Wikipedia talk:No original research

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Peacedove.svg The project page associated with this talk page is an official policy on Wikipedia. Policies have wide acceptance among editors and are considered a standard for all users to follow. Please review policy editing recommendations before making any substantive change to this page. Always remember to keep cool when editing. Changes to this page do not immediately change policy anyway, so don't panic.
Frequently Asked Questions (FAQ)
I disagree with the definition of secondary source.
Wikipedia mostly follows the definition in use by historians, which requires more than simply repeating information from some other source or rearranging information from the author's notes. The earliest definition of a secondary source in this policy was in February 2004 "one that analyzes, assimilates, evaluates, interprets, and/or synthesizes primary sources".
WikiProject Spoken Wikipedia
WikiProject icon This page is within the scope of WikiProject Spoken Wikipedia, a collaborative effort to improve the coverage of articles that are spoken on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
 
Shortcuts:

Definition of Research[edit]

Perhaps I'm missing something, but isn't all research by definition - unless plagiarized - original research? Further, if one relies on publishment as a means to legitimize research, what legitimizes a publisher? Surely not consensus - for myriad logical as well as ethical reasons. Just looking for clarification -- thanks. Amizzo (talk) 01:20, 7 February 2015 (UTC)

Original research is a term usually reserved for collecting, analyzing and interpreting primary data to gain answers to a thus far unanswered research questions. Almost all scientific research aims to be like that. Wikipedia is however not into primary scientific research but in aggregating existing knowledge.
A publisher can publish what they want using whatever quality standards they use. No consensus needed, and that is one of the reasons why scientific research is considered a primary sources in Wikipedia. Primary sources in themselves are fine, however without knowledge of the context it is hard to judge their contents at the right value.
However, the value of a journal depends highly on consensus in the academic field. Some journals are considered essential to a field, and those will be present in almost all scientific libraries, these are the journals scientists send their best work to, and these are the journals where publications contribute to the esteem of the author. Obviously such a journal gathers more revenue for a publisher.
If a respected journal lowers its standards, its desirability will drop over the years, number of good papers submitted to it will drop, etc. On the other hand new journals often need a number of years of uphill struggle to be accepted as a good journal. So there is definitely some consensus going on in determining what good journals are. (Of course an editor would need intimate knowledge of a field to know this). Arnoutf (talk) 09:03, 7 February 2015 (UTC)
On Wikipedia I think the term is more frequently used with respect to the collecting, analyzing and interpreting of secondary sources. The SYNTH section here doesn't mention "primary" even once, and most accusations of original research are accusations that someone has engaged in SYNTH. There is no absolute ban on using primary sources. There is on SYNTH. It follows that making out a SYNTH charge is more likely to get a decisive settlement in one's favour than accusing someone of using a primary source. Some people argue that there's a de facto ban on primary sources by making a big deal out of "secondary or tertiary sources are needed to establish the topic's notability" since when there are such secondary sources, there is no longer a need to use the primary source and the source could just be switched to the secondary. This notion of a double-cite RS policy (one cite for the material, and one cite for its "notability" even though the second cite would usually make the first cite redundant) is, however, not clearly and unambiguously WIkipedia policy (WP:RS doesn't indicate a need for two cites) unlike SYNTH hence a charge of SYNTH is the most common form of OR accusation.
That doesn't mean editors alleging OR won't try to run with whatever works, e.g. I add material to an article about an advocate of fringe theories like WiFi causes cancer, material I cite to secondaries that challenge the fringe theory, and I get accused of SYNTH because I've caused readers to think the advocate might have an axe to grind and the cited sources don't criticize the advocate directly, just the theory. I then try to accommodate by substituting in direct criticism of the advocate cited to a widely cited critic but use the critic as a primary. I'm then called upon to produce a secondary source in order to show that the controversy is "notable" (both because WP:OR says "secondary or tertiary sources are needed to establish the topic's notability" and the now more direct criticism means there's supposedly a BLP issue that in turn requires "multiple reliable third-party sources documenting the allegation or incident"). Well the reason why secondary sources covering the "controversy" are hard to come by is because the advocate is too fringe to get equal time with the critics or, more to the point, get the attention of critics! You end up with a Wikipedia with all sorts of article subjects who are notable enough to get an article but not notable enough to get secondary source coverage of their deviation from the mainstream that explicitly expresses it as such. People read these stub articles and reckon the subject is every bit the authority the subject claims to be, use of sources other than the subject being so sharply circumscribed. The way WP:OR is often used thus frequently enables the very quackery it's supposed to limit. WP:OR nonetheless makes perfect sense if one believes most of the dubious information in Wikipedia did not originate off-Wiki or with editors not following WP:RS and WP:NPOV but right here on-Wiki with WP:RS and WP:NPOV-following editors who synthesize secondary sources or use primary sources.--Brian Dell (talk) 03:13, 8 February 2015 (UTC)

At a university I was attending an assistant professor was dismissed because although he had published many papers over the previous few years, the university decided that they were only summaries of other people's papers, and that he had not published any original research of the quality the university expected its factuality staff -- in other words the diametrically opposite of Wikipedia's criteria. -- PBS (talk) 15:11, 4 April 2015 (UTC)

Is WP:Cherrypicking#Contradictions invalidated by WP:STICKTOSOURCE?[edit]

WP:Cherrypicking#Contradictions says, "As to contradictory information, if, for example, a source says 'Charlie loves all blue coats and hates all red coats', to report in Wikipedia that according to the same source "Charlie loves all ... coats" is cherrypicking. It is cherrypicking even if the source is precisely cited."

WP:STICKTOSOURCE says, "Take care not to go beyond what is expressed in the sources, or to use them in ways inconsistent with the intention of the source"

Suppose a source itself baldly commits exactly the example error by saying "Charlie loves all blue coats and hates all red coats; therefore Charlie loves all coats." To avoid cherrypicking, it would be necessary to report on the contradiction. However, the intention of the source is to prove that Charlie loves all coats, so it would not be sticking to the source to mention that Charlie hates red coats.

Sticking to a source belongs to WP:OR, so would seem to be more vital than the essay on Cherrypicking. More vital still, though is WP:V. Something that is verifiably false from a source is not verifiable from that same source. WP:EP also says, "a lack of information is better than misleading or false information". A source like this is unusable if the contradiction can't be raised.

"Charlie hates all red coats" might be an important claim for an article, though. It would be a different instance of cherrypicking to use it without mentioning the source's spurious conclusion of loving all coats, but I think as long as that's mentioned, the source's intent should not prevent the citation of Charlie hating red coats. STICKTOSOURCE just needs to be interpreted less encompassingly. While the source intended to conclude that Charlie loves all coats, it also intended to say Charlie hates red coats at the point that it said it.

