Wikipedia talk:No original research/Archive 47

Archive 40

←

Archive 45

Archive 46

Archive 47

Archive 48

Archive 49

Archive 50

RFC - Replace examples in WP:SYNTH with link to examples

Please comment on the following proposed version for WP:synth. It differs from the present version only by replacing the two examples currently in the section with a link to a page that contains the two examples. --Bob K31416 (talk) 18:42, 2 September 2009 (UTC)

---

== Synthesis of published material that advances a position ==

Do not combine material from multiple sources to reach a conclusion not explicitly stated by any of the sources. Editors should not make the mistake of thinking that if A is published by a reliable source, and B is published by a reliable source, then A and B can be joined together in an article to reach conclusion C. This would be a synthesis of published material that advances a new position, and that constitutes original research. "A and B, therefore C" is acceptable only if a reliable source has published the same argument in relation to the topic of the article. For examples see Original synthesis at Wikipedia:No original research/Examples.

Carefully summarizing or rephrasing source material without changing its meaning is not synthesis—it is good editing. Best practice is to write Wikipedia articles by taking material from different reliable sources on the topic and putting those claims on the page in your own words, with each claim attributable to a source that explicitly makes that claim.

---

Comment of requester: This is an improved version that has a size that makes for better reading and communication and has a link to the examples, instead of having them in the section. I think it is a mistake to have most of WP:SYNTH filled with examples since it takes up too much space compared to the other more important parts and obscures them. --Bob K31416 (talk) 18:42, 2 September 2009 (UTC)

I like Bob's idea. He makes the rule easy to understand: thus, easy to follow! :-) --Uncle Ed (talk) 18:45, 2 September 2009 (UTC)

• Oppose, the WP:NOR/Examples page was never meant to replace the examples in WP:NOR. As I stated here, the reason for that is because I believe it’s best to keep a couple of main examples of synth right in the policy. Unlike subpages, main Policy pages are watched closely so there's less chance of having poor or misleading examples added. I also think it's good to have a couple of examples right in the policy, because it helps give direct substance to the concepts outlined therein - subpages can help expand on those, but aren't necessary to the core understanding of the policy. Therefore, I oppose Bob's proposed changes. Dreadstar ☥ 19:05, 2 September 2009 (UTC)

Please note that the link is to a page in the history of the examples page so it can't be altered. The only way to change it is by going through WP:NOR to link to another page in the examples page history that contains the desired changes. This is somewhat serendipidity since I made this link because you recently removed one of the examples over there, so I linked to a page that has all of the examples. Thanks. --Bob K31416 (talk) 22:51, 2 September 2009 (UTC)

• Oppose. I oppose replacing the examples with a link. We need one page, the policy page, that people can look at to find out what OR is, without having to visit yet another page too. SlimVirgin ^{talk|contribs} 01:22, 3 September 2009 (UTC)
• Oppose. Some significant percentage of people who would not follow the link if the examples were removed, will read the examples if they remain on the page, and concrete examples are a teaching tool that makes accessible what abstract description may fail to convey. In addition to the see also link, the examples section can have a link at the the end not unlike: "See Wikipedia:No original research/Examples for additional examples of improper original synthesis", but the few on the page should remain. The examples are not filling the page, obscuring other points. It's a short list that I don't think places much strain on those visiting. Moreover, those who would find reading through them too much to handle are, in any event, exactly the types who aren't likely to visit a linked satellite page.--Fuhghettaboutit (talk) 02:04, 3 September 2009 (UTC)
• Oppose to moving examples to a sub-page as examples should add clarity to what may be a confusing policy point. I believe it's important for key examples to remain in-line rather than having user's jump to another page. For example, someone is likely to be confused by the A, B, C stuff in WP:SYNTH. The examples should then be clear enough that the reader then understands A, B, C. --Marc Kupper|talk 07:34, 3 September 2009 (UTC)
• Oppose. Examples like this are common and illuminating in policy pages. These examples are particularly good ones. Sub-pages are not well watched, and may fill up with poor examples. Jayjg ^(talk) 01:32, 4 September 2009 (UTC)

FYI, the link in WP:SYNTH is to a page in the history of the examples page, so it can't be altered without going through WP:NOR to link to another page in the history with the desired changes. What do you think of the alternate proposal below? --Bob K31416 (talk) 02:02, 4 September 2009 (UTC)

Alternate proposal (only reorganizes within section and keeps examples)

Re Fuhghettaboutit's comment: "Moreover, those who would find reading through them too much to handle are, in any event, exactly the types who aren't likely to visit a linked satellite page." - If they found the more complex example too much to handle and stopped reading, they would miss the important last paragraph of the section which comes after the examples. How would all of you feel about putting the last paragraph after the first paragraph and then have the examples follow?

Click on show to view the contents of this section

== Synthesis of published material that advances a position ==

Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources. Editors should not make the mistake of thinking that if A is published by a reliable source, and B is published by a reliable source, then A and B can be joined together in an article to reach conclusion C. This would be a synthesis of published material that advances a new position, and that constitutes original research. "A and B, therefore C" is acceptable only if a reliable source has published the same argument in relation to the topic of the article.

Carefully summarizing or rephrasing source material without changing its meaning is not synthesis—it is good editing. Best practice is to write Wikipedia articles by researching the most reliable published sources on the topic and summarizing their claims in your own words, with each claim attributable to a source that explicitly makes that claim.

A simple example of original synthesis:

The UN's stated objective is to maintain international peace and security, but since its creation there have been 160 wars throughout the world.

Although no conclusion is drawn and both facts are true, the sentence implies that the UN has failed to maintain world peace. If no reliable source has combined the material in this way, it constitutes original research. It would be easy to imply the opposite using the same material, illustrating how, when no source is provided, facts can easily be manipulated:

The UN's stated objective is to maintain international peace and security, and since its creation there have been only 160 wars throughout the world.

The following is a more complex example of an original synthesis. It is based on an actual Wikipedia article about a dispute between two authors, here called Smith and Jones:

Smith claimed that Jones committed plagiarism by copying references from another author's book. Jones responded that it is acceptable scholarly practice to use other people's books to find new references.

Now comes the original synthesis:

If Jones did not consult the original sources, this would be contrary to the practice recommended in the Harvard Writing with Sources manual, which requires citation of the source actually consulted. The Harvard manual does not call violating this rule "plagiarism". Instead, plagiarism is defined as using a source's information, ideas, words, or structure without citing them.

The first paragraph was properly sourced. The second paragraph was original research because it expressed the editor's opinion that, given the Harvard manual's definition of plagiarism, Jones did not commit it. To make the second paragraph consistent with this policy, a reliable source would be needed that specifically comments on the Smith and Jones dispute and makes the same point about the Harvard manual and plagiarism. In other words, that precise analysis must have been published by a reliable source in relation to the topic before it can be published in Wikipedia by a contributor.

--Bob K31416 (talk) 22:41, 3 September 2009 (UTC)

• Not only do I think that change does no damage, I think it reads better, i.e., I support that change.--Fuhghettaboutit (talk) 23:06, 3 September 2009 (UTC)

• Support alternate proposal to move the last paragraph up to just below the lead. The next line can then be changed from "A simple example" to "A simple example of synthesis". --Marc Kupper|talk 08:34, 4 September 2009 (UTC)

• It was like that [1] before someone moved it, as I recall without discussion (though maybe I missed the discussion, or have forgotten it). SlimVirgin ^{talk|contribs} 08:44, 4 September 2009 (UTC)

Does that mean you would like to implement the alternate proposal? --Bob K31416 (talk) 18:48, 4 September 2009 (UTC)

I don't see what difference it makes either way. SlimVirgin ^{talk|contribs} 22:07, 4 September 2009 (UTC)

Does that mean you oppose the alternate proposal? --Bob K31416 (talk) 23:30, 4 September 2009 (UTC)

The last small paragraph of WP:SYNTH is important because it helps avoid the misapplication of NOR, but unfortunately, it is located in a place where it can easily be overlooked. I recall having a hard time finding it when I needed to refer to it. When I came to the examples I stopped my search for it in that section because it looked like all that was left were the examples. --Bob K31416 (talk) 18:33, 8 September 2009 (UTC)

I took the liberty of copying the following comment of Blueboar from the section below: Are there other issues.--Bob K31416 (talk) 21:57, 16 September 2009 (UTC)

• I assume you are talking about the "best practice" paragraph... is so, I completely agree. At minimum, this paragraph lead off the section (similar to how I proposed above... but with the newly agreed upon language). Alternatively, perhaps it could be stated in its own section (suggest a header of: "Source based research" if we go this route). Blueboar (talk) 20:27, 16 September 2009 (UTC)

Editprotected request for the above Alternate proposal

{{editprotected}} Please note that this proposal involves no policy change. It moves the last paragraph in WP:SYN to the position of being the second paragraph, and changes the beginning of the resulting third paragraph from "A simple example:" to "A simple example of original synthesis:". The consensus consists of support from 4 editors including myself, and an unclear position from another. Otherwise, there have been no objections since the Alternate proposal was first proposed Sept 3 2009, almost 2 weeks ago. --Bob K31416 (talk) 22:08, 16 September 2009 (UTC)

This page isn't protected from editing in any way. Just do what you want done. Hersfold ^(t/a/c) 01:41, 17 September 2009 (UTC)

Full protection removed from WP:NOR 2 hours after editprotected request, so I have made the proposed change of the Alternate proposal. --Bob K31416 (talk) 02:15, 17 September 2009 (UTC)

Proposed SYN changes

From the earlier discussion, it seemed that two changes had potential consensus:

Current version

Proposed version

(First sentence of SYN section:) Do not combine material from multiple sources to reach a conclusion not explicitly stated by any of the sources. (Last sentence of SYN section:) Best practice is to write Wikipedia articles by taking material from different reliable sources on the topic and putting those claims on the page in your own words, with each claim attributable to a source that explicitly makes that claim.

(First sentence of SYN section:) Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources. (Last sentence of SYN section:) Best practice is to write Wikipedia articles by researching the most reliable published sources focusing on the article topic and summarizing their claims in your own words, with each claim attributable to a source that explicitly makes that claim.

For the reasoning behind these proposals, see the discussion above. I would like us to raise an editprotected request to implement these changes. Before we do so, are there any objections? --JN466 18:33, 13 September 2009 (UTC)

It's been a long time since I've had anything to say here, but I really like the phrase "most reliable published sources focusing on the article topic," I've seen too many cases where a published source that touches incidentally on the topic is used while ignoring sources that explicitly focus on the topic. Well done. --SteveMcCluskey (talk) 19:30, 13 September 2009 (UTC)

I would have hoped that both of those changes would be unnecessary because nobody would have to be explicitly told that, but unfortunately we really do have to spell things out for some of our less experienced and/or argumentative editors. DreamGuy (talk) 20:11, 13 September 2009 (UTC)

Seems fine to me. SlimVirgin ^{talk|contribs} 21:16, 13 September 2009 (UTC)

Well, I'll just say again that I prefer to switch things around and state best practice first... to make clear how a misapplication of best practice involving synthesis can lead to OR). But that is just a preferance and not a necessity. Blueboar (talk) 21:43, 13 September 2009 (UTC)

Fine with me to proceed. Seems like an improvement and will add clarity. N2e (talk) 21:51, 13 September 2009 (UTC)

I support this change too. "Imply" addresses juxtaposition, which had been a concern of mine for a long time. I think this version is overall better. Crum375 (talk) 21:57, 13 September 2009 (UTC)

Looks good to me too. Dreadstar ☥ 22:04, 13 September 2009 (UTC)

There's an ambiguity in the following part of the proposed version:

Best practice is to write Wikipedia articles by researching the most reliable published sources focusing on the article topic...

Is the editor doing the focusing or is the source doing the focusing? --Bob K31416 (talk) 23:26, 13 September 2009 (UTC)

Good catch Bob... Suggest: "Best practice is to Write Wikipedia articles by researching the most reliable published sources that focus on the article topic..." Blueboar (talk) 00:11, 14 September 2009 (UTC)

Except they may not actually focus on the article topic, only mention it incidentally, and still be valid and useful. How about "that relate to ..."? (or "relating to" instead of "focusing on"?) Crum375 (talk) 00:15, 14 September 2009 (UTC)

I disagree... remember, we are talking about "best practice" here, not acceptable because it is "valid and useful". The best practice is to use sources that directly focus on the topic. We may not always achieve the goal... but it is still best. Blueboar (talk) 00:38, 14 September 2009 (UTC)

I disagree. We need to build an encyclopedia. Best practice is to find the best possible sources that relate to the topic of interest. If you restrict "best practice" to ideal sources which focus on the article topic, you may well end up with absolutely nothing in many cases. So it doesn't make sense for a policy to focus on an ideal and often non-existent situation; we need to give editors real and practical guidance. Crum375 (talk) 00:50, 14 September 2009 (UTC)

(edit conflict) (Re Blueboar's remark) That's what I tried to convince myself of too, but it was wishful thinking. Once the ambiguity is removed, it implies that other sources are undesirable. --Bob K31416 (talk) 00:58, 14 September 2009 (UTC)

I would remove the word "focusing", as the text reads fine without it. — Carl (CBM · talk) 01:08, 14 September 2009 (UTC)

Seems like the same problem. --Bob K31416 (talk) 01:20, 14 September 2009 (UTC)

It is perfectly possible that the "most reliable published sources on the article topic" may not focus on the article topic. However, if there are sources that do focus on the article topic, those are more likely to be among the best sources. — Carl (CBM · talk) 01:30, 14 September 2009 (UTC)

Agree with CBM. Crum375 (talk) 01:37, 14 September 2009 (UTC)

Since I consider focused sources more significant than those that treat a topic tangentially, I agree with Blueboar's change from "focusing on" to "that focus on". --SteveMcCluskey (talk) 02:58, 14 September 2009 (UTC)

Focused articles are generally better than broader ones, but not always, and in any case are not always available. So being focused is a plus, but not a requirement. Crum375 (talk) 03:25, 14 September 2009 (UTC)

I don't think we are saying it is a requirement... as I see it, best practice is always more than just what is required (that's why it's 'best')

But as another suggested alternative, what about "...most reliable published sources that directly discuss the article topic." Blueboar (talk) 04:18, 14 September 2009 (UTC)

"Best practice" has to be a practical guidance, so it can't just cover a (possibly) small subset of real-life cases encountered. And the sources may not actually "discuss" the article topic, but only mention it or relate to it. I think the original wording is quite reasonable: "...different reliable sources on the topic...", as it allows the various possibilities, which is what CBM noted above. Crum375 (talk) 11:11, 14 September 2009 (UTC)

Let us remember that the WP:GNG notability guideline requires that there be sources that "address the subject directly in detail, and no original research is needed to extract the content." It adds that "significant coverage is more than a trivial mention but it need not be the main topic of the source material." So if there are no sources "directly addressing the subject in detail", we would not be having an article on it. JN466 11:30, 14 September 2009 (UTC)

"Sources directly addressing (or relating to) the subject" would be acceptable too. The point is that there may be a combination of sources. Some may be right on topic but not as reliable, others may address it tangentially but be highly reliable. And of course this may be just one small aspect of the main article topic. A policy has to cover them all. Crum375 (talk) 11:38, 14 September 2009 (UTC)

It seems that we are getting bogged down in something that doesn't even have to be mentioned. The following should work well,

"Best practice is to write Wikipedia articles by researching the most reliable published sources and summarizing their claims in your own words..."

I think that the main intent of this passage was to temper the prohibition on OR synthesis with the recognition that summarizing multiple sources is a type of synthesis that is allowed. The attempt at adding extra details here is a digression that leaves the main point less clear. Please note that this issue is already discussed, and discussed better, in the Sources section of WP:NOR. Here's an excerpt.

Take care, however, not to go beyond what is expressed in the sources or to use them in ways inconsistent with the intent of the source, such as using material out of context.

--Bob K31416 (talk) 12:24, 14 September 2009 (UTC)

Using sourced material out of context is exactly what SYN is, and I feel we should address this point in the SYN section. The basic concern here is this: if there are influential, widely cited standard works on a subject, then good writing requires that an editor familiarize themselves with the standard works and summarize the arguments and analyses made in these sources. Even for topics that are less well covered, it is essential that editors familiarise themselves with those sources that have made the topic notable, and access those sources first of all. This is expected of any professional writer in real life, and following this practice ensures that our articles are seen as competent. Acquiring such background knowledge will help an editor evaluate the significance – or lack of it – of a tangential mention of some fact related to the article topic. The problem we seem to be having often is that editors see something in a source which does not mention the article topic, but which they feel may be related. Just to recap some examples of this problem:

1. We had mentioned the example with the boycott call of a minor activist and the targeted company's drop in sales. No source connected the boycott call to the drop in sales, but it was still added to the article on the activist (actually by the activist himself, if I recall correctly). I would argue that best practice here would have been to look for sources that connected the sales drop to the boycott call, i.e. that focused, at least in some part of the overall text, on the activist that was the subject of the article. Absent such sources, the drop in sales should not be mentioned as a fact in the article on the activist (except perhaps to say that the activist claimed on his website that he had caused it).
2. Another typical SYN pattern goes like this: "X became president of Y in 2006.(ref A) Since he became president, unemployment in Y has risen by x%, and industrial output has fallen by y%.(ref B, not ascribing this to President X) It should be clear to everyone that given the recent world-wide economic downturn, statements like this could be made about almost any head of government on earth. Adding information like this from sources that do not focus on X's tenure as president degrades an article. Adding information like this sourced to a widely cited analysis of President X's performance, on the other hand, is "best practice", because we are citing a source that focuses on President X.
3. An article on event Y in Israel cites U.S. reporter X's analysis of the event. Another editor adds a sourced statement alleging that American reporters writing on Israel are generally biased. The cited source, however, mentions neither U.S. reporter X whose article we quoted, nor does it mention the specific event Y the reporter wrote about. Again, I would argue that editing like this is not best practice and does not result in high-quality articles. Best practice is to find and include reliable sources focusing on event Y that present alternative viewpoints of the event, and to include these per WP:NPOV.

These are the sorts of situations that this would address. In such situations, it is important to locate and summarise existing detailed analyses focusing on the article topic, rather than synthesising a missing perspective.

Of course, let's not forget that in many cases where editors are seeking to add something via synthesis, there is an underlying "rightness" to their concern. What this proposed wording should do is encourage editors to do a little bit more work, digging for sources that actually make the points they have synthesised themselves. That should ultimately benefit article quality. At any rate, that is the intent of the proposal as written. JN466 14:00, 14 September 2009 (UTC)

Except for some of those involved in this discussion, I don't think that readers will recognize all that from just the phrase on focusing. What the readers will think is that they need to use only articles that focus on the topic, where "focus" isn't defined. This is too restrictive in any case. Trying to clarify that with more verbiage just makes it even more awkward. Also, the ambiguity about the editor or source doing the focusing, mentioned previously, still remains. I stand by my previous recommendation. --Bob K31416 (talk) 17:15, 14 September 2009 (UTC)

Blueboar suggested a way to remove the ambiguity of who does the focusing. Failing that, I am open to suggestions on how we can fix the concern shared by several editors here, that the current phrasing of the "best practice" description seems to encourage SYN cherry-picking rather than source research. JN466 17:25, 14 September 2009 (UTC)

As I recall, the cherry-picking aspect was addressed by changing "taking" to "researching". --Bob K31416 (talk) 19:41, 14 September 2009 (UTC)

Trying to finish off the 2nd part of the proposal

Okay. The current policy wording is,

"Best practice is to write Wikipedia articles by taking material from different reliable sources on the topic and putting those claims on the page in your own words, with each claim attributable to a source that explicitly makes that claim."

Looking back on the comments made in the above discussion, would anyone have an objection to the following wording?

"Best practice is to write Wikipedia articles by researching the most reliable published sources on the article topic and summarizing their claims in your own words, with each claim attributable to a source that explicitly makes that claim."

--JN466 23:42, 14 September 2009 (UTC)

Consider an example of an article on Ulysses S. Grant. A biography of Grant would be a source on the topic of the article. Whereas a book on the American Civil War would not be on the topic of the article. I think best practice should include researching the American Civil War book for information on Grant.

Consider the following change instead:

"Best practice is to write Wikipedia articles by researching the article topic in the most reliable published sources and summarizing their claims in your own words, with each claim attributable to a source that explicitly makes that claim."

--Bob K31416 (talk) 02:35, 15 September 2009 (UTC)

If a book on the American Civil War is among the most reliable published sources on Ulysses S. Grant, then I agree it would be best practice to cite it. That sort of thing often happens with minor topics that tend to be covered in works with wider scope, rather than having entire books devoted to them. I think the proposed wording allows that. Note that "reliable sources on the topic" is part of the present wording of this sentence; it never seems to have caused a problem. Could you live with the text as proposed? A number of people above said they liked it. JN466 10:00, 15 September 2009 (UTC)

"reliable sources on the topic" is part of the present wording and your proposal has changed it to "reliable sources on the article topic" which is a bit stronger wording and more likely to put the book on the American Civil War out of the category of best practice. Since this discussion is about your changes and not other possible changes, your proposed changes would be acceptable to me if you didn't include "article". If you would like to include "article", it would be acceptable to include it in the way that I suggested in my version. --Bob K31416 (talk) 14:02, 15 September 2009 (UTC)

Thanks. I can live with that and will drop "article" from the proposal. JN466 21:24, 15 September 2009 (UTC)

• I support Jayen's proposal. Any information added to a Wikipedia article should, ideally, be based on reliable sources that present this information in direct connection with the article subject. Whether the source focuses on the specific article topic (Ulysses S. Grant) or a broader subject area (American Civil War) which includes the specific article topic, is irrelevant. --Phenylalanine (talk) 10:53, 15 September 2009 (UTC)

Actually, you seem to be agreeing with my point that using both types of sources would be best practice. --Bob K31416 (talk) 14:35, 15 September 2009 (UTC)

I think Pheny is saying that we should not distinguish between a book on the Civil War that includes a reasonably detailed discussion of Grant and (say) a biography that is specifically about Grant... both should be considered considered "reliable sources on the article topic" (Grant). However, a book on World War II that happens to mention Grant in passing (say to make a comparrison with Eisenhower), would not be a "reliable source on the article topic" (Grant). Blueboar (talk) 14:57, 15 September 2009 (UTC)

I tend to agree with Bob K31416's point. The revised rewording ("reliable published sources on the article topic"), implies that a book on the Civil War isn't as good as a book on Grant in an article on Grant. BTW, do we really need the word 'published'. That's covered elsewhere and isn't really relevent to the point that this is trying to convey. A Quest For Knowledge (talk) 16:33, 15 September 2009 (UTC)

Just to state my opinion for the record... Either version is acceptable in my book (I think they mean the exact same thing). I still think this paragraph should lead off the section... but we can discuss that later. Blueboar (talk) 16:55, 15 September 2009 (UTC)

My objection to this passage is not to the intent, but to the language, which makes little sense: writing that one should "research articles" implies that the person is doing research for the writing of said articles, not taking information from them. How about this:

"Best practice is to write Wikipedia articles by summarizing the claims from the most reliable published sources in your own words, with each claim attributable to a source that explicitly makes that claim."

(I'd revise this sentence further for fluency, but I think this illustrates the most important thing that needs changing.) -- llywrch (talk) 17:42, 15 September 2009 (UTC)

As an alternative to "on the article topic", we could use "in the relevent field". This would leave open using a book on the Civil War as a source for an article on Grant.

Best practice is to write Wikipedia articles by summarizing the claims from the most reliable sources in the relevent field in your own words, with each claim attributable to a source that explicitly makes that claim.

A Quest For Knowledge (talk) 18:13, 15 September 2009 (UTC)

Editprotected request for first part of proposal

{{editprotected}} I think we can implement the first half of the above proposal. It was previously discussed here and has attracted support from all editors who commented on it. The change involves adding the words "or imply" to the first sentence of the SYN section, so it would then read as follows: "Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources." (The second half of the proposal is still under discussion and should not be implemented at this time.) Thank you. JN466 17:03, 14 September 2009 (UTC)

I support this edit. Blueboar (talk) 18:06, 14 September 2009 (UTC)

I support it too. Crum375 (talk) 18:37, 14 September 2009 (UTC)

 Done - Jarry1250 ^{[ In the UK? Sign the petition! ]} 19:36, 14 September 2009 (UTC)

Thanks! --JN466 23:44, 14 September 2009 (UTC)

Editprotected request for second part of proposal

{{editprotected}} I believe we now have a version for the second half of the above proposal that enjoys broad support, per the above discussions. The change involves revising the last sentence in the SYN section, which currently reads as follows:

"Best practice is to write Wikipedia articles by taking material from different reliable sources on the topic and putting those claims on the page in your own words, with each claim attributable to a source that explicitly makes that claim."

to the following:

"Best practice is to write Wikipedia articles by researching the most reliable published sources on the topic and summarizing their claims in your own words, with each claim attributable to a source that explicitly makes that claim."

