Wikipedia talk:No original research

From Wikipedia, the free encyclopedia
  (Redirected from Wikipedia talk:NOR)
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.

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)

Just because something is published in a book does not make it reliable. I have seen many items by one professor in his books with no citation and can prove by citable sources that his "facts" are untrue. He does not cite to sources because he cannot; they do not exist!

I agree witht the person who has criticized this article on the basis of the “best evidence rule,” a legal principle. As an attorney, I have serious questions about the intellectual value of this policy. Surely, the best evidence is original research (unless we have a disagreement about the definition). Here is a specific example. I discovered a newspaper article about convicted pickpocket Abraham Greenthal, then using the alias of Myers, and attacking then Gov. (and President-elect) Grover Cleveland of New York for “pardoning” Greenthal. Here is but one of several citations: Wednesday, 10 December 1884, The Evening Journal, Jersey City, NJ, p. 4, col. 3. So I wrote to the New York State Archives for a copy of the “pardon.” What came back was a copy of his “commutation of sentence,” clearly so labelled, not a pardon, with the autograph signature of Gov. Grover Cleveland on it. See New York State Archives, Record Group A0597, Box 31, Folder 12. The folder contains an Application for Pardon cover sheet for Abraham Meyers. The pardon application was denied on Wednesday, 26 September 1883, but granted as a commutation on Friday, 16 May 1884. Meyers was one of Greenthal’s many known aliases. See, for example, Sunday, 26 November 1871, New York Times, “Grenthal Caught,” p. 3, col. 3, in which it states that he identifiied himself as Abraham Myers. Surely that is the best evidence of what happened and not the newspaper articles written by political enemies and ignorant copiers of what others wrote. There is a huge difference between a pardon and a commutation of sentence. When a pardon is issued, all citizenship rights are restored; when a commutation of sentence is issued, the convict remains a convicted criminal with no right to vote (at that time), and in this case, since it was at least his fourth felony conviction, the law enforcement community could simply stop, search, and detain the convict without a showing of probable cause. That was the law in New York State at the time. It was the Republican press that called the document a pardon, probably for political reasons. The rule, as expressed in the article shows a lack of understanding of what is the best evidence. Surely, it was a good thing that I obtained the original document, using “original research.” So far as I know, the "original document" has never been published, but no one can credibly deny that Gov. Grover Cleveland legally "uttered" it; it is the best evidence of the facts. Either the article is using words in a tortured manner or the concept is faulty altogether. The policy needs to be amended to take the best evidence rule into account. (talk) 11:45, 30 May 2015 (UTC), you didn't issue the commutation of sentence. You didn't interview Grover Cleveland and ask him about it. You relied on a source that is available to the public, so for Wikipedia purposes, is published. You didn't engage in original research, you engaged in source-based research. Also, old newspaper articles are often considered primary sources, so all the sources you found could be thought of as primary.
Also, your phrase "I agree witht the person who has criticized this article" is confusing. Wikipedia articles are the finished products we present to our readers. When you look at an article, the title will just be the title; it is in what we call the main namespace. The "Wikipedia:No original research" policy is not an article, it is a project page, and it is in the "Wikipedia:" namespace. I hope I have correctly understood which page you are referring to. Jc3s5h (talk) 12:14, 30 May 2015 (UTC)

Can we really state that x book doesn't contain Y statement with no source other than the book?[edit]

If we can't, I'd like something explicit in policy, as I'm being told it's basically ok as you can read the book. Thanks. Dougweller (talk) 15:19, 7 May 2015 (UTC)

Isn't this too specific/detailed an issue to include in such an high level policy. Arnoutf (talk) 17:42, 7 May 2015 (UTC)
No, it's not too specific. And yes, we can say that. Books (and TV shows, etc.) are their own sources, which anyone can verify, albeit not always without some difficulty (e.g. inter-library loan). Our sourcing rules require that sources be publicly available (you can't cite an unpublished manuscript, or something that was published in a small quantity and then all of them destroyed). Our facts have to be verifiable, but sources are often behind paywalls and otherwise not trivial to fact-check, and that's ok. The only quasi-problem with "X book does not contain Y statement" is that if someone wants to prove it beyond all shadow of doubt, they have to read the entire source book. In practical application, this is not really an issue, because books tend to be divided into chapters or other sections. It doesn't take much work at all to discover that a particular dictionary has no entry on a neologism like "sexting", that the Star Trek novel The Rings of Time does not end with Scotty's death, that W. H. Murray in Roby Roy MacGregor: His Life and Times does not say that Rob Roy grew up in Clan Campbell, and so on. Most such claims can be verified by reading a single section/chapter. If someone is attempting to game the system by frequently including dubious negative assertions of this sort in a bad-faith exercise intended to waste many hours of editors' time reading entire books to disprove the false assertions, this will become apparent quickly, and is a user behavior and disciplinary matter for WP:ANI, not a flaw in the WP:NOR guideline, which like our other rules assumes good faith about editorial approaches to it. All that said, it's obviously better if we indicate what a source does write rather than what it didn't. "Sources A and B say X, while source C indicates Y" is far more useful to the reader than "Sources A and B say X, while source C disagrees". It's just begs the question, "disagrees in what way(s)?"  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  09:20, 11 May 2015 (UTC)

I outline some exceptions to that last statement, in later discussion.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  05:31, 29 May 2015 (UTC)

For the claim that "x book doesn't contain Y statement", x book is a primary source. In the 4th paragraph of the section Primary, secondary and tertiary sources, which begins with "Policy: Unless", is the following,
"A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge."
An editor may be prohibited from stating that x book doesn't contain Y statement with no source other than the book, by adding to the above excerpt from policy the underlined text in the following.
"A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified in a straightforward way by any educated person with access to the primary source but without further, specialized knowledge."
--Bob K31416 (talk) 22:46, 12 May 2015 (UTC)
The addition of poorly defined words to a sentence that has problems already won't help the situation. Frankly I don't think a general rule can cover the complexities of this issue. Often it will be WP:SYNTH to write that a book doesn't contain something, but not always. It's easy to think of examples that would be perfectly fine and examples where it wouldn't be. Zerotalk 01:08, 13 May 2015 (UTC)
Would you care to give the details of your comment? --Bob K31416 (talk) 03:13, 13 May 2015 (UTC)
Imaging there is a book on woodworm in the USA. There is a list of cities where woodworm is a problem. The fact that the book omits a particular city from the list is citable (subject to WP:WEIGHT etc), but it could still be SYNTH depending on the circumstances. The fact that the book doesn't explain why most of the cities are in warm states is not citable, since it requires not only reading of the whole book but interpretation of the content. As for problems the sentence has already, think of our articles on quantum physics. Many of them rely on sources that need a higher degree in physics to understand. A literal enforcement of "without further, specialist knowledge" would essentially eliminate those articles to the great loss of the encyclopedia; same for advanced mathematics, etc. I tried to fix this problem several times, but getting consensus on changes to core sentences is very difficult. Zerotalk 03:30, 13 May 2015 (UTC)
Re "The fact that the book omits a particular city from the list is citable" — That would be allowable according to "that can be verified in a straightforward way” .
Re "The fact that the book doesn't explain why most of the cities are in warm states is not citable, since it requires not only reading of the whole book but interpretation of the content.” — That would not be allowable according to "that can be verified in a straightforward way” .
Re "As for problems the sentence has already, think of our articles on quantum physics. Many of them rely on sources that need a higher degree in physics to understand.” — But those are secondary sources, not primary sources, the latter being the type of source we are discussing. Recall that the sentence we are discussing begins with "A primary source".
--Bob K31416 (talk) 13:54, 13 May 2015 (UTC)
While I agree that it is not illegal to do this per the reasons SMcCandlish provided, I wonder if there are many legitimate cases in which it makes sense to do so without it ultimately being synthesis. While The Rings of Time may not end in Scotty's death, it doesn't make sense to include that statement by itself in the article. That book probably also doesn't include Cyndi Lauper or literally billions of other things, but we wouldn't include those omissions either. Now, if for some reason most Star Trek novels ended in Scotty's death and this one was notable in that it didn't, then that statement would require some sourcing as there are several assertions being made apart from the content of the one primary source. To give another example, an article could make two statements like "Dubai is a city. It is not included in Godot's Comprehensive Handbook of Important Cities of the World." Those two statements, while both true and neither requiring a secondary source, together imply that Dubai is not an important city. Regards, Orange Suede Sofa (talk) 03:49, 13 May 2015 (UTC)
The example I had in mind was the statement The Men Who Stare at Goats that the book, which is about the Stargate Project, doesn't actually name it: "This was the Stargate Project[1][2][3][4], which the book never mentions by name." How practically can that be verified? How many times would you have to read the book to be sure? I guess you could if it had an ebook version. The editor insists he can use the book as a source. Dougweller (talk) 14:31, 13 May 2015 (UTC)
  • One could argue that looking for the words "Stargate Project" in the book and concluding that it is not there, is analysis that reaches a conclusion not stated in a reliable source, and thus violates WP:NOR.
  • Regarding the part of policy that says, "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." — Because of the word "only", this is a necessary condition for using the book as a primary source for the statement, not a sufficient condition. So one could not argue from this that the statement should be included.
If you want to clarify this policy to exclude such a statement, then the change indicated by my first message would do that, i.e adding to the above excerpt from policy the underlined text in the following.
"A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified in a straightforward way by any educated person with access to the primary source but without further, specialized knowledge."
--Bob K31416 (talk) 17:22, 13 May 2015 (UTC)
One cannot reasonably 'argue that looking for the words "Stargate Project" in the book and concluding that it is not there, is analysis that reaches a conclusion not stated in a reliable source, and thus violates WP:NOW.'. To actually apply that logic would be precisely the sort of absurdist, ultra-literalist misinterpretation of policy, a form of WP:GAMING the system by distorting the spirit and clear meaning of WP:POLICY, that Zero0000 observes would also effectively disallow specialist sources to be used at all. Reading a book [or otherwise examining a source in the normal way, e.g. watching a film] is precisely "verifi[cation] in a straightforward way". There is no possible more straightforward way to verify what a book says or omits than by reading it. Of course, observing that a source does not cover something can be WP:UNDUE weight, and can involve WP:OR, per various examples above, like the Dubai one; but there are obvious cases where it does not. E.g., in an article on a classic film, it would be fine to include a sentence of the form, 'It was listed in the "Top 100 Films of All Time" published by the New York Times in 2013, the "Top 50 American Films" list of the Motion Picture Academy of America (as of 2015), the "Best Films of the 20th Century" by Entertainment magazine, ... [5 more such lists here], but did not make the American Film Institute's list of "50 Most Important American Films of the Last Century".' In fact, omitting that omission might well be a WP:NPOV problem, an example of cherry-picking critical coverage to artificially inflate the acclaim and importance of the subject. I'd almost be willing to bet money that we already have a number of film articles with exactly this kind of sentence/paragraph, with that sort of balance, and I'm 100% certain that I've seen song/album articles with similar constructions, noting that songs or albums charted in certain markets, but not others (i.e., they do not appear on certain Billboard and other charts, i.e. are not mentioned in specific sources).

As I observed earlier, but didn't spell out very clearly, the length/size of the source isn't relevant. While it's easier to see what is or is not in a list of 100 songs or films than whether a name appears in a long work like War and Peace, it's also easy to see what does and does not have an entry in a truly enormous work like the Encyclopaedia Brittanica. There is no principle in WP:NOR or WP:V (or other) policy, or the WP:RS guideline, that verification has to be "easy" or "convenient"; in fact, the opposite is true, and citing sources that are quite difficult to even obtain is still permitted, as is citing facts from academic journal articles that no one but a subject-matter expert is likely to be able to fully understand correctly (without engaging in applying an interpretation of the facts therein; that requires secondary sources).

Note that this sort of straightforward verification of what a source does or does not contain, and using this information in a neutral, non-synthesis manner, is entirely distinct from the fairly frequent error of doing some source research and concluding something like "No sources mention...", or "This idea is unknown to science", or "It cannot be verified that...". The problem with such patently OR negative assertions is that they presuppose that someone has verified every possible source that could imaginably exist. The fact that we can't do this can be frustrating at times (especially when dealing with pseudo-science and fringe material), but we just can't. We can cite an external authority stating that science can't prove something, or whatever (subject to WP:UNDUE, of course), but we can't come to this conclusion ourselves in Wikipedia's voice.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:13, 14 May 2015 (UTC)

Re "There is no possible more straightforward way to verify what a book says or omits than by reading it." — With regard to verifying that a book doesn't contain some material, yes there is. By having a reliable source that states that the book does not contain the material. For an editor to determine that the material is not in a book by reading the whole book, including all the footnotes, intro, forward, etc., or just checking the parts of the book where one thinks it can be, is subject to possible mistakes. Avoiding these kind of mistakes by Wikipedia editors is a reason for the policy No original research. --Bob K31416 (talk) 22:56, 14 May 2015 (UTC)
There's also this thing called a "search" tool. Get an eBook version (or Google Books version) and search for it. As Blueboar says, the problem here is WP:DUE, not NOR. WhatamIdoing (talk) 16:23, 24 May 2015 (UTC)
The book The Men Who Stare at Goats, apparently does not have an ebook version.[4] Even if there was an ebook version, you couldn't be sure that you or the editor who did the original research for the edit used the right keywords and sequence of keywords in the search to determine that the book didn't mention the Stargate Project because it could have been mentioned in the book by various other names, e.g. Project Stargate, Project Star Gate, Star Gate Project, or names that are synonymous but completely different. --Bob K31416 (talk) 23:13, 24 May 2015 (UTC)
Yep, and these are seriously lame semantic nitpicks. The odds that book A actually says something like "book B does not contain statement X" are pretty close to astronomically low. It's like responding to "I feel very safe with my new home security system" with "But what if your house was hit by a comet?" Aside from Bob K31416's later correct observation, the search thing is pointless argument. I didn't refer to the most expedient way to check book content, I said most straightforward. And besides, reading online book content pulled up by searching the e-book version is – guess what? – reading the book. I didn't say reading the paper edition of the book. JFC, people. If the best anyone could think of to try to refute, poorly, were these two bits of trivia, then I guess my detailed argument is pretty well-reasoned.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  05:20, 29 May 2015 (UTC)

I am not sure we are asking the right questions here... before we ask the question: Is it OR to cite Book X itself for the statement "Book X does not contain Y statement" ... I think we should be asking why we need to mention this fact in the first place - is it is APPROPRIATE to say it? That's more of an WP:UNDUE question than a WP:NOR question. Blueboar (talk) 16:13, 15 May 2015 (UTC)

It is in violation of spirit if not the expressed letter of both OR and UNDUE. Calling out that X isnt in the source is discussing X out of context of any source. " Similarly, do not combine different parts of one source to reach or imply a conclusion not explicitly stated by the source." when the source does not contain anything about X using it as a source to comment about the lack of X is utilizing a source to support a claim that is not explicit in that source. and if no sources have commented on the lack of X in source Y, for Wikipedia editors to do so on their own is clearly placing UNDUE emphasis on the fact that Y doesnt address X. -- TRPoD aka The Red Pen of Doom 22:19, 22 May 2015 (UTC)
I think this is true in most cases, but I can think of plausible exceptions. An article on the Crusades could list the main chronicles and indicate which of them cover the siege of Acre. This is not OR but just the normal editorial process of summarising the sources. (However it might be OR to draw conclusions from the list.) An article on a dead person suspected of murder might mention that that person's autobiography does not admit it. This is verifiable information that would interest readers. Zerotalk 02:15, 23 May 2015 (UTC)
It could be an UNDUE problem, but there are many cases where it would not be. Another trivially simple example: In describing the traits of an standardized animal breed, where the US fancier organization has one published standard and the UK one has a slightly different standard, that differs from the US one only in the omission of a couple of criteria among dozens, it's entirely reasonable to explicitly state that the breed standard WP is using in the article is the US one except where noted, and then note at each omitted trait that it's not in the UK standard. This is a real example; I did this just the other week, and I stand by it.

We already have a well-developed supplementary essay that covers this: WP:These are not original research#Works of fiction and non-fiction. Note carefully that it even includes an example addressing precisely the kind of case under discussion. I've been editing or 8 or 9 years or something with precisely the same interpretation as almost everything on that page, and only noticed the essay yesterday. I'd say it's pretty solid (well I did edit a few bits of it upon first read).

Anyway, yes, really pointed, POV-pushing abuse of the ability to mention that something isn't in a source would be problematic, but we have WP:NPOV policy for that, both generally and for undue weight in particular. An example, to direct contrast with one I gave earlier, might be trying to show that a positively reviewed album lauded by many critics wasn't all that great in your opinion, because Rolling Stone's "Top 10 Albums of the Year" didn't include it, and doing this not in the context of a paragraph mentioning the top-whatever lists, and other accolades, that it did receive, but just stuck in there as an isolated, doubt-casting negative. These kinds of things are really easy to detect, though. I.e., there is no problem here to address. If we really wanted to, it wouldn't be hard, though, to draft language distinguishing these kinds of cases, and we already have enough pro and con examples to work with. The simple act of citing what a source doesn't contain isn't problematic; it's a problem when it is done it in a non-neutral way, or a synthesizing or unverified assumption-making way ("this academic review article didn't cover paper X, so the theory in it must have been rejected by the scientific community...").

In closing, I agree that, in the original question of the thread, that The Men Who Stare At Goats is its own primary source for the statement that it doesn't name the Stargate Project. This sort of "negative citation" is totally routine. We have thousands of these, at least. But I also agree with Zero (as did Bob K, with an additional reason) that '"The fact that the book doesn't explain why most of the cities are in warm states is not citable, since it requires not only reading of the whole book but interpretation of the content.”' (Bob K added: 'That would not be allowable according to "that can be verified in a straightforward way”', which which I also concur).  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  05:20, 29 May 2015 (UTC)

Sorry, but I think your comments have been confused and I stand by my previous remarks. --Bob K31416 (talk) 13:23, 29 May 2015 (UTC)
Fallacy of proof by assertion. You demonstrate no such "confusion" on my part, and have not refuted any point I've made, in very step-wise logic.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  01:29, 3 June 2015 (UTC)

Reliable sources reliably published[edit]

We keep going from "reliable primary sources" to "primary sources that have been reliably published", and back again. I think we're confusing people here.

The probable interpretation for "reliable primary sources" is pretty straightfoward:

A primary source that meets WP:Reliable sources

The other phrase, "primary sources that have been reliably published", is more confusing. Does that mean "a primary source that is not self-published"? This would rule out all WP:SPS sources, definitely including everything on Twitter or a personal website. Does that mean "A primary source that has a reputable publisher"? That would rule out most WP:SPS sources and a good number of articles written by experts but published in weak publications (like corporate or advocacy websites, which do not have "reputable" publishers). Does that mean "A primary source that has a multi-person editorial staff"? That would rule out all SPS sources plus all tiny publications (probably anything with a single editor, and definitely any news outlet whose editor also writes articles).

I think we can figure out what to say, if we can agree on what we mean to say here. Here's what would help me: Give me a couple of examples of sources (real or hypothetical) that you want to include or exclude here. Once we've agreed on what we're trying to accomplish, we should be able to find words to express that. WhatamIdoing (talk) 16:53, 24 May 2015 (UTC)

It should definitely be "reliable primary sources". I've been meaning to fix this for some time, but keep forgetting. The "sources reliably published" thing is someone trying to avoid repetition, but incautiously, at the cost of radically changing the meaning. We don't need to avoid repetition at the cost of much of anything; our policy pages are not novels or journalism pieces, they're technical writing, and a page like this needs a high level of precision, even if it's not a joy to read. This "reliably published" confusion does in fact frequently lead to attempts at WP:GAMING the system, such as assertions that a source with known reliability problems is still fine, because the publishing company is generally reputable. I encountered this very argument today, spilling over into confusion about secondary sources. Someone asserted that two outlying sources must be authoritative (despite others having shown they were contradicted by all actually reliable sources on the topic), because they came from usually "reliable" publishers, and this somehow made them unassailable. Well, to hell with that nonsense. It's a logically bankrupt idea, no different from an assumption that because PBS has a reputation for producing quality documentaries, that everything they've ever aired is a reliable source. We don't even concede that everything ever written by a particular, usually reliable, author is necessarily a RS, nor even that everything in a generally RS is necessarily reliable, so it's logically impossible for us be believe that general positive regard for a publisher can magically bestow reliability on everything it publishes.

At any rate, it can't really mean "a publisher that is reliable" because that doesn't even make sense. A "reliable" publisher would be one that fulfilled its obligations, i.e. met deadlines, delivered books without missing pages, etc. There is no such thing as a publisher that is intrinsically reliable in the source sense. Source reliability is about content, authorship, independence, and editorial control (i.e. fact checking). Besides, WP has no real means of or criteria for assessing whether a publisher is "reputable", which isn't even what the wording says (NOR doesn't contain that word), though we go about this informally to an extent, sometimes. We mostly just care that the publisher is not (except where SPS is okay) the author (directly or through a self-published book outfit), and is not institutionally tied directly to the subject (house organ), or institutionally grossly biased (e.g. a fundamentalist religious organization publishing a "science" book that contradicts actual science). WP:RS has this covered well enough. RS does not contain "reliably published", nor do WP:V or WP:NPOV. NPOV does not mention publishers. V, at WP:SOURCE, suggests that publishers "can affect reliability", but does not elaborate in any way. V says the same thing and elaborates, but only by stating "Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both." But "reliable publication process" is a confused and essentially meaningless phrase in our context. The plain-English interpretation is that as long as their printing presses work properly and they manage to print books with the covers on straight, that this somehow increases reliability. Nuts. It's hard to be absolutely certain what RS is trying to convey there, but it's probably something like "editorial process with reliable fact checking". I guess we'll have to post over at WT:RS and see what the regulars there (to the extent they aren't the same as the ones here) say. Anyway, to come full circle, NOR should say "reliable primary sources", even if RS wants to make some point about publisher reputation lending some reliability points. How to determine reliability is a RS matter; what kinds of sources to use (i.e., reliable ones) and how to avoid OR is an NOR matter.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  04:28, 29 May 2015 (UTC)

Hi User:SMcCandlish. No source is reliable per se, but the reliability always depends on the context where you want to use the source. A source might be reliable for one purpose, but unreliable for another. But I disagree with your interpretation, though. For example, are you saying that "Oxford University Press" isn't reliable? Jayaguru-Shishya (talk) 12:15, 29 May 2015 (UTC)
@Jayaguru-Shishya: I never said any source is reliable per se, and of course reliability is context-dependent. I guess you simply didn't notice that I said: "We don't even concede that everything ever written by a particular, usually reliable, author is necessarily a RS, nor even that everything in a generally RS is necessarily reliable". As in our discussion at another thread on this talk page, you do not appear to be carefully reading what I wrote, and don't seem to understand my interpretation, so I'm skeptical that you actually disagree with it in any practical sense. Re: 'are you saying that "Oxford University Press" isn't reliable?' – That's not actually a meaningful question.

I'll explain in more detail why "reliable publisher" is a meaningless phase on Wikipedia, and a compound logic error. We have an internal definition of what a reliable source is. WP:RS is a whole guideline defining it and the process by which we evaluate source reliability. One aspect of that evaluation is whether the source is from a reputable publisher, i.e. one that is well-established and has a well-regarded reputation. (All three of those italicized terms are used to describe reputable publishers in WP:RS, while "reliable" never is.) We have no concept on WP that the publisher itself is "reliable". The wording confusion going on here, changing the idea that all reliable sources come from reputable publishers, to the nonsensical construction that all reliable sources come from reliable publishers, is exactly the same copyediting error as changing the construction "all tailless cat varieties arise from insular environments" to "all tailless cat varieties arise from tailless environments". The only difference is that "reputable publisher" isn't as laughably bizarre a construction as "tailless environments", so we don't notice the problem as quickly. The problem in this policy text has now been noticed, so we're going to correct it. The copyediting error (which arose probably just because both words are of the "r...le" form, and relate to trustworthiness, thus are easily confused) results in circular reasoning in both cases, compounded by the factual error of mislabeling the publisher/environment half of the equation with a quality that pertains to things that came from it (this is a variant of the fallacy of composition).