WP:NOTOR has some things to say about handling two contradictory sources that I think apply equally well to a single source contradicting itself: "We have a responsibility to present an accurate and factual overview of the topic addressed in the article. This may include indicating when a given authority may be wrong and presenting contradicting claims using proper weighting. A solution is to accurately and honestly cite the authority, while also citing the conflicting fact(s); point to the problem, but do not attempt to solve it with your own arguments." Rhoark (talk) 02:48, 14 March 2015 (UTC)

Suppose a source itself baldly commits exactly the example error by saying "Charlie loves all blue coats and hates all red coats; therefore Charlie loves all coats." To avoid cherrypicking, it would be necessary to report on the contradiction. No, it's not necessary. A contradiction conveys no useful information; any part of a source that contradicts itself is clearly unreliable. So simply don't use it at all. This isn't "cherrypicking" which is selective use of sources intended to mislead. Peter coxhead (talk) 09:10, 14 March 2015 (UTC)
To clarify, its not a priori necessary to report on the contradiction, but if any of the three prongs (loves blue, hates red, loves all) is reported, it would be cherrypicking unless all three are reported. Rhoark (talk) 13:45, 14 March 2015 (UTC)
Precisely. Peter coxhead (talk) 17:34, 14 March 2015 (UTC)

Can Cabinet Papers be used as sources?[edit]

Usually, after 30 years or so, some governments including the UK Government, publish the minutes of the Cabinet meetings and lots of Government papers. Are we allowed to use those Cabinet minutes and Government papers to feed into edits on Wikipedia? Or are we limited to whatever books or papers are published by others that refer to them and can't use them as sources without secondary sources referring to them? Frenchmalawi (talk) 14:11, 3 April 2015 (UTC)

@Frenchmalawi: They would be primary sources, so should be used cautiously - but they are not ruled out. You should present specific passages and the claims you'd like them to support at the reliable sources noticeboard. This doesn't really have anything to do with OR policy. Rhoark (talk) 16:55, 6 April 2015 (UTC)

Reliably published[edit]

See Wikipedia talk:No original research/Archive 59#Reliably published

I have reverted the change by user:Bob K31416 (05:58, 7 June 2014). There is a fundamental difference between a reliable primary source and a reliably published primary source. The whole point of reliably published, is to stop OR with primary sources that have not been reliably published. For example a letter from Winston Churchill in an archive that has not been indexed would still be a reliable primary source, but if it has not been published in a reliable source, then to use it on WP would be OR. -- PBS (talk) 14:57, 4 April 2015 (UTC)

The consensus of that thread (Wikipedia talk:No original research/Archive 59#Reliably published) seems to have been that "reliably published" is a "dreadful phrase...confusing...abuse of the English language". So why are we putting it back in?
Originally this sentence contained a link to WP:RS#Definition of a source, which says that "source" can refer to (a) the work, (b) the author, or (c) the publisher, and that each of these has a different kind of reliability. Exactly how they can be reliable is a complicated question, with criteria that are different from the criteria we use to tell whether a source is primary. So what I would suggest here is that we just leave the question of RS up to WP:RS and concentrate here on the distinction between primary, secondary, and tertiary. This is complicated enough without getting into arguments over reliability. For example, whether a blog post (self-published!) by Ben Bernanke (expert!) on interest rates can ever be an RS.
What I am suggesting is that we simply remove both "reliable" and "reliably published" from this section. So that's what I'm going to do – edit this sentence to read "Unless restricted by another policy, primary sources that have been reliably published may be used in Wikipedia;". That seems to me to be the simplest and easiest to understand.
If someone wants to insist that "primary sources" are unacceptable unless a reliable publisher is involved, could they please come up with a better phrase than "reliably published"? – Margin1522 (talk) 08:08, 5 April 2015 (UTC)
Blog posts by Bernanke are not, afaik, peer-reviewed - a criterion that is met by many other publications on the topic of interest rates. Samsara 08:33, 5 April 2015 (UTC)
Peer review is a kind of quality check for academic papers. Most other sources are not. The mentioned letter of Churchill is unlikely to be peer reviewed, a broadcast interview on a reliable network is not. So let's not bring in yet another layer of complexity. Arnoutf (talk) 09:32, 5 April 2015 (UTC)
About Bernanke, as the two-time chairman of the FRB he is a widely recognized authority on interest rates. As widely recognized as you can get. And there are many passages in our policies that allow for exceptions when the "author is an established expert whose work in the relevant field has been published by reliable third-party publications." (WP:USERG) It is also true that many editors refuse to recognize these exceptions and insist that a blog post can never, ever be an RS. This is an endless dispute that will never be settled. What I am saying is that we don't need to settle it. OR is original interpretation by a Wikipedia editor of any primary source. Can't we concentrate on that, and leave the disputes about what is or isn't a reliable primary source to some other forum? – Margin1522 (talk) 09:36, 5 April 2015 (UTC)
I wonder who comes up with utter nonsense like that - "author is an established expert whose work in the relevant field has been published by reliable third-party publications" is a hurdle that probably virtually all active researchers can meet. In fact, that definition would embrace the opinions of lab technicians if they've ever been granted co-authorship, which is not all that uncommon (most of them splendid people, but not generally considered notable). I would suggest that phrase be sent straight to the bin, which is where it belongs. Samsara 09:56, 5 April 2015 (UTC)
Samsara, your argument is a very valid argument why we should be very cautious with primary sources (which most empirical scientific papers are). The peer reviewed scientific journal is a reliable source for the primary evidence from the empirical study but not what it means in the larger scientific development, just as much as Bernanke's (or in fact anyone else's) blog is a reliable source of the opinion of that person (assuming the blog is not hacked of course). In my view it has little to do with reliability of the source, but with the use of primary sources. Can we try to disentangle those arguments. Arnoutf (talk) 14:32, 5 April 2015 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────@Margin1522 the addition of the link to WP:RS is relatively new and it was an incorrect addition as WP:RS is a guideline not a policy. WP:SOURCE links to policy and describes what reliably published means.

The question of whether or no a blog page is reliable is a question for WP:V and I think confusing this issue.

I am coming from this from the point of view of history and literature and the fact that this is the policy called "No original research", a primary source may not be a reliable source in any usual meaning of the word, for example an eye witness account of an event, in that it would never be considered a reliable secondary source. Without the criteria of reliably published it is far too easy for an editor to carry out original research in those fields.

Let me give you an example: There is a controversial British military officer called Orde Wingate, and as often happens with controversial men he has his detractors and admirers. Some years ago I was involved in a debate about some meetings Wingate attended with General (later Field Marshal) William Slim. Slim records the meetings in his military history Defeat into Victory, the admirer wished to add text to a Wikipedia article that contracted Slim's account based on an unpublished manuscript he [said (AGF) he] had found in the archives. There is no doubt that a manuscript by Wingate about the meetings is a reliable primary source and if this manuscript is ever published and a reliable secondary source concludes that the meetings were not as Slim describes then that of course should be added to a Wikipedia article, but as it is not, it is a classic example of OR. Allowing editors to use unpublished primary sources, is in many fields, to give cart blanch to editors to add "novel narrative or historical interpretation" into Wikipedia articles, something that is against the spirit of this policy.