The proposal was first discussed at #Proposed_SYN_changes and enjoyed broad support then; it has only been tweaked since then ("sources on the topic" rather than "sources focused on the article topic"). This is probably still not the final and best version, but I think there is a general sense that this adds value over and above what we currently have. The good is not the enemy of the perfect. ;) Thanks. JN466 21:35, 15 September 2009 (UTC)

 Done The final clause is awkward, but it is in the original anyway.  Skomorokh  21:48, 15 September 2009 (UTC)

Interwikilink Bengali wiki

Please add this [[bn:উইকিপিডিয়া:কোন মৌলিক গবেষণা নয়]] instead of [[bn:উইকিপেডিয়া:কোন মৌলিক গবেষণা নয়]] because our wikipedia namespace changed.- Jayanta Nath ^{(Talk|Contrb)} 16:17, 14 September 2009 (UTC)

• Done. Crum375 (talk) 16:43, 14 September 2009 (UTC)

Protection

Hi,

I see that this page has been protected for a while. I hate seeing policy pages like this fully protected, because then things stagnate. We're a wiki! Sometimes there will be disagreements over exact wording on the policy page. That's okay! The idea is to compromise through edits and discussion on the talk page. Hopefully we can avoid the revert-based edit wars that happened leading up to the protection. It's important to discuss things on talk pages, but it's also important for people to boldly make edits and for others to edit those contributions *thoughtfully*, not simply with a revert or an undo. Most of the history of this policy is people making thoughtful contributions in an honest attempt to improve things. Even the bit of reverting recently was done in good faith, but hopefully we can move on. Especially considering that many of the involved editors have been around Wikipedia for a long time.

So, unless there is a good reason not to, I'll be unprotecting this page soon.

Thanks! kmccoy (talk) 22:31, 15 September 2009 (UTC)

Keep it as is for now. Unfortunately, what went into WP:NOR was being decided by who had the greater reverting power. The present situation adds considerably more order to the process. --Bob K31416 (talk) 23:42, 15 September 2009 (UTC)

No, the present situation makes this not a wiki. And it makes it so that the page is basically unchangeable. You'd like to keep the protection longer. How much longer? Indefinite protection is simply not a solution for anything but a few exceptions like Main Page. And if you'd be okay with unprotecting it in a week, or a month, or a year, what's the difference with doing it now? kmccoy (talk) 00:18, 16 September 2009 (UTC)

Re "And it makes it so that the page is basically unchangeable. " - Would you care to explain this remark, considering the recent changes that were made during protection? --Bob K31416 (talk) 00:24, 16 September 2009 (UTC)

Three changes have been made since the protection was established. One was a minor edit to an interwiki. So, yes, two very small changes have been made since protection was put in place. I think "basically unchangeable" is a fairly accurate description of that situation. It also prevents anyone from changing the page by being bold, except admins who are willing to ignore the protection policy, and sets admins as the ones who may determine consensus (since they're the only ones with the technical means to implement changes discussed on the talk page.) This is not wiki-like. The protection needs to be removed, and editors need to figure out how to edit the page without revert wars. kmccoy (talk) 00:34, 16 September 2009 (UTC)

Regarding your edit summary "protection is evil" - apparently you have a POV re protection in general and you shouldn't be deciding on what should be unprotected. --Bob K31416 (talk) 00:47, 16 September 2009 (UTC)

Yes. You hadn't figured out that I have a point of view regarding protection from the text I wrote previously? Of course I do. I believe that protecting pages so that they cannot be edited is contrary to the nature of a wiki and should be used sparingly and for limited durations. This is an indefinite full protection in reaction to a fairly minor revert war, which I believe is improper. (This belief also happens to coincide with our policy on protection, though I'm usually more of a common sense over policy sort of guy.) Why would having a point of view about protection mean that I shouldn't make such decisions? Clearly the person applying the protection had a point of view that indefinite protection is an appropriate response to the reverts that were happening here. Why doesn't that disqualify that admin from imposing the protection? kmccoy (talk) 02:39, 16 September 2009 (UTC)

The problem with bringing one's prejudice into a situation is that one may misjudge the situation by only considering one's prejudicial generalities without considering the specifics of the situation. This appears to be the case here with you. For example, you wrote, "I hate seeing policy pages like this fully protected, because then things stagnate." There hasn't been stagnation, as evidenced by the discussion on this talk page since the protection and the resulting edits on the policy page. What has subsided is determining changes on this policy page by reverting, and forcing changes or obstructing changes based on greater reverting power. An administrator who is judging whether or not to unprotect a policy page, should approach this task in an unbiased, unprejudicial manner. --Bob K31416 (talk) 11:35, 16 September 2009 (UTC)

Nah. I can look at the page history and see the results of protection. And I can look at the page's protection history. It's often been protected and then unprotected a few days later. My opposition to protection isn't "prejudice", with all the negative baggage that word carries, it's simply an affection for the idea of a wiki. I'm approaching this task unbiased in the way that counts, in that I really don't care about which version of the page from the string of reverts is the one that appears on the page itself. Having a bias against page protection isn't a problem, and won't stop me from removing the protection. After the protection is removed, there is nothing preventing people from continuing to discuss on the talk page. In fact, it's encouraged. But so is being bold. If necessary, editors who revert without comment and without joining the discussion can simply be blocked. That's a better solution than protection. kmccoy (talk) 11:44, 16 September 2009 (UTC)

You're apparently acting as an advocate of unprotection in general, not as an unprejudiced judge of this situation, and you seem to be boldly admitting it. Clearly, there's nothing I can do myself to stop you or to convince you not to remove the protection. But note that when protection was implemented, there weren't objections to that action like there is now to what you are planning to do. --Bob K31416 (talk) 12:22, 16 September 2009 (UTC)

Kmccoy, if you look at the discussions above, we have been discussing things and good making progress towards reaching a consensus (one of the most wiki-like concepts there is). This is exactly what protection is supposed to promote. Give it a bit more time... as for how much? As long as it takes to either reach consensus or determine that we can not do so. Blueboar (talk) 01:26, 16 September 2009 (UTC)

You're seeking a "consensus" on what? The final state of the article? The next edit? If you can't reach a consensus on whatever it is you're trying to reach it on, then you'd have the protection removed, even though you couldn't reach a consensus? kmccoy (talk) 02:39, 16 September 2009 (UTC)

Have you seen what happens when someone just tries to add something to a guideline without discussing it first? It gets immediately reverted, and then a discussion starts whether or not to add the info, and consensus is reached whether to add, change, or ignore the new addition. Having the page protected just stops edit warring and goes straight to discussion. I agree that most pages should not be protected, but a guideline is more complicated than just an article on Wikipedia. It may not be what you consider to be "wiki", but it makes things go more smoothly and forces people to discuss before fighting. Angryapathy (talk) 05:24, 16 September 2009 (UTC)

Then the editors need to figure out how to do edits without reverting. The inability of a few editors to avoid clicking rollback or undo should not impact the ability on everyone else to edit. If you'd like permanent protection as a rule for policy pages, then go around to the village pump and other noticeboards and gain support for that. But until then, permanent full protection is only for the main page and logo image and the like. And I have yet to see anyone here give a time period for this one. I saw some decent incremental edits before the reverts, and I've seen other policy pages edited without this problem, even without being fully protected. kmccoy (talk) 11:17, 16 September 2009 (UTC)

So you acknowledge that edits on this page can turn ugly, but offer no solutions except that your standard belief that, "Policy pages shouldn't be protected?" Honestly, has this page been permanently been changed less since the protection than it had before it was protected? Angryapathy (talk) 14:08, 16 September 2009 (UTC)

Folks... I suggest that we drop this... it is sidetracking us from what is important... ie continuing our attempts to reach a consensus on our disputed language... that is the quickest way to see that the page gets unprotected. Blueboar (talk) 14:36, 16 September 2009 (UTC)

Update

• Editprotected request for Alternate Proposal made 22:08, 16 September 2009 (UTC) [2]

• Protection removed from WP:NOR 00:03, 17 September 2009 (UTC) [3] [4]

• Editprotected request no longer needed for editing WP:NOR.

• According to Alternate proposal, WP:Synth changes made 02:15, 17 September 2009 (UTC) [5]

--Bob K31416 (talk) 03:10, 17 September 2009 (UTC)

Are there other issues

OK... I think we have consensus on at least some of our issues regarding the SYNT section (see above)... are their other issues we need to discuss? Blueboar (talk) 14:39, 16 September 2009 (UTC)

Let's try to resolve the Alternate proposal that was mentioned in the RFC that is currently at the top of this page. Editors expressed support for the Alternate proposal, there was no clear opposition, although one editor may have reservations, but the respondents were small in number. Blueboar, I hope you and others would help that situation by contributing your opinions up there. Thanks. --Bob K31416 (talk) 16:07, 16 September 2009 (UTC)

May I add that it is quite timely since it involves moving the small last paragraph of WP:SYNTH with the sentence recently discussed and modified, to a more prominent place. --Bob K31416 (talk) 16:37, 16 September 2009 (UTC)

I assume you are talking about the "best practice" paragraph... is so, I completely agree. At minimum, this paragraph lead off the section (similar to how I proposed above... but with the newly agreed upon language). Alternatively, perhaps it could be stated in its own section (suggest a header of: "Source based research" if we go this route). Blueboar (talk) 20:27, 16 September 2009 (UTC)

Thank you. I took the liberty of copying your message to the Alternative proposal section and indicating that it is a copy in that section. --Bob K31416 (talk) 21:57, 16 September 2009 (UTC)

Consensus? I don't know. It surely seems that for years up until now there hasn't been. Perhaps consensus has been reachedas of 16 September. I suggest the participants make a mental note to re-examine the idea of consensus in three months to see if, based on the dearth of discussions on the topic appearing here, consensus has been reached. Three months being just before Christmas let's add a month and make the (personal) re-evaluate date be January 16, 2010.

Just as Wikipedia values verifiability over truth (which does get to be over-expressed as a concept at times) I value consensus over reasonableness. So if there's consensus, even though I disagree, that will suffice. (With luck I may never even look here again. Well, luck plus self-discipline. Mostly the latter.)

Currently (and for a long time) the discussion has been wrongly focused. It is new ideas that are to be shunned, and that is a good policy and a reasonable one. New ideas can appear, Wikipedia just is not the place for them. Great. The entire discussion of synthesis appears to have arisen from real or imagined cases in which some editor took (or might take) established ideas from one or more sources and combined them to create a new idea - or at least a new approach to a way of striking a blow in some long-lasting dispute (that extends far beyond Wikipedia but rages on Wikipedia.) As far as I can recall all the examples that have been presented would also fail to appear in the Encyclopedia Britannica (which doesn't have such a strict no-new-idea) rule. They'd fail to appear because they simply are not encyclopedic, because the EB likewise isn't a place to strike partisan blows. The examples are, then, to me bad examples: they'd not pass muster even for an encyclopedia that does allow synthesis.

As appears above the original example of forbidding synthesis has resulted now in an absolute prohibition of logic. That's excessive. The synthesis example started as an attempt to illustrate one possible way that new ideas might be introduced and (correctly) pointed out that this is against the intended nature of Wikipedia. But it (of course) got Wiki-lawyered. Forbidding logic also forbids the use of logic to amplify the ideas implicit to a topic but perhaps not anywhere explicitly stated. EB is, after all, limited in size in a way that Wikipedia is not: terseness has to apply somewhere lest the work grow to the size of a full library (even doubling in size would be problematic for the printed version.) A major strength of Wikipedia can be that it is unlimited size enough that a fuller explanation on a significant topic can be made. Wikipedia, too, could welcome proper synthesis.

The test is whether the synthesis creates a new idea. I maintain some ideas are so inherent in a subject that logically combining two ideas to create a 3rd does not actually create a new idea, it only creates a new way of viewing the topic. That can be useful.

But look above: I favor consensus over reasonableness. So if true consensus is reached it is time to celebrate. If, however, months more pass and consensus still eludes (as shown here by continued discussion on the same topic) perhaps it is time to reexamine the matter from a fresh start. If months pass and the discussion has not ended that appears to me to be prima facie evidence that there is not consensus. I would guess it is still indicated, in the article on consensus, that never-ending discussion is evidence of the lack of consensus. It would be a major misdeed to alter the consensus article to remove that thought. Minasbeede (talk) 19:20, 19 September 2009 (UTC)

Wikipedia:Conceptualization

I've just come across the essay Wikipedia:Conceptualization, which appears to me to be a complete contradiction of WP:NOR. I was thinking that it should either be userfied or deleted, but I was looking for other input first. Jayjg ^(talk) 23:51, 22 September 2009 (UTC)

Syllogisms and synthesis: Is Socrates mortal?

After careful study of the policies and examples, I still can't really tell where the line is drawn. Does the policy on synthesis actually forbid pure syllogistic reasoning? Does it make a difference if the major and minor premises are both contained within the same source? It would seem from some comments that "Socrates is mortal" would in fact be regarded as inadmissable original research if it was not stated as a specific conclusion. If this is the case, what about the limit case of two-valued logic? If a source states that "X is true", is the statement "Not X is not true" an original research synthesis, or simply a restatement? Personally I would think common sense would prevail in favor of commonplace logic, but wp:common sense tells me that there is no common sense, only existing agreements. Ben Kidwell (talk) 08:26, 4 October 2009 (UTC)

Look at the article Socrates and you will see that it does not state that he was mortal. There are three good reasons for this:

• I guess nobody tried to add this 'information'. (Not sure about this one, though.)
• It is a totally unremarkable fact, and not noteworthy at all. (Noteworthiness is a concept distinct from notability. It's about what to mention or not in an existing article. We don't have written rules about it except to cover some special cases that tend to be contentious, see WP:UNDUE. If information is noteworthy it should be added and should not be removed once it is present, although it may of course be modified or moved to a different article. Information that is not noteworthy may still be present for reasons including style, presentation etc., but it's likely to disappear if the article is rewritten.)
• It's borderline original research. For most people it would be original research without any qualification. For Socrates the case is different because this is a frequently cited example. But if we relied on an author to back up the statement "Socrates was mortal", then the author might well protest with words such as these: "I never said that Socrates was mortal. I merely explained how this follows from two other statements by way of a syllogism. I am a serious philosopher; I don't make ridiculous statements about the mortality of people I have never even met. And as far as I am concerned Plato has made Socrates immortal."

The first item is our most efficient defence against nonsense: The overwhelming majority of it never even comes up. The second item is sufficient, and quite effective, in cases such as an editor who insists on adding the sentence "The date of his death is not yet known." to Barack Obama. But since our handling of noteworthiness is not codified anywhere, the last item is our only strong argument against an editor who, say, tries to add "He was a mortal man." to every article on a dead philosopher. Facts that are not noteworthy tend not to be said by reliable sources.

A second reason why a prohibition of syllogisms and other synthesis makes sense is this: Even a 'valid' syllogism may well be wrong. Often the same word is used in slightly different ways in the two premises. A single source may well say that Sea Shepherd Conservation Society is an "eco-terrorist" organisation; that "eco-terrorism" is a form of "terrorism"; and that "terrorism" is a "crime". Yet if the same source does not say that Paul Watson is a criminal, then we have no reason to do so. The author may well know something that we don't. The author may have asked a prosecutor why Paul Watson is still free, and may have learned that he is not doing anything illegal and there isn't even a warrant. (I didn't make this example up entirely; there is a similar conflict at the Sea Shepherd article.)

So much for why this rule makes sense. And here is why it doesn't really hurt: Most of the time, when you practise "improper synthesis" you will get away with it because nobody really minds or doubts your conclusion. In the case of a blatantly correct and uncontroversial application of a syllogism you will even get away with it after it is challenged by a confused editor. We generally apply our rules inconsistently, to keep most of the lunacy, advertising etc. out of the project while not obstructing the writing of an actual encyclopedia more than necessary. Hans Adler 11:53, 4 October 2009 (UTC)

Thanks very much for taking the time to provide a detailed response, that does help clarify my understanding of the policy. I do see some danger that an editor looking to engage in POV wikilawyering could use this policy in a detrimental way, because secondary-source authors sometimes make their claims by presenting evidence to support premises, and allow the reader to make the obvious deductions and conclusions. Often a piece of investigative journalism may be structured by carefully establishing a set of premises, then introducing a stream of evidence carefully chosen to support those premises, leaving no doubt that the intention of the author is to support the conclusion that can be drawn. I hope that editors won't use this to justify deliberately 'missing the point' of an author who is assuming the reader is completing deductions on their own, without the author stating them after the premises have been established. Ben Kidwell (talk) 15:23, 4 October 2009 (UTC)

Well, whether synthesis is OK or not really depends on a lot of factors and nuances. An author may leave an obvious conclusion unwritten because he thinks it's more effective if the readers draw it on their own. Or because he knows it is false and he doesn't want to make the lie explicit. Or because he believes it but is afraid of a libel case. (In the case that we actually discussed on the Sea Shepherd talk page, it was more interesting: An FBI official insinuated to the US Congress, 5 months after the attacks on NYC, that Sea Shepherd is eco-terrorist. I am sure if he had been asked: "Why should we give you money for hunting down eco-terrorists if you are not even trying to arrest Paul Watson?", then he would have answered something like "Well, our experts aren't so sure that Sea Shepherd is actually breaking the law, so there is nothing we can do at this point. Incidentally, this shows that apart from funding we also need more practical laws to cope with this threat.") There are no perfect policies or guidelines that make both smearing and whitewashing impossible. We need to get the balance right to minimise both. We can only do this by experimentation and observation. Hans Adler 16:26, 4 October 2009 (UTC)

I'd say a case where these objections do not apply is where the reasoning fills in a logical argument between sourced premises and sourced conclusions. Only the connecting argument is not sourced. Of course, the argument could be wrong, or may be overly restrictive (a more general basis for the connection might exist), but little harm is done. Brews ohare (talk) 21:49, 5 October 2009 (UTC)

RfC at WP:Civil

A Request for Comment has been posted at WP:Civil concerning reversion using the one-line Edit Summary. It is suggested that such summaries that employ WP:OR require a Talk page back-up that provides specific indication to the contributing author of just what it is that makes the reverting editor believe WP:OR is applicable. Please take a look and comment. Brews ohare (talk) 22:38, 5 October 2009 (UTC)

Use of WP:NOR in one-line Edit Summaries

The proposal at One-line Edit Summaries found few supporters, as many felt that any mandatory requirement upon the one-line Edit summary was onerous. However, a modification of WP:Civil was made suggesting that on-line edit summaries be explicit. I have imported a version of the text added to WP:Civil modified somewhat to apply to this guideline. Brews ohare (talk) 14:52, 7 October 2009 (UTC)

It may be civil to write explicit edit summaries, but I see no reason to state a requirement to do so in this policy. The place for this is WP:CIVIL ... this policy is about dealing with OR, not about civility. Blueboar (talk) 15:06, 7 October 2009 (UTC)

Besides... I think what was added is wrong. While it might be all warm and fuzzy to write detailed edit summaries, sometimes a blunt "This is OR" is both acceptable and appropriate. I don't think it is automatically uncivil. Blueboar (talk) 15:46, 7 October 2009 (UTC)

Question: Where does one go if one wants to insert original research and speculation into an encyclopedia or wiki article?

Just wondering. Surely baseless troofing and fringe physics must have a home somewhere! — Rickyrab | Talk 18:54, 11 October 2009 (UTC)

Nowhere in Wikipedia... but our sister project, Wikiversity, accepts original research. Blueboar (talk) 19:50, 11 October 2009 (UTC)

Wikiversity won't accept that kind of original research. May I suggest http://conspiracy.wikia.com/wiki/Main_Page? Abductive (reasoning) 20:09, 11 October 2009 (UTC)

You will have your answer if you click this link, but someone whom you do not know will... --Bob K31416 (talk) 03:43, 14 October 2009 (UTC)

Primary/Secondary sources in BLP

I think we need a slight clarification of the policy in relation to WP:BLP. In practice we tend, rightly, to use information published on eg a university website for factual information about a person, even though in practice this information is usually supplied by the person in question. The logic I think is that a university would be very unlikely to publish this information unless it were believed to be reliable, and if anyone spotted an incorrect claim they could rapidly email the part of the university responsible. If people are listed in Who's Who or similar publications then their biographies, although published by this 3rd party, are invariably sent to the people involved for checking and amendment. Again it is very unlikely that there will be false claims made, because if there were the editors would be informed and this would lead to embarassment or possible de-listing. Thus I think for the purposes of this policy applied to factual BLP bio-information from reputable publishers where there is 3rd party editorial control should be considered a secondary source, even if the information is derived from the biographee. What do people think? NBeale (talk) 14:44, 13 October 2009 (UTC)

PS by "factual" I mean eg "Books: Smith on Toads (OUP 2008)" as opposed to "wrote the definitive work, Smith on Toads (OUP 2008)" NBeale (talk) 14:47, 13 October 2009 (UTC)

The caution on self-published sources in BLPs is really talking about materials that are self-published by someone other than the subject himself. In other words, if Professor X, in a blog attached to his university website, says something about Professor Y... we should not use that self-published source in the article on Prof. Y (but we might be able use it in the article on Prof. X) Blueboar (talk) 16:36, 13 October 2009 (UTC)

Yes but I'm more concerned about the basic bio-details. When I write/edit BLP articles about an academic I routinely go to their institute's website - and so I think do most sensible people. But according to the letter of this policy I should not because it is a "primary source". NBeale (talk) 18:30, 13 October 2009 (UTC)

No, primary sources are fine for non-controversial fact verification. ^Bongo_matic 03:51, 14 October 2009 (UTC)

So could we say something like; "Note however that primary sources which are published by reputable third paties may generally be used for factual information in WP:BLP" What I have in mind is that if Bloggs says on his website "I have a degree from Berkeley" we should probably not use it, but if his bio-details at an institute at Harvard says he has a degree from Berkeley this is fine. NBeale (talk) 06:02, 14 October 2009 (UTC)

Rather than "reputable", I suggest "reliable" sources! SNALWIBMA ( talk - contribs ) 08:42, 14 October 2009 (UTC)

Perfect. Shall we go with: "Note however that primary sources which are published by reliable third paties may generally be used for factual information in WP:BLP"? NBeale (talk) 12:27, 14 October 2009 (UTC)

OK I've put this in - let's see what feedback we get if any. NBeale (talk) 11:42, 15 October 2009 (UTC)

(outdent) I have reverted this because it incorrectly identifies BLP sourcing as being less restrictive than other articles (where in fact it is generally the opposite), and also because primary information is allowed in all articles, with proper care and appropriate restrictions. Crum375 (talk) 14:37, 15 October 2009 (UTC)

Thanks Crum375. The point is that partly because of BLP policies we are in practice required to use primary sources for biographical information, especially since newspapers often get basic details wrong. When writing an article about an academic, for example, we typically rely heavily on the bio-details at their institute. According to the letter of NOR this is banned because we are supposed to use secondary sources. Would adding "since special care is needed to ensure their accuracy." meet your concern? NBeale (talk) 16:17, 15 October 2009 (UTC)

No, I don't think we need to mention BLP in this case, except perhaps to say that BLP sourcing requirements are even stricter than for other articles. I also take issue with your points above. It is incorrect to say that primary sources are "required" because of BLP policies. If anything, we encourage secondary sourcing. We also don't recognize "wrong" or "right": we simply report what the sources say, with appropriate weighting according to their quality (e.g. degree of vetting and/or reputation) and prevalence. It is also incorrect to say that we are "supposed" to use secondary sources: we are encouraged to use secondary sources wherever possible, but primary sources are allowed with extreme care and appropriate restrictions. For example, primary sources should not be the main source for an article. Crum375 (talk) 16:31, 15 October 2009 (UTC)

Yes but in BLP primary sources often are the main source for basic bio-details in an article. For example most articles about academics refer to their university websites for their biographical information. This is why we want to talk about "used for factual information" though perhaps it should be "non-contentious factual information" or even "non-contentious factual biographical information" if you prefer NBeale (talk) 18:03, 15 October 2009 (UTC)

Academic Wikipedia subjects should generally pass the professor notability test, which would typically include multiple high quality references to their notable work published in prestigious secondary sources. See the prof criteria for more details. It would seem to me that if the sole reference to an academic is some primary source, he/she would fail WP:PROF by definition. And again, we do allow primary sources for all articles, including BLP, when they are used carefully and are generally supporting other secondary sources. Crum375 (talk) 18:14, 15 October 2009 (UTC)