WP absolutely, positively does not treat all output from a reputable publishing company as reliable sources. (An obvious proof of this is that an obsolete source from a reputable publisher is not a reliable source, even though the publisher is reputable; Q.E.D.) We do tend to presume they're more likely to be reliable than material published by less well-regarded publishers, but that's not the same thing. Publisher reputability is just a factor in determining source reliability, not proof of it. Secondary source reliability is principally a matter of the expertise and research competence of the authors/editors, with various other factors weighed, like age of source, biases, target audience, and depth of the work, among others, including public perception of the publisher's reputability. That last point is important. Publisher reputability is something external to Wikipedia, while (as noted earlier) source reliability is something we determine internally for WP's own purposes by WP's own criteria. It's logically impossible for these two things to be equivalent. Reliability (a WP concept about a specific source when used for specific facts in a specific article context, as you noted yourself) cannot magically transfer itself out into the world and adhere to an entire publishing company and everything it produces. PS: The exact phrase "reliably published" is even more confused and meaningless than "reliable publisher"; it's simply irrational gibberish, and it has to go.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:04, 30 May 2015 (UTC)
P.S.: Here's some further evidence that some editors just get "reliable" from "reliable sources" stuck in their heads and misuse it. I just encountered, without looking for any such examples, ""Domesticated rats are physiologically and psychologically different from their wild relatives, and, when acquired from reliable sources (such as a breeder), they pose no more of a health risk than other common pets.", in the article Fancy rat. I sanitized this (in the sense of sanity, not cleanliness) to "...acquired from a reputable breeder or pet shop".  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  18:23, 31 May 2015 (UTC)

Specific proposal[edit]

Per the above discussion, clarify and normalize the following:

  • "Unless restricted by another policy, primary sources that have been reliably published may be used in Wikipedia"

to read:

  • "Unless restricted by another policy, primary sources from reputable publishers may be used in Wikipedia"

so that the wording agrees with WP:Reliable sources.

Reputability of a publisher (how well-regarded it is) is directly addressed at WP:Reliable sources, but the terms "reliable"/"reliability" are never used there to refer to a publisher, and as detailed above, cannot logically apply. In short form: Sources may be reliable (a determination Wikipedia makes internally, with a whole guideline full of criteria); a publisher may be reputable, a quality that WP:RS indicates we observe about the off-WP world's regard for the publisher.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  01:57, 3 June 2015 (UTC)

  • Support but suggest as a perhaps preferable alternative that this and any similar text might read "Unless restricted by another policy, reputable primary sources may be used in Wikipedia"
  • This is pedantry on the basis that many primary sources may be reputable but not many will be publishers. Ping: SMcCandlish. GregKaye 15:44, 8 June 2015 (UTC)
    • That would be a "little big" change in the meaning in two ways. It shifts the external criterion (well-regardedness, a.k.a. reputability, a.k.a. respected editorial control) from the publisher to the source itself, and removes the internal criterion (reliability, determined by our WP:RS rubric) entirely. It would thus permit well-regarded self-published sources, whether we thought they were reliable or not. The reason we want reputable publishers for primary sources is that there's a huge difference between a peer-reviewed journal article presenting new research on MAO inhibitors, and some guy's movie review blog even if it has a zillion followers. >;-)
  • Does this have the effect of deprecating self-published primary sources? A corporate or even a personal website might be a very appropriate and reliable source for some facts (e.g., name of the CEO or products sold for a business, the birthdate for a BLP subject for a personal website [assuming it's the BLP's own personal website]) but a self-published source cannot really be said to have a "reputable publisher". WhatamIdoing (talk) 08:05, 20 June 2015 (UTC)
  • Oppose This would exclude legitimate WP:ABOUTSELF sources that are primary and not reputable publishers. As I said months ago, there's no reason to try to encapsulate so much of WP:V when the intent is just to say that you can use primary sources. WP:V and WP:IRS are right there to look at for all the details. This sentence just needs to say a lot less: "Primary sources are not excluded from use on Wikipedia". Rhoark (talk) 01:26, 1 July 2015 (UTC)

Flat statement in lead that NOR does not apply on talk pages[edit]

If this is taken literally, it allows talk pages to be used as forums. An example is Talk:Number of the Beast where we often have posts such as: subject the right hand[edit] "Ok!Here is some original research which maybe allowed in talk page in which it is allowed for truth sake as long it is validated.Base 18, first count with multiples of 18, second 17 as long the quotient remains the same and 17 can be multiplied by 18 and the result is 306,third 34 and the sum of of digits equals to 34 ,of course which is the quotient, and their multiple by 18 equal 612 and on the fourth term the numbers of sum for 37 multiplied by 18 give 666. quotient:

      :17*18 for first
      :34*18 for second
      :37*18 for third"

See also the editor's earlier posts. I think the "OR is ok on talk pages" needs qualifying, at least. Doug Weller (talk) 12:10, 26 May 2015 (UTC)

  • How would you suggest we qualify it? Blueboar (talk) 12:29, 26 May 2015 (UTC)
If no original research were allowed on talk pages, we would not be allowed to sketch an argument showing that some bit of information in the article is ridiculous and should be deleted. It would also be very hard to discuss contradictions between various sources; without such discussion it may not be clear which sources are fringe, outdated, or mistaken on a particular point. (Of course, once the actual situation is figured out, we would then want to find a source that describes the actual situation, or list all the non-fringe results that appear in reliable sources.) Jc3s5h (talk) 13:08, 26 May 2015 (UTC)
I understand all of that and agree. That doesn't help with editors adding, for instance, numerology to talk pages, or their own ideas which have no sources whatsover and that they are just trying to publicise. I'm saying that we need to qualify the statement. Maybe just adding "a limited". We can't be too prescriptive and shouldn't try to cover all bases, but at the moment policy allows an unlimited amount of OR on talk pages while at the same time we say they are not forums. Doug Weller (talk) 18:01, 26 May 2015 (UTC)
From the 1st paragraph of WP:NOR,
"The phrase "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published sources exist."
So just about everything on a Talk page is original research. To treat the present situation, editors can use the following from WP:NOTAFORUM,
"In addition, bear in mind that talk pages exist for the purpose of discussing how to improve articles. Talk pages are not for general discussion about the subject of the article, nor are they a help desk for obtaining instructions or technical assistance. Material unsuitable for talk pages may be subject to removal per the talk page guidelines."
and the following from the beginning of talk page guidelines,
"The purpose of a Wikipedia talk page (accessible via the talk or discussion tab) is to provide space for editors to discuss changes to its associated article or project page. Article talk pages should not be used by editors as platforms for their personal views on a subject."
and from the section Behavior that is unacceptable
"Please note that some of the following are of sufficient importance to be official Wikipedia policy. Violations (and especially repeated violations) may lead to the offender being blocked or banned from editing Wikipedia.
Do not use the talk page as a forum or soapbox for discussing the topic. The talk page is for discussing how to improve the article."
--Bob K31416 (talk) 21:56, 26 May 2015 (UTC)
Thanks. That's basically what I do when obviously appropriate. Doug Weller (talk) 18:19, 27 May 2015 (UTC)
It doesn't allow talk pages to be used as forums, since WP:NOT#FORUM policy and WP:TALK guidelines don't. If we already have a rule against something, "failing" to re-mentioning the rule on some other page doesn't undo the rule. Having something in another page that a bad-faith editor could try to exploit as a loophole to get around a general rule doesn't actually permit them to do so; that's WP:GAMING the system. That said, there is no problem adding some clarifying wording that points to that policy and guideline, to forestall misinterpretations that might occasionally lead people to treat talk pages as forums. I don't see an extant problem to address, though. People using article talk pages as general debate pits are usually new arrivals who haven't learned how WP operates yet, and we're usually pretty lenient with them. For inveterate POV pushers and edit warriors who already know better, we have WP:ANI and WP:AE. We can't strip NOR policy of the rule that it does not apply to talk pages, just to address an issue that doesn't seem to be an issue. Much of what goes on at talk pages is OR, and discussing it and working through it to tease out what is sourceable and what is novel synthesis, etc., is much of what we do. It's not necessarily instantly apparent what proposed change is and isn't proper in what exact wording. WP:NOTOR and WP:NOTSYNTH would not exist in such detail otherwise. PS: A clear problem with saying "a limited amount" of OR on talk pages, is it immediately raises the question, "where is the limit specified?" I.e., it wouldn't seem to solve anything, and would just be another thing to try to game. A simple parenthetical clarification can be cribbed and inserted: "Talk pages are for discussing how to improve articles, not for general discussion about the subject of the article, nor are they a help desk for topical questions or advice."  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  13:48, 28 May 2015 (UTC)

Closing lame loopholes[edit]

WP:PSTS had a silly loophole in it, which lead to some actual confusion (see Wikipedia talk:Use of tertiary sources#Interpreting NOR differently), and could in theory lead to someone actually attempting to WP:GAME the discrepancy, in any of at least four ways.

The policy section said the same thing twice, in different wording:

  • Somewhat precise: 'All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source [...]'
  • Vague: 'Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source.'

The vague version had two loopholes:

  1. It accidentally allowed "interpretive" and "synthetic" claims, while disallowing only "analytic" and "evaluative" ones, to be cited to primary and tertiary sources (I just also noticed the first one omitted "evaluative", so we have a two-way loophole.)
  2. It didn't actually require a citation, only the known fact or perhaps even deductively reasoned assumption that such a source "has been published".
    1. At the discussion I linked to above, someone seriously argued that any tertiary reliable sources can be used to source these kind of claims, because reliable tertiary sources summarize reliable secondary sources, so we can "assume" that the claim must have been published in a secondary source at some point.
    2. It could also be used to reason that if an ostensibly reliable primary source had such a claim in it, and itself cited a secondary source the editor doesn't have on hand, then citing the primary source which cites the secondary source would satisfy this policy, without WP:SAYWHEREYOUGOTIT being triggered.
    3. And, if an ostensibly reliable source was published, but later retracted an error, or was superseded by a later edition with a different claim in it, someone might argue to keep the obsolete version, because it "had been" published. I've actually encountered similar arguments to this, in the form of resistance to updating citations to refer to a newer edition of a source, even though I was the one doing all the work to check them, fix page numbering, etc. "The original edition was reliable enough" went the reasoning.

None of these are practical arguments of course, but people will waste a lot of editorial time arguing about this kind of thing. Just dealing with the one case linked above has cost plenty of time I would have rather spent on something more productive today.

The vague version also didn't make sense in another, less serious, way. Nothing is published "by" a source; sources are published by publishers, and claims are published in sources. A further, trivial, problem was the weird back-and-forth syntactic structure of "interpretive claims, analyses, or synthetic claims" in the first case.

I fixed all of this by normalizing the two passages (with some exceptions, pending further discussion):

  • Clearer than before: 'All analyses and interpretive or synthetic claims about primary sources must be referenced to a secondary source [...]'
  • Not vague any longer: 'Articles may make an analytic, evaluative, interpretive, or synthetic claim only if that has been referenced to a reliable secondary source.'

I'm not sure why the policy was saying the same thing twice anyway, but at least it's doing so more consistently and less confusingly now.

One remaining interpretational issue is that the first of these (in the original version and my version) limits its applicability to claims "about primary sources", while the second does not, and would seem to broaden its applicability to cover anything in an article, which I think is the actual intent. I shouldn't be able to make such a claim about secondary source material without being able to back it up either, nor about any series of isolated facts, whatever their disparate sources. If the broader meaning is intended, the narrower statement in the other passage is apt to be used to try to game it, and such attempts might even be successful, on the basis that the more specific wording intentionally "clarifies" the broader version.

Another maybe-issue is that the first of the two sentences does not state that the secondary source be reliable, but perhaps this is not problematic in the context.

Also for further resolution is whether the four-point consolidated list, "analytic, evaluative, interpretive, synthetic" should be compressed. Are any of these words redundant? I think I'd be in favor of dropping "evaluative", as essentially redundant with "analytic" but also suggestive of value-laden POV, which we're supposed to avoid. I left this word out of the sentence that didn't already have it. I also like "analyses" better than "analytic claims", because the former is broader. I left it in the passage it originated in, but did not add it to the second one.

 — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:31, 29 May 2015 (UTC)

Hi User:SMcCandlish. Just dropped by to leave a few comments (the numeration is corresponding yours):
  1. Does it? To me, neither of the passages seem "vague". They allow interpretive, analytic, synthetic and evaluative claims about primary sources only if they are referenced by a reliable secondary source. I can't see the two passages really disallowing "analytic" or "evaluative" ones. Besides, WP policies are to be interpreted as a whole, not as some separate fortresses fighting against each other.
  2. "...published by a reliable secondary source", in contrast to being self-published (WP:UGC) or being published by advocate group, or a group without any expertise on the field. I can't see a problem with that either.
    1. Let them argue, but it's not alone a reason for policy change. WP:TERTIARY explicitly states that: "Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources. Deciding whether primary, secondary or tertiary sources are appropriate on any given occasion is a matter of good editorial judgment and common sense, and should be discussed on article talk pages." So you could just have pointed out to the specific editor(s) that not just "any" tertiary source shall do, but instead it's a matter of editorial and communal judgement.
    2. How I've always understood this, is that the primary source should be referred to in a reliable secondary source, not the other way. In other words, we might sometimes want to have a more in-depth interpretation of a reliable secondary source by actually using the details provided in a primary source, but that's been already used in the secondary source we have.
    3. This is everyday reality with MEDRS articles, and if we have two high-quality sources making different claims about the same subject, we are inclined to favor the more recent one. If the more recent one is of a lower-quality, however, we might stick to the older one. I don't think it's good enough reason for a policy change if you have encountered some editors arguing against. Have you asked for WP:THIRDOPINION, or an opinion of some experienced editor/administrator perhaps?
Cheers! Jayaguru-Shishya (talk) 12:12, 29 May 2015 (UTC)
  1. One sentence says one thing, one says something considerably different in multiple ways. The one requires publication, the other requires citation of the publication. The first imposes limits on some categories of use, the other limits different, partially overlapping categories of use. I didn't say both were vague, I said one in particular was, obviously by implying that publication somewhere was sufficient, without an actual citation to that source (vs. a tertiary source that in turn cited it). That's the crux of the matter. I agree that policy should be interpreted as a whole, but I also linked directly to a case of this not happening (in good faith, not a gaming attempt). This example case is not the reason for the edit, just an illustrative case. If a policy is unclear enough, due to a minor wording self-contradiction like this one, to lead a logically-minded, experienced (and policy-experienced) Wikipedian to such an incorrect interpretation of policy, then the obvious and painless solution is to resolve the self-contradiction in the wording. It forestalls recurrence (and spread) of the misinterpretation, instead of blindly hoping it doesn't happen again, or not caring about the editorial time wasted when it happens again. It may be much worse than wasted time: It's highly probable that any editor with such an interpretation is actually making edits based on it, mostly undetected by anyone else – i.e., is improperly sourcing actual articles already, and will continue to do so unless and until the policy wording problem is resolved. So, this is not about an argument on a talk page. And the incorrect interpretation is neither crazy nor trivial to refute. It cost me several hours to research all the content policies in detail to refute it conclusively. The rubric that policy should be interpreted as a whole has real productivity costs when policy is unclear (and in this case unclear for no actual reason; it's just a gradual failure to keep two similar passages in agreement with each other).
  2. I've already addressed why the "published by" wording is faulty in multiple ways. I feel that you have not closely read what I posted. Moving on, nothing about normalizing the two passages to both refer to citation, not just alleged existence, of the source would do anything to any WP:UGC/WP:SPS analysis. (Since a secondary source is, by definition, published [among other qualities], and we must actually cite it, then restating here that the secondary source "has been published" is pointless. It's worse than pointless, because it's come at a cost, of failing to say that the source also has to be cited directly. UGC/SPS material is a primary source, by definition, so it's automatically excluded regardless.) I thus see no actual objection rationale here.
    1. Copyediting to resolve inclarities in wording is not a "policy change"; nothing about the meaning of the policy would be changed in any way. Do you revert every copyedit to policy pages? Surely not. Your point about tertiary sources is non-responsive to the actual discussion I linked to, in which the "published by" language leads to an interpretation that secondary and tertiary sources can be evaluated for NOR purposes as if they are equivalent. Again, it seems to me that you are justifying your revert by making assumptions about what has been posted, rather than reading it closely to see what it actually says.
    2. I agree, of course, but that's not related to what I said. Please read it again more carefully. I'm not talking about citation, as a primary source, of a primary source that cited secondary sources. I'm talking about policy inclarity leading to an interpretation that a tertiary source (in the earlier point) or even primary source (in this point) which cites a secondary source can be treated as if it were WP itself citing the secondary source, for purposes of making analytic/evaluative/interpretive/synthetic claims. Aside: Your "How I've always understood this" comment highlights why I included a couple of illustrative examples (the discussion I linked to, and the case of people reverting upgrades of citations to newer editions of a source). Editors tend to rely on their own interpretations of policy, and we want to narrow those to a consensus interpretation, or we might as well not have policy at all. A rule that means something different to anyone who wants their own interpretation (like, to allow them to cite tertiary sources as if they were secondary) is not a rule, it's just noise.
    3. I'm glad we agree that obsolete source citations should be replaced, and that the need to do so is commonplace. I'm wondering, then, why you reverted a copyedit that reinforces this, then posted this talk page message which has not provided a clear rationale for reverting, and is largely unresponsive to the rationale for it. This numbered point, about updating citations, is illustrated by an example of the kind of lame dispute that can be avoided by copyediting the wording, but the example dispute is not the reason for the copyedit. The rationale for the copyedit is that the two lines of near-identical policy are conflicting sharply in their plain-English wording, and you have not controverted this fact, or that rationale. There is no need for WP:THIRDOPINION; this isn't (as you point out yourself) about any particular disagreement with any particular editor. Both parties in that case already qualify as "some experienced editor". What this is about is copyediting the two sentences so WP:PSTS is no longer self-conflicting, and such time-wasting debates do not continue.
I closed with :'None of these are practical arguments of course, but people will waste a lot of editorial time arguing about this kind of thing.' Your response the entire thread seems to consist mostly of also pointing out that those policy-misinterpreting arguments are not practical, while ignoring the point of the copyedit. Most everything in WP policies and guidelines was added to stop editors from doing things we know are unwise. Simply re-observing they're unwise is not a rationale for not having these rules.
 — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:34, 30 May 2015 (UTC)

I'm putting the copyedit back in, having addressed the only issued raised.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  18:55, 31 May 2015 (UTC)

The "about primary sources" clause[edit]

I fully agree with the first of the two sentences SMC quotes [in the previous thread]:

  • 'All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source [...]'

The phrase I put in bold is important. When talking about a primary source, I agree that we do need a secondary source to support what we say (although I would expand it to also allow summaries of analysis and evaluative claims contained in and cited to tertiary sources)

I am less sure I support the second sentence:

  • 'Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source.'

This sentence could be interpreted as banning all mention of an analytic or evaluative claim that is contained in a non-secondary source. For example: if (for example) an old chronicle (such as that written by the Venerable Bede) includes an analytic or evaluative claim about the Saxons, someone might argue that we could not mention it because the chronicle is a primary source. I feel we should be able to mention it... we simply need to attribute it when we do. Blueboar (talk) 01:41, 30 May 2015 (UTC)

@Blueboar: :You didn't dispute my reasoning in the previous thrad, despite being the other party in the interpretational discussion I linked to, and have only responded to the "about primary sources" side point I brought up for later discussion. I'll take that as an indication the disagreement is resolved, unless you say otherwise.
In responding to your proposals, I'll number my responses here for easy reference. I apologize for the length, but this stuff is complicated to analyze, and it must be done very precisely.
Conclusions: A) Tertiary sources are not equated with secondary ones, including for analytic claims, for good reasons. B1) The addition of "about primary sources" to the interpreting sentence, absent from the policy-defining sentence, is an erroneous interpolation, and should be removed. It was probably added, in good faith, because when Wikipedians make this kind of synthetic error, it's most obvious (and probably most frequent) when they're working with primary sources. We certainly detect it more often, because we're more suspicious of primary source citations. B2) It would probably be reasonable to reword the interpretational sentence to say "especially about primary sources". I'd be in favor of that clarification, since it would interpret the policy correctly, while also still conveying the "use caution with primary sources" message. D) You can attribute Bede, where what he said is actually relevant, but not directly cite him (his work is manuscript material, not published). You have to cite a modern publication.
 — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  04:58, 30 May 2015 (UTC)
Perhaps we are concerned about different things... the key point I am trying to make is that if a source... any source (be it Primary, Secondary or Tertiary)... actually contains an analysis or evaluative claim, it is not Original Research for us to mention that analysis or evaluation in an article (for the simple reason that it is not original to us - we are taking it from the source). This is why I think the language of the second sentence you quote is flawed. It is actually not a WP:OR violation for an article to include an analysis or evaluation that has been made by a source (especially if you attribute it)... what we can't do per WP:OR is create and include our own analysis or evaluation based on what a source says. THAT is Original Research. However, this distinction between reporting on analysis or evaluative claim contained in a source, and creating our own analysis or evaluation based on sources holds true no matter what "type" of source we are using (creating our own analysis or evaluation based on secondary sources is just as much original research as creating one using primary or tertiary sources}. The issue isn't the nature of the source, but what we do with it. Blueboar (talk) 13:41, 30 May 2015 (UTC)
That simply isn't correct, though, and the policy says this twice in slightly differently-worded sentences. Many primary sources contain such claims, but we cannot use them as sources for such claims; in primary sources, they are hypotheses, unevaluated by other subject-matter experts in secondary sources. Yes, it is true that we can't create and include our analysis/evaluation/interpretation/synthesis based on what a source says; but this is not the only thing we cannot do. As a concrete example, many primary sources (e.g. opinional book reviews, and papers in academic literary and arts journals), have made the analytic claim that J.R.R. Tolkien's The Lord of the Rings is an allegory about England and its allies vs. Germany in the two World Wars. Reliable secondary sources pretty uniformly disagree with this assessment. If it were a brand new novel claim, unexamined in secondary sources, it would be WP:UNDUE weight to insert this untested, POV-pushing, questionably plausible interpretive claim into a WP article, and a form of original research (namely Wikipedians taking it on themselves to analyze whether this idea has merit as a literary hypothesis). Since it has been examined and been found lacking, in reliable secondary sources, and has been the source of plenty of reliably sourceable debate, it's perfectly fine to include mention of it as a hypothesis, and how well (actually poorly) that hypothesis has been received (in secondary sources). Doing so would be neither original research nor undue weight. [If you're curious, in actuality most secondary sources agree that Tolkien worked in a few allegorical elements, but that the work as a whole is not an allegory, unlike C.S. Lewis's contemporary Narnia series. The two authors frequently met to share and critique each others' new work, and Lewis's reliance on direct allegory was Tolkien's principle criticism of it.]

PS: Just because it's fun to use Tolkien for examples, here's an even simpler one. Tolkien himself (primary source) said that TLotR is not a trilogy, but a single work published in multiple volumes. Secondary sources nearly universally agree that this is true. There are many primary and tertiary sources that refer to it as a trilogy, but is wrong for WP to reanalyze on its own and agree with the primary and tertiary sources, and re-label it a trilogy.