There is nothing new in this phrase "reliably published" it was added to this policy with Revision as of 23:02, 12 March 2010 (The discussion linked to that is to he found here Wikipedia talk:No original research/Archive 50#Primary sources reliably published.). But it replaced similar wording that has been in the policy for many years. It was in this policy back in April 2008 (before the current reformatting of the section) as "Primary sources that have been published by a reliable source may be used in Wikipedia", and as early as April 2005 this policy stated "it is essential that any primary-source material used in an article has been published or otherwise made available to people who do not rely on Wikipedia". So if you want to remove it Margin1522 I suggest that you author a fully blown RfC to see if there is a consensus to remove a concept that has been part of this policy for at least 10 years.

If other editors consider "reliably published" to be difficult to understand, then we could go back to the older wording (pre-March 2010) "published by a reliable source". -- PBS (talk) 17:40, 5 April 2015 (UTC)

I don't think anyone is disputing the concept, but rather the way it is expressed. The phrase "reliably published" is unclear. If you want to use it, I suggest you explain its meaning with a footnote.
Re the suggestion, "primary sources that have been published by a reliable source" — It's confusing because it uses "source" twice with different meanings, and a publisher may not be reliable in general but reliable for presenting a particular primary source.
This idea has complications and may best be discussed at the WP:V policy. Then there can be a reference to that part of WP:V from WP:NOR. I think this is essentially the approach suggested by Margin1522. --Bob K31416 (talk) 20:51, 5 April 2015 (UTC)
Are you confused by the wording? If not what make you think others would be? It is generally not a good idea to footnote a concept that is defined in another policy, as the two tend over time to drift apart and that causes problems. The link WP:SOURCE defines what is and is not reliably published. -- I don't agree with the footnote -- but that is what it is. However as my major concern is to make sure that information in archives that has not been catalogued and/or is not available to the public is not used to introduce novel history or novel interpretations of literature, I can live with it: It is in WP:V in the section " Reliable sources": Source material must have been published, the definition of which for our purposes is "made available to the public in some form".This includes material such as documents in publicly-accessible archives, inscriptions on monuments, gravestones, etc., that are available for anyone to see. Unpublished materials are not considered reliable.. So one option rather than adding a footnote is to add WP:SOURCE (although I would have thought that anyone who read the paragraph in the lead "(NOR) is one of three core..." would make that association without a link). -- PBS (talk) 09:25, 6 April 2015 (UTC)
Regarding WP:SOURCE, are you OK with the following part of it?
"Source material must have been published, the definition of which for our purposes is "made available to the public in some form".[6]
and footnote 6 is
"This includes material such as documents in publicly-accessible archives, inscriptions on monuments, gravestones, etc., that are available for anyone to see."
The reason I'm asking is because of the example in your opening message, "For example a letter from Winston Churchill in an archive that has not been indexed would still be a reliable primary source, but if it has not been published in a reliable source, then to use it on WP would be OR." --Bob K31416 (talk) 15:30, 6 April 2015 (UTC)
But if the letter would be in a publicly accessible archive - it would be available to the public and thus be verifiable. Arnoutf (talk) 16:19, 6 April 2015 (UTC)
Most archives of unpublished works are not open to be public, but may be open to accredited scholars. But without an index (catalogue) an archive may theoretically be open to scholars but before the material is catalogued there is no easy way to find the material and citing the material (in such a way that others can access it in the future is not possible), so such archived material is not reliably published. Of course if the archive is catalogued and accessible to the public is by the criteria of Wikipedia reliably published (because an archivist has been through the papers and verified who wrote what). Stuff in "Grandma's attic", is also not reliably published, because even if a photograph is taken of a letter purportedly by Churchill is placed on the web, without an expert verifying that it is indeed a letter by Churchill, there is no way for Wikipedia editors to judge if it is. -- PBS (talk) 17:14, 6 April 2015 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Maybe it's because I don't edit history, but I've never encountered an editor who wanted to use an unpublished manuscript. But even if the manuscript was published in someone's collected papers by Oxford University Press, it's still a primary source. Calling it "reliably published" gives the impression that it's OK to cite that manuscript to make whatever point the editor wants to make. But it's not OK. Anything that is controversial or involves interpretation or judgment has to come from a secondary source. That's what we should be focusing on in this policy.

The other reason I want to avoid "reliable" is not because I don't think it's important for sources to reliable. Of course I do. It's because there are endless edit wars in articles, and endless disputes at WP:RSN and WP:AfD over the exact definition of "reliable". Edit warriors are constantly finding new and ingenious arguments to show why the other side's sources are "not RS". I just think it would be simpler for this policy to avoid getting entangled in these arguments and the 15 different interpretations of "reliably published".

Basically whether a primary source is reliable has nothing to do with whether using it is OR. It's OR if you want to use it to make a point, advance an argument, settle a dispute, or say anything that involves evaluation or judgement. All of that has to come from secondary sources. That's the policy we want to define here, and I think we should try to define it as clearly and as simply as possible. – Margin1522 (talk) 19:58, 6 April 2015 (UTC)

PBS, The phrase “reliably published” is not defined and is thus not useful for your purpose of excluding material from unindexed archives. If you want to exclude such material, you can do that by adding to footnote 6 in WP:SOURCE the word “indexed” so that footnote 6 becomes the following (where I have underlined the addition just for this discussion).
This includes material such as documents in publicly-accessible indexed archives, inscriptions on monuments, gravestones, etc., that are available for anyone to see.
I can’t say whether or not I would support that because I don’t know enough about archives or indexing them, but I would be willing to read arguments from both sides. --Bob K31416 (talk) 02:02, 7 April 2015 (UTC)
Margin1522 given the phrasing in the lead "Because these policies work in harmony, they should not be interpreted in isolation ...", reliably published is obviously referring to that part of WP:V in the section WP:SOURCE which is about the publication of "Reliable sources". As you will know, as you have been edition since 2008, the phrase has been in the policy since before you arrived on the scene and in its current form for just over five years, that it refers to just one definition at WP:SOURCE not another and certainly not one of "15 different interpretations of 'reliably published'". If you think it necessary (although I do not) I would not object to a specific link to WP:SOURCE. -- PBS (talk) 13:10, 7 April 2015 (UTC)
Bob K31416 I ought to have used the phrase "catalogued", rather than "indexed", I do not think that the change you suggest is necessary because in practice no uncatalogued archive will be open to the public. Also in practice it is reinforced with the bullet point in WP:EXCEPTIONAL "challenged claims that are supported purely by primary...". My objection to the definition in "WP:SOURCE" is the "inscriptions on monuments, gravestones, etc., that are available for anyone to see." which is beyond the scope of this conversation, as that is an issue for WT:V. -- PBS (talk) 13:10, 7 April 2015 (UTC)
PBS, Re your comment, "in practice no uncatalogued archive will be open to the public" — If that's the case, then the example in your opening message has a false premise because according to WP:SOURCE, material from an uncatalogued or unindexed archive is not "published" because it is not "publicly-accessible" and "Unpublished materials are not considered reliable." For reference, here's the false premise in the example from your opening message, "For example a letter from Winston Churchill in an archive that has not been indexed would still be a reliable primary source". So I have restored the phrase "reliable primary source" [1] by reverting your recent change. --Bob K31416 (talk) 14:04, 7 April 2015 (UTC)
It is not a false premise because of the assumption around WP:SOURCES is that reliability consists of three parts one of which is the author and their expertise to a particular subject, the issue is that a source that would otherwise be reliable is not reliable if it has not been reliably published. Without this association to reliably published (however that is defined in WP:V) we end up with OR and V in a position where they can be seen to contract each other. "primary sources that have been reliably published may be used in Wikipedia" is different from "Reliable primary sources may be used in Wikipedia". -- PBS (talk) 14:30, 7 April 2015 (UTC)
Another option is to alter the sentence to read "Published primary sources..." -- PBS (talk) 14:38, 7 April 2015 (UTC)
I don't think so. I'll be leaving the discussion now. The current status of the policy is that PBS made an edit that reverted to his previous edit that put in "reliably published" after this phrase had been replaced following discussion in June 2014. --Bob K31416 (talk) 15:18, 7 April 2015 (UTC)
I think that if the wording you prefer is to be used then there should be an RfC as the wording to which I reverted has been in the policy for a decade or more and it was changed by a handful of editors without wider input. I have only looked at it recently because I was involved in a conversation where reliably published is of far more use than reliable primary source because of the ambiguities involved in the a meaning that has three parts. -- PBS (talk) 15:28, 7 April 2015 (UTC)