I agree that we cannot (generally) establish WP:N from primary sources, but having established WP:N we can, and should IMHO, use suitable primary sources to provide factual bio-details for the article. NBeale (talk) 21:05, 15 October 2009 (UTC)

That's quite true: after establishing notability, for example for a professor, by supplying multiple secondary sources which show their contributions to their area of study meet or exceed the requirements in WP:PROF, you can supply additional sources, some primary, to add details or corroborate the secondary sources. But the primary sources may not be the main basis for the article or for the person's notability. This is not special to BLP, since primary sources may be used, with care and restrictions, for any Wikipedia article and subject, generally to enhance information supplied by reliable secondary sources. Crum375 (talk) 21:14, 15 October 2009 (UTC)

Here's a hypothetical situation: Let's say a Wikipedia editor wants to have his bio on Wikipedia, but unfortunately for him it has been deleted one, two, three, four times in 2½ years. Now, his first strategy for getting the article back in Wikipedia is to simply park it in his userspace, and appeal on his blog for all and sundry to restore it. But, despite the fact that several 5 or 10 edit Wikipedians restore it to article space, it keeps getting deleted. You see, as it turns out, his biography relies almost entirely on primary sources, or biographies written by the subject himself and published in tertiary sources (Debrett's People of Today and the Faraday Institute) without any fact checking. So, what's a fellow to do? Well, he might publish some nasty slander on his blog, but the blog route didn't get him too far before. Or, he could try to change the WP:NOR policy to include the sentence Note however that primary sources which are published by reliable third parties may generally be used for factual information in WP:BLP. Because, of course, once that is part of the policy, he can then use it to insist that his Debrett's People of Today and the Faraday Institute bios are now reliable, and thus prove his notability and worthiness of a Wikipedia article. So, in this hypothetical situation, would it be a good idea to actually create a special exception to a fundamental content policy so that this individual can finally stop his bio from being deleted? Or would it be better, perhaps, to simply leave policy as it is? Jayjg ^(talk) 23:27, 15 October 2009 (UTC)

"nasty slander" - come off it! Remeber WP:NPA NBeale (talk) 08:29, 16 October 2009 (UTC)

Your hypothetical is too far-fetched for consideration. We should keep speculation down to a minimum ;0 ^Bongo_matic 23:49, 15 October 2009 (UTC)

Quite. The whole point is that, although primary sources cannot establish notability, they are essental for actually writing a useful BLP article. In practice there is no problem with these, but it allows a loophole for wikilawyering, which, if actually applied consistently, would result in deletion of thousands of perfectly good BLP articles. (And FWIW WP:AUTHOR#3 is self-evidently verifiable without sources like Debretts & Who's Who). The fact that User:Snalwibma collaborated on the wording of this amendment is good evidence that it is not partisan from my POV. :-)NBeale (talk) 08:29, 16 October 2009 (UTC)

Well-spotted Jayjg. Following a recent, and hypothetical, 4th deletion, I did wonder what such a hypothetical editor's next move might be. I'm most impressed by the hypothetical depths to which such a clearly imaginary editor would sink to disguise such deception. But Bongo is, of course, correct: such a craven, self-serving, shameless and COI-blind editor could not possibly exist. --PLUMBAGO 08:41, 16 October 2009 (UTC)

Jayjg has beaten me to this, but yes. NBeale is a user who's hell-bent on having his own article on Wikipedia, and thus his suggestions for a policy change like this cannot be taken in good faith. The fact that this discussion has gone on so long already is more than enough. rʨanaɢ ^talk/_contribs 09:03, 16 October 2009 (UTC)

Comment to NBeale - the fact that I "collaborated on the wording" does not in any way suggest that I agree with it! I merely suggested replacing "reputable" with "reliable", in the (vain, it turns out) hope that you would get the point. Your fondness for concepts like "reputable" is nicely in line with your tendency to base your arguments on the number of Nobel laureates, knights of the realm and assorted windbags who have made pronouncements or said nice things about the topic under discussion, such as a certain book or a certain person. They are all extremely reputable, but they are not necessarily reliable. See argument from authority for a discussion of the style of argument you tend to adopt, and which you also resorted to in the discussion above. SNALWIBMA ( talk - contribs ) 09:57, 16 October 2009 (UTC)

All this ad-hominem stuff and references to clearly wiki-political deletions (have we ever deleted anyone else who has written a WP:N book?) is perhaps mildly amusing, but doesn't address the problem that Wikipedia has a bad reputation with some key people (the Chair of that v well known media group was scathing) because we don't do a good job on BLP. I'm not suggesting that we use OR or primary sources to establish notability, but the policy here as written just doesn't fit with what professionals, or most Wikipedia BLP articles, actually do. Those who care about building a serious encyclopedia, rather than playing RPGs, might care to consider a re-wording. I'll steer clear to avoid irrelevant distractions. NBeale (talk) 21:14, 21 October 2009 (UTC)

I think the simple answer is that there is, and should be, no distinction between biographical articles and other articles as regards the general rules on the types of sources that are acceptable. The emphasis on secondary sources is entirely correct, because it acts as a kind of notability, reliability and verifiability filter. Any failings in Wikipedia's BLP articles have nothing to do with the balance between primary and secondary sources. To shift the emphasis towards primary sources would be a very bad idea, especially for biographical articles. It would open the door to all sorts of non-notable people about whom nothing is written in reliable secondary sources, but who are the subject of extensive information (much of it self-generated) in primary sources. Such people fail the notability test, and so should not have Wikipedia articles. The use of primary sources for basic facts is already accepted practice, and is perfectly well described at WP:SELFPUB; to shift the balance as you appear to be suggesting would weaken the encyclopaedia by letting in articles about non-notable people, based on their own unreliable self-published statements. SNALWIBMA ( talk - contribs ) 08:26, 22 October 2009 (UTC)

Why are articles not written by the foremost experts?

When reading articles on geology, it is often painfully clear to me that experts have not written the articles. Why is this? Prohibiting original research is a very broad brush, and - though absolutely necessary - doesn't distinguish well-reasoned explanation from the incorrect ones. After a few decades of research or teaching, experts likely can't remember whether the reasoning that makes a difficult subject simpler was created themselves or borrowed. It doesn't matter to lectures or journals, as long as the explanation isn't a significant part of the presentation. But, it matters here.

As an example, I am able to write 'any theory of heterogeneous buffering in thermodynamics will likely be based upon equations relating the derivatives of a characteristic function', but only if I can find it somewhere in the published literature. This is not likely, and no good teacher feels 'facts speak for themselves'. Many students pay exorbitant tuition to hear what the Wikipedia prohibits. However, I agree that the dangers of allowing 'well-reasoned explanation' are too great: it invites opinionated, poorly-reasoned ones.

Ironically, allowing statements from current texts, which I was told was preferred to 19th Century, primary literature, allows knowingly false statements to make it to the Wikipedia. Statements in texts are often inherited, slightly changed, from earlier generations of texts. The primary literature's statement proves either absent, very different from current statements, or was itself just an opinion.

Texts written (not from love but) for mass sale to U.S. state colleges are commonly referenced, but are known to be, essentially, collections of obsolete and even 'false' statements. Experts just wince & stay away. Contributors do not. Experts have read the original, primary literature and often can easily list the best articles on the topic (some are chapters in early texts or monographs).

The Wikipedia had the opportunity of inviting the best experts to write articles of interest to them. These have likely already been written, and the owning of articles is a problem in itself. (In my years of making corrections or suggestions in Talk pages, not one has been adopted.

How to attract experts and understand why they are not contributing is undoubtedly something the Wikipedia has addressed. This contribution is to suggest that prohibiting 'original research' has, to expert teachers, thrown out their babies with original explanation.

Better people than I may think of a solution. The only objective modifications I can suggest is (1) limiting the use of introductory texts to stubs, and (2) reminding authors that science is made of theories, theories we all hope will be false tomorrow. Dogmatic explanations are not scientific ones. This problem deserves better thought than I can offer. Geologist (talk) 11:00, 14 October 2009 (UTC)

Most of this is already covered under the tertiary source provision. They should be used sparingly for the reasons you state above, while current but notable journal articles (covered or mentioned in the secondary literature) should be used whenever possible to represent the contemporary body of knowledge. As you made clear above, textbooks are already outdated by the time they are first published. What you are really talking about is authoritativeness and currency, two things I have tried to address in the past with little success. I would support the addition of those two things to this policy or further emphasis placed on their importance. Viriditas (talk) 11:10, 14 October 2009 (UTC)

And if you write a [Questions of Truth|book]] on a subject, co-authored with one of the world's foremost experts, with a foreword from one Nobel Laurate and an endorsement from another, then if you ref it in your edits you will be attacked for WP:COI Ah well :-) NBeale (talk) 12:35, 14 October 2009 (UTC)

Geologist, One of your points seemed to be that you feel too limited in how you are allowed to explain something in Wikipedia. In regard to that, consider the following two quotes for guidance:

1. From WP:SYNTH,

Carefully summarizing or rephrasing source material without changing its meaning is not synthesis—it is good editing.

2. From [6]

It isn't appropriate for us to try to determine whether someone's novel theory of physics is valid, we aren't really equipped to do that. But what we can do is check whether or not it actually has been published in reputable journals or by reputable publishers.

My interpretation of these two quotes, in the context of this discussion, is that you are allowed to explain something from published sources in your own way. However, you are not allowed to inject your own ideas, or the unpublished ideas of others, that go beyond the published sources that you are trying to explain. --Bob K31416 (talk) 14:18, 14 October 2009 (UTC)

And just for fun, I did a quick look for some support of your statement, "any theory of heterogeneous buffering in thermodynamics will likely be based upon equations relating the derivatives of a characteristic function". I found this abstract. Possibly this article might be the source that you need since the abstract mentions, "A necessary part of the definition of a buffer..."? --Bob K31416 (talk) 18:26, 14 October 2009 (UTC)

Illness has caused my absence; but I wish to thank each of you for your excellent suggestions (and nice reference). Distinguishing a new presentation from new content is difficult. My apology for not yet reading the fundamental papers of the Wikpedia. There is a problem with referencing both books & journals in that many books being referenced are not of high quality (and experts are not even aware of them); and some experts read only a paper's data (geologists go for the map), ignoring the author's personal interpretation of it. Authors here cite only the personal interpretation.

May I offer here a quick observation of two problems the principal editors may wish to address; problems that do not appear in highly edited encyclopedias. The first is that of audience, and the second is that of presenting science dogmatically.

My old 'Encyclopedia Britannica' wrote to two audiences: the beginning of the article summarized the following content for everyone; and the following content was often for specialists. It was usually a history, pointing them to classical papers in the literature. This made an encyclopedia the place to look for an introduction to a subject and a list of highly regarded books to continue one's study. A specialist need not learn his discipline from an encyclopedia.

Examples

Fundamental to all sciences is the the 'equivalence class', a collection of objects having in common a list of properties. This should, consequently, be one of the Wikipedia's best articles, simply written for varying audiences. Equivalence class

Second, science changes regularly, and empirical objects or phenomena should not be presented as theoretical objects, when they are not. It is common in good secondard references to present the object & question, then a history of explanations, ending with the current one. Every primary reference ends with a personal opinion of the article's importance. Scientific articles should name the theory used when offering an explanation. Older 'Encyclopedia Britannica' articles usually offered a history of theories. Being written by experts, they never offered definitive explanations. Volcanic Arc

What I could do, from bed, is possibly start a history (for some subjects) of the better articles & books (possibly annotated) that the reader would want to consult for details. All pretty much agree which these are, and they could offer the reader expert presentations on the different theories. I shall, of course, read the fundamental articles here before contributing anything (but these criticisms :-) Thank you both for the clarifications! Geologist (talk) 23:43, 23 October 2009 (UTC)

Creating only English source for information

At WT:N I have argued that if a topic is not sufficiently notable to have a reliable source in English, then it is probably not sufficiently notable to have an article in the English Wikipedia, but most others disagree. But that discussion has lead me to this argument: If a given topic has no reliable sources in English, then creating an article in the English Wikipedia on that topic means creating the only, the original source in English for that topic. That seems to me to be a violation of NOR, in spirit if not in letter. I mean, isn't doing research about a topic in non-English, and not using any sources that are in English, and then writing about it in English, original research by definition? Comments/ thoughts? --Born2cycle (talk) 15:01, 28 October 2009 (UTC)

Born2cycle... you really dislike non-English sources, don't you?... you have floated this at WT:RS, at WT:N and now here... forum shopping? To answer your question: no, it is not original research to read a reliable source and summarize what it says... no matter what language the source is in. Blueboar (talk) 16:13, 28 October 2009 (UTC)

As for your first comment: notability is notability worldwide. Wikipedia strives to be a global encyclopedia that happens to have articles in different languages but ultimately has the same standards of information (we are light-years away from reaching that, but anyway). Notability standards are unevently implemented across the different projects, but in theory something that's worthy of inclusion on one Wikipedia should be worthy of inclusion on all of them.

As for your second, that's just a fundamental misunderstanding of what Wikipedia writing is. Creating Wikipedia articles is not the same as doing original research; likewise, translating information from one language to another is not original research. Original research involves trying to make or prove new ideas that have not been published somewhere. I think you need to reread the policy, because it seems that you're interpreting it as something different than what it actually says. (You wouldn't be the first; I can't tell you how many people I've seen get peeved over other editors citing POINT in the wrong way.) rʨanaɢ ^talk/_contribs 00:04, 29 October 2009 (UTC)

Born2cycle, you have raised this at WT:V (edit: this thread was initiated by someone else), WT:N and WT:RS and now you are trying to open a fourth debate here? Do you think the overwhelmingly negative response to your suggestion will be different here? Please stop, opening up four discussions on the same topic, hoping that one of them will give you the result you desire is getting disruptive becuase debating the same thing over again in different places is a waste of time. Sjakkalle (Check!) 11:40, 29 October 2009 (UTC)

It might be helpful and informative if someone gave the excerpt(s) here from Wikipedia policy that discusses the use of foreign language articles. Thanks. --Bob K31416 (talk) 13:40, 29 October 2009 (UTC)

Start with WP:Verifiability#Non-English sources. Blueboar (talk) 15:03, 29 October 2009 (UTC)

From the link that Blueboar kindly provided, it appears that it isn't an NOR issue. However, the situation of having an english Wikipedia article dependent entirely on foreign language sources, would severely limit the editors to mainly those who understand the foreign language. This is a situation ripe for violations of WP:NPOV, since those who aren't fluent in the language will be at a disadvantage and outnumbered. For example, consider the Israeli-Palestinian conflict and a Wikipedia article whose only sources are in the Arabic language, or one whose only sources are in the Hebrew language. However, I can't think of an actual example, so I'm just putting forth this idea to see if anyone has an interesting thought regarding it. --Bob K31416 (talk) 20:28, 29 October 2009 (UTC)

First, Blueboard, the issue I've raise here is different from those other places. Related? Sure. Same? No way. The question at issue here, and relevant to this page, is whether creating the first and only source on a given topic in English is original research. This is a very different question and issue then what is being discussed on those other pages, and the answers here are very different.

Sjakkalle's answer is particularly helpful because it essentially points out that since WP article writing in general is not OR, then writing the first one that happens to be in English is not either. I think that is a very good point, and I accept it.

As to Bob's concern, that belongs in the dicussion at WT:V since it is about verifiability. --Born2cycle (talk) 00:25, 30 October 2009 (UTC)

Interviews as Primary Sources

I disagree with the assertion interviews are exclusively primary sources. Most interviews are conducted by a third-party (not directly involved with the subject) and often include an introduction. If that introduction were taken out of the interview and put by itself it couldn't it be considered a secondary source? Furthermore, interviews include comments as well as questions. Couldn't those comments should be considered secondary as well? -Stillwaterising (talk) 01:47, 5 November 2009 (UTC)

I think you are confusing an "interview" with an "article about an interview". The dialog between the interviewer(s) and the interviewee(s) is a primary source (AKA the actual interview), while the article about the interview (including the comments and introduction) would be considered a normal secondary news source. Angryapathy (talk) 14:31, 5 November 2009 (UTC)

What I'm talking about is the article that's placed before the interview. It is often written in advance of the interview by the interviewer about the interviewee. It is typically not a summary of the interview (an "article about an interview") but a background reference (article about the subject). -Stillwaterising (talk) 17:39, 5 November 2009 (UTC)

I take it that interviews are mentioned in order to give an example of something that's typically a primary source, not to declare a blanket rule. For instance, an interview with a soldier about his experiences in battle would be a primary source, but an interview with a historian about the same events would not be a primary source. The introduction from the interviewer could be a primary or secondary source, and could be a more or less reliable source, depending on the context and the identity of the interviewer. In most interviews, the responses and the introductory material are both drawn from information supplied by the subject about himself or herself. When that's not the case, you can exercise good judgement about how to use the source. --Amble (talk) 18:29, 5 November 2009 (UTC)

After reading, WP:PSTS, the wording is, "video or transcripts of surveillance, public hearings, trials, or interviews". So again, the policy states that the primary source is the dialog between the interviewer and the interviewee; the article which contains the interview might have portions that would be secondary sources, as most news articles are. Does that answer your question? Angryapathy (talk) 19:06, 5 November 2009 (UTC)

I would like to propose a change to the guideline: "Interviews can be considered a mixture of primary and secondary sources. Statements made by the person(s) being interviewed should be considered primary sources. Statements and comments from the interviewer may be considered either primary or secondary sources, depending on context." -Stillwaterising (talk) 23:39, 5 November 2009 (UTC)

I'd prefer to remove it entirely. It really shouldn't have been added in the first place. SlimVirgin ^{talk|contribs} 00:12, 6 November 2009 (UTC)

Which parts and how much do you propose to remove? --Amble (talk) 02:33, 6 November 2009 (UTC)

I'd like to remove interviews as examples of primary sources, because they sometimes are and sometimes aren't. It introduces unnecessary complexity to have them in the list. SlimVirgin ^{talk|contribs} 04:45, 6 November 2009 (UTC)

Interviews really deserve their own section, however I support removing it from the list (and footnotes) until a consensus can be reached as to what that section should say. -Stillwaterising (talk) 14:36, 9 November 2009 (UTC)

I think interviews are under primary sources because they can easily be abused. There is no secondary interpretation of the dialogue between interviewer and interviewee, and someone could easily insert a POV by selecting which quotes to include, and what parts to summarize from the interview. In addition, a new user could summarize a fact from the interview that is otherwise ignored by secondary sources, and give undue weight to what they think is important. So while an interview may straddle the line between primary and secondary, their inclusion in this section just lets users know that they should take the same care with summarizing interviews as they do with general primary sources. Angryapathy (talk) 14:57, 9 November 2009 (UTC)

The transcript (or audio/visual recording) of an interview is clearly a primary source... a news report discussing and commenting upon the interview is a secondary source. Both can be used appropriately... and both can be abused. Blueboar (talk) 15:05, 9 November 2009 (UTC)

Not necessarily. An interview with a world renowned historian who describes the content of his new book about WWII, is a secondary source, no less than the book itself or an article written by that historian. OTOH an interview with a witness to a crime is a primary source. So it depends on the situation and can clearly be both. Therefore, an interview is not good example as either secondary or primary source. Crum375 (talk) 15:15, 9 November 2009 (UTC)

Well... I have to agree that a lot really depends on what we are talking about... The raw transcript of an interview with a historian would be a primary source for a statement of opinion such as: "The loss of his crown in the Wash is considered by some historians to be an indication of King John's overall failure as a monarch". On the other hand it would be a secondary source for a statement of fact such as "King John lost his crown in the Wash". Blueboar (talk) 15:31, 9 November 2009 (UTC)

I don't see a primary/secondary difference between the transcript of a historian being interviewed, and his articles or books. There is the issue of the publisher vetting, perhaps, but that would impact the reliability, not the primary vs. secondary distinction. Many people don't realize that that distinction has to do with the distance from the event to the observer. In the case of the historian, he is clearly distant, and the format difference between an interview and a book or article does not change that distance. Crum375 (talk) 16:20, 9 November 2009 (UTC)

And to clarify, from WP's perspective, the hypothetical historian's books and articles would be much preferable to his interview transcripts, because the former would likely be better vetted and therefore more reliable. Crum375 (talk) 16:56, 9 November 2009 (UTC)

And let us not forget that, as far as NOR goes... it does not really matter whether an interview is considered Primary or Secondary... Yes, you should be extra careful with a primary source, but they can be used appropriately. No matter what you call it, as long as you stick directly to what the source says, and do not add your own commentary, conclusions or analysis based upon the source, you are fine. Blueboar (talk) 18:14, 9 November 2009 (UTC)

Yes, assuming it meets other criteria, such as WP:UNDUE and WP:SOURCES. Crum375 (talk) 18:21, 9 November 2009 (UTC)

Quote added to lead

I just added a quote to the lead and am wondering if it is formatted properly. I wanted to give credit to the author and used a diff as the reference. Would it be more proper for the author to make the edit so that there would be no quotations marks and the edit history would be the only credit given, or is using a diff to give credit okay? The reason I added the quote is because I don't see this very important aspect covered in the article. -- Brangifer (talk) 02:14, 8 November 2009 (UTC)

Anyone may edit Wikipedia, but we don't quote each other on policy pages, except in very unusual circumstances. Crum375 (talk) 02:25, 8 November 2009 (UTC)

That's why I'm asking. I am in doubt, but since policy isn't based on V & RS (unlike articles), but on reasoning based in other policies and designed to protect Wikipedia and enhance its quality, just using the words should be enough. I'll remove the quotation marks and ref, and leave a note below giving credit to the source. It's a very insightful statement and deserves to be used to enhance this policy. -- Brangifer (talk) 02:44, 8 November 2009 (UTC)

Now that I've read your edit summary for deleting the quote, I'll contact the author and seek their advice. Here is the quote with ref:

• "Wikipedia is behind the ball - that is we don't lead, we follow - let reliable sources make the novel connections and statements and find NPOV ways of presenting them if needed."[7]

Brangifer (talk) 02:48, 8 November 2009 (UTC)

I'm not sure if it's even needed but you're certainly welcome to use it. I'm sure it's been said in various ways far better than this by others prior. Also as it's Wikipedia I'm not even sure it has to be attributed. -- Banjeboi 02:48, 9 November 2009 (UTC)

Thanks for the permission. Your statement sums it up so well that it deserves to be used. I'll give it another try, maybe in a tweaked version, although it would be hard to improve on it. If someone else sees it, they may be able to do so, which would be fine. -- Brangifer (talk) 04:22, 9 November 2009 (UTC)

It uses an unfamiliar expression, and doesn't notably improve on the meaning of the lede. Policies should be edited with an over-abundance of caution and consensus, particularly their ledes. Every word should be examined, and then pared to the minimum. The comment was interesting and reasonably accurate, but not policy material. Jayjg ^(talk) 04:31, 9 November 2009 (UTC)

You are welcome to improve on it. It covers an aspect that is left unsaid and only implied. Policies should be worded clearly enough for newbies to understand. That's why the timeline needs to be clearly stated. Right now it's only implied. -- Brangifer (talk) 04:40, 9 November 2009 (UTC)

Logical deductions

I see the previous discussion on this has just been archived, prematurely in my view since this is still an ongoing topic as evidenced by the project page edit history. In order for this change to be accepted a much wider consensus is warranted because this does represent a significant change to a core policy and it is not without issues. Citing WP:NOTOR and presenting it as a simple cross referencing exercise between different policies is not valid: WP:NOTOR is not policy or even guideline and has no formal recognition. The original proponent of this change (Brews Ohare) was censured recently by ArbComm for applying these kind of "simple" logical deductions to controversial or simply plain wrong effect, and then arguing that these changes are above reversion since they are somehow reliably sourced: I suspect part of the motivation for this change is simply sour grapes.

The problem is that this kind of process can be used to mask problem areas or elide over huge areas of controversy. To take the very example cited one areas and sub-areas, consider Stanley. Stanley is in the Falkland Islands. I daresay we could find an official Argentine source that describes the Falklands as Argentine, and we will presume for the purposes of illustration such a cite has been found. Therefore we can infer that Stanley is in Argentina. This is clearly not consistent with NPOV.

For a more clear demonstration of a logical fallacy this allows, consider the number zero. [8] states "the positive integers are the same as the natural numbers". [9] states the natural numbers are "the set {0,1,2,…}". In combination this shows that natural numbers are positive and that zero is a natural number. Therefore, zero must be a positive number. This is obviously nonsense.