If you think these examples do not get at the kind of scenario you have in mind, then please give us some examples that illustrate why you think the policy must be changed in such a major way.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  18:10, 31 May 2015 (UTC)

PS: Actually, the policy makes the point three times, since it also says "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." An analytic/evaluative/interpretive/synthetic claim in a primary source is not a "fact", but a hypothesis.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  19:02, 31 May 2015 (UTC)

Re: Many primary sources contain such claims, but we cannot use them as sources for such claims. Yes... we can... although we are limited in what we can say about them. If primary source X contains claim Y, we can state as a descriptive fact that primary source X contains claim Y (citing primary source X itself). That is NOT original research.
What the policy bans is "any interpretation of primary source material" without a secondary source... and I fully agree with that. But what I am talking about is not an interpretation of the primary source material... I am talking about a simple description of the source's content. Hell, if necessary you can quote the primary source directly. The same is true for any source (primary, secondary or tertiary). We may not be able to assert that "TLotR is not a trilogy" based on what Tolkien said... but we can mention the fact that he said it (and cite where he did so). Blueboar (talk) 19:30, 31 May 2015 (UTC)
As I said in semi-related discussion at WT:TERTIARYUSE, I think we keep talking past each other. Yes, of course we can state as a descriptive fact that primary source X contains claim Y; that's not what's at issue here. That is not using a primary source to cite any analytic/evaluative/interpretational/synthetic claim we're making in an article, it's quotation (or perhaps paraphrasal, depending on how it's written), with attribution. To get back to the gist, earlier you said: 'I am less sure I support the second sentence:', and then quoted it: 'Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source.'. You continued: 'This sentence could be interpreted as banning all mention of an analytic or evaluative claim that is contained in a non-secondary source.' (boldface added). But that's clearly not true, by your own "state as a descriptive fact that primary source X contains claim Y" example, which we both agree is a valid approach. So, I'm at a loss for what we could actually be disagreeing about in principle, or more to the point, where anything about the policy wording of that point should change. People clearly aren't interpreting it that way, since surely there are at least hundreds of thousands of statements and attributive citations of this sort in WP articles.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:07, 3 June 2015 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── This is a bit TLDR for me right now, but User:Blueboar, I think a distinction between making a claim and mentioning one addresses your problem. In your example, the Venerable Bede says something about the Saxons. The Wikipedia article could "mention" the Venerable Bede's claim that the Saxons are a bunch of loutish ruffians; it should not "make" that claim itself. A primary source is reliable for the fact that the source says something; it is not (necessarily) reliable for a claim that the alleged fact is true. WhatamIdoing (talk) 08:11, 20 June 2015 (UTC)

Blueboar wrote "actually contains an analysis or evaluative claim, it is not Original Research for us to mention that analysis or evaluation in an article..." SMcCandlish wrote "That simply isn't correct, though, and the policy says this twice in slightly differently-worded sentences. Many primary sources contain such claims, but we cannot use them as sources for such claims", but we can and do. To use an example that has been used since at least 2007 in the archives of this talk page: It is used for statements such as Wellington's comment on the Battle of Waterloo that it was "the nearest-run thing you ever saw in your life" -- he was after all was one of the two leading experts on the battle -- and despite the 200 years of debate (they say more has been published on the battle than there are days since the battle), no naysayer has come forward to contradict it.
There is more to the quote but it is not misleading to take just the one sentence. The first sentence tends to be omitted because it does not say much and it is confusing to modern generation "Its been a damned nice thing." and Wikipedia does not mention the third sentence in the Battle of Waterloo article because it is not so neutral: "By God! I don't think it would have done if I had not been there." and expresses a personal speculative opinion (although almost certainly true).
As to a matter of presentation, as Wellington's opinion is a POV, as WhatamIdoing suggests such opinions need in-text attribution although all such POVs ought carry in-text attribution whether the source is primary, secondary or tertiary. If the information was factual, for example the taking of the weather measurements at Greenwich at noon on the day, I do not think it would be necessary to quote it and use in-text attribution, eg "The weather in London at midday on 18 June 1815 was ... (citation)".
-- PBS (talk) 12:47, 11 July 2015 (UTC)

Can I use an educational institute's website to say it isn't accredited?[edit]

Raising this as "Cambridge Kipp" has just been removed from List of unrecognized higher education accreditation organizations.[5] I'd normally look for a source, but the only mention of accreditation or in this case accredited on their website says "The purpose of this document is to provide evidence that Cambridge KIPP is eligible to be accredited as an institute that offer a wide-spectrum of researches in different fields."[6] 'Doug Weller 09:21, 30 May 2015 (UTC) — Preceding unsigned comment added by Doug Weller (talkcontribs)

Probably not in this case as (a) I do see no claim on their website they claim to be an "accreditation organization" so they would most likely not belong on that list in any case. (b) The document you quote is talking about research not education accreditation (i.e. not the accreditation to pass diploma, but the accreditation to put in grants for specific types of research funding bodies). So the contents of the document appears not to be relevant to its educational aims. (and (c) the document is a primary source, so utmost caution in used should be applied) Arnoutf (talk) 10:35, 30 May 2015 (UTC)
Oh I agree they shouldn't be on the list. And I agree about the document. They should be on List of unaccredited institutions of higher education. But I can't find an independent source saying they are unaccredited.
We need to be consistent. If a book can be the source for a statement that a certain thing is not in the book (see above), then an organisation's website can be a source for saying they don't mention something in their website. The problem I see is that our policy doesn't deal with this clearly. Doug Weller 10:41, 30 May 2015 (UTC) — Preceding unsigned comment added by Doug Weller (talkcontribs)
The problem is that not mentioning that you are accredited can only be interpreted as such (there is no mention of accreditation on their website). However an institution can be accredited without mentioning it. So in this case lack of evidence of accreditation cannot be interpreted as evidence of lack of accreditation. Arnoutf (talk) 12:01, 30 May 2015 (UTC)

The distinction that needs to be made here is the distinction between omission and commission. It is not Original research to leave an organization off a list (omission) based on lack of sources ... but it would be original research to include an organization on a list (commission) based on lack of sources... you need a source to actually support inclusion. Blueboar (talk) 15:49, 30 May 2015 (UTC)

User:Arnouff, are you saying that I can use the website as a source to state that the institutions website does not say that it is accredited? I take Blueboar's point about a list. Doug Weller (talk) 12:38, 31 May 2015 (UTC)
Euhm yeah, but in what unbiased context would you want to use that? Not mentioning stuff on websites hardly seems evidence for anything.
In any case makes sure no to use it for anything like implicit synthesis, that is, if no accreditation is mentioned something is going on. To avoid a non-neutral point of view you should probably look at all websites of all educational institutions and consider all that do not mention accreditation under the same header. I had a quick look at the website of University of Oxford and could not find any reference to them being accredited......... So in my estimate we can extract no relevant information at all from an institution omitting such information. Arnoutf (talk) 13:05, 31 May 2015 (UTC)
That's not a good example, as it is a state university (with in this case a Royal Charter I believe)> So you wouldn't find the term accreditation. You'd have to compare it with private institutions. — Preceding unsigned comment added by Doug Weller (talkcontribs) 14:02, 31 May 2015 (UTC)
Most non-US universities would be bad examples, because most countries other than the US have meaningful regulation of universities by the government. In the US universities and colleges are allowed to award degrees without meaningful regulation by the states, so private accreditors are needed to assure quality. Jc3s5h (talk) 14:37, 31 May 2015 (UTC)
Cambridge Kipp appears to be UK based. Arnoutf (talk) 14:59, 31 May 2015 (UTC)
I don't see the relevance as it's a private institution. If UK law applies to it, then it can't grant degrees. You need authorisation from the Secretary of State, a Royal Charter or an Act of Parliament to grant degrees in the UK. Doug Weller (talk) 20:48, 31 May 2015 (UTC)
Accreditation is a claim that must be reliably sourceable, as must "recognition" in whatever sense fits the inclusion criteria of the first list mentioned. I'm skeptical of the value or sourceabilty of negative lists like List of unrecognized higher education accreditation organizations and List of unaccredited institutions of higher education. They reverse the burden of proof, and would require proving a negative. That's not what WP is for.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  18:14, 31 May 2015 (UTC)
SMC is correct here... the problem lies in the fact that this is a negatively phrased list, and that makes verifiability difficult, if not impossible. Given the negatively phrased topic - "Not Accredited" becomes a claim that must be reliable sourced. If we can not find a source that says "X is not accredited" then we can not say "X is not accredited". The only option available to us is to remain silent on whether the institution is accredited or not... and silence would mean not including it on the list. Blueboar (talk) 19:44, 31 May 2015 (UTC)

Publication by a Wikipedian?[edit]

If a Wikipedian writes and publishes a book that contains previously unverifiable information and the information is true, can that Wikipedian then add the information to Wikipedia and cite it using their book? Is that allowed? Is it considered OR? — Preceding unsigned comment added by (talk) 12:51, 1 June 2015 (UTC)

It depends as OR is only one factor. Was the book published by a quality publisher (if it's self-published then OR basically applies)? What reception did the book have among experts in that field? Do the claims being cited differ greatly from the mainstream view? These points, along with conflict of interest guidelines, suggest the best course of action would be to first propose changes on the article's talk page. --NeilN talk to me 14:13, 1 June 2015 (UTC)
So if the book was published by a quality publisher and had decent reception and no outstanding claims and the Wikipedian wasn't in any financial or significant personal conflict of interest, then you'd think theoretically it may likely be allowable? — Preceding unsigned comment added by (talk) 14:54, 1 June 2015 (UTC)
I would say yes. It would be almost impossible to prevent this in any case, as most Wikipedians do not edit under their own name. We just have to use it as any other source if decently used (i.e. reliably published preferably secondary. No fringe views. No conflict of interest). Arnoutf (talk) 14:57, 1 June 2015 (UTC)
Then it should be treated like any other potential source. Point five in this section of WP:EXPERT has helpful advice. --NeilN talk to me 15:00, 1 June 2015 (UTC)
I swear one of these pages already covers this, in wording something to the effect that you can cite work that you've written and had reputably published, as long as you declare that it's yours, so others can evaluate whether you have a COI. I can't seem to find it, though. Anyway, WP:COI, WP:SOAPBOX, and WP:NPOV apply pretty clearly. WP isn't here for people to advance their own pet views on a topic. But we want subject-matter experts to edit WP, and they can't be banned from citing sources they're not entirely independent of, in a non-PoV-pushing way.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:16, 3 June 2015 (UTC)
@SMcCandlish: Did you look at point five in the section I linked to? --NeilN talk to me 02:36, 3 June 2015 (UTC)
Ah, that's it! I was searching for strings with "declare", and it was really "disclose".  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:44, 3 June 2015 (UTC)

Blueboar's History of Wikipeida Policies[edit]

The last few threads have made me think that it is time for another segment of "Blueboar's History of Wikipeida Policies" history lesson... Today's episode: A useful thing that the Policy used to say, but which got lost in rewrites.

Back in 2005 (or so) the policy noted that "Wikipedia should not be a Primary Source for information" (or words to that effect). I have always been sorry that this statement got lost in rewrites, because it did a lot to help explain the core concept of WP:NOR. the policy went on to explain that "Wikipedia should not be the first place to publish any information, because that would make Wikipedia the Primary source for that information."

This basic statement was actually the genesis of what is today the WP:PSTS section of the policy... it was felt that we needed to explain what a Primary source was... which in turn required explaining what a secondary and tertiary source was... And somewhere along the way, the explanations of primary secondary and tertiary took over, and the statement that they were originally written to explain got taken out. Unfortunately.

Would anyone object if I returned it (or something like it) to the Policy? Blueboar (talk) 00:53, 2 June 2015 (UTC)

Isn't that now effectively covered under Wikipedia:What_Wikipedia_is_not#Wikipedia_is_not_a_publisher_of_original_thought? -- TRPoD aka The Red Pen of Doom 18:48, 2 June 2015 (UTC)
Perhaps... but I think it also needs to be stated in this policy ... since it goes to the heart of why we don't want Original Research. Blueboar (talk) 02:08, 3 June 2015 (UTC)
Blueboar's understanding of the original injunction against "Primary (original) research" and its subsequent (mis)interpretation as a ban on the use of primary sources is perfectly in line with what I was taught in grad. school, that primary sources are always to be preferred to secondary (we never touched tertiary.) — Robert Greer (talk) 01:55, 3 June 2015 (UTC)
Different issues. It's a grad student's "job" to come up with novel syntheses of primary research, but doing so is inimical to WP's mission and purpose.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:49, 3 June 2015 (UTC)
I don't go quite that far... Remember, this isn't Grad School. In academic writing (unlike Wikipedia) original research is actually encouraged. Schools want original thinking about, and novel interpretations of primary sources. Here on WP, we don't.
Wikipedia is a tertiary source. Our job as writers/editors is to summarize what other sources say about a topic. We should, and rely mostly on secondary sources for this... but not exclusively. It is sometimes appropriate to describe what the primary sources say - they can help add context to an article. However, we should do so with extreme caution, because it is very easy to (unintentionally) misuse primary sources in ways that constitute original research. So.... I agree with all the cautionary language about the use of primary sources. I definitely would not want to remove that. But, perhaps we do need to make it clearer that this is a caution, and not a ban. I also agree that we should have a (different) caution on the use of other Tertiary sources (as they are often overly superficial, and do not include nuances that secondary sources may contain). But again this is not a ban. What we really need to make clear is that Original research lies not in which type of sources we use... but in HOW we use them. Blueboar (talk) 02:51, 3 June 2015 (UTC)
I'm wholeheartedly in agreement, and !vote to reinstate that statement. It's more relevant than ever now, since WP is in the top 3 search results for nearly everything, and often #1. It's the go-to source of quick, basic information for millions (billions?) of people, and we have a magnified responsibility, a decade later, to not foist off as fact anything that can't be verified or is pushing a point of view.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:46, 3 June 2015 (UTC)
OK... so where should we add it? Suggestions? Blueboar (talk) 02:51, 3 June 2015 (UTC)
In NOR's lead, as proposed below.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  21:58, 4 June 2015 (UTC)

For reference, here's what a primary source is, according to this policy WP:NOR.

  • Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, a period of history, a work of art, a political decision, and so on. Primary sources may or may not be independent or third-party sources. An account of a traffic accident written by a witness is a primary source of information about the accident; similarly, a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. Historical documents such as diaries are primary sources.[1]
  1. ^ Further examples of primary sources include archeological artifacts, census results, video or transcripts of surveillance, public hearings, investigative reports, trial/litigation in any country (including material — which relates to either the trial or to any of the parties involved in the trial — published/authored by any involved party, before, during or after the trial), editorials, columns, blogs, opinion pieces, or (depending on context) interviews; tabulated results of surveys or questionnaires; original philosophical works; religious scripture; ancient works, even if they cite earlier lost writings; tomb plaques; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos and television programs. For definitions of primary sources:
    • The University of Nevada, Reno Libraries define primary sources as providing "an inside view of a particular event". They offer as examples: original documents, such as autobiographies, diaries, e-mail, interviews, letters, minutes, news film footage, official records, photographs, raw research data, and speeches; creative works, such as art, drama, films, music, novels, poetry; and relics or artifacts, such as buildings, clothing, DNA, furniture, jewelry, pottery.
    • The University of California, Berkeley library offers this definition: "Primary sources enable the researcher to get as close as possible to what actually happened during an historical event or time period. Primary sources were either created during the time period being studied, or were created at a later date by a participant in the events being studied (as in the case of memoirs) and they reflect the individual viewpoint of a participant or observer."
    • Duke University, Libraries offers this definition: "A primary source is a first-hand account of an event. Primary sources may include newspaper articles, letters, diaries, interviews, laws, reports of government commissions, and many other types of documents."

It seems that the proposal would be understood just by editors who already understand this policy, so what's the point. For those that come to this policy page for help in understanding the policy, it would just make the explanation more complicated. Editors have a hard enough time understanding what a primary source is. I think we should be working towards making the policy more simple and easier to understand, than more complicated and abstruse. --Bob K31416 (talk) 03:23, 3 June 2015 (UTC)

It's routine for us to tie policies (and guidelines) to each other with cross references. So, the following would work fine, without adding any confusion, and help people find the relevant material in WP:NOT and WP:RS, which reduces rather than increases confusion: "Wikipedia is not a primary source for information." That's concise enough it can go right in the lead section of WP:NOR. I was tempted to put "for any information", but that's not quite true; WP can be a primary source for articles about WP itself, and pretty much nothing else. PS: WP:NOT#OR is a super-short summary of WP:NOR; having the point clearly stated in here isn't redundant with WP:NOT, but the other way around. And NOT is written in a summarizing, example-giving way, so that's fine.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  21:58, 4 June 2015 (UTC)

Peculiar question about SYNTHESIS[edit]

I do not know If I'm asking this in the wrong place but I'll ask anyway. I would like neutral opinions on this matter, from the people who did not participate in the discussion where this question is mentioned.

If the following claim SYNTHESIS?

  • A= "10 red apples are in a bowl, named a,b,c..etc"
  • B= "we know all apples are red"
  • A+B="apple named c is red"

Another example:

We know the following is correct.

  • A="all members states of EU voted on the subject S"
  • B="the subject S was sustained unanimously".
  • C="Germany is EU member state"

Does the following statement constitutes SYNTHESIS?

  • A+B+C = "Germany had voted in favor of subject S"

Can an article state the claim A+B+C based on separate claims A,B,C if we know those claims (sources) are correct. Note that there is not a single source that says anything specifically about Germany and its vote.

Furthermore, if we know A and B and we have a source that specifies the claim for apple c in the following was "apple c is red". Do the sources complement the source stating that apple c is red or are they completely irrelevant.

Furthermore, although we ( I presume, depends on the previous question) can't use the source B (together with A) to make an edit "apple c is red", because of SYNTHESIS. Can we use it to verify the other source that states: "apple c is green"? Can we, based on source B that says "all apples are red" ,which is undeniably correct, say that the source "apple c is green" (which has no footnote or any other clarification or argumentation in the new source) is not correct enough to enter the article, because it directly contradicts the stronger claim that "all apples are red". That would mean we have 2 sources and neither one can be used in the article, because they cancel each other. In the other case it would mean that source that states "apple c is green" can enter the article although the source expressing that claim has no footnotes or any other argument to sustain the claim, and it furthermore contradicts the undeniable claim "we know all apples are red". Asdisis (talk) 00:12, 2 June 2015 (UTC)

I also intend to expand this analogy to the concrete matter in hand, however I think that firstly this analogy has to reach a conclusion without any outside context and any knowledge to which discussion it is connected. Thus I urge everyone who know which discussion is connected to this analogy NOT to state it here until we firstly resolve the analogy. I also urge everyone not to look at the context and focus just to the analogy. We will come to the concrete matter after we reach the conclusion regarding this analogy, so do not worry. If you intend to participate you won't miss anything, just do not look ahead for the context yet. Asdisis (talk) 08:50, 2 June 2015 (UTC)

You may have to restate your example... because what you give us isn't actually an example of SYNTHESIS. Source A states that c is red (since it says that there are 10 red apples - one of which is c)... Source B agrees that c is red (in this case because it states that all apples are red)... so A+B is actually agreement between two sources, not a synthesis. Blueboar (talk) 01:21, 2 June 2015 (UTC)
The example is as I stated it. I do not need to change it because I haven't purposed any answer, although I , for the purpose of the second question said that I take the answer to be SYNTHESIS, since otherwise the answer to that question is obvious. However you had misunderstood the source A. It does not state that c is red. It states that all apples in the bowl are red. It does not state anything specifically about c, and that is the core of my question. Source A nor any other source states anything specific about c. I know that there is only one conclusion A+B gives, the the apple c is red, however we are not discussing the general logic, but a specific Wikipedia's rule. Also I would like everyone's opinion if a RfC is needed. This question is important regarding another discussion which lacks objectivity in my opinion. I can give another analogy. I will add it to the top one. Asdisis (talk) 07:51, 2 June 2015 (UTC)
Unless you are referring to a different apple named "c" Statement A does refer to apple "c" and describes it as red. "All" has no exceptions. • • • Peter (Southwood) (talk): 07:57, 7 June 2015 (UTC)
On second thought I thik this is a rather important question, and I am making a RfC out of this discussion. I apologize if I'm doing the wrong thing according to wikipedia's rules, however I'm not completely familiar with wikipedia's rules. I'm making a RfC in good faith to resolve this question. Asdisis (talk) 08:50, 2 June 2015 (UTC)
Perhaps a better example would be to ask what you really want to know. This is a great place to ask what you asked me on my talk page. Chillum 04:10, 2 June 2015 (UTC)
No. Please, I urge everyone not too look at the concrete subject in matter yet. I think this is a better way, because people are given an en example which does not bring any outside context. Of course I intend to bring the concrete example and we can then decide if my analogy is applicable or not to the concrete example. But for now I urge everyone not to look at your talk page and give their objective opinion regarding my question on this talk page. I would also like you restate your opinion and argument about the question on this talk page. I see that there is a disagreement even between experienced editors and I would like to resolve this question in the most objective way. That is why I had constructed an analogy and removed outside context. Also I do not know how to invite other editors, but I would like that a general consensus is reached regarding this question and more people are needed in this discussion. Asdisis (talk) 07:51, 2 June 2015 (UTC)
Wikipedia editors hardly ever read sources composed like logic exercises in a computer science book. The real authors of real sources will feel that if they wrote something that allows the trivial conclusion that the apple named "c" is red, then they have effectively stated that c is read, and there is no need to waste paper and ink explicitly giving a list "A is read, b is red, c is red, d is red, e is red, f is red, g is red, h is red, i is red, and j is red". Making obvious, trivial conclusions from material contained within a single source isn't synthesis, it's reading. Jc3s5h (talk) 10:54, 2 June 2015 (UTC)
I appreciate your opinion. However the real claim is more complex for your explanation to be applicable, but the analogy stands. Stay tuned, in an appropriate time I will go from analogy to the real case, which is as it always is with real examples, more complicated. Then we firstly we will have to decide (1) if the analogy is applicable, or I had constructed a wrong analogy (2) if the real example is the case of SYNTHESIS. Thank you again, stay tuned. I won't answer to each person who states his opinion not to clog the discussion, and I'll wait until an appropriate number of experienced editors states their opinion, then I'll introduce the real example. Asdisis (talk) 13:17, 2 June 2015 (UTC)
...and while we're waiting [7]. --Bob K31416 (talk) 15:39, 2 June 2015 (UTC)
Regarding your second example (Germany voting for S)... As you have presented it, without being able to review the sources and the context in which they make the statements, I would have to say that the conclusion is synthetic. We need a source to put the pieces together and affirmatively state that Germany voted for S because it is a member of the EU and all EU states voted unanimously in favor of S.
However, I strongly suspect that reality would never be as abstract as the example you give. In reality, all these statements would be made in some sort of context - and context is important. For example... I would find it highly unlikely that Source B would state "S was sustained unanimously" without context that made it clear who was doing the voting (the EU members, including Germany). And if such context is present, then Source B (on its own) is probably enough... we can state "Germany voted in favor of S <cite source B>".
This is why we keep asking you to be more specific. To give any meaningful answer to a question about synthesis, we need to examine the statements in terms of reality... a real article, and real sources... we have to look at the cited sources to see what else the sources say, and the context in which they say it. Blueboar (talk) 16:37, 2 June 2015 (UTC)
Imagine that we have a source stating that "the subject S was sustained unanimously by all EU member states". In fact, that will be the case in the real situation i intend to present. The real case will be presented in due time. I thought this is a better way to deal with the issue Asdisis (talk) 19:04, 2 June 2015 (UTC)

--Bob K31416 (talk) 17:04, 2 June 2015 (UTC)

Let's not jump ahead. I'm sorry you had seen the real example before you stated your opinion on the question asked above. You would not have been bounded to that question of you feel my analogy is not applicable. I purposed this way of dealing with the question to make a more objective discussion. Since I put it this way, people may be tempted to jump ahead, but please bare with me. ;) Asdisis (talk) 19:04, 2 June 2015 (UTC)

I apologize for the number of my comments. I'm just trying to be helpful, but I will try to reduce them to minimum so this topic does not get clogged. Asdisis (talk) 19:09, 2 June 2015 (UTC)

In my view Blueboar has a valid point. Complex socio-political issues (like EU voting) is much more context dependent than colours of apples. Your analogy between the two only works if statements A and B are universally accepted and not open to context or other interpretations in either of the cases. While I think we can safely assume this for apples, for EU voting this assumption is not trivial. Arnoutf (talk) 19:50, 2 June 2015 (UTC)
  • On the original question (I've not read all the back-and-forth about how it might apply to whatever case someone's having a dispute about), I concur with Blueboar that the first example, in basic logic terms, indicates agreement between two sources, not a synthesis. We basically could not write anything but single-source articles, otherwise. See also WP:NOTSYNTH. I don't agree with him that the case being expanded to include, e.g. EU voting, introduces so much interpretation or complexity that it constitutes synth (if I'm understand him correctly). Such determinations are going to be case-by-case. It's definitely not categorically synth. The EU example is presumptively valid encyclopedia writing; it would need a showing that the source is less reliable than thought, by using some idiosyncratic redefinition of "members states of EU". This sort of issue comes up every day, and is usually resolved through discussion of the sources on the article talk page. For a real-world example of how SYNTH can arise easily in the kinds of cases Blueboar is right to point out can be problematic, see Territorial claims in the Arctic. One source indicates that five nations "signed" the UNCLOS agreement. It would have been a mistake to reword this as "ratified", an easy mistake for a non-expert to make; two other sources demonstrate that one signatory nation hasn't actually ratified it yet. We're encouraged to write WP in our own wording, not plagiarize, but this has to be done carefully or it can accidentally result in SYNTH.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:53, 3 June 2015 (UTC)

I remind you that there are two questions asked above. Maybe it is better I give an example.