It seems to me that "published" is being used here as a synonym for "verifiable", rather than in the vernacular sense of being published. As such, I think the word "published" should be deprecated in favor of "verifiable" if it needs to be specified. However, I concur with Margin1522 that it would suffice to say only "Unless restricted by another policy, primary sources may be used in Wikipedia." This does not need to be an all-singing all-dancing policy that encompasses WP:V and WP:RS within itself. Situations like the opinions of a non-notable lab tech are addressed separately by considering due weight and WP:ONUS. Rhoark (talk) 19:57, 7 April 2015 (UTC)

WP:VERIFY discusses the reliability of published sources separately, so these two are different things. For example,

Attribute all quotations and any material [...] to a reliable, published source -WP:BURDEN

I don't really see how these two, "verifiable" and "published", could be used as synonyms. Jayaguru-Shishya (talk) 16:53, 8 April 2015 (UTC)
What is published? Is the inscription on a public monument a publication of that text? It is however verifiable... and a primary source.
Is a top secret report, produced in several hundred copies by a secret service a publication? As long as that remains top secret it is not verifiable.
Adding the text of the monument should be fine within Wikipedia standards (as it is verifiable); adding text from the top secret report not. So regardless of publication status, I go with Rhoark here that (at least for primary sources) verifiability is the key word, not reliability of publication. 17:24, 8 April 2015 (UTC)
Let us suppose that the Churchill letter is sold at auction, and that Christie's experts verify that it is a letter by Winston Churchill, but the content is not published and it goes into a private collection. I can verify that the letter is by a reliable source (Churchill) and therefore it is a reliable primary source, but it has not been published. So the part of verifiability we are addressing is its publication. -- PBS (talk) 02:14, 9 April 2015 (UTC)
I agree that an unpublished letter by Churchill should not be used. We should address that issue somewhere. But the number of editors who want to use unpublished sources is small, and the number who don't understand the difference between primary and secondary is large. From the cites on talk pages and RSN, most many Wikipedia editors seem to be under the mistaken impression that WP:PRIMARY means "having a prior opinion on a controversial subject" or "affiliated with the subject" and WP:SECONDARY means "neutral" or "third party" or "no prior opinion". This is wrong. Bringing in the publisher just confuses the issue that editors need to understand, by diverting their attention to other issues that they are keenly interested in and eager to talk about.
Especially since in ordinary English "reliably published" means that the act of publishing is reliable. Something like "no missing pages", or "delivered to your doorstep every morning, without fail, regardless of the weather". That's not what we're trying to explain here. Nor are we trying to explain WP:V or WP:NPOV. On this page, which is about the third major policy, we have already spent several paragraphs explaining the other two – WP:V and WP:NPOV. Including several paragraphs in the lead. Now can we please move on to explaining Wikipedia's third major policy, which is WP:No original research? – Margin1522 (talk) 09:51, 9 April 2015 (UTC)
"the number who don't understand the difference between primary and secondary is large" then that is something to do with the definition part that starts "Primary sources are original materials that are close to an event...". "the number of editors who want to use unpublished sources is small" but those who do want to use unpublished sources are often tenacious. "[in] ordinary English 'reliably published' means..." if you say so, but then in the same paragraph it says "a reliable secondary source for that interpretation" would you suggest removing the reliable in front of secondary source as well? If you do then people will argue, not unreasonably, that a unreliable sources is OK to use, as else where in the page the use of secondary sources are preceded by reliable. What makes you think that a person reading this page is going to use two different interpretations of reliable in the same paragraph? As I said above if that phrase is a problem for you we can link to WP:SOURCE which defines what is reliably published. -- PBS (talk) 14:39, 10 April 2015 (UTC)
Yes, well, there are disruptive editors on Wikipedia, people who don't want to follow the policies even when they understand them. Part of the reason why the policies sound so overbearing and distrustful of the reader is that they are written like an extended argument with people who simply don't want to listen. Hammering home the same point over and over. This policy uses the word "published" 28 times, and the word "reliable" 38 times. If editors still want to use unpublished materials, it's not going to help to tell them one more time. We do have a link to WP:SOURCE, in the "Reliable sources" section, plus a link to Wikipedia:Identifying reliable sources. My point is that the "Primary, secondary and tertiary sources" is supposed to be about something else. Namely that there are three different types of sources, and that the differences are important even when everyone agrees that the sources are published and reliable. As for deleting "reliable" from the rest of this paragraph, sure, I think that would help. Including it in doesn't get you very much, because basically "reliable" is an extremely broad category that includes every source that is safe to use on Wikipedia. It contributes nothing to understanding the difference between a source that includes analysis and interpretation and one that doesn't, and in fact hinders understanding by bringing up a different topic. – Margin1522 (talk) 10:49, 11 April 2015 (UTC)
As the discussion seems to be over, I went ahead and simplified this paragraph to focus on the primary/secondary distinction. There is no disagreement that all sources need to be reliable – this is purely a matter of writing style. So if anyone prefers a different style they can go ahead and do that instead. – Margin1522 (talk) 19:16, 16 April 2015 (UTC)
I am sorry, but I have to disagree with the last edit.[2] As far as I am concerned, "primary sources" are different from "reliably published primary sources", as well as "secondary sources" are different from "reliable secondary sources". See, for being "reliable", the source needs to be an independent third-party source, i.e. not just any source that is "published" will do. That's being explicitly pointed out in WP:V and WP:RS.
For example, it is very different if you keep record of the daily temperatures and publish your findings yourself rather than, let's say, a University Press will publish those. That's the case for primary sources. Also, if you conduct a study that summarizes several scriptures of Tibetan Buddhism, if you may, again it's different thing if the source is published by oneself rather than a well-standing third-party institution. As per WP:V and WP:RS, I think the "reliability" is of the utmost importance here. Cheers! Jayaguru-Shishya (talk) 16:36, 17 April 2015 (UTC)
Completely agree with you. New editors are frequently pointed at WP:NOR and is important that they fully understand that something like someone's blog is not a reliable source. Otherwise we'll have, "someone wrote this on the Internet!". --NeilN talk to me 16:45, 17 April 2015 (UTC)
User:Jayaguru-Shishya, as explained in his talk page edit of 16:36, 17 April 2015 (UTC), is not in agreement with policy. It is totally unacceptable to mash together the separate concepts "self-publshed", "third party", "primary", and "secondary". Policy and the WP:Identifying reliable sources guidelineallow the use of self-published sources in some situations. Primary sources are allowed in some situations. I reject User:Jayaguru-Shishya's edit. Jc3s5h (talk) 18:16, 17 April 2015 (UTC)
@Jc3s5h, Jayaguru-Shishya's edit was a revert of the edit made by Margin1522 immediately before it. -- PBS (talk)