In each case the logical step is perfectly valid but leaves out a central element. If that central element is problematic then the deduction is invalid. These are simple examples and so easy to see through - it is easy to see that the problem in the first case is the sovereignty of the Falklands which is disputed. In the second case it is the lack of consensus as to what constitutes a natural number. Spotting problems such as these in more obscure topics may not be so easy. That requires genuine expertise in the subject area - we need someone considered reliable to make the connection to establish that it is a valid one. This is why we have NOR in the first place. CrispMuncher (talk) 15:37, 29 October 2009 (UTC)

The previous discussion is now archived here. Hans Adler 15:00, 30 October 2009 (UTC)

I think we should mention that uncontroverial logical deductions are allowed. The complexity is not that relevant. What matters is that such logical deductions are accepted to be valid in the relevant topic area. Example: Suppose the wiki article on some topic related to quantum field theory is based on sources such as the book by Itzykson & Zuber. In the wiki article, it is almost inevitable that you would have to present the materal differently than the way it is presented in the book. But that requires you have the necessary understanding of the basics that Itzykson & Zuber expected the readers of their book to have.

So, the mathematical manipulations to get from what is written in the source to what is edited in the wiki-article may be highly non-trivial and cannot be considered a simple logical deduction like unit conversions. However, it is not controversial to people who are qualified enough to understand the book. Also such manipulations are necessary to be able to write a good wiki article on such techincal subjects. Count Iblis (talk) 17:46, 29 October 2009 (UTC)

I think that some good points have been made and here's a modified version, based on those points, for your consideration. I essentially borrowed a phrase from the section Routine calculations and added it to the original version.

Simple logical deductions

---

This policy does not forbid simple logical deductions, provided editors agree that the deductions correctly reflect the information published by the sources from which they are derived. For example, if A is in district B, and district B is in province C, then A is in province C. This is a simple syllogism. Included are all of the other simple deductions. More complex logical deductions should not be used unless cited to a reliable source.

--Bob K31416 (talk) 18:50, 29 October 2009 (UTC)

You realize how many editors have no idea what a syllogism is? Durova³⁴⁸ 20:46, 29 October 2009 (UTC)

27?   : )   There is a link for it, but perhaps the sentence with "syllogism" and the one following it can be deleted? --Bob K31416 (talk) 21:08, 29 October 2009 (UTC)

On average, 90% of people don't click on links. Having seen a lot of POV-pushing disputes, the potential for misuse is quite worrisome. Durova³⁴⁸ 21:34, 29 October 2009 (UTC)

I am inclined to agree with Bob K31416 - that is what the link is for. Or (shock, horror) the reader may consult a dictionary. CrispMuncher (talk) 22:06, 29 October 2009 (UTC)

To return to the proposal it is certainly heading more in the right direction, and it is broadly similar to the status quo (in reality if not on paper, since generally people do not reach for policy on matters they agree with). I wonder if it may be made more robust though. I suspect there are two distinct classes of case where this may be used - one is valid and the other questionable.

The first would be deductions that it can reasonably be expected that the author had in mind when writing a source, or if you were to ask him (at the time of writing) he would immediately agree agree with. This covers unit conversions (when units are quoted it is usually a quantity of interest rather than the system of measurement. It also covers an omission based on, for example, an assumed common knowledge base between author and target audience. So for example, if source A states "The BC547 is a transistor" and source B states "a transistor is a semiconductor" the combination should be uncontroversial, since if we asked the author of A whether the BC547 is a semiconductor he would in all likelihood respond "Yes, of course".

The problem is with the second class, where two sources are combined in a creative manner to produce a result that neither author would have anticipated at the time of writing. This is where we do stray into OR territory. The ultimate example of this would be Einstein's theories of relativity, which were based on nothing that was not already known but the known facts were combined in a new way. That is an extreme example of course, and no one would question the original nature, but it shows that the logical deduction argument can be applied to create what are fundamentally new facts and original research.

How you would codify this and distinguish between the cases in a simple and robust rule escapes me at the moment but I'm sure something will come to mind in time.

As a side note and point of order, Brews Ohare (the original proponent of this change) is currently on a temporary block. I don't see that we need to put the discussion on hold for him but it seems only proper that we allow him right of reply before any assertion of consensus is made. CrispMuncher (talk) 22:06, 29 October 2009 (UTC)

Hmmm... There does not seem to be a currently active block [10], though one appears to have recently expired. There is an active topic ban. Wtmitchell (talk) (earlier Boracay Bill) 23:57, 29 October 2009 (UTC)

CrispMuncher, How do you feel about the above version? Would you support or oppose it? --Bob K31416 (talk) 01:35, 30 October 2009 (UTC)

Some of this discussion seems to be about the adequacy of the premises. There is no argument that the premises have to be sourced, and may be challenged by any editor. Given the premises, however, it is very difficult to imagine a situation where the use of simple syllogism can be argued about, as any automaton could use the premises to reach the conclusion. I believe Russell and Whitehead showed this to be the case. Brews ohare (talk) 14:29, 30 October 2009 (UTC)

I agree, Brews, and it's why I oppose adding this. Where the syllogism is clear and simple, common sense alone will tell you there's no problem with the material. Anything more complex may or may not be okay, but the text being proposed says that anything more complex is not okay. I think adding a paragraph about this opens a can of worms unnecessarily. SlimVirgin ^{talk|contribs} 14:40, 30 October 2009 (UTC)

Unfortunately, in my personal experience, denial of simple syllogism has been used multiple occasions as a pretext to require verbatim statement of a result obtainable by simple syllogism using sourced premises, and I'd guess from the discussion above opposing this simple addition, there are many here that would do the same. Without such a statement, obstructive editors can exclude material they just don't wish to see expressed by requiring verbatim statements, or lazy editors simply can avoid thinking about what is wrong. I'd go further and support Count Iblis. Brews ohare (talk) 14:55, 30 October 2009 (UTC)

If you don't find consensus behind a particular deduction, that's a good sign that it needs more direct sourcing. We should reject any proposed policy change that puts the burden on the editor contesting a particular deduction. Although you believe your deduction was justified, it is possible that the rule against original research worked correctly in this case. --Amble (talk) 22:50, 30 October 2009 (UTC)

Amble, It appears that you support that aspect of the proposal since it contains the phrase "...provided editors agree...", so that it does not put the burden on the editor contesting a particular deduction. --Bob K31416 (talk) 22:02, 31 October 2009 (UTC)

I think your proposal is a reasonable way to handle uncontroversial cases, but I'm not convinced that it needs to be codified in policy. The beauty of "no original research" is that it cuts through what would otherwise be interminable (and usually unproductive) discussions with a simple and fair rule. Any time an exception is added, it gives editors the opportunity to argue that the rule doesn't apply to their specific case, and makes it a little bit harder to bring a discussion to a conclusion. There are some comments here that reinforce my suspicion that any exception will be used as another way to prolong arguments even against objections or consensus of other editors. --Amble (talk) 06:38, 3 November 2009 (UTC)

Examples requested

Before we modify a core content policy (with all the risk of unintended side effects, loopholes, and consequences, especially with respect to future wikilawyering and legalistic reading of policy), could the proponents – or anyone else – supply some examples of cases where uncontroversial logical deductions or syllogisms have been removed from articles? Wikipedia has managed for many years without this particular proviso written into WP:NOR, I see two possible explanations. First case — this has been an ongoing problem, and good, encyclopedic content has been removed on the basis of too-narrow reading of WP:NOR. If that is the case, there ought to be ready examples of this damage to the encyclopedia, and a modification to this policy is called for. Second case — there isn't evidence of a problem here. Common sense already generally prevails, and the proposed change would be an attempt to outlawyer potential future wikilawyers with more wikilawyering. Frankly, that never works, and it encourages the idea that editors must be slaves to the specific, codified, yet mutable wording of {policy} pages.

So, what specific instances of reasonable deductions and uncontroversial, logical syllogisms have been removed from Wikipedia articles based on a misreading of the existing policy? In those instances, has the beneficial material eventually been restored under the guidance of experienced editors? Briefly, can someone show me examples of the problem we're trying to solve, or are we fixing something that isn't broken? TenOfAllTrades(talk) 15:00, 30 October 2009 (UTC)

I would be happy to comply, but providing such examples would violate my topic ban. BTW, it is not removal alone that is an issue, but wikilawyering using the present form of WP:OR to prevent introduction of simple deductions from sourced premises. That wikilawyering, already evident, would be reduced by this insertion. Brews ohare (talk) 15:07, 30 October 2009 (UTC)

If the only examples you have to draw on are from an area where you are now topic banned for your disruptive approach to editing, they may not be sound examples. In any event, meeting wikilawyering with more wikilawyering is usually profoundly unproductive. TenOfAllTrades(talk) 15:20, 30 October 2009 (UTC)

As it is the equations here and here are inconsistent. Explanation here of why. Sourcing not available; requires logic to fix.--Michael C. Price ^talk 15:10, 30 October 2009 (UTC)

In the discussion that you linked, it appears that the editors commenting on the issue agree with you that your change to the article (to update the formula and explain that the L-L pseudotensor was originally formulated assuming a zero cosmological constant) would be reasonable and permissible under existing policy and practice. I don't see any sign of edit warring at the articles in question. I'm not seeing the harm to the encyclopedia here. TenOfAllTrades(talk) 15:20, 30 October 2009 (UTC)

Can only be updated by invoking WP:IAR, which isn't satisfactory since later someone will remove it. As has happened here which I've had to reinsert. --Michael C. Price ^talk 15:25, 30 October 2009 (UTC)

Oh God. Is all of this policy proposal just fallout from the speed of light arbitration case? As an uninvolved outsider, I give up. TenOfAllTrades(talk) 15:30, 30 October 2009 (UTC)

I am not familiar with Dicklyon. But this particular edit would be consistent with him worrying about various historical definitions of the mile, yard etc., one of which I believe is used for geodesic purposes in some US states even today. Historically the relation between the yard and the meter was not exact; then it was exact but not the same everywhere. Only in the middle of the 20th century it was standardised to a new value, but with some exceptions. Therefore this incident doesn't convince me that consensual IAR will lead to problems in your example. Hans Adler 15:39, 30 October 2009 (UTC)

The need for consensual IAR indicates that the rules should be modified. Our goal should be to eliminate IAR here, just as we do elsewhere in the real world. --Michael C. Price ^talk 15:47, 30 October 2009 (UTC)

There is a trade-off: Every new rule and every new exception to a rule has a cost because it needs to be discussed, often repeatedly throughout its existence. And it also has a cost in that it increases the overall complexity of our rule set and therefore contributes to making the other rules less effective. The problem as I see it is to find out whether a new rule/exception is a net positive when taking the costs into account. As a general principle we are not optimising our rules to fix every loophole. Hans Adler 16:20, 30 October 2009 (UTC)

(ec) If there are really no sources anywhere in this case that's a stupid problem with the real world that needs fixing. And I would think that one or more members of WP:PHYSICS should be able to fix it by putting a short preprint on their homepage. A more pragmatic solution is that in the subculture of WP:PHYSICS and WP:WPMATH we simply ignore the prohibition of logic thinking in such cases. Thus we avoid policy creep and strengthen awareness that common (con)sense trumps policies. Hans Adler 15:29, 30 October 2009 (UTC)

There are no sources for the two examples precisely because they aren't real-world problems that require fixing. But they need fixing here. And you're right, we get around it (most of the time) pragmatically, by simply ignoring the rules. Is that really the best solution, though?--Michael C. Price ^talk 15:44, 30 October 2009 (UTC)

Could someone give me an example, please, of a simple syllogism not being allowed in an article on the grounds that it's OR? SlimVirgin ^{talk|contribs} 15:49, 30 October 2009 (UTC)

I will paraphrase an example to avoid conflict with my topic ban that includes all physics. Source A said a system called "abra" was characterized uniquely by a certain set of parameters. A source B said a system called "cadabra" was characterized uniquely by the same parameters. I said "abra" was a synonym for "cadabra" The resisting editor said I had no source that said "abra" was "cadabra" and therefore this statement was [[WP:OR} and [[WP:SYN]. Obviously, the resiting editor understood the matter perfectly. However, their position was that it violated guidelines and could not be included without a verbatim quote from a reliable source. Brews ohare (talk) 16:08, 30 October 2009 (UTC)

Thank you. If there are two technical terms for the same thing, and it's common knowledge in the field that the two terms denote the same thing, there's bound to be a source somewhere that says "Abra and cadabra both refer to X." If there isn't, it suggests the issue might not be so straightforward. SlimVirgin ^{talk|contribs} 16:16, 30 October 2009 (UTC)

Brews, your example does not sound like a simple syllogism to me. For a logical deduction to be allowed under the "simple syllogism exception", it has to be bluntly obvious to both an expert and the average non-expert reader. Things like noting that if a town is in Belgium, it is also in Europe. Blueboar (talk) 16:25, 30 October 2009 (UTC)

Would noting that the town is in Europe be a violation of WP:NOR? --Bob K31416 (talk) 16:38, 30 October 2009 (UTC)

Allow me to request a bit of imagination here to envision a reasonable example of this sort, and not request very specific detail that I am not allowed to go into. Suffice it to say, in the case at hand, there was absolutely no way to misconstrue the identity of "abra" and "cadabra", and yet objection was raised because the terms appeared in different sources. Eventually I found a source that used both terms synonymously and I made a verbatim quote. However, I had to look high and low because most sources used one term or the other, and not both. Brews ohare (talk) 16:50, 30 October 2009 (UTC) The point here is not whether I am delusional or lack sophistication of mind, but whether such obstruction or confusion (whatever you want to call it) should be countenanced. Brews ohare (talk) 16:57, 30 October 2009 (UTC)

How about common sense: if other users cannot follow the syllogism, then it isn't simple. Angryapathy (talk) 17:41, 30 October 2009 (UTC)

Angryapathy: The point is that editors do follow the syllogism, but think that WP:OR denies its use, or enables its rejection. Brews ohare (talk) 18:03, 30 October 2009 (UTC)

Again, proving my point. If somebody challenges the syllogism, then find a ref for it. If it isn't simple enough for everyone to agree on, then it isn't a simple syllogism. Angryapathy (talk) 19:37, 30 October 2009 (UTC)

Seems like the proposal satisfies the points of each of you. --Bob K31416 (talk) 20:15, 30 October 2009 (UTC)

Modifications of proposal

How about removing the simple vs complex modifier? The emphasis is really on the editors agreeing. If everyone agrees on a change, doesn't matter how complex it is, does it? That would give us a much clearer proposal:

Logical deductions

---

This policy does not forbid logical deductions, provided editors agree that the deductions correctly reflect the information published by the sources from which they are derived. For example, if A is in district B, and district B is in province C, then A is in province C.

Better to argue about whether the syllogism is objectively a syllogism, rather than whether it is subjectively simple.--Michael C. Price ^talk 20:49, 30 October 2009 (UTC)

I'm with Michael on this. We can't definatively define what a "simple" syllogism is, and it basically comes down to consensus whether or not the information is logically deduced. Angryapathy (talk) 21:37, 30 October 2009 (UTC)

That is pretty much the existing situation without a provision like the one being proposed. I'm hesitant as to whether a new rule is needed but if one is to be added it needs a firm grounding - OR is still OR even if it widely accepted, and if your want to publish that Wikipedia is not the place to do it. I suggested something to this effect last night and went away and tried to codify it. I eventually came up with:

Logical deductions

---

This policy does not forbid simple logical deductions, for example, if A is in district B, and district B is in province C, then A is in province C. Editors must agree that the deduction is likely to reflect the intention and understanding of a source's author at the time it was written. Deductions are not acceptable where it is unclear that the deduction would be obvious to or even accepted by the source(s) on which they are based.

In the event of a lack of consensus over whether a source fulfils this criteria the presumption should be against reliance on that source as evidence for the deduction.

This is of course open to judgement in any particular case but the default case in favour of non-reliance would hopefully prevent disputes spiralling out of control - if it is controversial it is not allowed. The intention at the time of creation seems pretty fundamental to me too. I think the last thing we want is sources being combined to show points that were not apparent at the time they were written. CrispMuncher (talk) 22:50, 30 October 2009 (UTC)

Speculating on the intention and understanding of a source's author seems impractical. --Bob K31416 (talk) 22:56, 31 October 2009 (UTC)

Could we not replace :

"Editors must agree that the deduction is likely to reflect the intention and understanding of a source's author at the time it was written. Deductions are not acceptable where it is unclear that the deduction would be obvious to or even accepted by the source(s) on which they are based."

with just

"Editors must agree." !

or even:

"All editors must agree"

After if we all agree then there's no problem; by implication we will be agreeing about reflecting the original source's intention etc. So I propose:

Logical deductions

---

This policy does not forbid logical deductions, provided all editors agree. For example, if A is in district B, and district B is in province C, then A is in province C.

The cruical insert is the word all. This is a major change to OR, so let's build in a cast-iron safeguard and see how it goes. --Michael C. Price ^talk 23:28, 30 October 2009 (UTC)

Michael, Perhaps the example in your version is inappropriately simple since you seem to be proposing that more complex deductions be allowed, like the ones you mentioned in a previous message of yours, "As it is the equations here and here are inconsistent. Explanation here of why. Sourcing not available; requires logic to fix." --Bob K31416 (talk) 01:19, 31 October 2009 (UTC)

The example given is only an example; more complex deductions are permitted. But perhaps we should be explicit, saying that complex deductions require universal assent, simplier deductions require the normal wp:consensus and giving an example of each.--Michael C. Price ^talk 01:43, 31 October 2009 (UTC)

I think we should stick with just the simple deductions for now. I recognize that you have confidence in your abilities, and probably rightly so, but there are editors that are unjustly confident in their abilities at scientific reasoning, for example. Because a deduction is complex, a consensus of them may not recognize when a complex deduction is incorrect. Recall this quote that I'm sure you've seen before,[11]

It isn't appropriate for us to try to determine whether someone's novel theory of physics is valid, we aren't really equipped to do that. But what we can do is check whether or not it actually has been published in reputable journals or by reputable publishers.

Although you're not suggesting that Wikipedia include something like novel theories of physics, complex deductions may approach that as far as questionable reliability is concerned. I think we need to limit the deductions to the simplest kind, like ones similar to the example in the proposal. --Bob K31416 (talk) 02:15, 31 October 2009 (UTC)

I think limiting logical deductions to very simple cases only, does not work for articles on technical topics. Logical deductions and reasoning is needed very often when you read technical articles or books. This is then purely to understand the sources, not to do any original research. If editors were to not do this, then you can expect mistakes to slip into articles. That has been my experience here on Wikipedia, which prompted me to propose these guidelines.Count Iblis (talk) 03:08, 31 October 2009 (UTC)

<outdent>Count Iblis, I read the proposed guidelines that you mentioned and I could support them except for the part which says,

It does not constitute WP:OR to provide the logical connection between sourced premises and sourced conclusions, since “Carefully summarizing or rephrasing source material without changing its meaning is not synthesis—it is good editing.”

It appears that this part allows, for example, original proofs by Wikipedia editors of unsolved mathematical conjectures. This goes beyond "carefully summarizing or rephrasing source material". Paraphrasing the quote that I mentioned in a previous message, "It isn't appropriate for us to try to determine whether someone's [complex deduction] is valid, we aren't really equipped to do that." However, we are equipped to determine whether the simplest kind of deductions are valid, like those similar to the example in the proposal.

Re "Logical deductions and reasoning is needed very often when you read technical articles or books. This is then purely to understand the sources, not to do any original research. " - It seems that you are referring here to the normal process of editing that is allowed by the policy, "carefully summarizing or rephrasing source material", rather than including in an article a description of one's own complex deductions and thought processes. This of course is already OK, since editors don't describe in articles the thought processes that they use in summarizing and rephrasing sources. Also, along the lines mentioned in your guidelines, discussion of these thought processes on the article's talk page is OK. --Bob K31416 (talk) 13:55, 31 October 2009 (UTC)

I am opposed to adding loopholes to this policy, since they inevitably constitute instruction creep, and will lead to far more abuse than the "issues" they are intended to address. As a simple example, I remember having an editing dispute with one editor, who insisted that because a source said that a specific human population had an admixture of 0.5% per generation for 80 generations, that the total admixture was 40%. In fact, as the source itself stated, the total admixture was estimated at 12.5% What appears to be simple mathematical (or other) deduction is very often not that. Jayjg ^(talk) 00:55, 1 November 2009 (UTC)

It looks like it was simply a wrong deduction, right? So this isn't relevant. --Bob K31416 (talk) 01:13, 1 November 2009 (UTC)

How is the Wikipedia editor supposed to know it's a "wrong deduction"? Looks like simple math/logical deduction. Jayjg ^(talk) 01:37, 1 November 2009 (UTC)

It's not clear which editor are you referring to. The editor that puts the item in the article or the editor that evaluates the edit. Anyhow, in both cases the editors are faced with essentially the same situation they are faced with whenever a source is used, i.e. interpreting the source correctly. --Bob K31416 (talk) 02:05, 1 November 2009 (UTC)

Indeed and then it's wise to stick to WP:ESCA  :) Count Iblis (talk) 02:07, 1 November 2009 (UTC)

The editor that was doing the calculation thought it was simple math. It wasn't, but that wasn't obvious. Opening up the loophole would have made this an allowable one, without the source specifically contradicting it. As for WP:ESCA, that would have opened the loophole as far or farther, at least in the form in which you want to include it, Count Iblis. Jayjg ^(talk) 02:40, 1 November 2009 (UTC)

J, Whether the editor put it in with or without the math, it would have been misinterpreting the source without having the the other item you mentioned in the article too. --Bob K31416 (talk) 02:48, 1 November 2009 (UTC)

Exactly, you have proved my point. It would have been misinterpreting the source, but that's not obvious to the lay reader or editor. The current policy would not allow the editor to include the mistaken calculation (40%), but the proposals to loosen the NOR policy would allow that mistaken calculation (sans a specific source contradicting it). Jayjg ^(talk) 03:02, 1 November 2009 (UTC)

C, Your guidelines look like good work now. --Bob K31416 (talk) 02:48, 1 November 2009 (UTC)

You really should link to a specific version when you write that, so it's clear which version you mean. Can we assume you mean this version? Jayjg ^(talk) 03:02, 1 November 2009 (UTC)

Agree with Jayjg. In addition to previous problems, the current proposal substitutes two cans of worms for one. Under this wording the wikilawyers gain a wedge to claim that local consensus can trump policy. Durova³⁵¹ 03:01, 1 November 2009 (UTC)

Would you care to expand on what you wrote re cans, worms, wikilawyers, local consensus, etc.? --Bob K31416 (talk) 03:12, 1 November 2009 (UTC)

Here is a quick example of how your precise formulation [12] can be used for wikilawyering: "PETA are eco-terrorists and eco-terrorism is terrorism. Therefore PETA are terrorists. Terrorists kill people, therefore PETA kill people. This is just two simple logical deductions exactly as described in the exception. The fact that most media are afraid of PETA and don't mention the obvious truth is no excuse for censorship here. The article must state that PETA kills people to be NPOV."

Your proposal makes it harder to dispose of editors who argue in this way. If you doubt that there are editors arguing like this in the real world, have a look at this. The editor filing this request has used even worse logic in the past. Once people stopped responding to his nonsense he claimed consensus through silence since nobody had reacted to what he felt was conclusive proof. Hans Adler 09:19, 1 November 2009 (UTC)

The requirenment that "all editors agree" would block this illogic.--Michael C. Price ^talk 09:59, 1 November 2009 (UTC)

It appears Michael C. Price thinks a majority of editors are able to be logical, so a majority view would trump illogical thought. However, the advantage of logic is that it transcends a majority opinion. Why would it be hard to point out that this argument is invalid because "all A is B" does not imply "all B is A" , or possibly because "some terrorists kill people" doesn't mean "all terrorists kill people"? Are we to have guidelines simply so editors do not have to present their reasons for objections, but can simply "rubber stamp" or "pigeon hole" when reasoning is a nuisance? That is what seems to me to be the position behind denying inclusion of simple syllogisms in WP:NOR, and avoiding misuse of WP:NOR is the reason for the explicit inclusion of syllogisms. Brews ohare (talk) 15:52, 1 November 2009 (UTC)

It appears Michael C. Price thinks a majority of editors are able to be logical, so a majority view would trump illogical thought. It appears that Brews does not understand the difference between "a majority" and "all". --Michael C. Price ^talk 19:32, 1 November 2009 (UTC)

It's not hard at all to point this out. What sometimes turns out to be extremely hard in such situations is to make the people using the faulty logic understand that their logic is sufficiently non-trivial to make them not entitled to edit war or argue until everybody else has run away from the talk page.

For a real-life example of this problem see here. (Sorry for the length, which is actually a main point of this example. Be sure to read the following sections as well. This continued at ANI.)