We have a general source that says S is either A D (those are to words) , B D, or C D. We assume this is correct. Not we have a source that says s (specific case of S , like apple a from a bowl of all apples) is D. Can we use the first source to say that the second source is ambiguous and that we can not draw any conclusion from it. I'll again present the real case in this matter as well. Asdisis (talk) 08:17, 3 June 2015 (UTC)

Again, you are not really presenting us with a synthetic situation ... we simply have two sources that say somewhat different things... and you are asking us to choose between them - or at least to determine how much weight we should give to each source. That isn't a WP:NOR question, but one of comparative reliability and Neutral point of view. And for that we definitely need to look at the real life sources. Blueboar (talk) 11:26, 3 June 2015 (UTC)

It seems to me that you may be presenting an overly simplistic example when the real situation is more complex. For those that want to cut past the hypothetical and get to the actual matter at hand you can look at: User_talk:Chillum#Tesla's citizenship discussion.

The question of somebody's citizenship in a time of chaos is a much more complex issue than the colour of apples. I think it is an excellent example of synthesis. Chillum 18:04, 3 June 2015 (UTC)

Please, I know the approach may not be the best one, but you should respect it. I've added a spoiler. Asdisis (talk) 19:27, 3 June 2015 (UTC)

Please do not fold, spindle, mutilate, or hat my comments again. This is not your talk page. Chillum 19:49, 3 June 2015 (UTC)
Oh, I'm not allowed to do that? Well then I must say that your act has made me present the real example now, although I do not have time to do it properly now. I had other plans instead of putting this example together now, and I specifically plead not to reveal the real example yet, so people can firstly thing about the abstraction. I don't know why you did that, but it's not correct towards me who now had to stay home and deal with this matter now. Asdisis (talk) 21:00, 3 June 2015 (UTC)
I am sorry if you did not like that. However this is a community discussion board and you don't get to frame the discussion or decide in what manner it goes forward. I found your abstractions to be false analogies at best and a waste of peoples time at worst. I think an informed community is more likely to come up with a good answer and that your abstractions will only result in confused hypothetical answers that will likely be misapplied. Chillum 21:08, 3 June 2015 (UTC)
Ok, but I said that no one is bounded by the abstract example. Everyone else respected the process. Well, now it's over, but I just wanted to state that your act caused me a great deal of trouble. Asdisis (talk) 10:54, 4 June 2015 (UTC)
If the approach is not the best one, we should not use it. You keep trying try to make an analogy between a complex social issue - that heavily depends on specific context and point of view of the different sources - and a simple one. That does not work, and your unwillingness to accept that your analogy contains a fundamental flaw (or providing case specific evidence why it does not) makes this discussion going around in circles. Arnoutf (talk) 19:45, 3 June 2015 (UTC)
Aargh. Ethnicity is involved. That is one of the most ill-defined concepts around. Ethnicity ranges from everywhere between racial ethnicity to self-perceived identification with a specific cultural group. In addition nationality is involved in a late 19th century context, a time when the current concept of nationalism was still evolving, so using the current definition for people from that time (especially those from non-nation states like Austria Hungary) is hopelessly convoluted.
In this context your examples would better be represented as
We know the following is correct.
*A="all members states of EU voted on the subject S" (but we do not agree on the definition of states, membership of EU and what exactly voting is)
*B="the subject S was sustained unanimously". (but we do not agree on the definition of unanimous, and we are pretty sure that sustained is not exactly the same as voting in support of)
*C="Germany is EU member state" (but we probably use a different definition for state and membership compared to statement A)
In this case I would say the conclusion that Germany voted in favour is definitely synthesis. Arnoutf (talk) 19:57, 3 June 2015 (UTC)
No it's not about ethnicity, but nationality. Nationality is a strictly defined concept and a set of well defined rules was used to define someone's nationality, because it was not done for specific case separately, but it was presumed. This is not ill defined as ethnicity. Anyways we are not having a discussion about it, but we are discussing wikipedia's rule, so the subject in the matter is irrelevant. Look at the real example below, you will see. The read discussion regarding this question is not this one so please let's not engage into resolving the question of another discussion, but let's determine if my question is a case of SYNTHESIS. Regarding your answer, you changed the question. In the real example the question of citizenship is strictly and well defined. Asdisis (talk) 21:09, 3 June 2015 (UTC)
People can and did change nationality. Just because you found some rule somewhere about the default nationality of certain types of people does not mean you can claim a specific person had that nationality. Strictly defined is exactly what nationality was not at the time. You are making assumptions. Chillum
Let's not go into the discussion. If you go to the discussion you can see I already found 2 sources that speak specifically of Tesla. One states that Tesla was "of Croatian origin" and the other put him in the "Croat" group. I should make one more question. If one source claims he was Austrian and other Croatian, then I think the more specific source is more valuable. Croatian is as specific as it gets, and Austrian incorporates Croatian, as all people in Austro-Hungary had a common Austrian citizenship, and then further Austrian or Hungarian national and further Croatian, Hungarian,...etc local citizenship. I will make third question, and I hope you won't regard this as a disruption because I posted so many comments in this discussion. Asdisis (talk) 21:23, 3 June 2015 (UTC)
So, someone of Croatian origin / Croat cannot have been citizen of any other nationality? I am not speaking about Tesla here, but in general. Sure one can be, let's say, Albanian origin but still have been a German citizen for his entire life. Moreover, please notice that "Austrian" is different from "Austrian citizenship". Trust me, there are people "of Croatian origin" who have Austrian citizenship even nowadays ;-) Cheers! Jayaguru-Shishya (talk) 15:36, 8 June 2015 (UTC)

The real example[edit]

I see some people are anxious so it's maybe time to go to the real example. We have 2 questions here, both regarding synthesis. I still urge everyone not to look at the discussion yet, but first to think about this question and give an answer.

1) If we take the following claim to be correct. "from 1861, local citizenship was constituted and it was determined by belonging[birth, marriage, residing for at least 10 years..etc] to a certain municipality [retrospectively]". In another words "all people who belong to a certain municipality have local citizenship C".

And if we know (this is not questioned) that person P was born, and lived in a certain municipality from his birth to 1861 and some time after.

Does the claim "person P has local citizenship C" constitute a came of synthesis?

This is the question 1

2) If the sources prove that (we take that to be true for this discussion) 2 forms of citizenship existed, national and local citizenship. National citizenship can be Austrian or Hungarian (with the other name of Croatian-Hungarian). Local citizenship can be Austrian,Hungarian, Croatian-Slavonian...etc..Furthermore we know that all people from Austro-Hungary were viewed by foreign counties as the same political nation with the name Austrian. We all this take to be true for this discussion.

If we have a source that claims "person p has Austrian citizenship", without any further explanation or any footnote. Exactly like this. This I haven't simplified.

Can we state in the article that person p has Austrian citizenship, although sources claim that Austrian national, Austrian local citizenship exists, and furthermore that any person from Austro-Hungary (with Austrian or Hungarian national citizenship) is regarded by foreign counties as "Austrian".

So one sources tell the general story about the question of citizenship, and one sources state a specific claim about a certain person without explaining to which form of Austrian citizenship it is talking about or if it even is talking about citizenship or it is stating the "Austrian" in the sense foreign counties see both Austrians or Hungarians. So a person with Hungarian citizen would be stated as Austrian citizen in US documents and the source would say that person p declared himself as Austrian citizen in US documents. Is it wrong to state that person p has Austrian citizenship on the grounds of ambiguity? Asdisis (talk) 20:26, 3 June 2015 (UTC)

Both assumption 1 AND 2 are unlikely to be completely unchallenged, and the definitions may have evolved in the last 150 years. So we should stick with reliable secondary sources that make these combinations and not do so ourselves. Arnoutf (talk) 20:37, 3 June 2015 (UTC)
But please, let's concentrate on the question of wikipedia's policy. We take those claims to be true for this purpose. If you want you can join the discussion . It would certainly be helpful to have more objective people in this complex discussion. Asdisis (talk) 20:56, 3 June 2015 (UTC)
Okay. First, what would the sentence be like that you suspect to be WP:SYN? I guess you had some sort of construction in your mind. Second, can you provide the exact quotes for each three claims? Sometimes simplified analogies are good, but not always. When it comes to the complexity of real world, you might quickly run into "omitted variable bias", if you may. Cheers! Jayaguru-Shishya (talk) 20:44, 3 June 2015 (UTC)
If "person P has local citizenship C" a SYNTHESIS from the claim "from 1861, local citizenship was constituted and it was determined by belonging[birth, marriage, residing for at least 10 years..etc] to a certain municipality". This is the main question. Note that we are talking about Wikipedia's policy, and we are not engaging into determining whether the initial claim is correct. The reason is that some people on the talk page claim that this example constitutes SYNTHESIS. Their argumentation is, as I have understood is, "if Tesla is not mentioned specifically, it is a case of SYNTHESIS", regardless of the claim. Asdisis (talk) 20:56, 3 June 2015 (UTC)
Yes, it is SYN. If the sources do not specificly state "X is a citizen of Y", we cannot take content from multiple sources to make the claim. Its is clear SYN - particularly under a situation where citizenship rules are changing. It is the clear example of taking content from multiple sources to advance a claim that is not specifically made by any of the sources. -- TRPoD aka The Red Pen of Doom 21:01, 3 June 2015 (UTC)
Thank you for your answer, please don't forget the other question, although I did not construct it so the question is if it constitutes an example of SYNTHESIS. Unfortunately I was hurried so I did what I could. The other question is regarding the sources that speak of the general question of citizenship and their role in the verification of the specific claim. Asdisis (talk) 21:12, 3 June 2015 (UTC)

3) I will add a third question. How many specific sources are needed to prove the general case. The general case defined 3 kinds of citizenship. I already asked what we can conclude if one source says Austrian. This question is the same but it deals with the most specific citizenship, local citizenship. If I have a source that says Tesla was of "Croatian origin" is this source in clash with the source that states Tesla had "Austrian" citizenship taking into consideration the sources that explain in general the question of citizenship.

Imagine if we had several (only fer 2,3 sources) source that state "Austrian" (which is in my case ambiguous) and some sources that state "Croatian" (now only one source states "of Croatian origin"). What to state in the article. Nothing, Austrian or Croatian?

Note that sources are sparse and I haven't yet studied Croatian sources that may deal with the question more extensively than foreign sources that only mention "Austrian" or "Croatian" in one sentence without further explanations or any footnote. Asdisis (talk) 21:30, 3 June 2015 (UTC)

  • I have to apologize once again. I was hurried into constructing the real example yesterday so I forgot to present the most important example that the abstract examples (apples, Germany) were derived from.
(4)If we have a source that says "All people living in Austrian Empire up to 1867 had Austrian citizenship". We take that claim to be true. If we know Tesla lived his whole life up to 1867 in Austrian Empire, does the claim "Nikola Tesla had Austrian citizenship from his birth to 1867" constitute a case of SYNTHESIS? Asdisis (talk) 12:49, 4 June 2015 (UTC)
Yes, it is SYNTHESIS. I would only object to it if I wasn't reasonably sure that the synthesis was true. WP:IAR. I would let consensus decide whether or not to put it into the article. --Bob K31416 (talk) 14:40, 4 June 2015 (UTC)
(5)The other example is: if we have a source that says "All people living in the lands of Hungarian crown of the Austrian Empire in 1867 gained Hungarian(Croatian-Hungarian) citizenship". We take that claim to be true. If we know Tesla lived his whole life in the lands of Hungarian crown up to 1867 and several years after, does the claim "Nikola Tesla had Hungarian(Croatian-Hungarian) citizenship after 1867" constitute a case of SYNTHESIS? Asdisis (talk) 13:07, 4 June 2015 (UTC)
Yes, it is SYNTHESIS. I would only object to it if I wasn't reasonably sure that the synthesis was true. WP:IAR. I would let consensus decide whether or not to put it into the article. (same answer as above) --Bob K31416 (talk) 14:40, 4 June 2015 (UTC)
  • Notice that there are 5 questions here. I apologize for that, but I was hurried into constructing the case earlier that I had planned, because one editor had revealed the topic although I plead to deal with the abstract example first. It would be of great help to resolve this 5 questions. Note that we are not having a discussion about Tesla's citizenship. We are having a discussion about Wikipedia's rules. Some people already went in the wrong way stating that "People can and did change nationality. Just because you found some rule somewhere about the default nationality of certain types of people does not mean you can claim a specific person had that nationality. Strictly defined is exactly what nationality was not at the time. You are making assumptions.", or "Aargh. Ethnicity is involved. That is one of the most ill-defined concepts around. Ethnicity ranges from everywhere between racial ethnicity to self-perceived identification with a specific cultural group." or "Both assumption 1 AND 2 are unlikely to be completely unchallenged" This and comments like this are regarding the discussion about Tesla's citizenship, and we are not having that discussion here. Thus is why we take certain claims that speak of the general case of citizenship to be true, so we can determine if the claim about a specific case constitutes SYNTHESIS. Note that the user TRPoD aka The Red Pen of Doom is the only one that had the right approach. His answer that "If the sources do not specificly state "X is a citizen of Y", we cannot take content from multiple sources to make the claim. Its is clear SYN" does not go into the real discussion about Tesla's citizenship but deals with SYNTHESIS. Unfortunately, since a certain user has hurried the discussion we haven't heard this user's opinion on the abstract examples thus I would like to ask him to answer those questions as well. From this explanation I would conclude that, in the abstract example with apples, his opinion is that that case constitutes SYNTHESIS as well. I think it is obvious why I started with abstract examples, because people already went the wrong way, to the discussion about Tesla's citizenship which is not this discussion. Asdisis (talk) 13:07, 4 June 2015 (UTC)
I responded to (4) and (5) above, based on the premises you gave for those hypothetical cases. I think I chose those two items because they seemed to be the ones that were simplest. Looking at the others too, I'll try to make a statement regarding this type of situation in general.
Regarding adding material to a Wikipedia article, it should adhere to Wikipedia policy WP:NOR. In rare cases, material that does not can be added anyway according to WP:IAR, but only if it has consensus. On the other hand, regarding keeping out of an article questionable material that has a reliable source, that is possible per Wikipedia policy section Verifiability does not guarantee inclusion.
In the particular case you are concerned with, having to do with adding that Tesla has some citizenship prior to 1891 other than Austrian, you don't seem to have a reliable source that says that specifically about Tesla, and there doesn't seem to be a consensus that it should be included anyways according to WP:IAR. So I don't think you can put it in the article. Regarding Tesla's Austrian citizenship, that is supported by a reliable source that specifically applies to Tesla, and there seems to be a consensus that this should be in the article. I think this issue of Tesla's citizenship prior to 1891 is too complex and controversial for Wikipedia editors to try to determine themselves without reliable sources about Tesla, and we should go with the available reliable sources that are specifically about Tesla. In other words, for this case I think we should follow WP:NOR, rather than WP:IAR. --Bob K31416 (talk) 18:20, 4 June 2015 (UTC)
With due all respect, I think the apple example was rather a poor one. First of all, we would have to ask, "is the substance of apples permanent, or can it be substantiated?" For example, citizenship can be changed from one to another, but the substance of apple can be not. And I guess that's what you failed to take in account in your analogy. Cheers! Jayaguru-Shishya (talk) 15:40, 8 June 2015 (UTC)
Thanks for your opinion. However you had dealt with the other discussion. Here we only need to reach a conclusion regarding the 5 questions and Wikipedia's policy. One question asks if the sources that speak of general case can help us with interpreting the sources that speak of Tesla. If one source states that Tesla "states Austrian citizenship" and the general sources state that "he would be obligated to state Austrian citizenship even if he had Hungarian citizenship" and that the context of the word "Austrian" in general is not regarding citizenship but "common affiliation that would appear to foreign states". Also I asked if the following constitutes a case of "guessing". One source says "Tesla had Austrian citizenship", other sources say "there existed common affiliation[(not citizenship)] to the empire named Austrian the would be used in foreign affairs, Austrian and Hungarian national citizenship, and Austrian, Hungarian, Croatian-Slavonian...local citizenship". So, is it to ambiguous to state "Austrian citizenship" in the article. Furthermore the source does not say "had Austrian" but "stated Austrian" which is a significant difference if we take into consideration that other sources say Austrian was states in international affairs in the context of common affiliation to the empire, and not in the context of citizenship. I will advocate against such interpretations that go against what the other sources that speak of citizenship in general say. And lastly, we had again gone into the wrong discussion. Also I'm afraid I will be accused of disruption and banned because I posted too many comments in this discussion. I apologize for that, but I'm trying to be helpful. Asdisis (talk) 19:44, 4 June 2015 (UTC)
After all the discussion of these various issues, here and elsewhere, how would you assess the consensus of Wikipedia editors regarding what you are advocating? --Bob K31416 (talk) 21:28, 4 June 2015 (UTC)
The problem is that consensus in such discussions with high nationalistic tensions has to be determined among objective arguments and sources. As I found out it is mostly determined by voting and thus we have problems. Asdisis (talk) 12:46, 5 June 2015 (UTC)
The consensus system is what is used to determine what goes into Wikipedia articles. This doesn't mean that the consensus is necessarily right for a particular issue. If the consensus is against an edit, and there seems to be little or no hope of changing that decision with more discussion, I think it's best to leave the discussion and move on. --Bob K31416 (talk) 19:23, 5 June 2015 (UTC)

  • COMMENT - Why are we still arguing about Tesla's nationality? This is a debate that has been going on for almost 10 years.
Here is my solution to he whole debate... Omit it. It does not matter what Tesla's ethno-nationality was... or whether saying he was X or Y is original research. Unless his ethno-nationality was a defining characteristic for Tesla himself (and I would argue that it wasn't) then there is no need to mention it in the first place. So... I would suggest that we simply omit discussing his ethno-nationality... completely. If we don't state his ethno-nationality then there is nothing to argue about, It woun't matter whether claiming him as an X or Y is original research, because we won't include any claim. I think it is enough to simply state that he was born in X town ... moved to Y city in Z year ... etc? Those are uncontested facts that all sides can agree on... There is no need to put contentious entho-national labels on those basic facts. Blueboar (talk) 13:57, 4 June 2015 (UTC)
We are not. This is only the real example in this discussion about Wikipedia's rules. This is exactly why I started with an abstract example. I would appreciate if everyone could give their opinion of all 5 questions. Asdisis (talk) 16:47, 4 June 2015 (UTC)
abstract examples are meaningless... context always matters. Blueboar (talk) 17:48, 4 June 2015 (UTC)
That is why I always intended to present the real example. I thought we can start with abstract so not to get involved into the wrong discussion. I never wanted to show the real discussion. The real example would then be just like another abstract example but this time with context, however people spoiled that intent and introduced prejudice. That much is obvious from comments like "Aargh. Ethnicity is involved.". Let's concentrate on the 5 questions that are to be resolved by this discussion. Asdisis (talk) 17:56, 4 June 2015 (UTC)
Abstract examples only work if the translation of a word to an abstract term is uncontested. My "aargh ethnicity" comment was to underline that the term ethnicity is among the words where the abstract underlying idea is about most contested of all words I know. Hence abstraction of that specific term out of context is impossible. Basically this follows all the previous. Without the context we do not know whether the abstraction can be upheld. Now we know the context, we can say the abstraction most likely cannot be upheld Arnoutf (talk) 18:55, 4 June 2015 (UTC)
Concur. Using an abstract example when you're really talking about an active debate can only be interpreted as disruptive, as intending a "Ah-ha! Gotcha!" moment. That's not helpful anywhere and especially not here. Ravensfire (talk) 19:47, 4 June 2015 (UTC)
Ditto. I also concur with Blueboar that the obvious solution is to eliminate the conflict by not trying to shoehorn Tesla into some kind of ethno-nationalistic category.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  21:47, 4 June 2015 (UTC)
I agree only in the way that the questions that are ambiguous and not found in sources are to be removed from the article. For instance it is stated that Tesla is Serbian-American scientist and a bunch of sources state him as Croatian, or Yugoslav scientist. The Serbian-American also neglects that Tesla by nationality was Croatian-Hungarian. The discussion of high nationalistic tensions is hardly to be determined by consensus. Asdisis (talk) 12:46, 5 June 2015 (UTC)
  • FYI... Asdisis got himself blocked for POV pushing at the Tesla article. I think we can end the discussion now. Blueboar (talk) 23:14, 5 June 2015 (UTC)

Important question which has to be resolved[edit]

I made a separate section because this question is of at most importance to prevent similar discussion to the one about Tesla's citizenship.

  • The last question. I'm sorry but this is the last question about SYNTHESIS. It is rather important. I thank you all for your participation and I'm glad I had the fortune to participate in such objective discussion. We had reached some conclusions that will resolve another hard discussion.
(6) If source A says "Hungary, moreover, promised to help Croatia to obtain both Dalmatia and Military Frontier Province, still under Austrian control. The latter territory, after many delays, was incorporated in Croatia in 1881". Note the word "control", which is ambiguous by itself, because Military frontier could be in legislative sense territory of the Kingdom of Croatia, but under Austrian (military) control. This word brings context upon which all who participated had insisted. If source B says "Nikola Tesla was born in Military frontier". Is the conclusion "Nikola Tesla was not born in Croatia" a case of SYNTHESIS. I must note this is a rather important question because the answer to it can prevent another exhausting discussion like the one of Tesla's citizenship. Asdisis (talk) 22:50, 4 June 2015 (UTC)
I would suggest avoiding even the potential for OR by simply saying: "Tesla was born in what is today Smiljan, Croatia" (bolding added here to highlight my point). That turns the statement of his birthplace into a non-contentious fact of modern geography... without even hinting at the potentially synthetic issue of what his birthplace means in terms of his nationality or citizenship. Blueboar (talk) 14:01, 5 June 2015 (UTC)
But let's not get into a different discussion. For now the conclusion about the above question would resolve the problem of walls of texts and it would direct the discussion to come to presenting sources about Tesla. If majority of sources tell Tesla was born in Croatia, and if Tesla himself stated that, I do not see any problem stating Croatia in the article. I have no contempt towards Croatia, but others do. The formulation "what is today Croatia" is coined by them so they can negate Tesla had any connections to Croatia. If you think I'm wrong, just dare to mention Croatia in the discussion, you will see. However we again digress... Asdisis (talk) 14:32, 5 June 2015 (UTC)

SYNTH is not a rigid rule[edit]

Calm down everyone.

Pretty much by definition, situations which might be revert-able due to "synthesis" fall into 3 categories:

  • The obvious cases, where nobody except maybe the editor being reverted would argue otherwise
  • The "obviously pedantic" cases where it may technically be synthesis but keeping the edit (or finding a different reason to remove or edit it out) is clearly the better option (see WP:Ignore all rules if you are confused on this point)
  • The unclear cases, which, by definition, need to be discussed on a case-by-case basis.

davidwr/(talk)/(contribs) 20:24, 6 June 2015 (UTC)

Proposal on "rather than OR" wording[edit]

I see that one of my edits, in which I removed italics for emphasis, was reverted with the comment that the point being made was important because some people on the talk page "weren't getting it". My point was that overuse of italics for emphasis is generally regarded as poor writing style. Also they make the policy sound like an extension of a talk page argument. Ideally the policy should be based on consensus, not a king-of-the-mountain style struggle to wrest control of the policy from the other side of an argument.
Another thing that has gone missing is that this passage used to say that interpretive claims must be cited from secondary sources "rather than" being based on the editor's OR. That is something that everyone can agree about. After all, the name of the policy is "No original research". When presented with a choice between secondary sources and OR, everyone will choose the secondary sources. But currently the choice is gone and we have the assertion that all interpretive claims must be cited from secondary sources. Period. But what about tertiary sources? I realize that SMcCandlish has just written an essay arguing against the use of tertiary sources for such claims. But this seems to me to be a minority viewpoint, or at least one with less support that what we used to have. I propose that we put the "rather than OR" back in. – Margin1522 (talk) 10:44, 4 June 2015 (UTC)
Emphasis: I think we all agree that over-use is poor style; I'd been of a mind to remove the emphasis from the word "may" there because it seemed to serve no purpose. In the case of "only if", it should remain emphasized, because the rule is crucial (it's one of the most important content policy rules we have), yet frequently missed. There is no "other side of an argument" from which "control" must be "wrested"; I don't understand the point of imagining such a hyperbolic scenario. The policy is emphatic there because new editors (and some not-so new ones) often get this backwards. This is especially true of university and grad school students, and professional academics, all of whom are very used to seeking primary not secondary sources; see related other thread on this page that brings up this very point of confusion. Academic and encyclopedic writing have radically different purposes and bases.

