts, and PSTS is the underlying concept of sources, and hence a critical part of NOR. Crum375 (talk) 22:33, 3 January 2008 (UTC)

If you want to help turn Wikipedia:Evaluating sources into something that could ever be proposed as a guideline, please feel free. When I last checked, it contained gems such as "In the humanities, materials that are potential objects of academic research but do not themselves constitute academic research are considered 'primary sources'," and advice such as "sources must be evaluated in context," which reminds me of a sentence I once found in The Holocaust that said "children also suffered." SlimVirgin ^{(talk)(contribs)} 21:05, 3 January 2008 (UTC)

Basically, I rather like this idea, though as a complement to psts rather than as an immediate replacement. However, defining "contended*" could be a can of worms, and open articles on hot topics such as creationism or homeopathy up to endless demands for additional references commenting on every source cited. .. dave souza, talk 18:28, 4 January 2008 (UTC)