It's the most crass example I know. Even though it's about policy space, I am sure it's pertinent. This story continued yesterday with an absurd Arbcom filing of 20x the maximum length, the relevant passage being here. I think if you follow these links and read what you find there (it's funny enough to make this potentially rewarding) you will understand the problem: Good faith editors who think they are using flawless logic when they are just making assertions, and who think that their opponents are avoiding their arguments and just attacking them personally when they are actually using proper arguments. In other words: If we enable the use of logic explicitly (practically speaking it's no problem anyway), then we will have to argue against all the cranks who think they can win based on their superior logic. Hans Adler 17:05, 1 November 2009 (UTC)

Arbitrary break 1

Re Hans Adler's remark,

"For a real-life example of this problem see here. (Sorry for the length, which is actually a main point of this example. Be sure to read the following sections as well. This continued at ANI.) It's the most crass example I know. Even though it's about policy space, I am sure it's pertinent."

I don't see the pertinence. Perhaps you could point out the part on the policy page where an edit included a simple deduction that was a problem. Or are you suggesting that deductive reasoning should not be used on Talk pages??? Also please note that policy pages can be considered almost entirely "original research" or the unsourced original thoughts and guidance of the editors who contructed those policy pages. --Bob K31416 (talk) 17:39, 1 November 2009 (UTC)

A look at the issue from a different angle

WP:NOR evolved out of WP:V... and essentially says the same thing from a different perspective... ie: in order to say something in Wikipedia, you need to be able to cite a source for it.

Now... when it comes to WP:V, it is understood that not every single sentence in Wikipedia has to actually be cited. Everything has to be verifiable, not necessarily verified. However, WP:V does not actually come out and say this... it does not explicitly say: "it is ok to add some information without citing a source"... instead it implies this by using the phrase "challenged or likely to be challenged". The phrase implies that it is OK to add verifiable but uncited material that is unlikely to be challenged. And to make things clearer... WP:V goes on to say that if the material ends up being challenged, the burden of finding and supplying a source falls on those who wish to include the information; and, if a citation is not supplied, then the information can be removed.

So how does this relate to WP:NOR and our discussion of logical deductions... I would argue that we should take a similar line... We should use similar language as is used at WP:V... we should state that logical deductions that are challenged or likely to be challenged require a citation (and make it clear that, to prevent a Synthesis, the entire deduction must be cited to one source.) We should leave unsaid the implication that some logical deductions (ie those that are unlikely to be challenged) may not need to be cited, by focusing on where the burden lies when they are challenged.

In other words we should follow the path taken by WP:V... don't come right out and actually state that some logical deductions are OK... imply it by focussing on how to deal with the ones that are not OK. Blueboar (talk) 16:38, 1 November 2009 (UTC)

I think Blueboar is saying the same thing I was saying earlier (although he is saying it much more eloquently). Since we would technically be allowing information that is not directly verifiable, the burden on the syllogisms/logical deductions should be whether or not they are likely to be challenged. If an editor does not see the information to be easily deduced, then it should require a source. Angryapathy (talk) 17:43, 1 November 2009 (UTC)

I think we'd be opening a can of worms unnecessarily. The situation is that all material added to WP must be attributable i.e. there must be a reliable published source out there somewhere that says it. That source must be added if someone requests it, or if it's the kind of thing where someone is likely to request it. This is true of all material. We don't have to start adding, "and it's also true of syllogisms." SlimVirgin ^{talk|contribs} 18:02, 1 November 2009 (UTC)

(ec) A big difference when comparing V to NOR, is that it's fairly easy for virtually any editor to verify that "source X says Y", while in the case of unsourced logical deductions, judging the degree of difficulty, novelty or complexity in reaching conclusion Y from sources X1 and X2, can be very subjective and depend on one's education or professional background. Therefore, if we were to include the language proposed by Blueboar, it would have to be made very clear that the 'challenge' may come from any editor, not just an expert in the field, since we are not equipped to verify expertise. And then you could run into situations where someone with an opposing POV could use the 'challenge rule' as a roadblock against inserting something fairly obvious, but contrary to his POV. This is why I prefer to leave things like "logical deduction" to the common sense of the editors working on the article. Trying to define things further than they are, besides being an instruction creep, may introduce more problems (e.g. with POV pushers) than it will solve. Crum375 (talk) 18:08, 1 November 2009 (UTC)

Um... I don't think I have proposed any language... In fact, if I am proposing anything it is a lack of language. My intent was to simply suggest another way of looking at the underlying issue. I definitely don't want to open any cans of worms... in fact, I am trying to keep worm cans closed... which is why I stress discussing what isn't allowed, rather than outlining what is. By purposely not stating that uncited logical deductions might, in certain circumstances, be OK, you remove the chance that someone will squeeze their own OR into that statement. Essentially, I think we should stress that logical deductions should be cited like any other piece of information in Wikipedia... so if a logical deduction is questioned, the burden is on those who wish to include it to find and cite a source for it (and that source needs to follow the same chain of logic as is included in the article). However, to give editors the freedom to write good articles, we should remain completely silent on what happens when no one actually does question an uncited logical deduction. Blueboar (talk) 19:09, 1 November 2009 (UTC)

Crum375 puts it very well. This proposal appears well intentioned. With respect extended toward the editors who are putting it forward, in practice it would probably cause more problems than it solves. Durova³⁵¹ 19:11, 1 November 2009 (UTC)

Why is there an obession with leaving things unsaid? So that we can have endless squabbles? Why not be explicit? Is it beyond our ability to formalise what mostly already happens in a confused fashion? Would not clarity be welcome? --Michael C. Price ^talk 19:38, 1 November 2009 (UTC)

We have clarity already. For everything we add to an article, there must exist a reliable published source who says that precise thing in the context of the article's topic. If we're asked to produce that source, we must do so within a reasonable time, or the edit may be removed.

So far, no one has produced a real example of a truly simple syllogism not being allowed in an article. It seems we're discussing a problem with no evidence that it really exists. SlimVirgin ^{talk|contribs} 20:45, 1 November 2009 (UTC)

Have a look at megadose or orthomolecular medicine. I gave up on them awhile back because simple syllogisms would be reverted.--Michael C. Price ^talk 22:51, 1 November 2009 (UTC)

Michael, can you give us an example from megadose where a "simple syllogism [was] reverted" and the reversion stuck? This would have to be material which is based on published reliable sources which are acceptable to the other editors, and the 'syllogism' so simple that virtually all editors working on the article would accept as logically correct, but reject solely because the NOR policy is improperly written. Crum375 (talk) 23:53, 1 November 2009 (UTC)

Syllogism rejected on talk page. Orthomolecular medicine (OMM) advocates megadosing. I said:

If an OMM source says OMM implies taking loads of vitamin D then we can mention studies about vitamin D and leave it to the reader to form any conclusions (if any).

This was rejected with the conclusion..

.. without an independent source linking the current mania for Vitamin D to orthomolecular medicine, it's all WP:OR.

and the current OMM article contains not a single reference to vitamin D. --Michael C. Price ^talk 01:20, 2 November 2009 (UTC)

That's a different matter. That's not a dispute about a syllogism but (if correct as you state it) you have been a victim of bad wikilawyering. Of course it's up to OMM sources to decide whether vitamin D is relevant to OMM or not, and if it is studies about vitamin D can be cited to source specific claims. Within reason, of course, to avoid being misleading by giving undue weight to the vitamin D business.

I found nothing relevant at the first article, but at the second there were the usual surreal discussions that one would get into with OrangeMarlin. He never argued rationally and typically made rules and facts up according to his intuitions. The proposed change would not have helped you. (And as far as I know he stopped editing after I contacted a reincarnation of his.) Hans Adler 01:30, 2 November 2009 (UTC)

True, OrangeMarlin was a large part of the problem (glad we agree about him!). But the OR rebuttal (above) was made by someone else, and so the point still stands that difficult and irrational editors are able to use OR as a weapon to block progress. I take your point that their argument was flawed, but it was still presented. Changing the policy to be clearer or more precise would remove that obstructive cop-out. --Michael C. Price ^talk 01:38, 2 November 2009 (UTC)

Re example - There's a discussion that is currently in progress at WP:NORN. During the discussion, the editor who brought it up has simplified his question to the following issue, in order to get help with understanding what is and what is not allowed. I expect the editor will get back to the more complicated issue once he gets this simpler hypothetical example clarified. (I edited it a little to clarify.)

My question is if the source says that 64% of Turks are unwilling to live next to Jews and 75% of Turks are unwilling to live next to atheists, can we simply say that more Turks are unwilling to live next to atheists than to Jews?

I think the proposed addition and Blueboar's approach, would not prohibit this simple deduction. But it could still be excluded if it was objected to and didn't get a consensus. --Bob K31416 (talk) 00:16, 2 November 2009 (UTC)

Assuming the source mentioned this in relation to the article in question, and that NPOV and other potential issues are dealt with, we could simply report that according to source X, 75% of Turks are unwilling to live next to Atheists, and 64% are unwilling to live next to Jews. The word "more" is not needed, since the reader can see the numbers. And this would be a straightforward presentation of a reliable source per V and NOR, with no 'syllogism' needed. Crum375 (talk) 00:25, 2 November 2009 (UTC)

But this is an encyclopedia that summarizes where possible. If not for the prohibition, the author could have summarized the info without giving the statistics which are more cumbersome. Also, note that with the proposed change in policy, if anyone objects to it's inclusion, the burden is on the editor who wants it in to demonstrate that it has consensus. --Bob K31416 (talk) 00:34, 2 November 2009 (UTC)

We already allow routine arithemtic calculations, and stating that 75 is more than 64 would fall under that, assuming other issues are properly handled. I can see situations where statistics can be misused, such as the polls taken at different times, the margin of error could play a role, or the format of the question could have an impact on the results, and would have to be included, if controversial. But just saying that 4 is more than 2 does not require a change in policy, per above. Crum375 (talk) 00:46, 2 November 2009 (UTC)

Here's a quote from an editor that opposed it on the article's talk page.

"Can you quote the sentence that specifically says that "turks are less willing to live near atheists than jews"? If it did, there wouldn't be a problem. But it's just not there. You are assuming that this is the case from the numbers given, (and you may be right) but on WP you can't do that per No original research."

This is the usual refrain: You're probably right but we can't put it in because it's OR.

With the proposed change the editors wouldn't be limited by this but could include or exclude it using more reasonable criteria. --Bob K31416 (talk) 01:15, 2 November 2009 (UTC)

That's exactly the problem. These criteria wouldn't be more reasonable. We must live with people sometimes being probably right and not being able to put something into an article. Hans Adler 01:18, 2 November 2009 (UTC)

(ec) No, stating that 75 is more than 64 is not a routine calculation in this case. There are all sorts of potential reasons to make the "more" statement invalid. E.g. the following scenarios that I made up completely on the spot for illustration:

1. The Turkish language has two words for "Jew": One referring to the religion and one to the ethnic group. The distinction is relevant in Turkish society since there are many ethnic Jews who are atheists, Christians or Muslims, and also a number of ethnic Turks, Kurds and Armenians of Jewish faith. It turns out that while 'only' 64% of all Turks are unwilling to live next to ethnic Jews, a full 82% don't want to live next to people of Jewish faith. Therefore it would be more correct to say that more Turks a unwilling to live next to Jews than to atheists. (Since in this context "Jew" will be read as referring to the faith, so 82% applies.)
2. The Turkish language has only one word for "Jew", but the two numbers come from 1) a telephone poll conducted among 200 households in Istanbul, and 2) from the responses to a question that an East Anatolian radio chain asked to its audience. Therefore the numbers are simply not comparable.
3. The numbers are comparable in principle, but both come from the same telephone poll with 50 non-representative participants. Upon reading the full version of the report we learn that with a 95% confidence, 38-81% of Turks are unwilling to live next to Jews and 53-92% are unwilling to live next to atheists. Therefore it's impossible to say with the same confidence that more Turks are unwilling to live next to atheists than to Jews. Hans Adler 01:18, 2 November 2009 (UTC)

I agree with Hans. This is exactly why NOR is OK the way it is. Statistics is a very dangerous weapon in the wrong hands. It involves simple numbers, but to correlate them to reality, esp. in controversial public opinion polls, is very tricky if not impossible, and is often misused to pursue specific agendas. We can say 4 is bigger than 2, but we can't draw far-reaching conclusions about behavior or feelings of people based on statistical polls. Crum375 (talk) 01:26, 2 November 2009 (UTC)

Hans, your 3 objections may be valid, but were they mentioned in the source? If not then we shouldn't be concerned with them either, but just with the question of whether the syllogism /calculation is correct.--Michael C. Price ^talk 01:28, 2 November 2009 (UTC)

(edit conflict)Hans Adler, I think you're missing the point. All the arguments that you are bringing up are possibly the "reasonable criteria" that I mentioned in my last message. Please reread my last message with this in mind. --Bob K31416 (talk) 01:30, 2 November 2009 (UTC)

Yes, sorry, due to fear of edit conflicts (which freeze my browser) I was a bit hasty and misread what you wrote. Hans Adler 01:36, 2 November 2009 (UTC)

Michael: No, that's not how it works. There are millions of ways the argument may be wrong. Even ignoring the POV question whether 75% is really "more" than 64% or "on about the same level". The syllogism is correct if and only if none of these flaws applies. Finding out that all of these flaws don't apply is massive original research. We can't do this. We must rely on a RS to do this for us. And if the fact is noteworthy, the odds are a reliable source has done it for us. Hans Adler 01:40, 2 November 2009 (UTC)

Hans, I think I see the source of confusion here; I was making an implicit assumption. To be explicit: Suppose the question were rephrased slightly, so that the revised claim is attributed to the source:

My question is if the source says that 64% of Turks are unwilling to live next to Jews and 75% of Turks are unwilling to live next to atheists, can we simply say that, according to the source, more Turks are unwilling to live next to atheists than to Jews?

Is this now acceptable?--Michael C. Price ^talk 01:55, 2 November 2009 (UTC)

You are right that having both in one sentences really changes things. I would say it's a borderline case and may still depend on the context of that sentence. Depending on that we may be able to say "more" (most likely, I admit), "about the same number", or neither. Of course a single word like "only" in the source would make everything crystal clear. Hans Adler 02:18, 2 November 2009 (UTC)

But doesn't Hans' argument show that you need to allow for more than just simple syllogisms? Keeping an edit out of an article is in a certain sense also an edit (a negative edit). But this is then based on a more complex argument based on statistics that proves that the simple syllogism is not valid and hence one would need a direct quote from a source. Count Iblis (talk) 01:53, 2 November 2009 (UTC)

No. Just like the real world, Wikipedia has only few rules prescribing actions but plenty of rules proscribing actions. I can think only of two rules that tell us that we have to say something in article space: The principle that Wikipedia is an encyclopedia, and WP:NPOV, which tells us that if we tell one half of the truth, we must tell the other as well. Insofar as the other rules talk about content they are only meant to prevent us from creating hoaxes, whether intentionally or not.

But we are never forced to say something just because it's true and we can source it. Even the principle that removal of sourced information is disruptive isn't universally valid; it's not meant to trump encyclopedic brevity, for example. Hans Adler 02:28, 2 November 2009 (UTC)

Getting back to the example

There is a source that says that 64% of Turks are unwilling to live next to Jews and 75% of Turks are unwilling to live next to atheists. Michael raised the question of whether the following statement would be acceptable, " according to the source, more Turks are unwilling to live next to atheists than to Jews." Hans felt this was better, but still seemed to feel that NOR prohibits it. This is the crux of the issue. Editors believe that this statement is prohibited by the current form of WP:NOR. Futhermore, that appears to be the fallback position of those opposed to the statement on the article's Talk page. It doesn't matter whether it is right or wrong, it's a simple deduction and therefore prohibited. The proposed changes would not prohibit it but still require consensus for including it. With the proposed change in policy, the burden to establish consensus is still on the editor who wants to put it in, but the possibility of including it would not be eliminated solely by WP:NOR, which is currently the case. --Bob K31416 (talk) 14:35, 2 November 2009 (UTC)

Correction: The change that made it acceptable to me wasn't "according to the source". I wasn't even aware of that and might have objected to it. What made it acceptable to me was that in the revised example both numbers occurred in the same sentence, obviously inviting the reader to read between the lines and make the comparison. Hans Adler 19:14, 2 November 2009 (UTC)

Thanks for clarifying. But that makes the point even stronger. You feel that the above excerpt with or without the part in bold should be excluded, if it appears by itself. SlimVirgin, in a comment below, seems to think that WP:NOR allows statements like that. Perhaps you two should discuss that. --Bob K31416 (talk) 19:40, 2 November 2009 (UTC)

You are demonstrating how dangerous the proposal is. Deciding whether the statement in question really follows from the source needs much more competence than a syllogism does. And, by the way, even formally valid syllogisms do not always produce correct results in practice, since the real world isn't as simple as the ideal world modelled by formal logic. People like you misunderstanding the guideline in this way is exactly what we need to avoid. I don't think there is any disagreement with SV, since she is talking about syllogisms and I am talking about this complicated example. Hans Adler 20:10, 2 November 2009 (UTC)

Here's a simpler and less emotional example: If a source said 1/4 of all cultivatable land was grown in hay, and 1/2 in corn, it presently would be WP:OR to say more cultivatable land was in corn than in hay unless that claim could be found verbatim in some source. Opponents to a change in WP:OR will argue that inability to include this statement (or enabling obstruction of simple statements) is just collateral damage, because allowing it opens the floodgates. It is an impasse. Brews ohare (talk) 18:58, 2 November 2009 (UTC)

What I'm not understanding is why, if someone were to object to your summary, you couldn't just say "1/4 of this, 1/2 of that," per the source. There's no need to change a policy when a tweak of the writing will meet objections, rare as they doubtless are. SlimVirgin ^{talk|contribs} 19:02, 2 November 2009 (UTC)

Well, SlimVirgin, you are taking a simple example intended to illustrate the point, and making it into the point. To phrase matters more abstractly, "There are simple statements that presently are excludable, and in practice are excluded by citing WP:OR, that would become explicitly allowed with a modification of WP:OR; but changes to WP:OR that would allow explicitly such statements are opposed on the basis that such changes make the editing task of excluding bogus arguments too onerous." Brews ohare (talk) 19:12, 2 November 2009 (UTC)

They're not excluded. We're just not explicitly saying they're okay. We're leaving it up to the editors on the page, because circumstances vary, and some apparently simple syllogisms may not be so simple when looked at carefully. The editors on the page know the context and the sources, and are better equipped to judge. SlimVirgin ^{talk|contribs} 19:16, 2 November 2009 (UTC)

Exactly, and that is the problem. Take a look at this discussion. Brews ohare (talk) 19:17, 2 November 2009 (UTC)

It's too much to read. Can you point to the specific issue? SlimVirgin ^{talk|contribs} 19:19, 2 November 2009 (UTC)

I'm afraid that would border on my topic ban. Nonetheless, it is an example where WP:SYN is used to block a simple syllogism. However, to continue, no matter what examples are brought up, and no matter how clear-cut the problem may be in some particular instances, there is no way to combat the argument that explicit inclusion of such cases opens the floodgates and makes editing too difficult. What is "too difficult" for a particular editor is a judgment call beyond argument. Change of WP:OR is at an impasse. Brews ohare (talk) 19:24, 2 November 2009 (UTC)

Brews, throughout this entire discussion, no one has given a real example of a simple syllogism being disallowed with this policy being cited. I think you need to produce one, or let the discussion come to a close, because as things stand, we seem to be talking about a non-existent problem. If you can talk about changing the policy without violating your topic ban, and you can link to that discussion, I can't see what difference adding an example would make. SlimVirgin ^{talk|contribs} 19:28, 2 November 2009 (UTC)

You are upping the ante here to say that no clear-cut examples exist. I have pointed one out: you claim it's too much to read. The example is moot, because even if we ultimately agree that this example is clear-cut, the next step in your argument is that it is "only" one example. That makes the argument a matter of persuading you that "sufficient" examples exist, and a blow-by-blow construction of a list of what is mutually agreed upon are examples, and that their number is sufficient. That is going nowhere. It is an impasse. Brews ohare (talk) 19:33, 2 November 2009 (UTC)

A suggestion: perhaps you could leave the issue for now, and for the next six months or so, whenever you see an example of this, make a note of it—and perhaps some of the rest of us can do the same—then if you have sufficient numbers to suggest there's cause for concern, come back here and we can take another look. SlimVirgin ^{talk|contribs} 19:37, 2 November 2009 (UTC)

(edit conflict):Slim Virgin, Re your remark, "They're not excluded." - Perhaps you and Hans should discuss this since you two seem to have a difference of opinion on that. See my response to Hans above. --Bob K31416 (talk) 19:40, 2 November 2009 (UTC)

Slim Virgin: I have no intention of pursuing this argument along those lines. The matter really should not be decided upon this basis at all, but upon the basis that WP:OR never should be abused to exclude material that is the result of a simple syllogism. If the value of that principle cannot be established, there is no point in proceeding. Brews ohare (talk) 19:42, 2 November 2009 (UTC)

Brews, 'simple' means different things to different people in different contexts. My simple can be your complex, and vice versa. This is why Wikipedia does not allow editors to provide their own conclusions, and requires them to neutrally report and summarize what published reliable sources have said about issues, not create new material. This is what this project is all about. Crum375 (talk) 19:49, 2 November 2009 (UTC)

Crum375: The notion of simple in this discussion is used in the combination "simple syllogism" and refers specifically to the topic syllogism. Brews ohare (talk) 19:59, 2 November 2009 (UTC)

Brews, real life situations can rarely (if ever) be reduced to "simple syllogisms", acceptable to all sides, especially in controversial issues. Therefore we must stick to what the reliable sources have said, and present them neutrally. Crum375 (talk) 20:03, 2 November 2009 (UTC)

Crum375: This discussion is not about controversial issues: it is about syllogisms, regardless of whether they are common or interesting. Brews ohare (talk) 20:07, 2 November 2009 (UTC)

(edit conflict)Crum375, Re your comment "'simple' means different things to different people in different contexts. My simple can be your complex, and vice versa. " - We're not talking about novel theories of physics but the example of making a statement regarding more atheists than Jews when the source says 75% atheists, 64% Jews. On an article talk page, at WP:NORN, and here, this has been considered a violation of WP:NOR by various editors. As I mentioned numerous times, consensus would still be required for including the statement, where in the case that you brought up, the question of whether it is a simple deduction can be brought up and would need to be confirmed by consensus before the simple deduction is included in the article. --Bob K31416 (talk) 20:06, 2 November 2009 (UTC)

Bob, I think you are missing the point. In real life, esp. in circumstantial issues, it is very difficult to agree on reduction of the sourced information into a "simple syllogism", and the word "simple" is not a clear concept in itself. In the example brought above about Turkish attitudes towards ethnic groups, it relies on statistics, which are notoriously unreliable and variable, esp. when it comes to polling people. This is why it is much preferable to simply quote the source verbatim, or close to it, and not "editorialize" by adding our own conclusions, because they can introduce subtle (or not so subtle) POV. If the issue is simple arithmetic calculation, that's already covered; when you venture farther into "logical deductions", you open the door to POV and OR. In summary, stick to what the reliable sources said, esp. in contentious issues. Crum375 (talk) 20:23, 2 November 2009 (UTC)

Crum375: Again, this discussion is not about "real life" and is not about "statistical reliability". It is about syllogisms, which are symbolic manipulations of premises that find statements logically equivalent to the premises. These manipulations are purely mechanical, can be made by automatons, and have nothing to do with the validity of the premises, only with statements that are equivalent to the premises. Brews ohare (talk) 20:27, 2 November 2009 (UTC)

Crum375, Re your comment "I think you are missing the point. In real life, esp. in circumstantial issues, it is very difficult to agree on reduction of the sourced information into a "simple syllogism", and the word "simple" is not a clear concept in itself." - Again, as I mentioned numerous times, consensus is required for inclusion of the statement. If in the case you are bringing up, some editors believed it is not simple, consensus would not be reached and it would be excluded. --Bob K31416 (talk) 20:33, 2 November 2009 (UTC)

The main point is that in contentious cases there is rarely agreement on the reduction of the real issues to a "syllogism", and even more so to "simple" one. The only real example I see above is the Turkish attitudes poll and it was clearly not consensual that the issue was reducible to a "simple syllogism". In fact, as an uninvolved party I can see major issues in that reduction. If all sides agree that a conclusion is trivial and correct, they will use their common sense and consider it part of the allowed "summary and neutral presentation" process. If there is any dispute, then it's not really trivial or indisputably correct, and needs exact sourcing. Crum375 (talk) 21:09, 2 November 2009 (UTC)

Here's the proposal for reference.

Simple logical deductions

---

This policy does not forbid simple logical deductions, provided editors agree that the deductions correctly reflect the information published by the sources from which they are derived. For example, if A is in district B, and district B is in province C, then A is in province C. This is a simple syllogism. Included are all of the other simple deductions. More complex logical deductions should not be used unless cited to a reliable source.