"Rather than": We'd have to dig around in edit history and archives, with regard to the "rather than" bit. The problem with just shoehorning it back in is that the policy has been clarified in multiple ways in the interim. Analytic/evaluative/interpretive/synthetic claims require reliable secondary sources period, rather than everything else of any kind at all, including primary and tertiary sources, not just rather than an editor's own OR. (Another way of looking at it is that it's a form of OR to reinterpret primary sources' novel claims, or tertiary sources' uncritical regurgitation of data, as supporting an analysis or interpretation not found in reliable secondary sources, but it's much harder to explain this to editors than simply saying such claims require secondary sources, and moving on. Our policies aren't supposed to be essay-like.) So, if we want some "rather than" phrase, we need to come up with adequate wording that conveys that concisely. To me, it seems it would be redundant. Maybe it's okay for WP policies to be a little redundant, to get an important point across in multiple ways (see thread above about about adding back in something here that is kind of a restatement of "Wikipedia is not a publisher of original thought"). But that's basically just an argument for adding a different form of emphasis that some would regard as even worse writing than simply italicizing two words.

Tertiary sources: The idea that tertiary sources can be used for these kinds of claims (an idea that never seems to come with clear examples) is the minority viewpoint, and the policy already indicates in three distinct places that secondary sources are actually required. The very fact that you and one other editor commenting earlier (who appears to have come around, since he made a copyedit that clarified the matter even better) had been somehow interpreting "only if" as if it's not there, or doesn't mean what it says, is strong evidence in favor of the need to emphasize that "only if" phrase. See the other discussions, but the short version is: If an otherwise tertiary source is making such a claim then it is, for that claim, not a tertiary source, by definition. Tertiary sources just collect and re-present data from secondary and sometimes primary sources, without significant analysis/interpretation; secondary sources are those that engage in analysis/evaluation/interpretation/synthesis. The problem here is in misconceptualizing primary/secondary/tertiary. They are not innate or necessarily immutable qualities of a work, but descriptions of how they relate to a particular piece of information as it is being used in a particular context. There's an implicit "for" that comes after "tertiary source" (or primary or secondary). A source can be all three of these at once for different claims even in the same article. As one type of example, a topical encyclopedia written by a subject-matter expert is often all three kinds of source, depending on the details in question. A piece of editorial journalism, as another example, may also mix all three types (e.g. tertiary source for various gun-fatality statistics in a sidebar, secondary for analysis of pro- and anti-gun-ownership stances and their motivations, primary for the conclusion reached by the writer about how well a piece of legislation will address a gun-control issue). PS: the WP:TERTIARYUSE essay doesn't "argue against" their use for analytic claims; the policy says not to use them that way. The essay explains the policy and advises on how to apply it (and it's an in-progress draft, so it may not do so inadequately at this point). PPS: You said yourself earlier on this page: '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.' So, it's not clear to me where any real disagreement lies if we get the terminology straight.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:49, 4 June 2015 (UTC)

Here is an evaluative statement from the 1911 Encyclopedia Britannica (the classical example of a tertiary source), reproduced without change in the first version of our article on Anaximander:
Tradition, probably correct in its general estimate, represents him as a successful student of astronomy and geography, and as one of the pioneers of exact science among the Greeks.
The problem with an absolute statement like "only if secondary" is that it only takes one counterexample to prove the statement false. This example shows that it has never been true, since we have been including evaluative statements from tertiary sources since the very start of Wikipedia.
Nor do I think it's helpful to split hairs over "secondary statements" embedded in "tertiary sources", or engage in similar kinds of legerdemain. The only thing this policy needs to do is explain to editors that they can't say "probably correct" in their own voice. – Margin1522 (talk) 08:50, 5 June 2015 (UTC)
Aren't we much more concerned about it being said in Wikipedia's voice? It's usually easy to see when people insert their own random opinions, but much harder to detect when a reasonable-seeming evaluative statement is synthesized from a pile of sources but doesn't actually appear in any secondary ones. It can take hours of source verification to detect. A statement like the Anaximander one might be followed with 3, or 5, or 7 citations, all of which show that tradition says that about him, but none of which suggest the tradition is "probably correct" (or maybe one does, but it's a primary source). Cases of incautious use of material like Britannica 1911 has led to some inclusion of synthesis untraceable to a secondary source (if you want to view it that way) doesn't "disprove" that we have a rule that requires reliable secondary sources for it. The fact that a majority of people drive a little over the speed limit doesn't disprove that there's a law against it, it just demonstrates lax enforcement and a human penchant for minor transgression. A lot of Wikipedians probably don't even realize they're transgressing when they take the average view of the sources they've bothered to look at, and if those are reasonably consistent, say something like "The majority of sources agree that ..."; we need to keep saying this isn't the way to do it, and maybe say it more clearly. The essayists in many of us don't like writing of the form "Many sources, including X, Y, and Z, agree that ..., while others, such as N, conclude instead that ..."; but it's often necessary here, when the views in the RSes conflict and there's not a WP:UNDUE problem in reporting the disagreement. I don't see a way around that. PS: The "old sources become primary" theory, toward which WP seems to lean, would at some point (perhaps already) treat a century-old encyclopedia with laxer editorial standards than ours as a primary source anyway, no longer tertiary. It makes me wonder how much Britannica 1911 material we still have around, not yet re-worked and re-sourced. There aren't many topics, even figures of classical history, about which absolutely no scholarship has advanced in over 100 years.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  19:43, 8 June 2015 (UTC)
I have to say that I see no benefit whatsoever in redefining the EB1911 as a primary or secondary source as convenient, depending on whether we want to keep the passage in question or throw it out. Of course research into the presocratic philosophers has advanced in the past 100 years. Recent sources may be better. For that reason, we may want to say that the EB is no longer a "reliable" source. In that case, I would suggest taking that question up in the appropriate forum. The goal of this policy is is not to offer a watertight definition of "reliable", or to close every "loophole" that might allow editors to include content that someone doesn't like in the encyclopedia. This policy is supposed to be about original research. There is no sense in which citing the 1911 Britannica for a statement like the "probably correct" one above is original research, and this policy should never be used to prevent someone from doing that. – Margin1522 (talk) 02:37, 11 June 2015 (UTC)
It has nothing to do with whether we want to keep it or throw it out, only what kind of claim is being made. Well, there is one other factor too: Age. By the "100 years" rule observed by most serious researchers, Britannica 1911 has to be treated as a primary source, because it is too close the events about which it was writing (on then-modern topics), and too much new research has been done in the intervening century=plus. And that 100-year cutoff (often much shorter, more like 20 years) actually is extended to secondary sources, not tertiary ones; those actually date much more quickly. The 1911 edition has been regarded as primary for several generations now. It's tertiary only in tone and approach, not for any reliability analysis. Cf. #RfC: Should "news articles" be added to WP:PRIMARY?. Unanimously against treating all news sources as primary "the way historians do"; as various commenters pointed out, it's older news sources that have to be treated as primary.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:14, 23 July 2015 (UTC)

Television template encouraging OR?[edit]

I wasn't certain what to think about this issue, so I thought I'd bring it here for some input! My apologies in advance if my question is a bit nebulous! Basically, I feel in the context of TV/film, there is a tendency to privilege primary sources when Wikipedia generally privileges secondary sources.

We begin with an edit dispute over Humans (TV series), an 8-part TV mini-series starting this Sunday: see Talk:Humans (TV series) for details. It seemed to me and other editors that the infobox should say there will be 8 episodes, as reported in numerous reliable secondary sources. However, other editors feel differently. It appears that common practice with this template (Template:Infobox television) is that the number of episodes is incremented only as they air: see Template talk:Infobox television#num_episodes where discussion moved to. Likewise, the template says the infobox can only include the first-aired date after the episode has aired, not even a few days before when the date has been reported in numerous reliable secondary sources.

Now, there are a number of issues here around, for example, which number is more interesting and useful to the reader, the number of episodes there will be or the number that have been broadcast to date. I may be quite wrong about my feelings over those! But we don't need to get into them here. What strikes me is a more fundamental issue about Wikipedia.

As far as I understand, common practice among some TV-interested editors has become to watch a show and then increment the episode number (and first aired date) in the infobox having done so. I feel the infobox should be a summary of the article content (WP:IBX) and should reflect what secondary sources say. It seems to me very odd and un-Wikipedia-like to privilege this direct reportage of Wikipedians watching a show over what is reported in secondary sources. Someone watching an episode and incrementing the episode number is, it seems to me, at best using a primary source and, at worst, tantamount to original research. The bias in Wikipedia should always be to secondary sources. High profile content in an infobox should be for well-sourced, encyclopaedic content. If someone wants to tally each episode of a TV series in real time, I feel they can do that on a blog/Twitter/show-specific wiki/&c. (I can see WP:BALL is relevant here. I understand events some time in the future are a different matter, but I am talking here in the context of a show starting in 4 days that has been described by numerous reliable sources.)

I have seen this tendency to privilege primary sources and waiting for a show to air in related contexts. At Talk:Jenna_Coleman#New_Name and Talk:Doctor_Who/Archive_24#Jenna_.28Louise.29, you can see an old discussion over the name an actor uses where some editors were arguing we should wait until the first episode with the new name in the credits is broadcast. Likewise, there have been arguments over whether The Matrix should refer to the Wachowski Brothers (as they were credited in the primary source) or the Wachowski siblings (noting Lana Wachowski's current gender identity; see Wikipedia:Gender identity). Thinking about it, I'm more concerned about those examples and that's probably why I'm bothered by the Humans (TV series) example...!

So, what do others think? Is there some existing guidance on primary/secondary sources when dealing with such matters? Should an understandable desire to keep Wikipedia bang up-to-date be encouraging reference to primary sources? Do we wait to see what the credits say, or do we report what others have already said? Am I worrying over nothing? Any thoughts welcome. Bondegezou (talk) 18:06, 10 June 2015 (UTC)

WP:V / WP:CRYSTAL / WP:NOTADVERT all seem to be relevant. If we use the sources to reference all of the episodes before they actually air, we would need to couch it "Scheduled for 8 episodes" not as "8 episodes" -- TRPoD aka The Red Pen of Doom 18:43, 10 June 2015 (UTC)
I concur. Way too many TV show infoboxes state as fact things that are conjectural/projected/planned/announced with regard to the forthcoming or still-ongoing season/series. Happens with films a lot, too.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:10, 23 July 2015 (UTC)

RfC: Should "news articles" be added to WP:PRIMARY?[edit]

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.
The answer is no, unanimously, to both questions. Kraxler (talk) 03:48, 20 July 2015 (UTC)

Question: should WP:PRIMARY be updated to define news sources as primary, aligning more closely with the definition used by historians?


Our policy limits how primary sources can be used, particularly regarding living people. Exactly what constitutes a primary vs secondary source is subject to some debate; for example, the current version of our policy defines secondary sources as those that contain an author's interpretation, analysis, or evaluation of the facts, evidence, concepts, and ideas taken from primary sources. Primary sources are defined as personal accounts such as witness statements. Historians, in contrast, define contemporary news sources as primary - discussed, for example, here.

At Talk:Dennis Hastert (permalink), the latter definition is being applied. Because of the substantial impact to our sourcing practices if news sources are now considered primary at Wikipedia, our new practice should be updated in WP:NOR. Proposed revised wording of WP:PRIMARY follows:

VQuakr (talk) 15:27, 18 June 2015 (UTC)


  • Oppose & Speedy Close If news articles (from reputable pubs like the NY Times, Telegraph, BBC News, etc) from the time period can not be used as secondary sources, then a massive amount of content and many entire articles which have been live on wikipedia for years without any issues would need deleting. I'll go further and say that this RFC should be speedily closed since there's pretty much zero chance of this becoming policy - the issue here seems to be that an admin has closed an RFC in a way that is wholly inconsistent with existing wikipedia guidelines and policies, not that there is any actual need for a policy change. The best solution is to leave the policy as-is and overturn the erroneous RFC close. Fyddlestix (talk) 15:48, 18 June 2015 (UTC)
    • I think that if you've actually built whole articles on century-old newspaper articles (especially of the "current events" or "society column" types), then you've done something seriously wrong, even if the publishers were reputable back then. WhatamIdoing (talk) 08:27, 20 June 2015 (UTC)
  • Oppose, because the proposer repeatedly fails to understand what's going on. Some news articles are secondary sources, while others are primary. In this context, these news stories are primary sources, because they're news about the event that's going on (try taking a historiography class and telling your professor that these are secondary!), but the idea that all news articles are primary sources is a basic failure to get the point. Nyttend (talk) 15:57, 18 June 2015 (UTC)
@Nyttend: probably "continues to fail to understand" would be more accurate than "repeatedly." I thought I had addressed this with the "contemporary" phrasing in the proposed addition. Is there an alternate formulation of the proposed policy revision that would help me (and future readers) understand without the need to supplement our policy with external sources such as GSU? VQuakr (talk) 16:03, 18 June 2015 (UTC)
@Nyttend: No one questions that historians would define news articles as primary sources. The question is whether wikipedia defines them as such (and, more importantly, whether it wholesale rules out their use in BLP articles like Dennis Hastert). Assuming that we're talking about reputable news sources (ie, newspapers or record), I have never seen anyone suggest that the latter is true. In fact, sources like this (ie, news articles "about events that are going on") are used in wikipedia articles constantly (like, hundreds of times, every day) without anyone having an issue with it. Fyddlestix (talk) 16:14, 18 June 2015 (UTC)
@Nyttend: You are incorrect. Newspapers from the time of the event are not primary sources from WP's perspective. If you want to change the rules to forbid them, you can start a RFC here with that wording specifically, but I can't see it having a serious chance of passing. --Aquillion (talk) 01:14, 19 June 2015 (UTC)
Nyttend was referring to (among other things) the Wikipedia article Primary source, not to the Wikipedia policy WP:Primary, as I understand.Anythingyouwant (talk) 01:22, 19 June 2015 (UTC)
Nyttend is correctly describing the policy as written on this page. I'm not quite sure why this is such a hot button with this group though: there is no prohibition on using primary sources in an article (not even a BLP). The overall article should be largely based on secondaries, but primaries are acceptable, including primary news reports in reputable newspapers. WhatamIdoing (talk) 08:39, 20 June 2015 (UTC)
  • Oppose defining news sources as primary, because many news sources are not. Incidentally, I do not see how opposing here will affect the legitimacy of the Hastert RFC close.Anythingyouwant (talk) 17:12, 18 June 2015 (UTC)
  • Oppose & speedy close - Newspaper and media sources from reputable media organizations are not primary sources from WP's perspective. If we change the policy as suggested, thousands of articles will become no longer compliant. - Cwobeel (talk) 19:08, 18 June 2015 (UTC)
  • Oppose Although I can see where it comes from. Press agency releases that just list some facts are probably close to primary reports, while a thorough journalistic article that analyses interprets and discusses multiple angles is definitely a secondary source. Arnoutf (talk) 19:40, 18 June 2015 (UTC)
  • Comment There is an obvious consensus against the specific phrasing I proposed. Rather than this being a slam-the-door, yes-no vote I think it would be more effective to discuss whether specifically addressing news sources in the policy would be worthwhile. We have done something similar with the extensive discussion of medical literature in WP:MEDRS. News is probably simpler but such sources are very, very commonly cited on Wikipedia. To formulate a more explicit problem statement: The present wording of our policy does not differentiate between news sources that review other sources, and current-event sources that lack historical perspective. VQuakr (talk) 00:56, 19 June 2015 (UTC)
  • Strong oppose. Newspapers from the time of the event are not primary by Wikipedia's standards. --Aquillion (talk) 01:04, 19 June 2015 (UTC)
  • Oppose & speedy close. I agree with Aquillion, Fyddlestix, and Cwobeel, for the reasons stated above. Neutralitytalk 03:51, 19 June 2015 (UTC)
  • Oppose and speedy close Unless someone comes along and wants to open an RFC to propose to change wikipedia's policy on this (and I would oppose that change as well) then we should not have this RFC just because of one article RFC closure. This would be unworkable for wikipedia without huge changes in what we cover, with practically any current events impossible to cover. Davewild (talk) 06:52, 19 June 2015 (UTC)
  • I recommend that the people here who are making claims about "Wikipedia's standards" actually read this policy, paying special attention to the third footnote. This policy has listed news articles as a type of source that is (typically) primary for years. WP:PRIMARYNEWS deals with the nuances better than this policy (Anythingyouwant is correct that there are different types of "news" sources, and that not all are considered primary), but this policy has called news sources "primary" for a very long time – and it is "Wikipedia's standard" for what constitutes a primary source on wiki. WhatamIdoing (talk) 08:35, 20 June 2015 (UTC)
  • Oppose as an incredible simplification of the issue. Grognard Extraordinaire Chess (talk) Ping when replying 00:12, 25 June 2015 (UTC)
  • Oppose. Newspapers are mostly secondary sources, despite what the RfC question says (which is not neutral, because it begs the question). A primary source is written by someone very close to the event, usually someone involved or an eyewitness. For example, a newspaper report by a journalist who witnessed 9/11 is a primary source for material about 9/11. But most newspaper reports are not written by involved people or eyewitnesses, and therefore most of the newspaper articles used as sources on Wikipedia are secondary sources.

    A confusing factor, and perhaps this is what the OP means when referring to historians, is that contemporaneous newspaper articles about an event a long time ago – articles written at the time of the event – might be regarded by historians as primary sources, because the authors were close to the event relative to us. So contemporaneous newspaper articles about World War I might be primary sources about the war, because they tell us how the war was reported at the time. But for us, alive in 2015, an article by a journalist who watched 9/11 on television is a secondary source because the author wasn't close to the event relative to us. Sarah (talk) 02:44, 25 June 2015 (UTC)

    • I agree with you that decades-old newspaper articles are (almost always) primary sources, either for what people thought was happening or for how they interpreted it. But a bit of physical distance isn't sufficient. Secondary sources require some sort of transformative thought. Merely being a thousand miles away, but writing about what you see happening through the miracle of telecommunications, is not the kind of mental separation that's necessary. WP:Secondary does not mean independent, and primary doesn't require being personally involved. Also, a good deal of them actually are written by journalists who are also eyewitnesses. I just flipped through the local news for the biggest newspaper in my area. I found a report of a local government meeting (the reporter was an eyewitness to the meeting), some kid talked to a politician (eyewitness), took pictures at a sports event (eyewitness), a photogenic political protest (eyewitness), student sports events (eyewitness), more government meetings (ditto), a drug bust (wasn't clear), routine court news... I don't see any reason to think that any of those are secondary sources, although they're all WP:INDY sources (with no COI). WhatamIdoing (talk) 03:08, 25 June 2015 (UTC)
      • @User:Aquillion I agree with SlimVirgin (that makes a change Sarah!), the only thing I question is whether 9/11 is still current (there may be people reading this who weren't even born then). The question of when an event slips into history is difficult to pin down, but the 30 years rule for government papers is probably as good a point as any. I have contributed to a number of English Civil War articles some of which mention contemporary newspaper articles. There is no way that such newspaper articles can be considered anything but primary (see for example The Moderate Intellingencer). On the other side of that equation there is a Wikipedia page Wikipedia:Press coverage 2015 and those articles are either reliable or unreliable secondary source, but one day they will be primary sources. Those in the page Wikipedia:Press coverage 2001 are moving that way, because they tell you more about what people thought about Wikipedia in 2001 than they do about what Wikipedia is now (eg see the [MIT Technology Review a linked on the Press coverage 2001 page). An interesting area for discussion in the future will be what happens when a Wikipedia articles grow old. For example I helped to write the article Eritrean–Ethiopian War (1998–2000). It was started in 2005 and the core of the article was completed by the end of 2006. There were next to no books on the subject, so it was written using contemporary news sources. In 2005–2006 they were less than 8 yeas old. Now 10 years on more and more of the sources are only available via the Wayback machine. So far that is not too much of a problem, but in 20 years time, unless those sources are replaced with later ones, then the article that was initially supported with secondary sources will become supported with primary ones without one citation being altered! -- PBS (talk) 14:13, 6 July 2015 (UTC)
        • I disagree that all of this year's media output is secondary. That's just plain wrong, and you are unlikely to find a reliable source that actually claims this. Eyewitness news reporting, including writing an article about your personal experience of editing Wikipedia, is a primary source, full stop. So are routine articles like "There was an edit-a-thon last night". This purported "press coverage" in Wikipedia:Press coverage 2015 is nothing more than a tutorial on how to type Special:Random with an explanation that the author believes you'll have more fun reading Wikipedia if you look at random articles.
          However, I'd like to expand on an interesting point that you make: The classification depends upon how you use it, not just on its inherent qualities. All sources are primary for something. A meta-analysis of original scientific reports is always, inherently, a secondary source, no matter how old/biased/unreliable it is. It takes original material (the original reports) and transforms that original material into a new thing (the results of the meta-analysis) – that is, even according to the most stringent and exclusive definitions, a secondary source. However, there are two important caveats:
          1. That meta-analysis is also a primary source for things like the authors' names, affiliations, the journal, the number of words contained in the abstract, or whatever else isn't part of the "transformed" thought.
          2. When that meta-analysis is many decades old, or if it comes from a disreputable publisher, then you shouldn't use it like a secondary source. You should only use it like a primary source, for "what people thought about (the subject of the meta-analysis) in" previous decades (or not at all, per WP:DUE), not directly for the subject itself. WhatamIdoing (talk) 17:51, 6 July 2015 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────WhatamIdoing I think that you are making a difference between primary and secondary sources that may be a case of style rather than substance. Take for example Brian Hanrahan's report from the Falkland's war that included the famous line "I counted them all out, and I counted them all back." The BBC states that this was done as "a clever ruse to get round reporting restrictions, so that he could say that all British Harrier jets had returned safely." If instead a BBC news editor back in London had edited the piece and simply reported that "All the British Harriers returned safely" (and used stock footage of a Sea Harrier) it would no longer have been a personal observation so presumably it would then be a secondary source. If the BBC had instead had written a contemporary piece and the report that "John Doe [not a BBC staff member] stated last night 'I counted them all out, and I counted them all back'.", then it is no different from a book quoting a primary source 17 years later (page 245 Rolling Thunder: Jet Combat from World War II to the Gulf War by Ivan Rendall (1999)). -- PBS (talk) 18:08, 7 July 2015 (UTC)

No, I'm making a difference based on what reliable scholarly sources say. You might want to read WP:LINKSINACHAIN. Merely quoting a primary source does not make your document become a secondary source. WhatamIdoing (talk) 01:58, 8 July 2015 (UTC)
I am familiar with links in a chain, there is a practical example over date of the knighthood of Sir Oliver Cromwell. However I think that your conclusion is wrong, and also that if a change of presentation alters something from a primary source to a secondary one, then the line between the two is more fluid than you are allowing for. As always the devil is in the detail. -- PBS (talk) 13:25, 12 July 2015 (UTC)
The line is fluid, and that's one of the problems with people going about deleting good content with the anti-WP:PRESERVE excuse that "But it's primary, so it's WP:NOTGOODSOURCE!" Whether a source is primary or secondary (or tertiary) depends on both its inherent qualities and how you use it. And whether it's technically reliable for a given piece of material is almost entirely unrelated to its classification under historiography. "Secondary" is not some fancy spelling for "reliable". WhatamIdoing (talk) 22:19, 12 July 2015 (UTC)
  • Oppose Big oversimplification, it is impossible to find all news stories as primary. The change would turn the policy on its head. AlbinoFerret 00:15, 9 July 2015 (UTC)
  • Oppose. WP is already burdened by a lot miles of red tape, no more of that crap please. The Yeti 14:15, 9 July 2015 (UTC)

Secondary RFC[edit]

Question: should WP:PRIMARY be updated to define all news sources from soon after an event as primary, without exception?