────────────────────────────────────────────────────────────────────────────────────────────────────Margin1522 you wrote "As the discussion seems to be over," The discussion was over but that did not mean that I agreed with your proposed change, which you then implemented. For example you totally ignored the point I made with a question "would you suggest removing the reliable in front of secondary source as well? If you do then people will argue, not unreasonably, that a unreliable sources is OK to use, as else where in the page the use of secondary sources are preceded by reliable." So what made you think there was a consensus for the change you made with Revision as of 19:11, 16 April 2015? -- PBS (talk) 19:25, 17 April 2015 (UTC)

Adding a citation to a secondary source requires adherence to all policies, not just the one given before the phrase "...secondary source for that interpretation." If you say "Any interpretation of primary source material requires a reliable secondary source for that interpretation" (emphasis added) I could argue it's OK to add the reliable secondary source that I wrote myself, even though my primary motive was to sell more copes of my book, because you forgot to write "Any interpretation of primary source material requires a reliable secondary source for that interpretation which is cited by an editor who does not have a conflict of interest." Jc3s5h (talk) 19:56, 17 April 2015 (UTC)
I am going to copy this down to the new section you have created at the bottom of the page. As we are now splitting this second issue over two sections. -- PBS (talk) 20:28, 17 April 2015 (UTC)
@PBS: I thought I made it clear that, in my opinion, it would indeed be unreasonable for editors to argue that it's OK to use unreliable sources. Very unreasonable, given the amount of space we have already devoted to driving that point home. But it seems that my objections to the phrase "reliably published" still haven't gotten through. Let me explain one more time, and sorry if this sounds pedantic.
In a word, nowhere except in Wikipedia jargon does the phrase "reliably published" mean "published by a reliable source". In normal English, "reliably published" means that the publishing is reliable. This is because "reliably" is an adverb. Adverbs modify verbs, and in this case is the verb is "published". Therefore it is the publishing which is reliable.
How can publishing be reliable? Here are three examples from the first few pages of Google results (excluding examples from Wikipedia), which show how.
In each of these cases, the publishing is reliable because the material appears on schedule. So if we say that sources should be "reliably published", we are saying that it's OK to use the National Enquirer, because however unreliable it may be as a source, it is at least "reliably published", in that it appears in the magazine rack every morning, without fail. You can count on it.
Now, you may say, "No no, that's not what I meant." But as an editor (in the normal sense of the word) my job is to have no patience with that and not care what what you meant. The only think I care about is what you said. If the text that you wrote says something other than what you meant, you should fix the text. We are writing an encyclopedia here and are supposed to be literate people. So we should at least be able write our policies in language that doesn't do violence to the normal rules of English. – Margin1522 (talk) 20:41, 17 April 2015 (UTC)
Margin1522, you said: "I could argue it's OK to add the reliable secondary source that I wrote myself, even though my primary motive was to sell more copes of my book". No, I am afraid that's not possible. For those we have WP:SELFPUBLISH and WP:UGC, according to which: "Anyone can create a personal web page or publish their own book, and also claim to be an expert in a certain field. For that reason, self-published media [...] are largely not acceptable as sources." Cheers! Jayaguru-Shishya (talk) 15:55, 18 April 2015 (UTC)
Jayaguru-Shishya, your dismissal of a book written by a wikipedian incorrectly presumes the wikipedia published the book through a vanity press; I know of at least one wikipedian who has published trough well-known publishers, and at least one other who has published in scholarly journals. Jc3s5h (talk) 16:05, 18 April 2015 (UTC)
Reliable has more than one meaning when used as an adjective and according to the OED you are using one from a statistical derivation. It use in "reliably published" is 1.a of the OED definition:
A. adj.
1. That may be relied on.
a. Of a person, information, etc.: able to be trusted; in which reliance or confidence may be placed; trustworthy, safe, sure.
b. orig. U.S. Of a product, service, etc.: consistently good in quality or performance; dependable.
2. Statistics. Originally: accurate; free from error. In later use: (of a method or technique of measurement) that yields consistent results when repeated under identical conditions.