--Bob K31416 (talk) 20:59, 2 November 2009 (UTC)

The problem, again, is that some editors may disagree about the deduction, for example the ones favoring a small minority view. But because they may be overidden by consensus, the large majority will ignore them, and decide that their view can be discounted. This is exactly what V and NOR try to avoid: we should be able to source any statement when challenged. Adding this language will allow a majority of editors (able to achieve a "consensus" by their sheer number) to bulldozer their personal views using "simple logical deduction" as argument. Crum375 (talk) 21:16, 2 November 2009 (UTC)

Pandora's Box

I think one fear here is that while allowing some simple syllogisms might improve the encyclopedia, it could potentially be misused to introduce WP:OR and WP:SYN into WP. Brews has seen a "hole" in the guidelines that he feels should be plugged, but he is making the same mistake I have made a number of times: the policies/guidelines of WP cannot cover every possible set of circumstances. Some debates and arguments come up that are totally germaine to that conversation, and don't apply outside of that realm. Other times, the editors at a certain article must make decisions based on current policies (and the spirits of policies) which may only work for that article. That's the nature of consensus and WP:IAR. Adding the section on syllogisms/logical deductions is a bit like this: You have a mild headache, so the doctor prescribes drug A. Drug A causes a rash, so the doctor gives you drug B. Drug B makes you nauseous, so he gives you drug C, ad nauseum. Sometimes the attempts to fix a problem will cause more and more problems, so it's best just to deal with the inital headache. Angryapathy (talk) 19:57, 2 November 2009 (UTC)

I believe your comment is what I referred to as the "open the floodgates" argument, which is essentially an argument that it is too much trouble to deal with illogical editors, and so it is better to allow abusive editorial use of WP:OR to exclude the obvious, on the basis that abuse is easier to deal with. Unfortunately, such abuse actually cannot be dealt with: one is forced to find exact verbatim statements or have materials excluded. No-one can argue with you about what you find to be "too much trouble". The subject is at an impasse. Brews ohare (talk) 20:05, 2 November 2009 (UTC)

Brews, do you understand that "simple" and "obvious" to you can be totally different to another editor? That's where this issue lies. Angryapathy (talk) 20:15, 2 November 2009 (UTC)

Angryapathy: I don't agree that this is a subjective matter. Everyone agrees what a syllogism is, and a syllogism can be established by simple symbolic manipulation that is totally non-subjective. That is what the Russell-Whitehead link is about. Brews ohare (talk) 20:20, 2 November 2009 (UTC)

Angryapathy, Re your comment " 'simple' and 'obvious' to you can be totally different to another editor" - See my last message in the above section and here's a link to it. --Bob K31416 (talk) 20:25, 2 November 2009 (UTC)

Brews, the problems usually happen when we formalise a statement in order to apply a syllogism. We can't do this without simplifications. Some people are unable to see that the simplifications they are applying aren't valid in the particular example. When we tell them simple syllogisms are allowed, they will feel wronged when "their" syllogism isn't allowed. Hans Adler 20:41, 2 November 2009 (UTC)

Hans: Undoubtedly some will not understand that they are not using a syllogism, or have incorrectly stated the premises, or have incorrect premises. A discussion will follow, which may converge or not. The issue is how much is gained by a practice that allows abusive use of WP:NOR compared to how much is lost by excluding statements logically implied by sourced premises. Brews ohare (talk) 20:49, 2 November 2009 (UTC)

No movement towards including the proposal

So far, the only movement by editors opposed to the proposal, towards including it, was long ago in this discussion by the editor who originally removed it from the policy page, and that editor left the discussion long ago. I've done about all I can so I'll exit now. Thanks for the discussion. Regards, --Bob K31416 (talk) 21:45, 2 November 2009 (UTC)

Actually, I'm still here - simply because I haven't felt I have anything meaningful to contribute for the past couple of days does not mean I'm not around, and I certainly am not "long gone". As for the proposal, I find myself siding more and more firmly against it. I was willing to entertain a change so long as it was subject to robust restrictions and safeguards and received a very broad consensus - in my opinion this needs much more agreement than usual simply because this policy is so central to the project. In my view, the sheer amount of discussion this proposal has generated makes it very clear that such a consensus is not going to be reached. In that case we should adopt the cautious route and not accept the proposal. CrispMuncher (talk) 16:38, 3 November 2009 (UTC)

Alternative approach

Here is the original proposal:

Simple logical deductions

---

This policy does not forbid simple logical deductions, provided editors agree that the deductions correctly reflect the information published by the sources from which they are derived. For example, if A is in district B, and district B is in province C, then A is in province C. This is a simple syllogism. Included are all of the other simple deductions. More complex logical deductions should not be used unless cited to a reliable source.

This proposal seems fine to me, but objections have been raised. Therefore, below is an alternative. Brews ohare (talk) 21:18, 2 November 2009 (UTC)

Inasmuch as it already is part of WP:NOR that routine mathematical operations are allowable, I'd suggest that the statement be added that all syllogisms that can be related logically to routine calculations are acceptable, and are not considered to be WP:SYN or WP:OR. Perhaps an article can be written that demonstrates explicitly the equivalence of certain syllogisms to mathematical operations, and Routine calculations can be linked to this article. Brews ohare (talk) 20:43, 2 November 2009 (UTC)

Why do we need to say this? Blueboar (talk) 20:52, 2 November 2009 (UTC)

The discussion on this page is evidence that such a statement is considered non-empty, as many have expressed reservations about it. It is my view that the trivial nature of this statement is not evident to all, and that statements have been excluded from WP based upon a misunderstanding of its nature. Brews ohare (talk) 21:01, 2 November 2009 (UTC)

I agree with Blueboar. What will be gained by doing this? Angryapathy (talk) 21:04, 2 November 2009 (UTC)

Statements have been excluded from WP based upon a misunderstanding of the equivalence of syllogisms to routine mathematical manipulations. Other statements of this type have been included only after long discussions and the locating of exact verbatim statements, where simple logic based upon premises should suffice. These costs would be eliminated by adding this statement. Brews ohare (talk) 21:08, 2 November 2009 (UTC)

Please show us one such example, Brews. And please focus on one bite-sized atomic issue. Crum375 (talk) 21:19, 2 November 2009 (UTC)

This is getting out of hand. For the last few weeks small number of editors have been proposing changes that would constitute major change to longstanding policy; as each subthread fails to gain consensus the discussion gets a jump start with a minor variation on the same basic proposal--none of which addresses the substantive objections. In an ideal world the proposed changes might be reasonable, but they fail to safeguard against frequent and longstanding types of wikilawyering. Sometimes a minor inconvenience really is necessary for the good of the project. Please let it go. Durova³⁵⁵ 21:17, 2 November 2009 (UTC)

While there are many critiques of WP, I think WP:Verifiably gives WP credibility. You are advocating a system where someone can use "simple logic" to derive a conclusion, which I am sorry to say you fail to see the future implications. This will cause more problems than the very few that have happened or will happen without the proposed policies. Just find a source for the information, or leave it out. That's how it works, and has worked for years before this. Angryapathy (talk) 21:24, 2 November 2009 (UTC)

Durova: It would be a useful contribution to this discussion if you could list the substantive objections, which in my mind boil down to the single "floodgate" or "Pandora's box" argument. That is the one argument you have raised as well. It's natural for a discussion of this kind to take some time, even though it is not a "major change", or even a change at all, but simply an explicit statement of what is already implicit in the Routine calculations subsection. Let's see what people think about that. Brews ohare (talk) 21:36, 2 November 2009 (UTC)

Brews, our goal is to keep policies short and sweet and avoid instruction creep. To add or modify a policy we need justification, i.e. a clear example where the current policy wording fails. I have yet to see such example. Crum375 (talk) 21:43, 2 November 2009 (UTC)

I agree. Brews ignores the implications of the addition because it justifies why he fought so hard and eventually got a topic ban from physics. Angryapathy (talk) 21:48, 2 November 2009 (UTC)

Angryapathy: That remark of yours is a violation of civility , and an indication that you want to drag this discussion into the mire of hypothetical motivation (entirely erroneous, BTW) instead of addressing the topic. It also is a misstatement of the facts. Brews ohare (talk) 21:54, 2 November 2009 (UTC)

Well Brews, you have on a few occasions obliquely mentioned in this discussion how you would like to mention the topic in which you have been banned from discussing. I guess I'm using a syllogism to connect this proposal with that topic. Angryapathy (talk) 22:06, 2 November 2009 (UTC)

Like many WP policies, if this is applied with common sense, as it ususally is, it's OK. Sometimes Editors get too hung up on the literal wording and then it can be un-helpful, but almost any change to the policy would have the same objections, and would in practice lead to all kinds of wikilawyering problems alas. NBeale (talk) 21:47, 2 November 2009 (UTC)

I'd suggest that a perfectly acceptable argument for Wikilawyering a statement that uses a syllogism would be that the policy already allows routine mathematical calculations, the syllogism is logically equivalent to such a routine calculation, and therefore WP:NOR is inapplicable. What will be your response to such an argument, supposing you were the reverting editor? Brews ohare (talk) 21:54, 2 November 2009 (UTC)

The problem is that, as I and others noted elsewhere, routine arithmetic calculations are not the same as logical deductions. The reason is that in the case of logic we first need to reduce reality into a logical formulation. That process is very open to abuse by POV pushers, and is OR by itself. Crum375 (talk) 22:02, 2 November 2009 (UTC)

And Brews, I am still waiting to see one example which clearly demonstrates there is a problem with the current policy wording. Not a link to a discussion page, but an atomic bite-sized example, reduced to its essentials. Crum375 (talk) 22:07, 2 November 2009 (UTC)

User:OMCV wanted to put this article on AFD, just because it contains unsourced logical deductions/mathematical derivations. Count Iblis (talk) 22:23, 2 November 2009 (UTC)

Crum375: There is a semantic difficulty here: routine mathematical calculations are equivalent to certain logical syllogisms. See Russell and Whitehead. The "reduction of reality" to a logical formulation is a process that is error prone, as you note, but is not a part of logic. That process is always subject to editorial review, and may be a valid use of WP:NOR. However, that process is not part of the subject of including syllogism explicitly in WP:OR. As for examples, several have been provided above, one is pointed out immediately above by Count Iblis. However, as indicated, the existence of examples will not resolve the issue. Brews ohare (talk) 22:27, 2 November 2009 (UTC)

Glancing at the talk page of that example, I see other points being made by editors, and they are not related to an objection to syllogism. One editor makes a point that a source for the derivations would be helpful to students, another makes a point that the article is just a list of equations and derivations and thus does not merit its own page. I see nothing on that talk page that would support changes to any existing policy. Crum375 (talk) 22:34, 2 November 2009 (UTC)

Crum375: Again, the goal here is not a change in policy. It is the making explicit of the application of the section on Routine calculations to the logically equivalent use of syllogism. This application is already implied by this section of WP:NOR, and making it explicit clarifies inapplicability of WP:NOR as a grounds for excluding simple syllogisms. Brews ohare (talk) 22:53, 2 November 2009 (UTC)

Brews, if we are not here to change policy, then why not go somewhere else? And if we are here to change policy, it requires good justification. At this point, I don't see a single example to justify any change, however small, in this policy. But I am open-minded, and any good example could sway me. Crum375 (talk) 23:00, 2 November 2009 (UTC)

Crum375: Here is a possible formulation:

Routine calculations

---

This policy does not forbid routine calculations, such as adding numbers, converting units, or calculating a person's age, provided editors agree that the arithmetic and its application correctly reflect the information published by the sources from which it is derived. Likewise, simple syllogisms may be used that are logical equivalents of routine mathematical calculations, as described by Russell and Whitehead. Care must be taken to observe no synthesis takes place.

Brews ohare (talk) 23:09, 2 November 2009 (UTC)

I think what is missing here is an explanation of why we allow editors to perform routine calculation, but do not allow syllogisms... We do not call a routine calculation (such as coversion of standard US measurments into metric measurments) "Original research" because the calculation involved does not originate with the editor who performs the calculation. The formula for the calculation could be cited (to a grade school math text book) every time someone performs the calculation, but we agree they don't have to do so. The same is not true with syllogisms... an unsourced syllogism is Original research, in that the entire logic chain originates with the editor who creates it. Blueboar (talk) 23:11, 2 November 2009 (UTC)

I think the articles symbolic logic, mathematical logic and propositional calculus may be unfamiliar to the editors, and should be consulted. The concept of an "unsourced syllogism" is a bit mind boggling. Brews ohare (talk) 23:25, 2 November 2009 (UTC)

Brews, the way I see it, your wording opens the door to a lot of OR. I think we should stick to the principle that made Wikipedia a success: all material should be directly attributable to a reliably published source. Those "simple logical derivations" will allow POV pushers to create new material supporting their view, or frustrated authors to use WP as a publishing house for their original material. And I am still waiting for a clear example why this change is needed. Crum375 (talk) 23:20, 2 November 2009 (UTC)

As pointed out above, the articles symbolic logic, mathematical logic and propositional calculus may be unfamiliar. The use of a syllogism is a symbol manipulation, just like algebra, and cannot be thought of as original research. Brews ohare (talk) 00:59, 3 November 2009 (UTC)

A quick status check (straw poll)

While I'm generally strongly disinclined to recommend or encourage the use of straw polls as part of the policy-making process, it seems to me that (per Durova and others above) we've reached the point where the discussion is becoming circular, and no significant advancement is likely.

I would like to pose the following questions, and get a yes/no wherever possible. TenOfAllTrades(talk) 23:16, 2 November 2009 (UTC)

Evidence of problem

Has there been a clear demonstration that simple, unambiguous syllogisms are incorrectly removed from Wikipedia articles on a regular basis?

Yes, evidence has been presented that syllogisms are often removed inappropriately.

No, evidence has not been presented to demonstrate the problem.

1. TenOfAllTrades(talk) 23:16, 2 November 2009 (UTC)
2. SlimVirgin ^{talk|contribs} 23:21, 2 November 2009 (UTC)
3. Hans Adler 23:25, 2 November 2009 (UTC)
4. Crum375 (talk) 23:32, 2 November 2009 (UTC)
5. With slight distaste for straw polling as a method of problem solving, basically valid. Durova³⁵⁵ 23:33, 2 November 2009 (UTC)
6. I think the key here is not whether syllogisms have been removed... but whether they have been incorrectly removed. So far, the few examples of syllogisms seem to have been removed correctly. Blueboar (talk) 23:38, 2 November 2009 (UTC)
7. Angryapathy (talk) 00:38, 3 November 2009 (UTC)
8. Finell (Talk) 01:36, 3 November 2009 (UTC)

Comment - This is a rather poorly put question. "Clear demonstration" sets the bar high and would require a considerable amount of research into the histories of articles. "Unambiguous syllogisms" again unnecessarily narrows the subject more than it is since the proposal applied to simple deductions in general. "Often" in the yes part, and "on a regular basis" in the question, again makes the question difficult. Consider how this type of a question, narrowed in the way it has been, would be answered when it is applied to each existing part of WP:NOR. For example: Has there been a clear demonstration that there are situations like the Smith/Jones example that occur in Wikipedia articles on a regular basis and often?

Please note that when an example was requested, I found an example of a simple deduction that was excluded from an article, by giving a quick look at the current discussions at WP:NORN. --Bob K31416 (talk) 03:13, 3 November 2009 (UTC)

It shouldn't require any research at all. Those who propose or support this policy change ought to have instances in mind that justify it. We should not complicate a policy with an exception unless the lack of the exception has caused problems, or waste time trying solve a non-existent problem. As for "clear", the proposed policy change is to make an exception for "simple syllogisms"; if it is simple, it should be clear. If it isn't a clear instance, it probably wasn't a simple syllogism in the first place. —Finell (Talk) 03:35, 3 November 2009 (UTC)

I think you misread the question regarding what the adjective "clear" modifies. Also, you either misread the proposal since it included the more general "simple logical deductions", or got it mixed up with the ones in the section for modifications, which is understandable considering all the discussion. Regarding what someone ought to have in their mind and that it shouldn't require any research at all, is pure opinion and unsupported speculation. Apparently you ignored the fact that I found an example at the current discussions at WP:NORN with just a quick look. I'm getting the same feeling I had when I exited the other related discussion, so that's it for me here too. Bye. --Bob K31416 (talk) 04:00, 3 November 2009 (UTC)

The question is perfectly worded. The only syllogisms presented were marginal at best. If we can't show a list of examples in which valid, logical information is squashed by the weight of WP:NOR, then obviously we don't need to complicate the policy by adding the proposal. Angryapathy (talk) 15:02, 3 November 2009 (UTC)

Support for changes

As proposed above (in any section) would you endorse a change explicitly exempting syllogisms and/or logical deductions to this policy?

Yes, I would support the proposed addition (specify which).

1. Clearly, there are problems regarding OR, Synth, verifiability, "NotTextBook", etc. etc. etc. when editing technical scientific articles. It usually doesn't make "wiki-news headlines" at AN/I because there is usually no conflict between editors. Wiki policy making is driven by the need to resolve conflicts. But that has led to policies that are very good at dealing with issues you typically face on politics pages where editors edit with some agenda. But these policies leave no room for explaining things from first principles in wiki articles (it is e.g. almost unavoidable that you'll violate Synth). Count Iblis (talk) 01:43, 3 November 2009 (UTC)
2. I thought the proposal that I suggested, which was a modification of a previous proposal by Brews, was worthwhile, and I felt I successfully defended it against criticism. I stopped participating in the discussion because it appeared that the opposition couldn't be swayed, so that it was a waste of my time to continue. --Bob K31416 (talk) 02:56, 3 November 2009 (UTC)

No, such an addition to this policy is not necessary. Obvious syllogisms are adequately covered by existing Wikipedia policies.

1. TenOfAllTrades(talk) 23:16, 2 November 2009 (UTC)
2. SlimVirgin ^{talk|contribs} 23:21, 2 November 2009 (UTC)
3. Not exactly by the policies/guidelines themselves. They are just there to speed up the process of finding consensus. Where there is a consensus before people even start to think about guidelines, the guidelines are not needed. Hans Adler 23:27, 2 November 2009 (UTC)
4. Agree with Hans. Crum375 (talk) 23:32, 2 November 2009 (UTC)
5. Durova³⁵⁵ 23:34, 2 November 2009 (UTC)
6. I see no reason to specify an exemption. Syllogisms and other logical deductions are no different than any other statements in Wikipedia... If a syllogism or other logical deduction is created by an editor, and has never appeared in a source it is original to Wikipedia... the very dfinition of Original research.Blueboar (talk) 23:39, 2 November 2009 (UTC)
7. There is no evidence that this proposal would improve WP. Angryapathy (talk) 00:39, 3 November 2009 (UTC)
8. Finell (Talk) 01:37, 3 November 2009 (UTC)
9. I suggest that the proposal is, at best, not yet ripe for adoption. Perhaps a number of us can intuitively feel that there is a problem to be solved. But the drafting of any proposed solution probably needs much more consideration. Otherwise there is a risk it would turn out to be a two-edged tool, and even an obstacle against improving Wikipedia -- especially when, not if, the wikilawyers and other disruptive editors learn how to exploit it. Terry0051 (talk) 23:02, 6 November 2009 (UTC)
10. No valid reason given for special exemptions to the policy. Jayjg ^(talk) 22:26, 8 November 2009 (UTC)

Table the matter

As it appears that the discussion here has largely run its course, it may be worthwhile to table the matter for the time being. New proposals regarding syllogisms should not be brought forward again for a period of not less than 12 months.

Yes, the discussion is essentially complete. We can take at least a 12-month break from this issue.

1. TenOfAllTrades(talk) 23:16, 2 November 2009 (UTC)
2. SlimVirgin ^{talk|contribs} 23:21, 2 November 2009 (UTC)
3. Hans Adler 23:28, 2 November 2009 (UTC)
4. Crum375 (talk) 23:32, 2 November 2009 (UTC)
5. Durova³⁵⁵ 23:34, 2 November 2009 (UTC)
6. It is clear that there is strong opposition to any overt "exemption" for syllogisms... to continue to float proposals in this area is simply an example of WP:IDIDNTHERETHAT and is waisting everyone's time.Blueboar (talk) 23:39, 2 November 2009 (UTC)

   And we should mention that "table the matter" means to end the discussion (for those in the UK). Angryapathy (talk) 00:41, 3 November 2009 (UTC) While I do think that the proposal is dead in the water, and that Brews has been beating a dead horse (he mentioned that we were at an impasse, then proceeded to start a new thread with the same proposal, slightly reworded), on second read I think the "12 month" tabling seems a little silly. While I doubt consensus is suddenly going to flip in one year, this section is unnecessary. Angryapathy (talk) 14:46, 3 November 2009 (UTC)

7. Finell (Talk) 01:39, 3 November 2009 (UTC)
8. All done here. Jayjg ^(talk) 22:26, 8 November 2009 (UTC)

No, there is additional important ground which much be discussed on this issue. (Describe briefly.)

1. To place a time limit before further discussion of this topic is permitted seems bizarre - it simply stifles discussion, and not only for the present participants. If some editors do not wish to discuss this topic further, they have a simple, obvious option: don't discuss it. Go ahead and bring up other topics on different threads. There is no basis for preventing those that do wish to discuss from continuing to do so. If there is no interest, the thread will die all by itself. Brews ohare (talk) 23:35, 2 November 2009 (UTC)
2. As I explained above, there are problems that need to be discussed. There mere fact that the problems with the polcies usually do lead to edit wars is not a good reason to ignore them. Indeed, as I explained above, the lack of such conflict between editors is the very reason why such a problem could arise. Count Iblis (talk) 01:48, 3 November 2009 (UTC)
3. I made the personal decision not to continue the discussion because I didn't think those opposed could be swayed. I think everyone should make that decision for themselves. If you would like to end discussion, try encouraging the editor to form a serious proposal and then take a poll on whether or not to include it in policy. --Bob K31416 (talk) 02:47, 3 November 2009 (UTC)
4. The additional point (which has only emerged tangentially so far) that needs discussing is the problem of editors inappropriately using WP:NOR to block improvements. Discussions of valid updates are sidelined and dismissed out of hand; the current formulation of WP:NOR lends itself to such blocking behaviour (e.g. in the vitamin D example I cited above, where the response here was that this wasn't really a violation of OR, yet the update was blocked). Rewording OR would make this inappropriate blocking behaviour more difficult. This aspect of the discussion has not run its course. --Michael C. Price ^talk 11:02, 3 November 2009 (UTC)

Comment - Please note this excerpt from WP:Consensus:

"polls are regarded as structured discussions rather than votes. Both during polls and discussions, opinion has more weight when you provide a rationale."

Please note that when I posted this message, one side on these issues is just voting, without giving their personal rationale for their opinion, which reduces the weight of those opinions, according to WP:Consensus. --Bob K31416 (talk) 12:05, 3 November 2009 (UTC)

There are plenty of comments above this straw poll from other editors regarding the opposition to the proposal. Angryapathy (talk) 14:38, 3 November 2009 (UTC)

Not to put too fine a point on it, but the reason I'm suggesting that it might be time to let this discussion go is that it has become repetitive, and the participants positions have solidified. Repeating the same arguments over and over again is precisely what I'm hoping to avoid. Editors who don't agree with the straw poll options exactly as written seem to have had no difficulty expressing nuanced meaning.