The background is covered above, but Nyttend has criticized the wording of the RFC above, so I feel we should have one on their actual interpretation. As I understand it, they believe that all news sources from soon after an event are automatically considered WP:PRIMARY sources (they haven't specified what they consider the cutoff, but they feel that there is a cutoff clear enough and strong enough to close an RFC against overwhelming consensus by citing WP:CONLIMITED based on their personal assessment of particular news sources as too recent and therefore primary.) While I agree that sources from soon after an event should sometimes be used with caution, I think it's clear that both the wording of the policy and the way it has been applied throughout history does not agree with the interpretation that they are WP:PRIMARY, so we need an RFC to settle this question specifically. --Aquillion (talk) 01:30, 19 June 2015 (UTC)


  • Strong oppose this change as well. While we always have to be careful of WP:RECENTISM, sometimes topical events happen to BLPs which require immediate coverage from reliable sources; sometimes it will require saying something potentially negative under BLP. Those sources should be approached with care, definitely, but Nyttend's proposal to define all contemporary news sources as WP:PRIMARY is a hamhanded approach would make it impossible to write complete articles about many recent subjects, and I feel it's important to shoot it down decisively. (Consider: Nyttend used this argument to say that we cannot use the term 'sexual abuse', but in fact simply accusing someone of abuse based purely on primary sources would be unacceptable -- and according to their logic, no primary sources on the recent abuse scandal exist; therefore we would not be able to cover it at all until, I suppose, someone writes a book on it.) --Aquillion (talk) 01:30, 19 June 2015 (UTC)
You might be better off starting another RFC at some point. I'm not aware that a sub-RFC like this is workable.Anythingyouwant (talk) 01:32, 19 June 2015 (UTC)
It's not unusual to add an additional question to an RFC if someone objects to the wording of the first one, suggests an alternative resolution, or the like. --Aquillion (talk) 01:39, 19 June 2015 (UTC)
I object in this particular instance because it makes things much more complicated, and I don't think I would be entitled either to add another RFC question at this point. I don't object to starting another RFC once this one is closed.Anythingyouwant (talk) 01:46, 19 June 2015 (UTC)
WP:DROPTHESTICK - This has gone long enough already. No more RFCs on this aspect as we have very clear and long standing content policies. We have an RFC that was poorly closed and that will likely either re-opened or overturned. - Cwobeel (talk) 03:09, 19 June 2015 (UTC)
You guys drop the stick. Just let the admins decide on the first RFC. Why have a second, and then essentially a third, at a policy page to advance your position in a content dispute? Oy vey.Anythingyouwant (talk) 03:12, 19 June 2015 (UTC)
  • Strong oppose - Nyttend needs to acknowledge his mistake, and move on to avoid further drama. Let's go back to editing please. - Cwobeel (talk) 03:06, 19 June 2015 (UTC)
  • Oppose and speedy close Unless someone comes along and wants to open an RFC to propose to change wikipedia's policy on this (and I would oppose that change as well) then we should not have this RFC just because of one article RFC closure. This would be unworkable for wikipedia without huge changes in what we cover, with practically any current events impossible to cover. Davewild (talk) 06:52, 19 June 2015 (UTC)
  • Oppose and speedy close per Davewild. This only came up because a single RFC close misinterpreted existing policy. There's no need to reconsider the actual policy (for which there is a broad and durable consensus), and no need to let this tempest-in-a-teacup drag on. Fyddlestix (talk) 15:51, 19 June 2015 (UTC)
  • Comment - The basic problem seems to be that historians have a different definition of primary and secondary sources than Wikipedia does, because Wikipedia does have articles both on news and on history. If Wikipedia were to adopt the historical definition, we would be impoverishing our ability to cover news. We need to use common sense as to when newspaper journalism is a valid secondary source (for news) and when historiography is required. Robert McClenon (talk) 17:41, 19 June 2015 (UTC)
  • Oppose and speedy close - agree with above. Additionally, this change would lead to meandering and meaningless (and likely endless) discussions on what "soon after" means. Neutralitytalk 02:54, 20 June 2015 (UTC)
  • Suggestion: We should consider linking to WP:PRIMARYNEWS (or maybe just mentioning it in the /FAQ on this page, rather than putting it in the policy). The short answer is that most newspaper-type news articles are primary sources, but it's more complicated than that. Linking people out to a {{supplement}} is going to give us more room to explain (that section's running around 900 words at the moment) and therefore increase the odds that editors will find an accurate answer. WhatamIdoing (talk) 08:22, 20 June 2015 (UTC)
  • Strong Oppose as blanket declarations shouldn't be applied here. Grognard Extraordinaire Chess (talk) Ping when replying 00:12, 25 June 2015 (UTC)
  • Oppose. No, because not all news sources from soon after an event are primary sources. Most aren't until quite a bit of time has passed. (See my post above). Sarah (talk) 02:48, 25 June 2015 (UTC)
  • Oppose I think Robert McClenon's suggestion of adopting the historian's definition is unworkable (like changing which side of the road we drive), however, 'We need to use common sense as to when newspaper journalism is a valid secondary source (for news) and when historiography is required' is spot on. Is that thought, or some improved variant in guidelines? Pincrete (talk) 20:10, 5 July 2015 (UTC)
  • Comment - I didn't suggest adopting the historian's definition. I opposed it because it would impoverish our ability to cover news, and we do cover news. I was trying to point out that Wikipedia and historians have different definitions of primary and secondary, and we have to remember the distinction. Robert McClenon (talk) 19:43, 7 July 2015 (UTC)

Why does it matter?[edit]

Once again, PSTS classifications seem to be the tail wagging the dog here. We seem to be missing the forest for the trees... so let me cut through the bullshit and ask the important question: Does it really matter (in terms of "original research") whether a news source is classified as PRIMARY or SECONDARY? I don't think it does.
Original research does not stem from whether a source is primary or secondary... original research stems from what we do with the source. Are we merely passing on what the source says in our own words? That is not original research. Are we engaging in analysis or interpretation of the source, or drawing our own conclusions based upon it? That is original research. NOR really is that simple. All the navel gazing about defining what is and is not a Primary or Secondary source ... all the argument about whether a source is Primary or Secondary is unnecessary. In both cases... it's not OR if we stick to the source ... it is OR if we go beyond the source. Blueboar (talk) 21:14, 25 June 2015 (UTC)

"Original research" is not part of the normal English language, it is special term defined in the original research policy. This can be proven by writing an article that isn't terribly insightful but defies the original research policy, adding it to Wikipedia, and observe it get speedily deleted as "original research". Then submit it to a scholarly journal and wait about 15 nanoseconds to receive a rejection notice because the article is too derivative.
Having established that "original research" is a Wikipedia term of art, we then see that the original research policy says "Do not base an entire article on primary sources." So either newspapers are potentially secondary sources, or all articles on current events must be deleted. Jc3s5h (talk) 21:37, 25 June 2015 (UTC)
Some pieces that run in newspapers certainly are secondary (just not the majority, or even close to it, judging from the paper I'm looking at).
I think that Blueboar is on his way to a proposal that PSTS be put into a separate page. WhatamIdoing (talk) 23:10, 25 June 2015 (UTC)
Been there... done that. Sure, I continue to have serious concerns about how PSTS continues to be misunderstood and how it has lost its original purpose (it's original purpose was to explain that Wikipedia is a Tertiary source, and should not be turned into a Primary source - which is what happens when you add OR), but I fully understand that there is no consensus to remove it from the policy. That said... I still don't get why people get so worked up over defining sources. The policy explicitly states that we are allowed to use Primary sources. So it does not really matter if we classify news source as Primary or Secondary... they can be used as long as we use them appropriately. Blueboar (talk) 00:25, 26 June 2015 (UTC)
I couldn't agree more. The core problem seems to be that an unwillingness on the part of some of the community to accept the importance of editorial judgement, implicit in determining whether sources have been used "appropriately". Peter coxhead (talk) 07:26, 26 June 2015 (UTC)
I think that part of the problem is people not believing that we can use primary sources. I looked up the dispute that spilled over to the above sections, and the closing statement said that the information couldn't be included because the sources were primary and PSTS banned the use of primary sources. Now that the community has sort of figured out that WP:Secondary does not mean independent (which took about three years of steady work), maybe we should try to work on "you can WP:USEPRIMARY sources" as the theme. WhatamIdoing (talk) 00:11, 30 June 2015 (UTC)
Primary sources are often the best sources to use. You just have to be careful because it's easy to get things wrong. This policy has never said or implied that they can't be used at all, so I've always wondered where that idea came from. Sarah (talk) 01:59, 30 June 2015 (UTC)
I think it may be a cumulative effect... there is a heavy emphasis on using secondary sources throughout PolicySpace (multiple policies and guidelines stress the need for secondary sources) and this adds up... eventually editors start to think they should never use primary sources. Blueboar (talk) 13:59, 30 June 2015 (UTC)
The problem — or one of the problems — is the failure to distinguish between original research and primary sources, which may not involve any research at all. — Robert Greer (talk) 00:48, 1 July 2015 (UTC)
WP:BLPPRIMARY certainly discourages use of primary sources in BLP's, even if it falls just short of outright banning them. VQuakr (talk) 00:56, 1 July 2015 (UTC)
This thread is a revelation to me, I've been told so often that primary sources are verboten! The trouble with that essential quality 'judgement' (as to when to use them), is that it is inherently undefinable. Which is a problem for a project dependent on guidelines.(my 2d.) Pincrete (talk) 20:26, 5 July 2015 (UTC)
It's only a problem if instead of "guidelines" you want "rules", and you don't accept that the really important feature of Wikipedia is consensus among editors based on the reasons behind the guidelines. There are some rightly rigid rules, such as no original research. For historical topics, it's almost always the case that using what historians call primary sources involves OR. In other topic areas (e.g. recent scientific discoveries), using journal articles – which some editors call "primary sources", unhelpfully in my view – usually doesn't involve OR, but does raise notability questions (an accurate report in Wikipedia based on a single scientific journal article isn't original research, but will often not be notable). What's important is that editors are free to discuss their use of sources openly and to reach informed decisions, and not have the discussion closed down by slogans masquerading as Wikipedia policies. Peter coxhead (talk) 10:47, 6 July 2015 (UTC)
As the policy says, Primary sources can be used (without slipping into OR) ... you simply have to understand when and how it is appropriate to use them, and when it is not. Blueboar (talk) 11:09, 6 July 2015 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Sarah wrote above, "Primary sources are often the best sources to use." This is undeniably true. What do people think about adding those words to the policy? We might say "sometimes" rather than "often", but the effect might be positive. (We could give direct quotations as an example: it's usually better to cite the original than to play the telephone game with quotations.) WhatamIdoing (talk) 18:04, 6 July 2015 (UTC)

  • I think a lot of this goes back to notability and those endless debates at AfD. A secondary source carries more weight as evidence of notability than a primary source does, but once the topic has been established as a notable one, the primary source is often more valuable. For example, if someone discovers an important exoplanet and your sources are (1) a scientific paper and (2) the BBC's announcement of the discovery based on the scientific paper, then any conflict between those sources should be resolved in favour of the astronomer rather than the journalist who semi-understood his paper.—S Marshall T/C 11:13, 7 July 2015 (UTC)
    • I think you're right about a lot of the 'of course it's secondary!' stuff being related to notability. WhatamIdoing (talk) 03:53, 8 July 2015 (UTC)
  • WhatamIdoing, I'd support adding something along those lines. Sarah (talk) 18:51, 7 July 2015 (UTC)
In the case of articles on history topics, primary sources are often not the best sources to use. For a example contemporary sources about war are often very biased and as well known "The first casualty of War is Truth". Or take the case of the breaking of codes any strategic history of any theatre that the Western Allies fought is severely flaw if ULTRA and MAGIC is not factored in. As one historian has commented "When [Monty] put Rommel's picture up in his caravan he wanted to be seen to be almost reading his opponent's mind. In fact he was reading his mail" (Andrew Roberts 2009). One of the structural problems about this particular policy (as it is about some guidelines) is that it is written as one shoe fits all. This causes problems for different project areas, because advise that is useful in one is just confusing in another. -- PBS (talk) 19:22, 7 July 2015 (UTC)
PBS, I wonder if you could live with a statement like, "Primary sources are sometimes the best sources to use, such as for supporting direct quotations or describing recent events." I could live with a "but not for proper history" clause, if that might help. Or substituting "most authoritative", rather than "best". That would address S Marshall's issue about the journalist vs the scientist. WhatamIdoing (talk) 03:53, 8 July 2015 (UTC)
Again... it all depends on what you are trying to support with the primary source. A WWI era poster depicting a German soldier with a baby impaled on a bayonet would be perfectly appropriate for supporting a statement that "During WWI, some propaganda posters depicted German soldiers as savage murderers"... I would say it is excellent source for that statement... precicely because it is a primary source. It's not a direct quote... but a direct description. Blueboar (talk) 01:30, 9 July 2015 (UTC)
I agree. Once the extent of such propaganda has been established by reliable secondary sources, primary sources are the best way of providing examples. Peter coxhead (talk) 07:22, 9 July 2015 (UTC)
Blueboar doesn't your example involved an explicit breach of WP:PSTS? Surly, as Peter coxhead points out, one would need a secondary source that explicitly make the statement that "[some Allied WWI] propaganda posters depicted German soldiers as savage murderers"? So while such poster may be appropriate as a eye candy the statement would need a secondary source to support it. If a contemporary primary source was to be used it would need to be in the form of a statement eg "[we are commissioning posters] depicting German soldiers as savage murderers to ...". -- PBS (talk) 10:33, 11 July 2015 (UTC)
Nope... the policy states: "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge." In my example, there are no evaluative or analytical claims being made, no conclusion being drawn... it is simply a descriptive statement of fact about what the posters depict... a statement that anyone viewing the poster can verify. No original research involved. It's no different than citing a basic plot summary of a book or movie to the book or movie itself. Blueboar (talk) 12:42, 11 July 2015 (UTC)
In the example, without the support of a secondary source, I think you have analysed the content, "depicted German soldiers as savage murderers" how do you know that the depiction is of savage murder? To know if the killing were mean to represent savagery implies a knowledge of what the people of the time considered savage. For example Cromwell in his description of the massacre at Droghada describes his ordering the torching of a church tower in which some royalists had retreated, to our sensibilities such actions are savage (but Cromwell and the members of Parliament in London (to whom it was addressed and read), did not), but he notes that one of the victims said "God damn me, God confound me: I burn, I burn." to Cromwell that was shocking, because it was taking the Lords name in vain. To know if the killings in the poster represent murder involves a knowledge of the laws of war, which is an historical legal matter (and also a knowledge of what a contemporary audience would think they represented), and so I do not think it is a straightforward description of the content. -- PBS (talk) 13:12, 12 July 2015 (UTC)
The main value of the secondary source here is in determining whether mentioning this at all is WP:DUE. One could equally well pick out a propaganda poster that used the color orange, and write that some posters used a lot of orange. But why would anyone care, unless the use of that color were in some way significant? WhatamIdoing (talk) 22:25, 12 July 2015 (UTC)
Some of you might be amused to learn that in a discussion on another page, an editor with 35,000 edits has just said to me: Since it is a secondary source, it's clearly reliable.S Marshall T/C 21:54, 13 July 2015 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. No further edits should be made to this discussion.

Seeking out answers by contacting primary sources[edit]

Hi there, I think I'd like to see some clarification in the policy about whether or not users can take to Twitter or other social media streams and contact primary sources to get answers to questions for which there might be no reliable sources. For example, content was removed in this edit that used this Formspring query as a reference. This strikes me as original research because presumably someone has gone out and chased down a lead and manufactured an interview, albeit a brief one. Similarly in this article I find references like "I didn't record Sin this time, but the other guy yes. I think.", "If it's the old one, that's me as Sin. Just not the new one.", "Sorry, I thought I answered yes to that. Yes, I played Bedman. But not Sin.", etc. Some of these queries came directly from a Wikipedia editor. I suppose that some of this might be covered at WP:SELFPUB, but since this is the article on original research, it seems reasonable to include something here that clarifies whether this is an exception or not. Thanks, Cyphoidbomb (talk) 21:37, 10 July 2015 (UTC)

I think its laudable if editors try to expand the verifiable knowledge available to the project. They should probably make a edit request on the talk page rather than personally adding any social media content they themselves prompted. Rhoark (talk) 00:55, 11 July 2015 (UTC)
I have always understood that an OR inquiry (such as you talk about) can be used in a talk page discussion to gain consensus to omit (ie remove) inaccurate information from an article... but should not be used to add information to an article or "correct the record". Blueboar (talk) 01:11, 11 July 2015 (UTC)

"an analytic, evaluative, interpretive, or synthetic claim"[edit]

Scientific papers are primary sources and they usually do contain analytic, evaluative, interpretive or synthetic claims. For example, our article on HIP 56948 (a solar twin) is based largely on papers by Jorge Melendez and others published in 2007 and 2012; there are journalistic sources but all they do is parrot the primary source in dumbed-down language. Professor Melendez and his colleagues gathered the data, evaluated the results and interpreted them for us. And that's as it should be when your source is a full professor at a major university who's the world's leading authority on a narrowly-focused subject.

I can't square this with the bit of WP:NOR that says "Articles may include an analytic, evaluative, interpretive, or synthetic claim only if the claim has been published in a cited reliable secondary source." Actually I think that articles can and should repeat the primary source's evaluative claims ---- provided the primary source is by a scholar or academic who's acknowledged as a leader in their field.—S Marshall T/C 00:13, 11 July 2015 (UTC)

FYI... this has been discussed before. The problem is that the OR policy was mostly written by editors familiar with the humanities... and in the humanities a paper such as you mention would be considered secondary ... the paper is analyzing and interpreting original data... and that date would be the primary material. the data is the equivalent of original documents used in a history paper... primary material which is then analyzed and interpreted in the secondary academic paper.
That said... remember that we are explicitly allowed to describe a primary source as long as we don't interpret or analyze it ourselves, or draw our own conclusions from it. So it does not really matter whether we call Professor Melendez's paper primary or secondary. As long as we don't analyze it ourselves, it's not OR to state what a source says. This is why I keep raising on the concept of "appropriateness"... it does not matter if a source is primary or secondary if we use it appropriately. Blueboar (talk) 01:22, 11 July 2015 (UTC)
  • Discussed before but not fixed, I see. How about "An analytic, evaluative, interpretive, or synthetic claim not found in the primary source itself"?—S Marshall T/C 13:44, 11 July 2015 (UTC)
  • See our article District of Columbia v. Heller. The article cites the US Supreme Court decision in the case, as well as secondary sources. The decision itself has many characteristics of a secondary source; the Supreme Court never conducts experiments or visits the scene of incidents, nor does it take testimony. It reads the decisions of lower courts, and considers written briefs and oral testimony from the parties' lawyers about how to interpret the law and lower court decisions. Then it decides. It is the fact that the Supreme Court makes a binding decision that makes it a primary source; otherwise it would be a secondary source. And of course the decision is chock full of evaluations. I like S Marshall's formulation at 13:44 UT. Jc3s5h (talk) 14:03, 11 July 2015 (UTC)
I, too, think that you make a valid point and approve of your suggestion, S Marshall. –Prototime (talk · contribs) 21:26, 11 July 2015 (UTC)

Why it's become so confused for Wikipedians: It's about the content not type or name of publication[edit]

Wikipedia's own definition of primary and secondary (and tertiary) do not change with the topic. It's simply not possible, but I keep seeing people repeat it as if it were true. WP has a standard for how to determine source reliability level. It is WP:RS. Various academic, journalistic, etc., fields, and individual organizations, have their own standards. Their standards do not overturn each other. They don't overturn ours when it comes to how we evaluate sources. Ours doesn't overturn theirs in their home court, either. There is no primary vs. secondary difference between a humanities journal article or a physics journal article for Wikipedia. If you ask a humanties academic, they'll classify them both the same as each other for their purposes, and a hard-sciences professor will do likewise, and when you compare the two academics' assessments of both works you find that the two individuals categorized them differently. If you asked a panel of experts in their two disciplines to repeat this process, you'd get the same result. BTW, see the RFC a bit above where the pretty substantial !voter turnout unanimously rejected the idea than external classification preferences from a particular discipline (all news sourcing as primary, in history and related fields) can affect how WP chooses to classify sources for our own internal purposes. It can't happen because our criteria are different. We're mostly asking "where did they get this claim from?", and also making sure that the publisher is reputable enough to bother with. In science, the prestige of the journal and its editorial selectivity is paramount; they want to see the most amazing and best-vetted primary sources, and don't care so much about secondary literature reviews; the groundbreaking papers are the "sexy" party. WP wants the reviews, and we don't trust "groundbreakingness" until secondary sources tell us we can.

Next, a source, as a object, is not immutably and always either a primary, or secondary (or tertiary) source, for all time, in all contexts, for all facts. Just the opposite. RS is poorly written with regard to this and it has led to years of confusion. We need to fix it. When we say "this is a secondary source" [or whatever - don't fixate on "secondary" here] this means "this is a secondary source, in the present, for the following claims, in this particular context". It cannot sanely mean anything else. Everyone in academia (and several other forms of publishing) understand this. A reputably-published newspaper article (not op-ed – we can't verify independence!) is a secondary source for the material it has WP:AEISed from various primary and other secondary sources and run through a quality editorial process. If the journalist closes with a personal conclusion of her own, it is a primary source for that view, which must be individually attributed to or directly quoted from that writer. Where it presents "factoids" that were not subject to any AEIS, but reguritated (e.g. the fact that scientific name of the domestic cat is either Felis catus or Felis silvestris catus depending upon the source consulted, something she got from an encyclopedia, her article is a tertiary source for that information. News journalism is a tertiary source for virtually all "hard science" material (including news); it is summarizing and glossing over many nuances in the actual secondary (and maybe some primary) peer-reviewed material, if it even consults any, rather than abstracting from other tertiary science sources (there are many other tertiary factors at work, too, but it would be lengthy to cover them right here).

A serial publication is not uniformly secondary, either. Movie reviews, columns (at least typical ones), op-eds, and much of the content of editorials, along with any other opinional material, in news sources is primary. Subjective opinion does not magically transmogrify from opinion into fact just because of who churned it out. But it's staggering how many Wikipedians really do believe that every single word published by the New York Times or whatever must be secondary sourcing, because of how terribly this guideline is written.

Same goes for journals. There's this counterfactual belief that secondary material in a paper is somehow transubstantiated into primary material, if elsewhere in the work it's also presenting primary material (the point of all that secondary research they did). Peer reviewed AEIS is peer-reviewed AEIS; if it's from a reputable journal (and not outdated), it's high-quality secondary sourcing, as long as the secondary and primary material are distinct and not commingled. Bracketing it with a proposition in the intro and a conclusion at the end that give the publisher's position on all that researched material does not "degrade" the entire work into "primaryness". Any new data, results, hypothesis, model, description of methodology, etc., that is not AEIS of other high-quality sources, but has come from the author(s) of the paper (is their "original research") is necessarily primary. (Exactly the same thing is true of a journalist injecting their own views into an investigative piece.) Often, the same generally high quality source is mixture of all three; any topic-specific encyclopedic compendium by subject matters experts, for example, will mostly be tertiary, but contain plenty of the other two categories. Obviously a publisher is not a source, and isn't "reliable" (they do a good job printing?) but reputable [or not]. And age of the source reduced what classification level a source can be held at; all sources eventually become treated as if primary, no matter how secondary or tertiary they were at one time. You can't cite a 1941 article on particle physics or the causes of a riot as a secondary source, but attribute and use with caution as primary.