As I said if you think that is ambiguous phrase we can link it to WP:SOURCE which defines for better or worse what "reliably published" means in Wikipedia policy. --PBS (talk)
Again, sorry for being pedantic, but "reliable" is an adjective. "Reliably" is an adverb. If you want a word that modifies a noun (a source) use the adjective. If you want a word that modifies a verb, use the adverb. We are not arguing about policy here; this is grammar. – Margin1522 (talk) 21:30, 17 April 2015 (UTC)
OED reliably, adv. "In a way that may be trusted or relied on; in a reliable manner; dependably." -- PBS (talk) 22:13, 17 April 2015 (UTC)
Fine. But as the examples I gave above show, "dependably published" means that they meet their deadlines. In normal usage it has nothing to do with fact checking or independence. It is the publishing that may be relied upon, not the publisher. If we want to say that the publisher may be relied upon, let say that. I bring this up because in an earlier discussion the normal English wording "reliable, published sources" was suggested and rejected in favor of the grammatically incorrect "reliably published". I don't know why. Perhaps because it left the publisher's fact-checking process out? If that's the reason I can understand that. The problem is that readers won't understand it. – Margin1522 (talk) 23:17, 17 April 2015 (UTC)
If you think that is ambiguous phrase we can link it to WP:SOURCE which defines what "reliably published" means in Wikipedia policy, this will allow editors, who are not aware of the [WP:SOURCES]] section in WP:V to follow the link. -- PBS (talk) 20:00, 18 April 2015 (UTC)
@Jayaguru-Shishya: I completely agree that a "reliable primary source" is not the same as a "primary source". However, that is irrelevant in this paragraph, because in this paragraph we are not about talking about how one primary source is different from another primary source. We are talking about how all primary sources differ from all secondary sources. Do you see the difference?
I am especially eager to make this clear because it is a very common misconception to think that "primary" means "not independent" and "secondary" means "independent". No. That is not what it means. It says in the policy "Primary sources may or may not be independent or third-party sources." (My emphasis). Is that OK? Do you agree with that? And do you still think that we should insist that independence is a key attribute of "primariness"? – Margin1522 (talk) 21:16, 17 April 2015 (UTC)
The paragraph you altered is not just about how primary sources differ from secondary sources, the paragraph is also defining how to extract information from a primary source, and that involves using reliable secondary sources which is a a subset of all possible types of secondary sources. -- PBS (talk) 22:13, 17 April 2015 (UTC)
Greetings Margin1522! I never said that "independence is a key attribute of primariness". I clearly made the distinction in my post earlier[3], distinguishing between self-published primary sources and independent primary sources, as well as self-published secondary sources and independent secondary sources.
As you agreed, a "reliable primary source" is different from a mere "primary source"; the previous is consistent with the other WP policies, whereas the latter creates confusion over the matter. We should use the concise one that is less ambiguous and communicates the WP policy better. Cheers! Jayaguru-Shishya (talk) 16:00, 18 April 2015 (UTC)
So you so want the Original Research policy to depend on the Verifiability and Neutrality policies. Among policies, a second-class citizen, so to speak. Can I give an example? Suppose we have an editor who wants to interpret Buddhist scriptures. We say, no you can't do that because that's original research. You need to get your interpretations from a secondary source. So he quotes from a book written by a scholar of Buddhist theology. Is that OK? No, it's still not OK because we have edit warriors on Wikipedia who object to the fact that this book was published by the publishing arm of a Buddhist new religion, and they disagree with the theology. That is, they have Neutrality concerns. This is the kind of problem that comes up when you mash all the policies together. The editor thought he was quoting a secondary source, which should be OK, and all of sudden we're embroiled in disputes over Buddhist theology. – Margin1522 (talk) 23:40, 17 April 2015 (UTC)
If the scholar's book is published by an independent publisher (e.g. some university press), then it's OK. If it's published through some religious affiliation movement, then we should look for a reliably published source instead. Cheers! Jayaguru-Shishya (talk) 16:20, 18 April 2015 (UTC)
Jayaguru-Shishya's position is untenable. If religious ideas can't be mentioned unless they are published by publishers with no religious affiliation, Wikipedia becomes hostile to religion. Jc3s5h (talk) 16:25, 18 April 2015 (UTC)
This seems to be more a matter of neutral point of view or even not accepting fringe theories than of original research. If we start to prefer some publishers over others, I can also argue that any US source talking about the US is almost certainly biased and should for that reason not be used. Let's keep the policies separate. Arnoutf (talk) 16:45, 18 April 2015 (UTC)
Sources with some religious affiliation movement can be only used to a limited extent, and independent scholarly sources should be preferred whenever available. I've been editing Buddhism-related articles quite a bit, and I am not hostile to religion. Trust me, there's plenty of academic research on religious issues available.
Ps. There was a discussion at WikiProject Buddhism tangenting this subject. Anyway, I believe we are getting off track here, and this sort of discussion would belong to WT:IRS. Cheers! Jayaguru-Shishya (talk) 16:48, 18 April 2015 (UTC)
This is really less straightforward than you assume. What do we do with Christian Universities? Should we exclude them and label the work of their scholars as unreliable? That would be weird, we would e.g. be forced to disregard any work of VU University Amsterdam which is ranked 144 best University in the world (!) in the Times Higher Education index. Arnoutf (talk) 16:57, 18 April 2015 (UTC)
Exactly. This discussion belongs at WT:IRS. I completely agree with the comment by the closer of the (discussion at WikiProject Buddhism): "Where the source makes analytic or evaluative claims of other primary sources, it is a secondary source. It may or may not, upon its own merits, be a reliable secondary source." You can call it unreliable for any reason you like, but if makes analytic or evaluative claims it is Secondary. Period. Among editors who understand it, this policy is not controversial. It becomes controversial only when Secondary and Primary are misused as synonyms for "reliable" and "unreliable". – Margin1522 (talk) 19:17, 18 April 2015 (UTC)
@Arnoutf: Ooops...! I did mean to speak about "religious movement", not a "religious affiliation"! I wonder how I was so inconsiderate as I strongly emphasized that "the religious affiliation of a scholar should not matter" in the RfC at WikiProject Buddhism.
Anyway, here is my answer in short: If there is a scientific piece of work that has been published by some reliable institution, such as an university press, I don't really see a problem there. But if the same work is released through some revivalist movement's own publishing house, if you may, that's a whole different case already. Cheers! Jayaguru-Shishya (talk) 08:52, 19 April 2015 (UTC)

Twitter and other online posts as a primary source.[edit]

One thing I've seen come up several times recently is people using quotes from Twitter and other online forums where an article's subject has posted in order to make implicit statements about their beliefs, history, and so on. For instance, they might dig through someone's twitter feed until they find a politically-controversial statement, then put it in the article with the quote on Twitter itself as the only cite. I feel that this is likely to become common enough in an internet age that we should have a specific rule or guideline referencing it -- as a general rule, I think that quotes like that should require a reliable secondary source to establish relevance. (Ideally one qualified to be used as a source for statements of fact, although opinion pieces could be cited when they're high-profile and the opinion is attributed in text.) The issue is that without this, Wikipedia becomes a place for people to perform opposition research about figures they disagree with, to use quote-mining to make their own argument about how some popular figure holds a particular position, or similar things -- the sheer volume of quotes available in an electronic age (and the amount of searching that can be done with crowdsourced efforts) means that it is relatively easy to find quotes to support your own personal reading on a public figure, even when that reading has no coverage in reliable secondary sources. If a quote is genuinely significant, it should always be easy to find a reliable secondary source reporting on it, shouldn't it? --Aquillion (talk) 22:43, 5 April 2015 (UTC)

Selective quoting is certainly a problem when it occurs. I think the relevant policy is really WP:NPOV rather than this one. I wouldn't want to make a blanket rule against sourcing something to someone's twitter account. If it's uncontroversial, why not? (There's an example of reasonably appropriate sourcing to twitter in Anthony Head's article.)—S Marshall T/C 23:24, 5 April 2015 (UTC)
I agree with Aquillion "If a quote is genuinely significant, it should always be easy to find a reliable secondary source reporting on it". So I do not thin that quotes from "Twitter and other online forums where an article's subject has posted in order to make implicit statements about their beliefs, history, and so on." should be used, unless they are republished in a reliable source. If this seems unreasonable could someone give some examples of where it is unreasonable from examples in Wikiepdia articles (I don't think the example given helps improve that article). -- PBS (talk) 09:39, 6 April 2015 (UTC)
I totally agree. Another factor is that someone could be posting ironically (or drunkenly, or whatever), or making what they hope is a joke—such comments should not be used in an encyclopedic article. A secondary source is required to interpret comments and provide context. Even something as apparently simple as "I was born on April 6, 1994" could be a joke or even a mistake due to a typo. Johnuniq (talk) 10:03, 6 April 2015 (UTC)
Note that the tweet above if used properly can only be included as something like "X once stated to born on April 6, 1994 REF TWITTER" as anything beyond that would most likely be synthesis. So I am not sure we should write a policy on this, although we may emphasise that statements on social media should be considered with caution. On the other hand we would not want to exclude a well thought through blog post by an expert out of hand either..... Arnoutf (talk) 15:32, 6 April 2015 (UTC)