The point which I find most telling is that no one has offered any affirmative response to the first poll question — there still isn't any clear expression of what problem we're trying to fix. The 'clear-cut' examples in the earlier discussions have, on reflection, turned out to be somewhat ambiguous, and examination of those cases in light of WP:NOR, WP:SYNTH, WP:V, and other policies has been worthwhile. Attempting to stifle article-space discussions (which, after all, is the aim of these proposals) by creating a specific loophole in WP:NOR seems likely to be counterproductive. TenOfAllTrades(talk) 14:49, 3 November 2009 (UTC)

(ec with several comments) I think I can speak for most of the others here: The reason we did not give rationales is that we are sick and tired of this nonsense. No matter what the policy says, it will be abused by one way or the other. There is no magic bullet version of the text that solves all problems. A number of very experienced editors (meaning they have spent hours if not days fighting in both directions, depending on whether someone was trying to push their original research, or whether someone was trying to censor well-sourced content) think that the current text should not be changed. A smaller number of less experienced editors, some of whom are in current disputes relating to this policy or had such disputes recently, disagree. After about 250 comments (100 KB) it became clear that you are not going to stop. Perhaps in the case of some of you that is exactly what happened in your other conflicts for which you now want to change this policy; perhaps you are simply not able to accept that sometimes consensus is against you. In any case I think I can be forgiven if I think that if everybody who disagrees with you just gradually dropped out of this discussion to stop wasting their time, you would simply claim a consensus to change the policy and then a new round would start. So all those who are tired of this discussion had to leave at the same time and make it clear that they continue to disagree. That's the main purpose of this poll. Hans Adler 14:51, 3 November 2009 (UTC)

Another comment: This poll seems fundamentally flawed to me in that it rolls together two distinct points in a way that artificially discourages separation again. The first part seems agreeable to me - it should be clear that the requisite consensus for this proposal is not going to be reached and therefore it is obviously dead in the water. However, that position is inextricably linked to to the 12 month moratorium in the poll. In my view that is fundamentally improper since it leverage support for the first part to create false support for the second part which had received no prior discussion before the poll. CrispMuncher (talk) 17:28, 3 November 2009 (UTC)

Is it a good idea to block discussion for one year

A motion has been made on Wikipedia talk:No original research to prevent further discussion of a topic by setting a time limit of one year before anyone is allowed to bring up the subject. This action appears to me to be a violation of the spirit of WP, as well as being unnecessary. What is your opinion? Brews ohare (talk) 00:52, 3 November 2009 (UTC)

Reason for request

To place a time limit before further discussion of this topic is permitted seems bizarre - it simply stifles discussion, and not only for the present participants. If some editors do not wish to discuss this topic further, they have a simple, obvious option: don't discuss it. Go ahead and bring up other topics on different threads. There is no basis for preventing those that do wish to discuss from continuing to do so. If there is no interest, the thread will die all by itself. Brews ohare (talk) 00:52, 3 November 2009 (UTC)

Comments

This is ridiculous. The straw poll was an attempt to prevent further time-wasting. Now you are wasting even more editors' time by interpreting it as a formal decision. What the outcome of the straw poll really means is that most participants are going to walk away now, and even if you find one or two more editors who agree with you, you can't claim there is a sufficient consensus to change the policy without getting wider input first. Hans Adler 00:57, 3 November 2009 (UTC)

Per Hans Adler. Durova³⁵⁵ 01:05, 3 November 2009 (UTC)

Hans & Durova: I filed this RfC at the suggestion of TenOfAllTrades. There is no suggestion that this is a formal decision, but that it is a motion under discussion. We'll see if others agree that the idea is simply one of defining how many editors constitute consensus, or whether the idea is to have no more discussion, period. Brews ohare (talk) 01:08, 3 November 2009 (UTC)

My suggestion to you, Brews, was that you add a RfC for the existing poll, rather than to start yet another discussion about whether or not we can have a discussion about finally reaching some sort of resolution for this interminable discussion. I advised you that opening an RfC would be a good way to bring in neutral parties, in lieu of the selective canvassing you had been doing. I assure you that my intent wasn't to encourage you to soapbox on this talk page. I'll further note that while I did allude to my opinion in the prefacing remarks, my straw poll attempted to present the options neutrally. Your RfC singularly fails to do so. Finally, there would be no attempt to limit the discussion (so we can all get some peace and quiet) if you dind't keep bringing this issue up again and again despite a failure to attract consensus for a change. TenOfAllTrades(talk) 01:34, 3 November 2009 (UTC)

Thanks for the good laugh everyone. A small group of editors trying to restrict discussion on a subject they don't agree with by all the other editors on a talk page for a year! LOL! Looks like a mass block of editors is coming if those editors try to enforce it. --Bob K31416 (talk) 01:24, 3 November 2009 (UTC)

See strawman argument and push poll. Durova³⁵⁵ 01:25, 3 November 2009 (UTC)

Whoops. I misunderstood the type of poll in the previous section.[13] Sorry. --Bob K31416 (talk) 02:38, 3 November 2009 (UTC)

Thanks for the strikethrough; returning the favor. Durova³⁵⁵ 02:54, 3 November 2009 (UTC)

When I put forward the idea of closing discussions per consensus in the ARBcom case about the problems on the speed of light page, this was dismissed because it was not grounded in wiki policies. But in that case, you had very long discusssions that were seen to be a problem (not by me, but by most other editors). So, I am sympathetic to the idea that you could do this. However, in this case, the discussions have barely started. While a lot has been written, it has been a "rapid fire" discussion. If you visit this page every few days, the discussion could have been declared "over" before you would have had a chance to take your time to think things over and give your opinion.

Perhaps we should accept that this can be long discussion that can take some time and act accordingly (i.e. slow down, there is no need to reply immediately, posting once per day is more than enough). Count Iblis (talk) 02:02, 3 November 2009 (UTC)

There is nothing to comment on. Brews asks for comment on a "motion". There was no motion, and I am not aware that Wikipedia has a procedure or practice for motions on a policy talk page. The preceding section was very clearly labeled a straw poll. Brews misinterpreted that straw poll as a "motion". More needless drama. —Finell (Talk) 03:43, 3 November 2009 (UTC)

Finell: On the contrary, there is no misinterpretation of the straw pole. This pole's aim is to browbeat the responders who do not favor limiting discussion. So obtaining a wide response to this tendentious proposal is important. For this reason, an RfC was suggested by TenOfAllTrades, apparently a suggestion not greatly favored by those contributing to the pole so far. Brews ohare (talk) 07:47, 3 November 2009 (UTC)

It seems you missed the comment by TenOfAllTrades above, where he explains that you misunderstood him. But let's just wait for the feedback to this RfC. Hans Adler 10:22, 3 November 2009 (UTC)

Saying that a year must pass before further discussion takes place may be wrong... but so is continuing to push for discussion of an idea that clearly does not have consensus. It smacks of WP:IDIDNTHERETHAT. Blueboar (talk) 15:02, 3 November 2009 (UTC)

To build on Blueboar's comment, Brews stated, "The subject is at an impasse," and then about a hour later started a new thread on a slightly modified proposal. Blueboar is right, the 12 month thing is not right, but the intent was to prevent someone from started a new thread as soon as the discussion was finished. Angryapathy (talk) 15:10, 3 November 2009 (UTC)

A(nother) query

I have been accused of various nasty and unpleasant things as a result of my attempt above to assess the sense of the community on where (or whether) these proposals are going. Everyone (I hope) can agree that it's unnecessary, wasteful, and distracting to encourage or allow a policy proposal to drag on forever with no resolution. There must be an eventual end to the process. So now I'm going to ask the proponents of this proposal:

• Under what conditions would you accept that this proposal is not going to become part of WP:NOR? or, alternatively,
• How and when is a final decision to be reached?

I'm looking for a fairly specific framework here. 'After the discussion concludes' is not going to cut it. There's been significant input from a number of experienced, long-term Wikipedia editors. While I hesitate to draw policy conclusions from a counting of heads, I also don't think that policy change should be brought about by war of attrition. It appears that a substantial majority of participating editors are not persuaded that the proponents of these changes have made their case effectively. Moreover, many editors have made comments to indicate that they're not interested in expending much more effort in a futile endeavour. I have a suspicion that the patience of many editors has already been exhausted, and absent a clear indication that proponents are willing – or at least able – to accept a negative decision then those editors will disengage and there will be no change to the policy regardless of how much more text is spilled on this talk page. Without clear and reasonable answers to the above question(s), I know that I can see no further reason to discuss this issue. TenOfAllTrades(talk) 15:17, 3 November 2009 (UTC)

• I'd say it's clear there's no consensus to add this to the policy. Not one example of a simple syllogism has been offered where it was excluded from an article on the grounds of this policy. Until we have some solid examples, there's no reason to believe the policy is causing a problem in that regard, and we don't extend policies without good reason. There's therefore nothing to discuss until we have some clear examples. SlimVirgin ^{talk|contribs} 15:31, 3 November 2009 (UTC)

(edit conflict)TenOfAllTrades, Seems like you're not reading and understanding my comments. I pulled out of advocating my proposal and I think I was clear when I mentioned that in the section No movement towards including the proposal. I repeated that in response to your poll Support for changes. Also, I suggested in my response to your poll Table the matter how you might end the discussion with other editors. You seemed to be ignoring these remarks of mine. Please note this excerpt from Straw poll, "Straw polls provide important interactive dialogue..."

Frankly, I try to avoid prolonging discussion when I don't feel someone is thinking rationally or fairly. I probably should have done that much sooner in the recent episodes of discussion over the last few days. --Bob K31416 (talk) 15:46, 3 November 2009 (UTC)

Yes, thank you for showing good sense. I think the escalation may have happened because a lot of editors felt this was good timing because it was really time to stop the discussion, and then someone else tried to reanimate it. That's certainly the point where I lost hope that it's ever going to end. But it would be very unfair to blame this on you. Hans Adler 16:15, 3 November 2009 (UTC)

A review of the proposal history

It is expected that a proposal for possible changes to WP:NOR will undergo some evolution as discussion proceeds and editors express their views. The present form of the proposed change is that below:

Routine calculations

---

This policy does not forbid routine calculations, such as adding numbers, converting units, or calculating a person's age, provided editors agree that the arithmetic and its application correctly reflect the information published by the sources from which it is derived. Likewise, simple syllogisms may be used that are logical equivalents of routine mathematical calculations, as described by Russell and Whitehead. Care must be taken to observe no synthesis takes place.

Examination of this form of the proposal was interrupted by the posing of a straw poll, and a resulting discussion of procedure instead of the proposal. I'd suggest the proposal itself be discussed further. Below are some comments and responses from earlier discussion:

1. It is my understanding of many comments on earlier versions of this proposal that some editors believe the syllogism to be a debatable concept that requires sourcing for each syllogism at each occurrence. That notion is met by the above proposal by pointing out that what is allowed is homomorphic with the "routine calculations" already allowed. The proposed change is therefore one of making explicit what already is agreed to in the present form of WP:NOR.

2. Some comments express concern that making the acceptance of syllogisms clear will encourage extended debate with contributors that presently can be abrogated by simply stating that reversion of certain material is WP:OR, when the real reason is that the material is illogical (violates a syllogism). I'd suggest that the real reason be given in such cases: the logic is faulty. The above clarification makes clear the applicability of WP:NOR to cases where syllogisms are not used, leading to illogic.

3. Some comments suggest that making the acceptance of syllogisms clear will encourage contributions based upon false premises, and it will be more difficult to persuade authors that their premises are false. I'd suggest that the argument over premises is not made easier by dragging in an inapplicable argument that the logic is false. The correct use of WP:NOR in such cases applies directly to the formation of the premises, and the proposed clarification above has nothing to do with such debates over premises.

For this discussion to proceed, I'd suggest those who are opposed to the clarification above re-formulate their views taking the above three points into consideration. Brews ohare (talk) 17:29, 3 November 2009 (UTC)

Is this deja vu? Are we actually starting all over again? This is getting absolutlely ridiculous. Angryapathy (talk) 17:34, 3 November 2009 (UTC)

Angryapathy: The suggestion here is not to start all over again, but to build upon what has been done, and to make objections clearly address the issues now that they are clearly framed. Brews ohare (talk) 17:36, 3 November 2009 (UTC)

It is clear from previous comments and the straw poll that the consensus is that there is no need to include syllogisms in NOR. You can reword it all you like, the jist of the proposal is what is opposed. I'd rehash the many previous arguments, but that isn't necessary. You have provided no real examples of syllogisms. Angryapathy (talk) 17:42, 3 November 2009 (UTC)

Angryapathy: It is not clear that previous arguments still are applicable to the present formulation. Perhaps you could indicate your view of the points 1 - 3 as they apply to the new formulation of the proposal? Brews ohare (talk) 18:01, 3 November 2009 (UTC)

(belatedly interjecting) This is exactly what it was like with Brews for about a year at Talk:Speed of light, which was the subject of the recent arbitration. He just doesn't HEAR the consensus. This is what led to the arbitration finding that he edited disruptively. —Finell (Talk) 18:14, 3 November 2009 (UTC)

I was curious what Brews was trying to get into that article. Could you direct me to what he was trying to include? --Bob K31416 (talk) 18:19, 3 November 2009 (UTC)

It had to do with a particular interpretation, or giving additional emphasis to, the relationship between the speed of light and the definition of the metre in 1983. Look at the talk page's edit history from around March to November 2009. User:David Tombe joined in around July. —Finell (Talk) 18:34, 3 November 2009 (UTC)

Reply to Bob K31416: Please see here for details. Brews ohare (talk) 19:49, 3 November 2009 (UTC)

Thanks both of you! As far as the physics is concerned, I think I got the idea of the problem. Because it's a subtle issue it can cause problems in communicating and one or both sides may not understand the other side's point. It's a tough situation. --Bob K31416 (talk) 20:34, 3 November 2009 (UTC)

Brews, re "Likewise, simple syllogisms may be used that are logical equivalents of routine mathematical calculations, as described by Russell and Whitehead." I didn't understand what part of that page of R&W that you are using. Also, when a link to a part of a book is given and accessed a large number of times, the website might not give further access to the particular page because they would rather have people buy the book. --Bob K31416 (talk) 17:45, 3 November 2009 (UTC)

Bob K31416: Perhaps a link to symbolic logic or propositional calculus or mathematical logic would be better? I put in the link to Russell & Whitehead primarily to indicate that there is a long history connecting syllogisms and routine calculations. It is not essential. Brews ohare (talk) 17:51, 3 November 2009 (UTC)

I appreciate your attempt to clarify for the general Wikipedia editorship who might read it, but I don't think those links would work. Still not clear enough to be useful in this context, in my opinion. --Bob K31416 (talk) 18:04, 3 November 2009 (UTC)

Criticism of this section

I feel, from the consensus above, that the fundamentals of this proposal, regardless of the wording, should not be added. No amount of discussion rehashing the same arguments is going to change that fact. The issue should be dropped. Brews, please take a lesson from your Arbcom case to see that continually trying to argue the same point will not allow you to get your way. Leave the horse caracass alone. Angryapathy (talk) 18:43, 3 November 2009 (UTC)

Angryapathy: I would be happy to drop this proposal if I thought it actually had been responded to. However, an actual response would address the points 1 - 3 as stated above, which have not been addressed yet in this form. Please attempt such a response, which is not a request for a "rehash", as carefully pointed out above. Brews ohare (talk) 19:45, 3 November 2009 (UTC)

Did it ever occur to you that perhaps you have exceptional difficulty judging objectively whether your points have been responded to adequately? There are some editors here with this problem. Not many, but they do exist. It appears you are one of them. Hans Adler 19:50, 3 November 2009 (UTC)

Likewise, I'm sure. Why not attempt to address the points 1-3 (once more, in your view)? My answer to that question is that you will not respond to points 1 - 3 because they eviscerate your arguments. Brews ohare (talk) 19:53, 3 November 2009 (UTC)

No, we refuse to respond to the points because they have been answered. There is no more reason to discuss because you have not brought up any new arguments. Angryapathy (talk) 20:03, 3 November 2009 (UTC)

And rewording opponents' arguments in your own words (and thusly putting a negative spin on them) is an easy way to make it seem like you've "eviscerated" the arguments you invented for that purpose. However, it isn't working here. Angryapathy (talk) 20:06, 3 November 2009 (UTC)

Good point. I don't recall anyone saying anything remotely like 1, 2 or 3. All three are silly misunderstandings. If this is how what we say arrives in Brews' mind it makes sense that he feels we are not responding adequately. But that's not our fault, since the communication breakdown is clearly on his side. Hans Adler 20:12, 3 November 2009 (UTC)

My objection to adding the proposed language is that it encourages original thinking... and making Wikipedia the primary place of publication for conclusionary statements (the core issue behind NOR). The proposal allows editors to add thoughts and conclusions based on their own syllogistic reasoning rather than that of a reliable source. I don't care how valid, in valid or logical the syllogism is... The reason why an exception should not be made is because the syllogism originates with a Wikipedia editor. It is ORIGINAL research. We should not include conclusions or thoughts that originate with an editor. We should only do so if the syllogism originates with a reliable source that we can cite.

The reason why we allow routine numerical calculation is not because they are simple; it is because the algorithms involved in performing these calculations are not original research... they can be found in basic math text books and do not originate with the editor who does the calculation. The formula for converting Centigrade to Fahrenheit, for example, is well known and could be cited if someone challenged the calculation, we simply agree that there is no need to cite a basic grade school level text book every time you perform the calculation.

In other words... for me, the issue is (and always will be) the originality of the material: ie who has reached the conclusion stated?... not what the conclusion is, nor how logical the conclusion is. It is a matter of staying true to the basic intent of WP:NOR. Blueboar (talk) 20:49, 3 November 2009 (UTC)

Originality in the sense of a syllogism is not in any way about "who" said it. It is about the implications of premises, which cannot be avoided in any way, regardless of who says it. Only the premises have to be valid; that requirement is different from the syllogistic reasoning. Brews ohare (talk) 23:45, 3 November 2009 (UTC)

Originality in the sense of this policy is all about "who" said it. That is the problem. Blueboar (talk) 00:12, 4 November 2009 (UTC)

I have thought about this quite a bit. I think that the proper advice is that syllogistic reasoning is permitted in practice when the narrative it creates is not materially different from the narrative in the literature. However it is very difficult to express this in a clear and precise way, as would be required to try to convince people to add it to a policy document. — Carl (CBM · talk) 21:22, 3 November 2009 (UTC)

Syllogistic reasoning does not "create a narrative": what it does is make statements that are ineluctable conclusions based upon the premises. Given the premises (that is the big if) the result of a syllogism cannot be avoided. Thus, given A > B, inevitably B < A. That is the kind of thing being talked about. Brews ohare (talk) 23:45, 3 November 2009 (UTC)

Every new text creates a new narrative; the question relevant to us is how similar that narrative is to the existing literature. This point is one reason it is difficult to express my point succinctly in a way that is immediately obvious. — Carl (CBM · talk) 02:07, 4 November 2009 (UTC)

Summary

Due to an inability of several editors here to grasp what a syllogism is, and because they are continually arguing about matters totally unrelated to the proposal and to syllogisms; and due to their inability to see that syllogisms already are accepted by WP:NOR as it is presently written, this discussion cannot proceed. Unfortunately, should a future contribution employ a syllogism, and should the contributor claim that WP:NOR is not a valid reason to reject it because only a syllogism was used, exactly this same unending argument will recur for exactly the same invalid reasons. Brews ohare (talk) 23:45, 3 November 2009 (UTC)

You really don't understand the concept behind WP:NOR do you... if I am the first person to draw a specific conclusion based on given premises, and I include my conclusion in Wikipedia, then that conclusion is original to Wikipedia. It is OR by Wikipedia's definitions. It does not matter if the the premises are valid or not... it does not matter if my logic is sound... what matters is that I am the first person to draw the conclusion and that I chose Wikipedia to publish it. In other words... while saying "A>B thus B<A" may be a valid syllogism and flawless logic, but the fact remains that it would still be original research for me say this in a Wikipedia article... unless it were stated in some reliable source. If I am the first person to draw the syllogistic conclusion, even though it is obvious, it remains OR. Blueboar (talk) 00:10, 4 November 2009 (UTC)

I think that if you try to define what you mean by "specific conclusion" you will find it very difficult. Do you mean the literal sequence of words? The idea underneath the words? — Carl (CBM · talk) 02:16, 4 November 2009 (UTC)

I think this illustrates very well the need to have a different OR policy for technical scientific articles as I proposed in the section below. In practice these sorts of logical deductions are never a problem if the topic is, say, classical mechanics. To the contrary, if you can explain something better by using some logical deduction then that is always a good thing, regardless of whether or not that logical deduction can be found in any textbook at all. If we still call that OR, then it has to be made clear that this sort of OR is ok. Count Iblis (talk) 02:22, 4 November 2009 (UTC)

In practice we use such deductions all the time, so I don't see that there is much of an argument that this policy actually prevents them. On the other hand, there are also many original derivations that we would not accept, and it seems very difficult to explain in a general way which ones are OK. — Carl (CBM · talk) 02:25, 4 November 2009 (UTC)

Different policies for science articles?

Since there is lack of consensus to modify the policy w.r.t. logical deductions, I think that there is something else one can do. Editors on any page can always decide by consensus to ignore certain policies if they don't work in their opinion. Therefore, what we can do is write up alternative versions of the core wiki policies and then editors on any page can decide by consensus to stick to these alternative policies instead of sticking to the official policies.

A good example of this is the practice on the Global Warming page to only allow in peer reviewed sources for scientific claims. This practice was ruled to be in violation of the Reliable Sources policies in a debate on the Reliable Sources Noticeboard. There was a strong consensus there in favor of allowing in an edit citing the BBC saying that the Earth hasn't warmed in the last decade. However, the editors on the Global Warming page rejected this ruling and they stuck to their own rule regarding reliable sources. Count Iblis (talk) 19:13, 3 November 2009 (UTC)

Um, not familiar with the example there but that seems to go against the general principle that local consensus cannot trump policy. See can o' worms no. 2 from my objections above. Remarkably easy to game such a notion: get a few dozen Stephen Colbert fans together on Talk:Elephant and suddenly WP:RS takes a strange permutation that allows the population of elephants to triple within six months. Durova³⁵⁵ 19:28, 3 November 2009 (UTC)

Count, editors aren't allowed on any given page to ignore the core content policies. SlimVirgin ^{talk|contribs} 20:56, 3 November 2009 (UTC)

On the other hand... Local consensus can always invoke WP:IAR if people honestly believe that ignoring a policy or guideline in a specific situation is in the best interest of Wikipedia. Just note that litigimate invocations of IAR are rare... You can expect to be challenged, and will have to make a very good case for it at the article talk page... but it can be (and has been) done.

We understand that there may be an exception to every policy or guideline we write... we created the IAR policy to deal with those exceptions specifically so we didn't have to go and rewrite each policy and guideline every time we came upon one.

Not saying that invoking IAR would be appropriate in the Global Warming page (nor that this is what occured)... just pointing out the fact that we actually have a rule that covers when to ignore the rules. Blueboar (talk) 21:05, 3 November 2009 (UTC)

In practice, the way that technical articles are edited (science and math among them) is indeed different than the way that humanities and history articles are edited, which is in turn slightly different than current events articles are edited. In particular, there is a much more frequent use of uncontroversial reasoning, axiomatic deduction, and writing examples in science and math articles than in current events articles.

No one-size-fits-all policy is going to describe all these practices correctly, and so it is no surprise if the present policy wording describes one set of articles better than another. The current wording of NOR is more focused on humanities and current events than on science and math. But I don't think it's worth bothering to try to change the wording here, since we manage to deal with everything in practice despite the poor wording. For most purposes the wording here is good enough, when accompanied by a knowledge of the general practice in articles on the fields one edits. — Carl (CBM · talk) 21:15, 3 November 2009 (UTC)

I don't accept that the current policy is best suited to particular classes of articles than to others: the principles are general ones rather than detailed specifics. I think having multiple set of rules in operation concurrently is only going to fragment the project and be a Wikilawyer's paradise. In any case, how would it be determined which set of rules applies to a particular article? Since we are discussing science articles specifically here it would seem natural that the elements would fall under the remit for science articles. Usually this would be fine but consider gold for instance, which in addition to its status as an element has significant cultural and economic importance beyond its scientific description. How do you decide just which rules apply to it? CrispMuncher (talk) 22:01, 3 November 2009 (UTC)

I agree that there is no need to codify different rules. The present NOR text is intended to be very general, and mostly succeeds, except that different people interpret particular sections, such as the one on logical deductions, very differently. Anyway, the way we decide what to do on each particular article is by discussing it on the talk page. But in general the pattern is clear if you are familiar with other articles in the same field. — Carl (CBM · talk) 22:04, 3 November 2009 (UTC)

It should be pointed out here than much normal science consists of nothing more than drawing new, logical deductions from well-accepted statements. (My two unpublished dissertations did nothing but this.) This is original research. One problem is that deduction from both accepted observations or accepted theoretical statements can be wrong - it may conflict with new observation. In this case, either the observations are replicated or the theory is modified or discarded for a better one. This is how science grows. Scientist do miss important deductions, but such deductions are original research. Geologist (talk) 00:12, 4 November 2009 (UTC)

Actually, that is a very good point... it isn't that the policy applies to some topic areas and not to others... but different topic areas interpret it in slightly different ways, because their subject matter is different. Blueboar (talk) 00:15, 4 November 2009 (UTC)

In science, all logical deductions that are not explicitly stated in the literature are original research. The key prohibition is Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources. However, multiple sources aren't always needed. One paper I read (a single source), contained a conclusion that could have been far more important if a simple syllogism had been seen. (I wrote the author, but received no reply. Perhaps he thought I wanted to co-publish. :-) Scientists miss little. If it isn't explicitly stated in any of the sources, it's probably original research.