This isn't as simplistic as it's been made out to be by WP:RS's poorly written bits at WP:PSTS, with consequent related inclarities in WP:V and WP:NOR. But it's easy to figure out if you just it off and start fresh. For the claim in question, no matter where it is reputably published and in what form: Did the writer come up with idea/data (or can we not be sure)? Primary. Did they work at producing a complete picture from primary and/or previously published secondary material, without a serious bias? Secondary. Are they just summarizing, collecting, or repeating previous material without AEIS? Tertiary. Is it clearly opinional, hypothesizing, reaching? Primary. Was it produced close in time to the events it is about? Primary. Is the source categorically obsolete, or factually or culturally surpassed? Primary. For the source classification analysis, it doesn't matter one bit who the authors were, who the publisher was, or what the topic is (unless these interact to produce a conflict of interest (primary); those are otherwise factors for the quality level analysis (low to high): the reputations of the publisher and writer, and their competence/expertise in the topic.

I've gone into some other factors at WT:V and WP:RS (with some overlap).  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  01:42, 23 July 2015 (UTC)

In a cited reliable secondary source[edit]

There have been a flurry of edits to a sentence in this policy that I now think have changed the meaning in a way which I do not think has consensus:

The edit:

  1. Revision as of 19:14, 31 May 2015 SMcCandlish
  2. Revision as of 19:16, 31 May 2015 SMcCandlish
  3. Revision as of 19:20, 31 May 2015 SMcCandlish
  4. Revision as of 20:31, 31 May 2015 Blueboar
  5. Revision as of 14:12, 1 June 2015 Margin1522
  6. Revision as of 14:14, 1 June 2015 Margin1522
  7. Revision as of 14:19, 1 June 2015 Margin1522
  8. Revision as of 07:55, 3 June 2015 SMcCandlish

have changed:

  • "Articles may make an analytic or evaluative claim only if that has been published by a reliable secondary source."


  • "Articles may include an analytic, evaluative, interpretive, or synthetic claim only if the claim has been published in a cited reliable secondary source."

I am going to revert to the sentence as it was after the second edit listed above "Revision as of 19:16, 31 May 2015" by SMcCandlish. Personally I think the changes up to number two comes down to angels dancing on pin heads, so I have no objections to that change. But I do have an objection to what was the end result of the further changes.

The reason for this is that I suspect that many, as I do, now will understand it to mean that not only must one be able to prove that the "analytic, evaluative, interpretive, or synthetic claim" have appeared in a reliable secondary source that secondary source has been cited in another source. If that is not what is meant, and I suspect that it is not, then more clarity is needed.

Also with the revision as of 19:20, 31 May 2015 SMcCandlish wrote in the history "Normalizing these two near-identical statements of the policy to require actual citation to the source, per WP:V/WP:RS (see talk). There is no reason to reiterate here that a secondary source is "published"; that's part of its definition."

[I presume "see talk" is referring to #Closing lame loopholes].

I think that SMcCandlish is forgetting about primary sources when reviewing (and altering) this sentence. My reading of the first one is that "analytic or evaluative" or if one wants "analytic, evaluative, interpretive, or synthetic claim" is that these may be made in a primary source, but that primary source must have been published in a reliable secondary source. For example Wellington's claim that the Battle of Waterloo was "the nearest-run thing you ever saw in your life" is an evaluative statement, and it has been published in numerous secondary sources. Had it not, then it could not be published here in Wikipedia for the first time. However I do not take it to mean only those parts published in a secondary source of a primary source can be used, because that opens a NPOV issue. For example there is a real problem in the raid on Dresden with authors/historians selectively quoting parts of a briefing by Bomber Command to make a point about whether the bombing was a first step in the Cold War. To get around the POV it is necessary if the briefing is to be quoted, to quote more of the primary source than the partisans of the argument usually publish (to support their academic POV).

-- PBS (talk) 13:13, 11 July 2015 (UTC)

re: "...but that primary source must have been published in a reliable secondary source." ... I disagree... that would mean we could never cite a published autobiography. An autobiography is a primary source... but it is not OR to descriptively mention what it says. To give a hypothetical... let us suppose that General X commanded the winning forces in the Battle of Cumquat. Later, he publishes an autobiography in which he says: "I think the key to winning the Battle of Cumquat was the use of cell phones to coordinate the artillery"... this is an analytical,evaluative claim made by a primary source... but it is a claim that can be mentioned in the article on the Battle of Cumquat (and cited to the autobiography)... we simply have to do so with proper attribution so the reader knows that we are relaying the opinion of General X, and not necessarily stating it as accepted fact. Blueboar (talk) 11:44, 12 July 2015 (UTC)
A fair point about the publishing and perhaps further discussion is in order. This policy unlike some of the universities mentioned in footnote 3 does not list autobiographies as primary sources. This was discussed several times using Field Marshals Slim's book "Defeat Into Victory", and hence autobiographies are not mentioned in primary sources and it explains the sentence "A book by a military historian about the Second World War might be a secondary source about the war, but where it includes details of the author's own war experiences, it would be a primary source about those experiences." So I would argue that autobiographies are both primary and secondary sources, and it depends on what is being extracted from them. Also one has cases such as Wellington's despatches, and the general opinion seems to be such publications are secondary sources (as unlike an archive, there is editorial control of what goes into them, and they were published for a wider ordinance than an archive), but the individual dispatches ought to be treated as primary sources. I would suggest that the same is true for scientific papers published in Nature. The initial scientific paper on a new discovery may be a primary source, but it is a primary source published in a reliable secondary source (Nature). -- PBS (talk) 13:53, 12 July 2015 (UTC)
You hit the nail on the head when you say "it depends on what is being extracted from them"... Original research is all about what we (Wikipedia editors) extract from a source... and less about the primary/secondary/tertiary nature of the source itself. Use a published primary source appropriately and there will be no Original Research involved... use the same primary source inappropriately and there probably will be. Of course the same can be said of secondary and tertiary sources. It's simply easier to misuse primary sources (either intentionally or unintentionally) than it is to misuse secondary or tertiary sources... hence our caution (which is not a "ban") about primary sources.
It is not OR to descriptively report on the contents of a published primary source (and that includes reporting on including any analysis or interpretations that may be contained within the primary source)... as long as we do not interpret or analyze them ourselves. To put it another way... If a published compilation of dispatches contains an analysis or an evaluation of events, we can mention the fact that the dispatch contains such (it is a verifiable fact that it does), and we can describe it (blunt description is not OR). What we can not do is insert our own analysis, interpretation or conclusions based on the dispatches. We can not go beyond blunt description of what is directly contained in the dispatch. That would be OR. Blueboar (talk) 15:14, 12 July 2015 (UTC)
To give an example about how much use matters, here are two different ways to include Wellington's statement:
  • YesY Shortly after the battle, Wellington described it as "the nearest-run thing you ever saw in your life" and believed that the British might have lost if he had not been in charge (source: the original primary source (which appears to be the 1903 Creevey Papers).
  • N The battle of Waterloo could have ended with either side winning. The British victory depended upon Wellington's personal presence (source: original primary source).
As Blueboar keeps saying, the issue is whether you're using the source appropriately. It is almost never appropriate to use a primary source to write in Wikipedia's voice. WhatamIdoing (talk) 22:41, 12 July 2015 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────Well we all seem to be in agreement about sources. Lets move on to the sentence that has been subject to recent alterations, and see (1)if there is a consensus about what it is trying to say, (2) if it currently says it (given that it is read within the context of the whole section), and (3) if it needs to be changed what is the most appropriate wording to meet the consensus about what it is trying to say. -- PBS (talk) 08:24, 13 July 2015 (UTC)

I would say both versions under discussion are flawed. The flaw is in the word "make". Articles don't make claims ... they include claims. What makes a claim OR (or not) is whether we took it from a published source or came up with it ourselves. If the claim has never been published in a source... if it originates with Wikipedia... then it is OR and should not include it in one of our articles. If the claim is taken from a source... if it originates outside of wikipedia... it is not OR to include it in one of our articles. the Primary vs. Secondary nature of the source is actually irrelevant... what matters is whether the claim itself is (directly) stated in the source. Blueboar (talk) 11:14, 13 July 2015 (UTC)
I would correct Blueboar's statement of 11:14, 13 July 2015 (UTC). It is original research if the claim was not taken from a reliable published source AND it would be non-trivial to find the claim in a reliable published source. This formulation allows for our policy of not requiring a citation for Paris being the capital of France.
I view the case of an editor trying to include a claim from an unreliable published source as being original research on the part of the unreliable publisher; Wikikpedia editors aren't the only ones whose OR we reject; we also reject OR from unreliable publications that reaches us through Wikipedia editors. Jc3s5h (talk) 11:42, 13 July 2015 (UTC)
Oh, there are lots of reasons why we might not include something stated in a source... the author of the source may not be qualified to make the claim (ie not reliable)... mentioning the claim may give UNDUE WEIGHT to a minority viewpoint.... it may be too trivial to mention... etc. etc. WP:NOR is not the be-all-and-end-all of inclusion. My point is simply that it is not ORIGINAL RESEARCH for Wikipedia to repeat a claim that a published source external to Wikipedia has made.
This is an important... our sources are allowed to conduct original research. In fact, that is what we rely on them to do for us. What WP:NOR bans is conducting original research ourselves. WP:NOR is NOT about the original research conducted by our sources... it's about the original research conducted by our editors (whether intentionally or unintentionally). Blueboar (talk) 12:22, 13 July 2015 (UTC)
As for the issue of not requiring a citation for a statement like "Paris is the capital of France"... that has nothing to do with Original Research... the reason we don't require a citation for such statements isn't that it would be "non-trivial to find the claim in a reliable published source"... the reason is that there are so MANY reliable published sources that actually do say it that we consider it unnecessary to require a citation to any of them. It's easily verifiable without a citation. Blueboar (talk) 19:34, 13 July 2015 (UTC)
  • User:WhatamIdoing says: It is almost never appropriate to use a primary source to write in Wikipedia's voice. I'm a bit disappointed to learn that I've been doing so many inappropriate things during my time on Wikipedia... from the lithium abundance on HIP 56948, through the timelines on List of European dinosaurs, to the service record of Lisa Jade Head, I've repeated countless claims from primary sources in Wikipedia's voice. I'm rather afraid that I intend to carry on being inappropriate in this way.—S Marshall T/C 21:42, 13 July 2015 (UTC)
    • The bit on the star uses WP:INTEXT attribution, so it's fine (i.e., it asserts that it's the product of a single study, rather than asserting it as an unassailable fact). The dinosaur timeline is completely unsourced. I hope that you do not intend to carry on adding large quantities of unsourced information.  ;-) WhatamIdoing (talk) 22:36, 13 July 2015 (UTC)
      • I'm duly chastened.  ;-)—S Marshall T/C 11:47, 14 July 2015 (UTC)

Primary source – example of scientific paper[edit]

In the section Primary, secondary and tertiary sources, an example of a primary source is given as, "a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment". Is this referring to just the raw data from the experiment and its data reduction, or to the whole paper including the author's interpretation and conclusions regarding the data? --Bob K31416 (talk) 06:42, 20 July 2015 (UTC)

It appears that the example doesn't refer to the conclusions in the scientific paper because that would be of no purpose for this policy NOR, which prohibits using conclusions by editors, rather than conclusions by reliable sources. Perhaps this should be clarified in the example? --Bob K31416 (talk) 21:22, 23 July 2015 (UTC)

It needs to be clarified in more than just that example. Repeating a conclusion that is explicitly reached in a source (be that source primary, secondary or tertiary) is NEVER original research... because the conclusion did not originate with Wikipedia... it comes from a source.
Original research occurs when we take what a source says, and reach our own conclusions from that... conclusions NOT explicitly stated in the source. And, while I support our cautionary statement about using primary sources with care (because it is so common to see editors creating their own OR conclusions based on primary sources)... we should also note that it is just as wrong to create your own OR conclusion based on secondary and tertiary sources. The flaw is creating a conclusion, the type of source used to do this is actually irrelevant. Blueboar (talk) 23:32, 23 July 2015 (UTC)
The problem is that there is an unclear example of a primary source which needs clarifying. Here's an improvement which simply changes "outcome" to "data".
"a scientific paper documenting a new experiment conducted by the author is a primary source on the data of that experiment".
Blueboar, Would this change be acceptable to you?
--Bob K31416 (talk) 02:51, 24 July 2015 (UTC)
  • It includes the authors' interpretations and conclusions, of course. Some material in journal papers is secondary, sometimes, when it's presentation and analysis of previously published material. A typical paper might be structured like this:
Analysis of why it's primary
    1. Abstract
    2. Introduction - what the research goals are, methodology is, etc.
    3. Background - what the prior research is (may be several pages, a tightly-focused literature review of the state of the field on the questions central to the paper)
    4. All the details of the work (experiments, whatever)
    5. Analysis of the results
    6. Conclusion
All of that is primary, except #3. The question to ask is "Is this particular fact coming straight from the minds of the author(s) based on their own work, or is it from synthesis of previously published work, before it reaches the peer review committee?" The authors summarized their own work. They described their own goals and methodology, then all the details of their work. They did the analysis of their own work, and came to their own conclusions. That's all primary by definition; it's single-party material. Their review of the literature in #3 is secondary, being analysis/evaluation/interpretation/synthesis of prior sources, like a regular literature review in the same journal (though the quality level of the latter might be a little better), or an journalism piece (not op-ed!) in a major newspaper, or a history book from a reputable publisher, etc. In all those cases, the underlying facts are coming from someone external to the writer, and the writer is making contextual sense of them in relation to other facts from other external sources.

Another way of looking at it is: Each brain (or set of them - a research team counts as one) that a claim of fact passes through is a sanity check, and a secondary source has had at least two of them, that's why it's secondary. Because WP:RS requires reputably published sources, i.e. those that have been through an editorial review process (another set of brains) and isn't just self-published, a reliable secondary source requires at least three [sets of] brains saying it's not nonsense. A reliable primary source, which we can use "with caution" for something things, has had two (the authors and the peer review committee, or the op-ed writer and the newspaper's editorial board). An unreliable primary source, like a celebrity's Twitter post, has come straight from the original brain, and we usually can't use it except for WP:ABOUTSELF.

PS: When a journal or some other source is the one producing an abstract of something, not its authors, the abstract is a tertiary source; book reviews in science journals, and in some other publications, tend to be of this character, being summarizing rather than judging. Those in newspapers and such tend to be opinional, so are primary sources (the value judgements are coming straight from the mind of the reviewer).

Oppose, therefore, the proposed change. 'a scientific paper documenting a new experiment conducted by the author is a primary source on the data of that experiment' is terribly misleading and suggests, completely incorrectly, that the conclusions a paper's author draws about their own research is a secondary source. There is no difference at all, for WP:CORE purposes, between the data from the author's experiment, the description of the methodology used to gather it, the conclusions the drew from it, or even their abstract of the whole paper. It's all one author saying "what I tell you in this paper is true", and one level of sanity checking, the peer review committee, saying "this doesn't seem to be nonsense". Many short journal papers that do not include any secondary WP:AEIS material are primary from their first word to their last (other that the references list, which is tertiary).  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  06:43, 24 July 2015 (UTC)

SMcCandlish, Your comments seem to neglect the purposes of this policy No Original Research, which is to prohibit editors from using their own conclusions or information. It doesn't further the purposes of this policy to define the conclusions in the example scientific paper as a primary source. It furthers the purposes of this policy to define the data of the paper as a primary source because editors need to exercise caution that they don't make their own conclusions from the data. --Bob K31416 (talk) 08:01, 24 July 2015 (UTC)
@Bob K31416: It doesn't "un-further" the purposes either. It's just as much a case of OR (of a different kind) to take two unconnected studies that, in independent conclusions about different kind of research, suggest "cat hair causes cancer", and as a Wikipedian conclude that this has been proven, and repeat it as fact in WP's voice citing those two studies. This kind of synthesis is rampant in WP articles. It would be legitimate to just state this claim and cite a source, if it's a quality secondary one, like a literature review. But if it's just cited to the two original studies, not representing research that has been reproduced and accepted, and repeated in various secondary sources, it needs to be directly attributed: "According to a 2014 study ... In 2015, a new study ...".

Do we need a specific example of only taking raw data from a paper and coming to one's own conclusion about it? If so, how do we construct that example to not imply that the other primary parts of the paper aren't primary? They still are. The conclusions of the paper are a higher quality primary source that the raw data, but it doesn't magically transmutate into a secondary source. A secondary source is one that engages in analysis/evaluation/interpretation/synthesis of other primary and sometimes secondary sources, not of raw data. The text you proposed doesn't look like an example of editors doing OR by taking data from a table in a journal paper; it looks like a statement of policy that raw data, and by implication only raw data, in journal articles is primary sourcing, and by further implication that the rest of them are secondary source material.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  09:54, 24 July 2015 (UTC)

I'll just give you some feedback here that your message seems incoherent to me and leave it at that. FWIW, I think you're acting in good faith. --Bob K31416 (talk) 22:16, 24 July 2015 (UTC)

Unless other editors respond, it looks like this suggestion is going to be smothered by a wall of irrelevant information from one editor. --Bob K31416 (talk) 08:38, 24 July 2015 (UTC)

I collapse-boxed the bulk of it. I hope my followup makes it clear why it's not irrelevant.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  09:54, 24 July 2015 (UTC)
I entirely agree with you, Bob, as to the meaning of "primary" in the context of a scientific paper. The issue with scientific papers is not that they are "primary sources" in the historian's sense (the raw data is the nearest equivalent), but whether they are notable. A report in a single scientific paper is not a suitable source in an encyclopedia. However, if notability is established via other sources, whether multiple journal articles or secondary summaries, there's no reason not to use journal articles for specific information. Otherwise you have the absurd situation that it's ok to paraphrase an already paraphrased report of scientific research, but not to return to the original papers. In practice, if you look at science articles in Wikipedia, editors rightly make wide use of journal papers and will continue to do so. Peter coxhead (talk) 08:42, 24 July 2015 (UTC)
Please see my response just above to Bob K31416. I'm saying exactly what you are with "A report in a single scientific paper is not a suitable source in an encyclopedia". I'm not using the historians sense; I'm using Wikipedia's sense. In the historians sense, "primary", etc., tend to be labels for entire classes of works, not of specific content (e.g. all news media is "primary" to many historians, even the kind we'd call secondary). I agree with you, more or less, on the notability point, except we should probably call it reputability. (I'd be cautious of applying the term "notability" because it's satisfied, on WP, with coverage regardless what the reason for the coverage is; e.g., a fraudulent paper that generates controversy may be very notable for the controversy, as its reputability falls through the floor). A WP problem is that the average editor doesn't know how personally reputable any particular researcher is. Professional in science are looking for journal articles; that's where the meat is. But WP isn't; it's looking for "has this hypothesis tricked down into the science book publishing market, or at least appeared in a literature review?" I certainly agree that it's okay to use reputable journal articles and I hold that the "use with caution" rule about primary sources is more and more clearly satisfied the more reputable the authors, the more reputable the journal, and the more frequently the work has been cited in professional literature.

I want to be clear that I'm a defender of using journal papers in articles here, and I cite them all the time, just not using Wikipedia's voice to present their claims; attribution is important. There are long-term, hard-core Wikipedians who believe wrongly that nothing in a journal paper (as distinct from a literature review) can ever possibly be secondary source material. If you still think my interpretation is overly strict, you should hear those folks. The workaround for a case where one of them says you can't use a paper for the claim at hand, and wants to revert you, would be to cite the literature review or whatever, then add the citation to the real paper along with it. If there's no lit. rev. or other secondary source, they might make WP:UNDUE claims, and other might support them, especially if it's new research. But usually carefully attributing the claim mollies people. Anyway, maybe the concerns I've been raising on this thread weren't resonating for you because of the article space you've been mostly working in or something, and you're not seeing people synth'ing up PoV nonsense from unconnected research. One case I remember was a "drinking alcohol is actually good for you" thing I could give journal paper source-abuse details about, but it would add another couple of sentences.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  09:54, 24 July 2015 (UTC)

If we disagree (and we may indeed not), then it may be about the value of using the labels "primary", "secondary" and "tertiary" in the context of scientific articles. As far as I can see, they have become harmful to Wikipedia, because they are constantly misused. What matters is the kind of material a source is being used to support, and why. If I want to support an explanation of why the author of a taxon name gave it that name, the most reliable and reputable source is the original paper. As I know to my cost, authors of compendia of scientific names and their origins and meanings copy endlessly from one another and errors get propagated. If I want to know what order the APG III system puts a particular family in, the most reliable and reputable source is the original APG III paper. But in each case wider sources are required to construct the articles in which the piece of information appears. A taxon name has to have some solid degree of acceptance in the biological community, appearing in textbooks or major biological databases. A system like APG III has to be shown to be widely used by botanists before being used in Wikipedia, and we need to know whether later revisions have been accepted. In scientific articles, routine use of "primary" = bad, "secondary" and "tertiary" = good is just plain wrong. Peter coxhead (talk) 10:13, 24 July 2015 (UTC)
I do agree that use of the terms has been confusing here, very much so, because the hard sciences, social sciences, humanities, law, history, etc., uses of these terms are all different. I do believe this is fixable by having a sentence at WP:PSTS saying as much, and then explaining that a) WP has its own definitions of these terms; b) they do not vary by subject area; c) they are descriptive of particular content in a particular context, not of entire classes of publications: the kind of material a source is being used to support, and why is key; and d) they are: [definitions here, 1, 2, 3]. In the course of explaining how to apply "the kind of material a source is being used to support, and why", we'll have to categorize, and we already have these categories, so why not just fix our terribly-worded approach to the matter? If after a year or so (it would take that long for the clarification to propagate at least, see if it's sufficient, and debates with newly arrived people go something like "I know you want to prevent this citation as a 'primary source' in historians' terms, but this is a Wikipedia secondary source", or "no, this is not a secondary source under Wikipedia's definition, it's tertiary (a category that doesn't exist in your field); as such you can't use it for this kind of claim". If there was no improvement, then see about introducing new terminology. We already have so much jargon, I think that's a last resort. People from editorial tradition A tend to be smart enough to figure out what people in tradition B mean when it's explained to them, and to understand that there are multiple traditions N. They shouldn't have any trouble grokking a tradition W. They haven't been because we're not writing it clearly. I have no opinion on APG III, and am only addressing the issue in the broadest cross-WP sense.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  23:48, 25 July 2015 (UTC)
WP:PSTS has always been the "tail that wags the dog" in this policy. Whether a source is Primary or Secondary or Tertiary is often irrelevant to the issue of whether the source is being misused to support Original Research. All three types of sources can be used correctly, and all three can be abused (used incorrectly) in ways that constitute Original Research.
What is frustrating to me is that we spend so much time and energy trying to define what constitutes a Primary, Secondary or Tertiary source, that we keep keep loosing our focus on the point of this policy... defining what constitutes Original Research. That's what this policy needs to focus on. Blueboar (talk) 12:07, 24 July 2015 (UTC)
Actually I think that what constitutes OR is well understood and, in my experience, almost never an issue – those who add OR to an article generally know perfectly well what they are doing, they just feel it's justified, usually because what they've added is "true", not because it's not OR. But I absolutely agree that the time and energy spent on trying to define types of source is wasteful and frustrating. Peter coxhead (talk) 12:45, 24 July 2015 (UTC)
Yes, we all already get that point. No one in this discussion doesn't know that making up your own theories and posting them on WP is verboten no matter what sources you use and what their classification, and few non-noobs seem to fail to understand it; it's mostly problematic PoV-warriors who are trying to do that. That does not solve the problem I, among many others, are seeing all the way across WP:CORE, and issue that exists independently of Blueboar's concern. It is the idea, directly forbidden by this policy but very badly confused by many (see thread above and recent alteration to the policy that has to be reverted) that publication of primary source material (as defined by WP) in a mostly secondary (ditto) publication magically makes the primary source material into secondary and thus an acceptable place to get WP:AEIS material from. Blueboar personally doesn't care about the PSTS distinction. We get that. Others do, because misconceptualizing it is directly leading to WP:GAMING and serious source abuse. That is, there are [at least] two different OR cases to cover with regard to primary sources, and its a serious mistake to cover why the one is wrong in wording that directly encourages more of the other.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  23:48, 25 July 2015 (UTC)

What is and is not OR[edit]

To illustrate why I think it is a distraction to define Primary, Secondary and Tertiary sources in this policy... let's examine what constitutes Original Research in each case:

  • Primary sources
    • Original Research - Analyzing, evaluating, interpreting, or synthesizing material found in the source (or multiple sources) yourself to state (or imply) a conclusion not directly stated in the source.
    • Not Original Research - Reporting on the analysis, evaluations, interpretations and conclusions directly stated within the source. (However, this should always be done with in text attribution to the source)
  • Secondary sources
    • Original Research - Analyzing, evaluating, interpreting, or synthesizing material found in the source (or multiple sources) yourself to state (or imply) a conclusion not directly stated in the source.
    • Not Original Research - Reporting on the analysis, evaluations, interpretations and conclusions directly stated within the source. (Best done with attribution, but conclusions can be stated in Wikipedia's voice if multiple sources conduct the the same or very similar analysis, evaluations, and interpretations to reach the same conclusions)
  • Tertiary sources
    • Original Research - Analyzing, evaluating, interpreting, or synthesizing material found in the source (or multiple sources) yourself to state (or imply) a conclusion not directly stated in the source.
    • Not Original Research - Reporting on the analysis, evaluations, interpretations and conclusions directly stated within the source (Any conclusions can usually be stated in Wikipedia's voice).