I'm from Argentina, and the president usually post things on Twitter or Facebook, to avoid press conferences. When this happens, the newspapers report the things she said, what other politicians think of the things she said, etc. It is perfectly possible to make an article "Cristina Kirchner on social media", without a single direct link to Twitter or Facebook. Cambalachero (talk) 16:51, 6 April 2015 (UTC)

True, but the source would be the newspaper; a secondary source; not the tweet itself, the primary source. Arnoutf (talk) 17:02, 6 April 2015 (UTC)

Synthesis is synthesis, whether or not it uses Twitter. As to the expectation that anything worthwhile has a secondary source, the same theory could be wielded against any primary source on any topic, so it should be plainly clear that that is an unworkable guideline. There are many facts of encyclopedic interest that are not repeated in a newspaper or academic journal, because those things are not encyclopedias. As always, it is the role of Wikipedia editors to form consensus on what belongs in Wikipedia - that is all the defense that is needed against the types of abuse or error described above. Rhoark (talk) 17:08, 6 April 2015 (UTC)

All policy must fit in one sentence[edit]

The title of this thread is a bit facetious. But in this edit User:Jayaguru-Shishya reinserts words into the policy in yet another spot that sources must be reliable. Apparently Jayaguru-Shishya feels every sentence must mention every rule that applies to the subject matter of the sentence. The logical conclusion of this approach is that we must get rid of all our current policies, and instead write one sentence that contains every bit of Wikipedia policy. Jc3s5h (talk) 17:13, 17 April 2015 (UTC)