Have I left anything out? If not, then definition of what constitutes Original Research and not is exactly the same for all three "types" of sources. The only difference is whether the conclusions being reported are stated in Wikipedia's voice (as fact) or reported with attribution (as opinion). But that is not really an OR issue ... it's more a NPOV / DUE WEIGHT issue.
Now for the tricky part... I think we would all agree that it is far more likely that people will misuse primary sources in ways which constitute OR (whether they are doing so intentionally or unintentionally)... so some sort of cautionary warning is appropriate. The question is whether we need to define what constitutes a primary source in order to give that cautionary warning. I'm not sure that we do. Especially if doing so distracts the reader from understanding what constitutes Original Research (as I think our current definitions do). Blueboar (talk) 13:41, 24 July 2015 (UTC)

I agree, but also think there's no chance of a consensus for this view. It would be nice to be proved wrong! Peter coxhead (talk) 14:57, 24 July 2015 (UTC)
I agree with both of you. Sadly also that implementation will be difficult. Problem is that it is much easier to use /proscribe a black or white heuristic (primary source = bad, secondary source = good) than to weigh the sources on their own value; in the relevant context. Arnoutf (talk) 15:56, 24 July 2015 (UTC)
The implementation difficulty will be in fixing WP:PSTS at WP:RS, which is where it needs to be done. I agree that WP:NOR does not need to define "primary source"; etc. This material could be merged to PSTS where it belongs, and then NOR just address these (pointing to PSTS for definitions). That would go a long way to fixing the problems with both pages, though the exact wording in the extant definitions will need some help at some point.

I strongly agree with Blueboar that "it is far more likely that people will misuse primary sources in ways which constitute OR (whether they are doing so intentionally or unintentionally)... so some sort of cautionary warning is appropriate." What's triggered my concern in this thread is the flip-side of that: We also should not construct any language that implies that the only way to abuse primary sources for OR is with raw data! Some of the reasoning in the top subsection of this thread is faulty. This statement: "Repeating a conclusion that is explicitly reached in a source (be that source primary, secondary or tertiary) is NEVER original research... because the conclusion did not originate with Wikipedia... it comes from a source." is not correct. One of the most frequent NOR noticeboard dispute types is repeating (and properly citing and attributing) a claim from a paper that is unrelated to the material cited before and after it, but which, with that juxtaposition, pointedly leads the reader to a conclusion the editor believes is true but for which there is no source whatsoever. I think that's the clearest way to phrase what I've been getting at all along.

The obvious way out of this is: A scientific paper documenting a new experiment conducted by the author is a primary source on the data and outcome of that experiment. That will address both forms of OR we're talking about there. Then let's focus on getting these P/S/T definition stuff into the right page. Then on making it make sense to everyone, but clarifying that these are WP's definition, and we are not importing definitions from history, law, or any other field, much less randomly applying different definitions on-the-fly depending on what field someone things can claim scope over the topic (this view is more common than most of us thought, but it's utterly unworkable).  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:34, 26 July 2015 (UTC)

SMC I think we actually agree. Note the key words I put in bold... I stated that: "Repeating a conclusion that is explicitly reached in a source (be that source primary, secondary or tertiary) is NEVER original research".
You then state... "repeating (and properly citing and attributing) a claim from a paper that is unrelated to the material cited before and after it, but which, with that juxtaposition, pointedly leads the reader to a conclusion ...that juxtaposition, pointedly leads the reader to a conclusion the editor believes is true but for which there is no source whatsoever."
My response is... our two statements are not in disagreement... the synthetic conclusion that the juxtaposition points the reader to is Original Research because it is NOT explicitly reached in a source. The OR is not in repeating the claim... The OR consists of repeating the claim in conjunction with other material so that it reaches an implied conclusion that is not explicitly reached in a source. Blueboar (talk) 14:09, 26 July 2015 (UTC)
To put this another way... if we have Conclusion A (explicitly stated by a source), and "Conclusion B (explicitly stated by a source) and Conclusion C (not explicitly stated by a source)... it is not Original Research to repeat Conclusion A, nor is it OR to repeat Conclusion B... what is Original Research to state (or imply) Conclusion C. Why? Because C is not stated in any source. State A or B separately (ie not in conjunction) and they are fine as far as WP:NOR is concerned. Of course, one or the other will often become irrelevant or tivial when you separate them, but that is not an OR issue.
by the way... One very good way to highlight and explain synthesis to someone who does not get it is to simply switch A and B around. This often completely changes the synthetic Conclusion... this can help someone understand that A and B were being used improperly. - A similar test can be used to highlight POV pushing... if an article states A + B, and switching them to B + A changes the POV... you know that the article is not being neutral. Blueboar (talk) 15:44, 26 July 2015 (UTC)
Good "trick" at the end there. I'll remember that one. I can think of a specific place to use it already. Anyway, I agree that our views on what is OR are not in conflict. But many editors have great difficulty with this stuff. All I want is ... is a primary source on the data and outcome of that experiment so it doesn't mislead people into thinking that only the data in the paper is a primary source. The conclusion is, too, so it can be used "with caution" and directly attributed; but we mustn't imply that it can be used without attribution for the claims the conclusion makes, because it is not actually a secondary source. We don't need to explain that it can't, here, because this is the NOR page, not the V or RS pages. It's simply intolerable to have NOR appear to contradict RS. You and I might know that it it wouldn't really conflict, if you know your policy really well, but all the evidence before our eyes across Wikipedia is that it would be widely misinterpreted/gamed.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:18, 27 July 2015 (UTC)
From the beginning of WP:NOR
"Wikipedia articles must not contain original research. The phrase "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published sources exist.[1] This includes any analysis or synthesis of published material that serves to reach or imply a conclusion not stated by the sources."
If a conclusion appears in a reliable source that also has new data from an experiment, that conclusion is not original research according to this policy. Furthermore, I see no reason why that conclusion is any more susceptible to being used for synth than any conclusion in any other reliable source. --Bob K31416 (talk) 14:30, 27 July 2015 (UTC)
That's not the point at issue. And it's more susceptible because it's a primary source, presenting something brand new to the world (at least in part). WP trusts reputably published secondary sources because they are not doing this, but are the output of a multi-layered editorial "brain filtering" process that tells us that the real world is giving this claim some credence. I covered in this in detail in material that you denigrated as "irrelevant". Maybe you should have read it instead.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  21:21, 27 July 2015 (UTC)
Hmmm... could it be that we are not quite in sync as to what the actual issue under discussion is? Bob's comment certainly goes directly to the issue that I am concerned about. Blueboar (talk) 21:34, 27 July 2015 (UTC)
Retry: It's not the point at issue because the segment in question isn't defining what is or isn't OR, but what is or isn't primary (in relation to OR, theoretically, but worded as a broadly defining statement that will be WP:GAMEd as one). A primary-sourced conclusion is more susceptible to OR than one found in secondary sources (the kind from which we can repeat a WP:AEIS claim) because it's a primary source, presenting something brand new to the world (at least in part). WP trusts reputably published secondary sources because they are not doing this, but are the output of a multi-layered editorial "brain filtering" process that tells us that the real world is giving this claim some credence. I covered in this in detail in material that you denigrated as "irrelevant". Primary sources are highly prone to contextualizing/interpretive original research, especially via assumptions of acceptance of claims/conclusions (i.e. of veracity), by misinterpretation of precisely worded conclusions during recasting in a Wikipedian's own words (a don't-plagiarize process we're supposed to engage in but which is more "dangerous" when working with primary sources), correlation–causation errors, mistaking topically similar conclusions as correlative at all, etc., etc., etc.

Let's turn this around: Why is it seemingly very important to you to prevent WP:NOR from saying that anything but the raw data in a journal paper is primary sourcing, when there's no consensus that it is anything but? No one objects to adding the specificity that the data is primary, in addition to the conclusion (I agree wholeheartedly!). No one's even saying everything in a paper is primary; many papers contain a large amount of secondary material, where they engage in small-scale literature reviews on a particular matter; I also already covered this. [Actually I shouldn't say "no one" here; I know of two editors who should know better who keep trying to tell people that this secondary material is primary.] If this passage were leading directly into an example of abuse of raw data to engage in one kind of OR, it would be understandable to limit that wording to "data" (and then follow it up with a "conclusion" version and an illustration of the other kind of OR I'm talking about). But there are no such examples. And it's been proposed here to relocate the bulk of this material to WP:PSTS, anyway, which might just moot the matter entirely.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  21:47, 27 July 2015 (UTC)

1) It seems like the issue you are concerned with doesn't have to do with original research but rather with the reliability of sources. Am I understanding you correctly?
2) The definitions of the three types of sources are introduced with the following statement at the end of the second paragraph of WP:PSTS],
"For the purposes of this policy, primary, secondary and tertiary sources are defined as follows:[2]"
--Bob K31416 (talk) 23:39, 27 July 2015 (UTC)
1) Nope. 2) So? Does not address the issue I raised: The change you want to make is directly misleading. There is no rationale for removing the one word and substituting the other instead of just adding the word you want to be covered there. I'm highly skeptical that re-re-re-re-re-re-re-stating this another dozen different ways is going to be productive. Please see WP:NOTGETTINGIT.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:22, 28 July 2015 (UTC)
1)The reason I thought your issue had to do with reliability was your comment, "WP trusts reputably published secondary sources because they are not doing this, but are the output of a multi-layered editorial 'brain filtering' process that tells us that the real world is giving this claim some credence."
2) Regarding your comment "A primary-sourced conclusion is more susceptible to OR than one found in secondary sources (the kind from which we can repeat a WP:AEIS claim) because it's a primary source, presenting something brand new to the world (at least in part)." – When you say "primary-sourced conclusion", are you referring to a conclusion made by a Wikipedia editor, or a conclusion that appears in a reliable source that a Wikipedia editor is using in an article without going beyond that conclusion? --Bob K31416 (talk) 00:28, 28 July 2015 (UTC)
1) Ok ... I don't see where the inference comes from, though. 2) In that particular construction, I'm referring to the conclusion reached by the authors of a journal paper, which NOR defined unquestionably as a primary source (but we have used "conclusion" in some other ways in this discussion). Whether it's reliable enough to cite (WP:RS) at all (WP:V) doesn't relate strongly to how we can cite it, for what (WP:NOR at WP:PSTS).

I'll re-ask my above question, in different terms, since staying "on-track" seems to be a goal (and I agree with it): Why do you want WP:NOR to state that data specifically, and by implication only data, in a journal paper is primary source material? How could this be jusified, when the policy clearly says otherwise, and we know for certain that many editors are already confused, despite the clarity of the policy's definitions, into thinking that a paper's conclusion somehow isn't primary [for WP purposes]? Why make this confusion worse, with a substitution of "data" for "conclusion", instead of helping clarify by adding "data" to "conclusion"? We even know that the confusion comes from differing external definitions of "primary". What end could possibly be served by perpetuating the incorrect belief that WP uses external, randomly changing field-by-field, definitions of "primary", when the policy concretely defines what that term means for WP editing purposes?  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:57, 28 July 2015 (UTC)

But Bob's contention (correct me if I have this wrong, Bob) is that the conclusion shouldn't be defined as Primary. That this definition is inaccurate. Bob's proposal is to change the definition (so it no longer defines the conclusion as Primary... only the Data). SMC, Your argument above seems circular... you seem to be saying: you can't change the definition contained in the policy because the policy contains that definition. I suspect (and again correct me if I have this wrong) that the real reason for your objection is that you disagree with Bob and think the conclusion is primary. To be honest, I don't know which of you is correct... so... to help me make up my mind, let me ask you both to lay out why you take the stance you do. Why is a conclusion primary?... or ... why isn't it primary? Blueboar (talk) 01:31, 28 July 2015 (UTC)
You're understanding me correctly that the conclusion shouldn't be defined as primary. My reason is that the definition is for purposes of this policy No original research, as stated in the introduction to the definition, and that using any conclusion from any reliable source, without going beyond that conclusion, is not original research. And there's no reason to believe that new conclusions in a scientific paper are any more prone to being used for synth than are any conclusions in any other reliable source. --Bob K31416 (talk) 12:20, 28 July 2015 (UTC)
OK... you seem to be saying that if Primary = OR then the flip side must also be true: Not-OR = not-Primary. My reaction is that your argument is flawed from the beginning, because Primary doesn't = OR. A conclusion can be Primary without any OR involved. And OR can occur without the sources being primary.
That said... what I really wanted to know is: are conclusions in scientific papers considered primary or secondary in the real world (outside of Wikipedia)? Blueboar (talk) 12:59, 28 July 2015 (UTC)
I didn't say that primary=OR, or mean to imply that. But that's OK, sometimes editors are misunderstood. I have the same position as this policy regarding primary sources, i.e. "primary sources that have been reputably published may be used in Wikipedia; but only with care, because it is easy to misuse them". For example, the data in an experiment is prone to different interpretations and Wikipedia editors might offer their own conclusions about the data. Whereas the conclusions in a paper about an experiment, aren't any more prone to Wikipedia editors using them to make their own conclusions than are the conclusions in any other paper.
Re "are conclusions in scientific papers considered primary or secondary in the real world" – It seems to vary. Also note that the definition of primary source outside of Wikipedia is not for the purpose of discouraging the use of primary sources for original research. --Bob K31416 (talk) 14:24, 28 July 2015 (UTC)

Quick aside... time to give some history. If you look back to the development of the PSTS section... At one point, the policy included a statement that said (essentially): Wikipedia is a Tertiary source... adding Original Research turns it into a Primary source, so don't add OR. At the time, it was felt that it would be helpful to explain the terms used in that statement. The original intent was not to discourage the use of primary sources, or to say "using primary sources causes OR"... but to explain that adding Original Research turns Wikipedia into a primary source. I know that's quite different from what we say now ... a lot of instruction creep has occurred since those early days... I just wanted to remind everyone why we included the definitions in the first place. Blueboar (talk) 15:53, 28 July 2015 (UTC)

Status of proposed change[edit]

The proposed change in the subject sentence of Primary, secondary and tertiary sources was to change "outcome" to "data", so that the sentence becomes,

"a scientific paper documenting a new experiment conducted by the author is a primary source on the data of that experiment".

So far the responses suggest,

If an editor's intentions have been incorrectly interpreted in the above, that editor is invited to give their correct position. --Bob K31416 (talk) 22:32, 25 July 2015 (UTC)

P.S. The other editor in the above discussion, Arnoutf, and any other editors are invited to give their positions on the proposal. --Bob K31416 (talk) 23:15, 25 July 2015 (UTC)

  • Weak support – Only weak support. The data is definitely primary, but also the discussion section often contains primary arguments of the authors (especially in social sciences. Making that explicit would create a lot of confusion so I think the proposal is an improvement overall). Arnoutf (talk) 10:04, 26 July 2015 (UTC)
    • @Arnoutf: I'm curious why you think the conclusion reached by a paper in a journal is not also primary, when the WP:PSTS clearly says that it is, and WP:RS#Scholarship reiterates this in summary form. As discussed elsewhere in here (cf. comments by Peter Coxhead), there's obviously been perennial confusion going on between different external definitions of primary vs. secondary source in different fields, but WP has its own definitions, and they're quite detailed.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:37, 28 July 2015 (UTC)
      • If your read my rationale for weak support you could see I fully agree that conclusions and discussions of scientific papers are primary insofar that they reflect on the data collected by the same author. If we do, however, rigorously apply this all empirical papers are primary (and so would be all journalism based on interviews), and to take it even more into the extreme, even the conclusion/discussion of review papers would be primary as that would give the ideas of the authors. So to take this argument into the extreme would nullify the difference between primary and secondary sources. The world is not a simple black and white. Arnoutf (talk) 08:32, 28 July 2015 (UTC)
  • I have crossed myself off of the list of those supporting. I would say I am still undecided. Blueboar (talk) 14:15, 26 July 2015 (UTC)

"Status of proposed change" is "under discussion still". Wikipedia is not a vote and it's inappropriate for an involved party in the discussion to try to "assess the consensus" in a steering way, when the discussion is still on-going. I also object to your special pleading above that amounts to 'help! my pet proposal is being derailed by someone whose objection I don't want to address; please come help me shut him up'. WP:Consensus is formed by actually taking the time to understand what people's objections are and try to resolve them; not by trying to make the objector go away, or drum up an entourage to drown out the objection. Looking above at the discussion, the concerns I've raised have yet to be addressed. Some participants want to make one particular kind of point about OR and primary sources, while I am pointing out that the proposed wording implies that a completely different kind of primary-source-based OR is not OR. Some appear to understand my point. Everyone agrees that WP:PSTS has wording problems (though this is naturally not the place to try to fix them). No one is objecting to the idea of the original proposal (using raw data from a paper to draw your own conclusions is OR), only the wording. So WP:DONTPANIC. The discussion is ongoing. We don't need to count anything yet, and no consensus is going to be reached by refusing to address concerns of some of the participants. You can't "WP:WIN" that way.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:05, 26 July 2015 (UTC)

Out of order. This section is out of order because the alleged proposal for a change in wording described at the beginning of the section either does not exist at all, or cannot readily be found. The concerns about an involved editor trying to summarize an unfinished discussion also apply. Jc3s5h (talk) 15:24, 27 July 2015 (UTC)

Here is the the full paragraph of WP:PSTS, where I have underlined the subject phrase.
  • Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, a period of history, a work of art, a political decision, and so on. Primary sources may or may not be independent or third-party sources. An account of a traffic accident written by a witness is a primary source of information about the accident; similarly, a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. Historical documents such as diaries are primary sources.
The proposal is to change "outcome" to "data".
Regarding giving the status, there had been a lot of discussion digressing from the proposal and this was a way of getting it back on track. Also, it helped clarify editors' positions by requesting any corrections. For example, Blueboar clarified being undecided instead of supportive. --Bob K31416 (talk) 18:31, 27 July 2015 (UTC)
The title if this subsection, "Status of proposed change", implies there is already a proposal, and this section is giving the status of it. But in fact, the proposal is introduced in this subsection. This method of introducing the proposal fails to give interested editors reasonable notice that the discussion is occurring. Thus this discussion is invalid and cannot create a consensus to edit the policy. Jc3s5h (talk) 19:20, 27 July 2015 (UTC)
Jc3s5h... for the record... the proposal was actually introduced in Bob's comment at 02:51, 24 July 2015 (above ... it's the 4th paragraph of the original thread - Wikipedia talk:No original research#Primary source – example of scientific paper). OK... perhaps Bob could have marked it better, so others realized that he actually had made a proposal (But then again, do we really need to mark every proposal with the word "PROPOSAL" in flashing neon so no one can miss it?)... but my point is that it wasn't introduced in this sub-section. I don't think Bob was trying to improperly influence the outcome of the discussion by reporting on the "status" of that original discussion... I think he was trying to get us back on track... reminding us what his original proposal actually was, after we had gotten somewhat side-tracked away from it. Blueboar (talk) 20:37, 27 July 2015 (UTC)
I don't think it is fair to claim consensus for any change if the proposal is introduced after the discussion has lead editors to think the thread will not be of interest to them, causing the uninterested editors to ignore the thread in their watchlist. Jc3s5h (talk) 21:47, 27 July 2015 (UTC)
And "getting us back on track" doesn't require misleading vote-counting; even half those whose votes he counted disagree with his reckoning, anyway. We're also not "off-track"; at least one person in this overall discussion does not appear to have understood the quite central point I'm making, despite it having been explained in clear and detailed terms repeatedly. The deletion of "conclusion" and replacement with "data", instead of addition of "data", is misleading, would be WP:GAMEd without doubt, and is based on a misapprehension of what "primary source" means on (and for) WP and for this policy. (This misapprehension is reflected in the thread immediately above this overall one; I haven't even addressed that yet; it turns the meaning of that policy on it's head). I don't think there's a bad-faith attempt to improperly influence the outcome; It seems to just be heel-digging in response to the proposal being challenged in terms less easy to dismiss than they seemed to him at first. I have little direct interaction with the proponent and have no reason to assume bad faith about him.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  22:04, 27 July 2015 (UTC)
I think my previous comments about why I gave the status should be sufficient and I think the comments in the last two messages about giving the status are unreasonable criticism and divert the discussion from the substance of the proposal. --Bob K31416 (talk) 00:01, 28 July 2015 (UTC)
But this entire subsection, which you created, is a diversion of the discussion from the substance of the proposal. And I'm defending you on WP:AGF grounds in the wake of an uninvolved editors' criticism, even if also critical about the idea (expressed by someone else) that the discussion is "off-track", and of your judgement in launching the diversion this way to begin with. Given the immediately previous exchange (in edit history, not this subthread), it's clearly not off-track, because the central objection to the substitution-instead-of-addition edit you want to make has not been addressed. No amount of circular argument gets around that.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  00:29, 28 July 2015 (UTC)

Suggestion: Since we are already back on track in the thread above (ironically done by going off track in my "what is and is not OR" sub-thread)... may I suggest that we simply halt this sub-thread, and continue the "on track" discussion. Let's just pretend that this sub-thread never happened, and concentrate on trying to reach consensus. Blueboar (talk) 01:21, 28 July 2015 (UTC)

Comment - Having read through the discussion above, and thinking over the text a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment; is there any reason why we should need to specify which sections of the experiment are considered primary?

Would it not be simpler to simply remove the underlined section; giving a scientific paper documenting a new experiment conducted by the author is a primary source on that experiment? NB: Please feel free to move this comment to the appropriate place in the various threads above. I could not work out where was best. - Ryk72 'c.s.n.s.' 11:05, 28 July 2015 (UTC)

First off, thanks for joining this discussion. (If anyone moves Ryk72's message, please move my response along with it and keep them together.)
In an experiment, data is taken. Then a paper is written about the experiment and the paper includes the data from the experiment and conclusions about the data. The point of contention is whether or not for the purposes of this policy of no original research, to include the conclusions of the paper as a primary source. Did you mean for the term "experiment" to include or exclude the conclusions of the paper? --Bob K31416 (talk) 12:02, 28 July 2015 (UTC)
Hi Bob K31416, My inclination is that the paper, as a whole & its component parts, are primary; but I am keen to see the answer to Blueboar's question, below - perhaps it is not the categorisation of primary or secondary that is the part of policy that wants changing. - Ryk72 'c.s.n.s.' 13:57, 28 July 2015 (UTC)
Bob... does it matter whether the conclusions are considered primary or secondary? If so, why does it matter? Blueboar (talk) 13:30, 28 July 2015 (UTC)
It matters because Wikipedia discourages the use of primary sources. So material using a source that is labeled primary is more prone to being deleted. --Bob K31416 (talk) 14:30, 28 July 2015 (UTC)

It seems that I've said all I have to say on the subject in this long discussion, so I'll leave now, still supporting the proposed change from "outcome" to "data" after reading all the arguments against it. --Bob K31416 (talk) 15:02, 28 July 2015 (UTC)

It sounds like your underlying concern isn't so much with the definition... but with the discouragement and inappropriate deletion. Is this an accurate assessment? Blueboar (talk) 15:11, 28 July 2015 (UTC)

When to fact tag vs. just delete[edit]

FYI: Pointer to relevant discussion elsewhere.

This topic has had a lot of churn on various pages lately, so I thought people here might be interested in this discussion: Template talk:Citation needed#When to remove unsourced info vs. when to add this tag?.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  05:23, 23 July 2015 (UTC)