@Jc3s5h: And apparently you think it's okay to introduce a loophole that can be easily fixed. If someone objects to a change in policy wording, the solution is not to revert it back in but either join the discussion (one exists above) or start a new one. --NeilN talk to me 17:57, 17 April 2015 (UTC)
There is no loophole; sources must be reliable, even though that requirement is not repeated in every single sentence of the policy. Jc3s5h (talk) 18:05, 17 April 2015 (UTC)
If the policy page uses the term "reliable secondary source" in most sentences and paragraphs, but does not use it in others, then a person reading it for the first time will assume that there is a significant difference and for those sentence where "secondary source" is not proceeded with reliable will not unreasonably assume that in those situations all secondary sources, both those that meet the level of reliability as defined in WP:V and those that do not, are acceptable. It is therefore necessary to always qualify "secondary source" with reliable so that this policy is non-ambiguous with regards to sources. Far from getting rid of all our current polices, it is reinforcing the concept that WP:NOR relies on WP:V for what is an acceptable source instead of redefining it in this policy. -- PBS (talk) 19:45, 17 April 2015 (UTC)
Take for example this sentence "Any interpretation of primary source material requires a reliable secondary source for that interpretation." which was changed to "Any interpretation of primary source material requires a secondary source for that interpretation."the alteration states that any secondary source will do it does not have to meet the requirements of WP:SOURCE (a subsection of WP:V) -- PBS (talk) 19:50, 17 April 2015 (UTC)
Copied the following comment by Jc3s5h from above into this section as this section is more specific to this issue. -- PBS (talk) 20:30, 17 April 2015 (UTC)
Adding a citation to a secondary source requires adherence to all policies, not just the one given before the phrase "...secondary source for that interpretation." If you say "Any interpretation of primary source material requires a reliable secondary source for that interpretation" (emphasis added) I could argue it's OK to add the reliable secondary source that I wrote myself, even though my primary motive was to sell more copes of my book, because you forgot to write "Any interpretation of primary source material requires a reliable secondary source for that interpretation which is cited by an editor who does not have a conflict of interest". Jc3s5h (talk) 19:56, 17 April 2015 (UTC)
Wikipedia:Conflict of interest is a guideline not a policy, so there is no reason to include it, nor any other guideline (other than to help explain policy). This is one of three content policies and they ought not to contradict each other. One way of helping to keep possible contradictions to a minimum is by using clearly defined phrase such as "reliable secondary source". Also one has to read it as a reasonable person would, and if "secondary source" and "reliable secondary source" are used inconsistently, it would be reasonable to assume that the inconsistency is significant. Your hypothetical inclusion would only become significant in this respect if its use was inconsistent in this policy which is not the case. -- PBS (talk) 20:43, 17 April 2015 (UTC)
No, when we have just spent an entire section insisting that all sources must be reliable, I don't think that a reasonable person will assume that unreliable sources may be used because we don't insist one more time. Absolutely not. Sorry, but I'm not going to give up on this because another editor is more stubborn tenacious than I am. I really, really want this policy to have a clear explanation of the difference between primary and secondary sources, and I want it to be very clear that the lazy and very common misconception "secondary = reliable" is wrong.
How about if we compromise? If you allow me to delete the distracting and irrelevant "reliable" from "reliable secondary source", I will insert the following, from the essay Wikipedia:Identifying and using primary and secondary sources, which we cite in this section of the policy: "Secondary does not mean that the source is independent, authoritative, high-quality, accurate, fact-checked, expert-approved, subject to editorial control, or published by a reputable publisher. Secondary sources can be unreliable, biased, self-serving and self-published.". And then we can insert a statement, do not use those sources. It can go anywhere you like, in the text or in a footnote. But I would like it to be separate from the explanation of what makes a secondary source secondary. – Margin1522 (talk) 15:52, 18 April 2015 (UTC)
Are you deliberately using rhetorical trick with this statement If you allow me to delete the distracting and irrelevant "reliable" from "reliable secondary source", or is it an accident? I do not thing that the use of "reliable" is irreverent. So do you mean "If you allow me to delete what I think is the distracting and irrelevant 'reliable'..." my answer is that is I think it is a fundamental alteration to policy and that you need to see if there is a consensus via an RfC for such a change. If you were to include such a sentence in this policy, then what you would do is set up an alternative to WP:SOURCE as the definition of what is a reliable secondary source. I think such a step is a bad idea as the policy definition of a reliable source ought to remain in WP:V alone. As to your statement "I really, really want this policy to have a clear explanation of the difference between primary and secondary sources, and I want it to be very clear that the lazy and very common misconception "secondary = reliable" is wrong". (1) I do not think it is a common misconception (how large is your sample size). (2) The whole point of using the formulae "reliable secondary sources" is to differentiate between "reliable secondary sources" (per WP:SOURCE) and "unreliable secondary sources". Reliable ones can be used to interpret primary source, unreliable secondary sources may be used not. I think that the removal of reliable would do exactly what you say you wish to avoid and cause the "misconception [that all] 'secondary = reliable'" or at the least that any secondary source (not just reliable ones) can be used to interpret primary sources. -- PBS (talk) 19:53, 18 April 2015 (UTC)
WP:SOURCE has a lot to say about the difference between reliable and unreliable sources. That is fine. Reliability is the topic of that section of WP:V. The WP:SOURCE section of WP:V does not use the phrase "reliable secondary source" because defining "secondary" is not the responsibility of that policy. Defining "secondary" is the responsibility of this policy. Before we can say whether a secondary source is reliable or not, we have to define what a secondary source is. Can we start at the beginning? A very basic question. Since you are so sure that you know what a secondary source is, could you please define it for us? – Margin1522 (talk) 22:40, 18 April 2015 (UTC)
The definition of what a secondary source is here. The definition of what is reliable is there (WP:V). If you see a problem with that then you are 8 years too late see WP:ATT and Wikipedia talk:Attribution/Archive 14#This merger is a really bad idea and Wikipedia:Attribution/Poll. With nearly 800 participants expressing an opinion for and against and another 102 undecided that was probably the biggest attendance of a poll ever held on Wikipedia, and I suspect few want to go through that again (I was in favour of a merge, but not the way it was done, so I did not express an opinion either for or against it). This is why this policy last sentence after defining what a secondary source is states "Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source" (my empasis).-- PBS (talk) 17:12, 20 April 2015 (UTC)
What? I am not in favor of merging V and OR. They are different concepts. Merging them would be a terrible idea. You are the one insisting that if we define OR without mentioning V, editors will feel free to ignore V. Which I think is wrong and certainly not helpful.
Earlier you asked for example of this policy being misunderstood. Here is an example from a just a few hours ago. An editor claimed that a source was biased and therefore not a "reliable secondary source". So he assumes that since it's not a "reliable secondary source" it must be a "primary" source. Which is wrong. This is exactly the kind of misunderstanding that I would like to fix, and it can only be fixed by emphasizing that "reliable" does not mean "secondary", and "unreliable" does not mean "primary".
Since "reliable" is a different concept from "secondary", it would be helpful if we could keep it out of the definition of "secondary". That's all I'm asking. Later, once this is understood, we can insist that secondary sources must also be reliable before they can be used in Wikipedia. We can go on at great length, I'm fine with that. But let's keep it out of the definition. – Margin1522 (talk) 21:48, 20 April 2015 (UTC)
I think you are doing NorthBySouthBaranof a disservice as I do not think that NBSB makes the mistake you claim. Besides a primary source can be used to express an opinion a serach of this talk pages's archive on "Wellington" brings up long conversations on this very point.
The definitions of Primary Secondary and Tertiary do not mention the word reliable. It is how the are applied in policy were the word reliable is added. There are two good reasons for this. The first is that in discussion on talk pages such as the one to which you link, it helps if the wording in the section is unambiguous and can not be read in such a way that it contradicts another policy (as all polices are equal). It is no good arguing that the overall policy does not contradict another, because people often discuss by the linked section not by the total content of the policy. The second reason is because all polices are equal if one can be read to contradict another then much time is wasted arguing which is the correct interpretation. Can we go back to the sentence "Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source" (my empasis) and can you explain why you think the word reliable is superfluous because I think that the meaning of the sentence is altered by its removal. -- PBS (talk) 09:23, 21 April 2015 (UTC)
The fact that the source is biased is not what makes it a primary source, in my estimation. Rather, it is the fact that the source is written by someone who is effectively directly involved in the dispute - it is the corporate perspective of the company involved in the argument, and there is no detachment from the argument or competing analysis of whether or not this perspective is honest, fair or well-founded in fact. These are the hallmarks of a secondary source — a newspaper, magazine, journal article, scholarly book/monograph, journalistic website/blog, etc. NorthBySouthBaranof (talk) 15:13, 21 April 2015 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────I think the sentence "Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source." [emphasis added] has a bigger problem than the word reliable. PBS is concerned that if we leave out "reliable" people will think this is an exception to the other sections of the policy that say all sources must be reliable. I think it's pretty obvious that the sentence must be interpreted in the context of the rest of the policy, and it's obvious the secondary source must be reliable. After all, the sentence requires the reader to make a more subtle contextural inference: since the preceding paragraph describes what secondary sources do, the sentence in question is telling editors they are not allowed to make the sort of analysis and evaluation that authors of external secondary sources may do. But an editor who failed to make that inference would think that a primary source must not be used if it makes analysis or evaluation of the author's own work. So the editor who is incapable of reading a sentence in context might think that since the National Geodetic Survey performed a variety of experiments and reported "a new architectural height of the Washington Monument using the CTBUH standard as: 554 feet 7 and 11/32 inches +/- 1/32 inches" because the NGS report is a primary source, the NGS made an analysis of their own measurements, and analysis can't be put in Wikipedia unless it comes from a secondary source. Jc3s5h (talk) 16:04, 21 April 2015 (UTC)

I'd like to remind NorthBySouthBaranof that the name of this policy is "No original research". Wikipedia editors can't make their own original analytical claims. Not original. That's all. There is no requirement in the "No original research" policy that claims be honest, fair or well-founded in fact. The claims can be dishonest, unfair, and factually off the wall. The "No original research" policy doesn't care, as long as they aren't original.
But, but, but... you protest. What's the use of a policy so spineless that it doesn't care if a claim is honest or not? Very simple, it prevents editors from writing essays that describe their own opinions, their own discoveries, their own interpretations of the world. There are many editors who want to do that and we need a policy to tell them why they can't.
So what about honest, fair or well-founded in fact? Don't we care? Yes we do care, and that's why we have the other two core policies, namely Wikipedia:Neutral point of view and Wikipedia:Verifiability. Those are separate policies, and they have their own words. If you want to talk about those concerns, you should use the words from those policies, not "primary" and "secondary", which are words from this policy. – Margin1522 (talk) 22:19, 21 April 2015 (UTC)
On third thought, I think I am going to have to concede that the author of this blog post is a primary source insofar as he is writing about himself and his own Twitter feud. NorthBySouthBaranof was correct about that, so I should concede that much. Sorry. I still don't want to concede that a source can be labelled primary because of the source's affiliation, political or religious beliefs (e.g. the Dalai Lama's beliefs about Buddhism), perceived accuracy, or general "reliability". Those things probably account for 90% of the usage of "primary" on talk pages and notice boards, and my position is that this usage is wrong. Most of the blog posts that editors attempt to cite are not talking about the author's personal experiences. Usually they are commentary on something else. They should not be labelled "primary" simply because they are self-published and not fact checked. They can be rejected as unreliable for those reasons, but not labelled "primary". – Margin1522 (talk) 17:17, 22 April 2015 (UTC)