Implicit in the job description of good teachers is to formulate clear, illuminating explanations of phenomena, using theories. The above prohibition of original statements likely discourages some from writing here. However, I see no way around it. Distinguishing a new syllogism, when explaining, from original research is subjective. Geologist (talk) 01:52, 4 November 2009 (UTC)

Unless our articles are just a sequence of direct quotes, we cannot merely repeat what has been literally stated by sources. So we have to (further) interpret and rephrase the points made. The question is, how far can we go? That is where practice differs somewhat. For example, in my field, some texts use "computable function" and some use "recursive function" for the same thing. But in other texts, "recursive function" can mean several other things. So we have to interpret and merge what the sources say, as an inherent part of writing an encyclopedia article (and all such writing is interpretation). There is actually a very well established set of acceptable practices for how to get information out of specialist texts into our articles, it's just that those practices are hard to describe briefly, and they may be less applicable to articles in very different fields. — Carl (CBM · talk) 02:11, 4 November 2009 (UTC)

When a post-graduate student, I was asked by others to teach a complicated procedure to them. I glanced at the definitive presentation, but misread it. My presentation was correct, but it was a simple algorithm. (The treatise was not.) I later published it in 'Journal of Geological Education', for the only other place for it, 'Studies for Students' in the Journal of Geology had been discontinued. The point is that experts likely have presentations & explanations better than those published, for there is no place to publish lectures. These explanations must be considered original research, discouraging some faculty from writing articles. Safer, I agree, to discourage such explanations (sigh). However, researchers can also not write articles on their most understood subjects, because they would necessarily reference themselves: this could be loosened a bit, though only other experts could evaluate the neutrality of one's POV. Here I was drawing attention to a more fundamental problem of admitting deduction but prohibiting original ideas.

English is more ambiguous than most languages, so articles published in all fields are written with great care. Our interpretations of the writer's thoughts are subjective, but practice & expertise allow us to understand and thus express the writer's original thoughts, not our own. (An historical perspective is essential, I believe.) Permitting deductive reasoning without restriction can allow one to add new, original thoughts to a subject. I doubt anyone objects to deduction, if the results are old ideas that have been evaluated in the literature (and this is acceptable): it's the new thoughts that have not been expressed before by anyone that comprise original research (and this is not acceptable). It's not the words: it's the ideas that matter. This distinction should not prove a problem, I agree. (I should be interesting in 'how to get information out of specialist texts', for this is one of the goals of a postgraduate degree in that specialized field. Only then can a student be let loose.-) Geologist (talk) 07:33, 4 November 2009 (UTC)

I am surprised that WP:ESCA hasn't been mentioned here: it suggests that premises and conclusions be sourced, and the connection can be filled in with logical reasoning that is not necessarily sourced. Brews ohare (talk) 16:49, 4 November 2009 (UTC)

As for the universality of WP:NOR, the lengthy discussion above regarding simply making explicit the allowance of simple syllogisms shows that WP:NOR will be applied by different editors differently, some using WP:NOR to exclude syllogisms in some cases and some not. Where differences of view occur, no resolution will be possible, although those denying the inclusion of statements built upon syllogism actually will be misapplying WP:NOR, denying what it already includes.

An example is found here where a prolonged debate over the use of classical vacuum as interchangeable with free space was settled only by finding a verbatim quote to this effect, despite the use in texts of either term (but very rarely both in the same text) to describe the reference state of classical electrodynamics. This is an example of A = C in some sources and B = C in others, but difficulty in finding a verbatim source saying A = B. Brews ohare (talk) 16:54, 4 November 2009 (UTC)

I hope that the Admins reviewing this paragraph to determine whether it is a topic-ban violation, will realize the physics topic itself is not being discussed here; it is being used only as an example of the type of quandary under discussion here. Brews ohare (talk) 17:21, 4 November 2009 (UTC)

It looks like the issue there is that there was a disagreement over whether the terms are equivalent. If two terms really do mean the same thing, one can usually get several editors to agree to that on the talk page, and that settles the matter. But there were only two editors in the discussion, so if they don't trust each other then of course some sort of reference is going to be necessary. This is one reason it is often better to ask for other members of the relevant wikiproject to comment. Having numerous knowledgable people comment is also the best way to resolve spurious applications of WP:V (but I have not researched the linked discussion enough to tell if that three really were any verifiability concerns there). — Carl (CBM · talk) 18:36, 4 November 2009 (UTC)

Yes, and it's the reason we ask for sources for anything that's challenged. Even if editors feel supplying a source is unnecessary, it's faster to do that than to argue. SlimVirgin ^{talk|contribs} 18:44, 4 November 2009 (UTC)

Yes, of course. The idea that we would weaken WP:NOR for science articles is a non-starter. Science is one of the subject areas where the problem of original research has been most acute, either because of fringe advocacy or because some editors try to push their own analyses into articles. As a practical matter, no issue arises unless an editor contributes disputed content. Where content is disputed, requiring sources is the way to avoid lengthy arguments (unless an editor argues on and on and refuses to get the point). Speaking of which, why are we still discussing this? Consensus is clear, and there is practically no support for any of the proposals to change this policy. Can we please close this discussion and get back to productive work? —Finell (Talk) 19:18, 4 November 2009 (UTC)

Well, Finell's point is that I am a pain in the ass. That may be so. SlimVirgin's view is: Why argue if you can find a source? In this example, finding a source that said verbatim what was literally equivalent to accepting the syllogism took several hours. CBM's view is that the way to go was to involve more editors, which naturally means an RfC to avoid charges of canvassing. That might resolve the matter in three or four days (frankly, I doubt it). However, the objecting editor in this case agreed that some sources said A = C and some said B = C. Nonetheless, the claim was that WP:SYN was at work in stating A = B, and that WP:V required a verbatim statement from a single source saying in so many words that A = B. Apparently, Finell, SlimVirgin and CBM agree that such obstruction is a "good thing". I choose to think it is nonsense, and in fact, unnecessary under the present form of WP:NOR. Brews ohare (talk) 19:48, 4 November 2009 (UTC)

Posting at the relevant wikiproject talk page is not going to be construed as canvassing. I said that I do not know if the issue there was merely formal (just about about WP:V) or had some substance with the topic at hand, and I would need to do a lot of work to determine it. But experts in the field, which I am sure we have plenty of for physics, would be able to tell pretty easily if the matter was truly routine. I'm just a mathematician. And, I have little sympathy with formal pushing of WP:V when it does not actually benefit the article. — Carl (CBM · talk) 20:10, 4 November 2009 (UTC)

Can we close this discussion? Absolutely not! When did I start this thread? 19:13, 3 November 2009 (UTC). And now it is 19:47, 4 November 2009 (UTC), which is just one day later! I have a few things to say here regarding how to proceed, taking into account the replies by others. But I can only do that later today or tomorrow. Count Iblis (talk) 19:46, 4 November 2009 (UTC)

Yes... this specific thread was started a day ago... but we have been going on and on about the issue for almost a week, and it is clear that it is unlikely that a consensus to add a section on syllogisms will form. There comes a point when we have to say ... Enough is enough. Blueboar (talk) 20:28, 4 November 2009 (UTC)

Indeed, there is no consensus to add such a section or to make other similar changes. That's the whole point of this thread. I suggested that one could create unofficial policy pages that do not have the consensus to become official policies. SlimVirgin wrote that this would be illegal, Durova wrote that this may lead to other editors overturning a local consensus on a page to adhere to the alternative policies. You wrote that WP:IAR could be invoked and that the fact that this rule exists means that we don't need to bother.

I think that creating alternative policies that some editors would want to adhere to on some articles would be a good thing or Wikipedia, not just for the articles in which they would be used. It would lead to competition between different policies. The consensus here and on the other oficial policy pages won't stop other policies from becoming de-facto polcies if the alternative policies are seen to be working better. This may lead to a split on Wikipedia. But you can also imagine that the consensus here could change if the unofficial policies become successful. Count Iblis (talk) 20:55, 4 November 2009 (UTC)

Competing policies?... a good thing?... I could not disagree more. I think it would be a very very BAD thing, and would oppose it strongly. I hope you are joking. Blueboar (talk) 21:07, 4 November 2009 (UTC)

They would not compete with each other like two boxers compete. They would compete simply because editors can choose by consensus to adhere to the alternative unofficial policies or the official policies. Per IAR that is possible. Count Iblis (talk) 22:50, 4 November 2009 (UTC)

Count, IAR is used in extraordinary circumstances, to cover unique situations which the policies and guidelines didn't foresee. It is not a tool to dismantle or ignore existing rules in generic situations, such as the application of NOR or V to science articles. Having "competing" policies would cause chaos and complete anarchy. We got to where we are by building sound policies and generally conforming to them, not by going berserk. Crum375 (talk) 23:42, 4 November 2009 (UTC)

I don't forsee any chaos at all. What I suggest is to more or less build on WP:ESCA. Thing is that while the issue of simple logical deductions is seen to be too controversial to even mention here, on science pages they are essential. So, people do it anyway. But this can lead to terrible mistakes, I pointed out a few on the talk page of WP:ESCA to motivate why these guidelines are necessary. If you look at the recent reversions of WP:ESCA, you see that there is conflict between a phrase on the basis if this NOR policy. But the whole point of the guidelines is to prevent errors, not to create trouble. Count Iblis (talk) 00:29, 5 November 2009 (UTC)

Which is why you shouldn't be promoting, in ESCA, something which directly contradicts policy. Jayjg ^(talk) 01:39, 5 November 2009 (UTC)

Either the policies are modified to take into account the way some scientific articles are edited or we need to have alternative policies. It is similar to a policy that only talks about abstinence. The policy says: "To prevent STD, observe abstinence". What I'm saying is that this is not realistic, we do also need to talk about using condoms. Then, since that turns out to be taboo here, we have to propose alternative/parallel policies. Count Iblis (talk) 15:24, 5 November 2009 (UTC)

Count, you are repeatedly stating there is a problem with science articles, which requires a modification of NOR policy, yet I haven't seen a single example of such an article, with a clearly demonstrated 'problem'. Let us see a single science article where we cannot achieve an excellent result by reflecting the best published reliable sources, per V, NPOV and NOR. And don't say "there are many": pick just one example, and let's examine it closely. Crum375 (talk) 15:35, 5 November 2009 (UTC)

let's look at this core thermodynamics article in the state it was in on 26 July 2006

the same article in the state it was in on 18 April 2008, almost 2 years later. For almost two years Wikipedia was saying that:

$dU-TdS+pdV\leq 0$

Islolated error? I don't think so, the error here dates back from 10 November 2005.

The point here is not that an excellent result could not have been achieved in principle using the current versions of V, NPOV and NOR. The point is that for a few years experienced editors who knew all about V, NPOV and NOR did not correct the major errors. It is no good pointing out that some reliable source could have been used to correct the errors. If you do not note the presence of a possible error, you are not going to look up things in the first place.

The old versions of the articles contained logical/mathematical deductions, so do the new versions of the article. NOR or no NOR, that's how things are. What is needed are guidelines that give advice on how to edit such articles in a way that minimizes the likelyhood of error. Count Iblis (talk) 16:51, 5 November 2009 (UTC)

I'm somewhat confused, because I don't think that NOR as presently written prohibits the sort of deductions you are talking about. As you say, deductive passages have been used in many articles both before NOR was developed and after it was developed, and so it requires a sort of myopia to think there is a conflict (also, the present text of WP:NOR is far from ideal in its explanation of the actual policy). It's hard to explain concisely which deductions are OK and which are not, but the NOR policy has never meant that every deduction must be explicitly stated in a pre-existing source. — Carl (CBM · talk) 17:08, 5 November 2009 (UTC)

Ok, but you can see in the editing history of WP:ESCA that there is some dispute about OR here. So, if you want to have guidelines that address how to avoid problems like in the two examples I gave above, then it seems that such guidelines will come in conflict with NOR. Count Iblis (talk) 17:14, 5 November 2009 (UTC)

Actually, when I read WP:ESCA I never thought it was about OR at all; most of it was about editing talk pages. Which part of the current version is about OR? — Carl (CBM · talk) 17:19, 5 November 2009 (UTC)

The problem seems to be guideline 4 part 2. SlimVirgin and Jayjg objected to a previous versions, it was reverted a few times. The current wording is new, I edited that a few hours ago. So, perhaps this has solved the problem, but we'll have to await SlimVirgin's reaction to this change. Count Iblis (talk) 17:27, 5 November 2009 (UTC)

But older versions were OK, too. My impression was that the original ESCA essay that you wrote was not in conflict with general practice or WP:NOR. At least, when I read it back then, I didn't see anything objectionable. But more recently there was an incident in which Likebox used ESCA as a justification for adding OR (or perhaps added OR in an attempt to make a test-case for ESCA; I am not sure what the priority of ideas was). My impression was that this was a misreading of ESCA, which did not say anything about ignoring the literature.

The real issue that is under discussion there is, how much explanation should we provide in our articles? That question does not have any clearly written answer, and there are disagreements over it all the time. It is very related to the question of when we should include proofs in our articles. Some people would like proofs for everything, some would like no proofs at all, and we have arrived at a reasonable middle ground in practice. — Carl (CBM · talk) 17:59, 5 November 2009 (UTC)

Count, you keep saying very emphatically that NOR is broken, at least for use in scientific articles, and I asked for a single example demonstrating the problem, showing how NOR is broken. Instead you seem to be focusing on some essay you are writing, not the NOR policy, and the example you gave above was about simple errors which were missed in a wiki article, nothing to do with NOR, AFAICT. What am I missing? Crum375 (talk) 17:43, 5 November 2009 (UTC)

Count Iblis is correct that the actual practice regarding the use of explanations in scientific articles does not agree with the most narrow reading of WP:NOR nor the most narrow reading of WP:NOT. This is a relatively well-known issue among people who edit scientific articles, we usually just ignore it. The wording of the present WP:NOR page is just not aiming at the actual OR issues that arise in technical articles. Although the policy was started to keep crackpot scientific theories away, the focus moved over time to novel polemic arguments in humanities and current events articles. — Carl (CBM · talk) 17:59, 5 November 2009 (UTC)

So would you mind showing us one such example where the current "narrow reading" of NOR fails? Thanks, Crum375 (talk) 20:59, 5 November 2009 (UTC)

It fails here, here, here and in many other places. Count Iblis (talk) 22:48, 5 November 2009 (UTC)

Re Crum375: many technical articles include "examples" to illustrate the concepts. The fact that the recently-promoted FA Euclidean algorithm includes one is evidence that these examples are not completely deprecated. However, the narrowest reading of NOR would require that all such examples be literally copied from references, which is not our actual policy regarding examples. Unfortunately the NOR page does not describe the actual policy about examples at all.

Example 2. The article on Turing degree includes this passage:

Simpson (1977) showed that the first-order theory of ${\mathcal {D}}$ in the language ⟨ ≤, = ⟩ or ⟨ ≤, ′, =⟩ is many-one equivalent to the theory of true second-order arithmetic. This indicates that the structure of ${\mathcal {D}}$ is extremely complicated.

Now, I don't have Simpson (1977) in front of me, but I doubt he uses the term "many-one equivalent" at all. The term is use here because it is more broadly understood than the more precise terminology I believe Simpson actually used. Also, the second sentence of that passage is very unlikely to be challenged, but it is also unlikely that someone has made that claim in exactly those words in print.

I have been looking at math articles here for a long time, so I am confident in my belief that such things are perfectly accepted on Wikipedia. These articles do meet the NOR policy, just not the most narrow and pedantic readings of it. But a hypothetical editor who only edits current events articles might be surprised by the way that NOR is flexible enough both to handle current events articles and to handle technical articles. — Carl (CBM · talk) 23:21, 5 November 2009 (UTC)

Carl, I looked at your examples, and read your explanations, and I have yet to see a single problem in applying the existing policies to them. You say that these sample articles "meet ... NOR policy just not the most narrow and pendantic version of it." I disagree on this key point. NOR, as well as all WP policies are not ever meant to be interpreted "in a most narrow and pedantic way". They collectively tell us that we must adhere to sources, but we may use our own words, as well as the collective common sense of the editors on the page, to summarize what the sources are saying. As long as we don't introduce bias or original material, there is no problem. Note that "synthesis" per se is only prohibited when it "advances a position", not when it helps explain a topic conformant to the relevant sources in a non-controversial manner. If we take your example above of the term "many-one equivalent", I see no problem with using modern terminology for older terms as long as no editor sees a problem. If someone does and points it out, then we can always add a footnote, for example: "X used the term A, which is equivalent to the more modern term B, per source Y." And in general, any item in an article which some editor feels requires a source (or a better source), should be provided with such source, and/or footnote explanation as needed. So as bottom line, I feel the current NOR, V and other policies are in no way impeding good scientific articles. Crum375 (talk) 00:00, 6 November 2009 (UTC)

Yes, I agree the existing policies can be applied OK. The fear that some editors of science articles have is that some editor, unfamiliar with the general editing practices in a subject area but full of fervor for the NOR policy, will arrive and browbeat editors into removing valid content solely on pedantic grounds. This fear is not entirely paranoia; I have seen similar things actually happen. They are very draining for everyone.

For an example of what I mean by pedantic interpretation, if I tell some people about the sort of footnote solution you proposed just now, they will give a knee-jerk response that I need a source that specifically refers to the paper in question (Simpson 1977), and that a source that is more general cannot be used. These editors are often motivated by issues in current events or with biographies, issues that are much less relevant to articles on technical subjects. — Carl (CBM · talk) 00:16, 6 November 2009 (UTC)

If the issue is that "some people" are excessively pedantic, that's not specific to science — you'll find such editors everywhere, who insist we dot our i's etc. I don't necessarily see that as a bad thing, because they may balance the other extreme, those who are very sloppy and don't care much for any policy or guideline. The solution, as in most such situations, is to find a reasonable middle-ground. So if someone seriously insists on a source for something you feel is "trivial", it's often better to just get that source. After all, if it's so trivial, it should be relatively easy to find it. And typically, it will take less time to find that source than to argue endlessly about it. The point is that these issues are common to all article topics, not just science. And again, all policies and guidelines have to be applied with common sense, not blindly. Crum375 (talk) 00:51, 6 November 2009 (UTC)

You misunderstood my examples. They are examples where one cannot, as far as I know, satisfy a pedantic request for sources that literally parrot what our article says. But the examples also do not violate the NOR policy. There are many things that the NOR policy permits us to put in articles but for which there is no single source that we can point to and say "look there". Most of your comment seems to address things that I have not said. However, since I have already said what I want to say, I am going to exit the conversation at this point. — Carl (CBM · talk) 04:47, 6 November 2009 (UTC)

break

The current discussion of this subject began 08:26, 4 October 2009 (UTC); this particular subsection is just the latest incarnation. Further, the prospect of ever achieving consensus for adopting an exception to WP:NOR along the lines discussed on this talk page is nil—not because we're stubborn ("resistant to change") or stupid ("an inability of several editors here to grasp what a syllogism is"), but because the policy is sound, and adding in loopholes is a bad idea. But here's a compromise: let new participants join the discussion and say what the have to say, but those of us on both sides who have said what we have to say should stop recycling the same arguments, because it just makes the discussion TLDR. And if, perchance, someone slips, let's not respond to the same arguments by the same editors, so the rest of us can get back to productive work. —Finell (Talk) 21:14, 4 November 2009 (UTC)

Exactly. The policy does not need any loopholes, not for articles on science, mathematics, pseudoscience, history, popular culture, language, politics, or anything else. Jayjg ^(talk) 01:39, 5 November 2009 (UTC)

I think the bit about "reasonable educated person without outside knowledge" kind of implies that simple logical deductions are ok, unless something can be found wrong with it. Based on the experience I recently had with the polls about Turks, Jews, and atheists, I think it would be nice if there was a link to an explanation of the ambiguity of polls included in the policy. In other words that even the simplest of comparisons is arguable depending on the way in which the study is done. AzureFury (talk | contribs) 04:41, 5 November 2009 (UTC)

Some observations:

• If a reduction is obvious and mathematically correct, it will typically go unchallenged, whether or not it violates the letter of policy.
• If a reduction is mathematically incorrect, it will typically be reverted.
• If a reduction is mathematically correct but too complex to be obvious, it may rightfully be reverted by someone who doesn't see the math, resulting in a discussion, an explaination, and a consensus decision to keep or drop the edit. Or, all editors who see the edit may be great at logical reasoning and treat it like a simple reduction.

These apply in heavily-watched articles where nobody is pushing an agenda. If there are lightly watched articles, things slip through. If the article is "controlled" by a faction that has an agenda - and don't kid yourself, at any given time a number of topics on the wiki are under the "control" of dominant editor or group - then edits hostile to that agenda are much more likely to be reverted than edits friendly to it, everything else being equal. Where possible, such politically-minded editors will use Wiki-policy to defend their actions.

Based on this, I recommend people use common sense: If a mathematically correct reduction is simple enough that most people with a 100 IQ will understand it within seconds, then let it pass. If it isn't obvious, and you can explain it in less than a couple hundred words on the talk page so most people with a 100 IQ will understand it, make the edit and explain yourself. If it takes more than that to explain it, either drop it altogether or open a discussion. Obviously, in highly technical articles with a highly technical audience, you can assume the average reader has an IQ much higher than 100 and edit accordingly.

In no case should any mathematical reduction which is either disputed or, if original, likely to be disputed by anyone who understands the math, be accepted. Likewise, if the math is too hard to understand by the average reader of that type of article, it should not be added without discussion. davidwr/(talk)/(contribs)/(e-mail) 23:55, 5 November 2009 (UTC)

This is sensible, and also consistent with my experience of how things actually work on Wikipedia, in math and science. —Finell (Talk) 00:06, 6 November 2009 (UTC)

Sources

I suggest that part of the problem behind all this is the nature of sources in science or at least some parts of science. I say some parts of science because this whole discussion is really only about mathematics, physics and some areas of chemistry. Some popular textbooks take some topics and simplify them so much that they are actually wrong. To get an accurate description needs a source in the primary literature or in original books or perhaps in advanced textbooks which are not readily available. Different texts give different stories. It is not OK to say in these circumstances, unlike perhaps in say history, "according to A this is the case, while according to B that is the case". We have to get the science right. Science editors who know the material well, possibly because they are Ph D students in the area, will give the advanced explanation of the topic. This is then open to challenge on the basis of a larger number of editors who have read a common first year university text. Of course, the text does not say it is simplifying the explanation, and we are unlikely to find a reliable source that says that X's first year text on physics simplifies the problem. If the simple, possibly wrong, explanation is given in an article, we have to say where it simplifies the problem, but we do not have sources for that. Perhaps we should be giving the simpler explanations as many of these articles, such as those in thermodynamics, are incomprehensible to the average reader. --Bduke (Discussion) 02:20, 6 November 2009 (UTC)

Hmmm, the problem is really that in high school physics is taught in a flawed way. If the correct explanation is too difficult, then the reader should go to Simple Wikipedia, a.k.a. "Wikipedia for Dummies". The best thing we can do here is explain everything from first principles. The last thing we should do is dumb Wikipedia down. Count Iblis (talk) 03:40, 6 November 2009 (UTC)

I think you have misunderstood me or perhaps I was less than clear. Actually I am partly on your side. Some of the policies and guidelines fit science less well than other areas and give us problems. Sources are one of them. I am firstly saying that when you, quite rightly, explain everything from first principles, it is difficult to understand for the average reader and therefore perhaps less useful as an encyclopedia article. Second, I am saying that it is quite likely that an undergraduate will come along and replace it with details from his first year text book (I am not talking about High Schools). This gives you a problem as both are backed with reliable sources. The problem is that the university text may be wrong in the sense that it over simple. Finally, I guess I am saying that it would be nice to say what the texts are saying when they simplify things and explain why they are wrong or where the simplification is. However, we are unlikely to find sources for that. Very finally, I am disgusted that you refer to the Simple Wikipedia as "Wikipedia for Dummies" and that you do not see that articles that are too difficult for the average reader have a problem. All wikipedia articles at least should have an introduction that is understandable to the average reader. Regrettably many science articles do not, including, even more regrettably, some I have written. --Bduke (Discussion) 04:17, 6 November 2009 (UTC)

I agree that sources are a major issue for science articles and I think a guideline to help editors give a context to their sources would be really helpful. The guidelines need to go beyond WP:RS and need to be more generally applicable than Wikipedia:Reliable sources (medicine-related articles). I agree with Bduke that we need to make material as accessible as possible. I think some of the best articles are written when there is an editor that knows a great deal and another that knows very little. For example, I offered an explanation of competing electrochemical half-reactions on the electrolysis talk page and another editor who knew less about the material dragged the explanation onto the page itself. After some communal editing it turned out to be a really nice understandable addition covering content that is often fudged in college level classes.--OMCV (talk) 16:28, 8 November 2009 (UTC)

I hope you found good sources for everything that was dragged from the talk page to the article... you may have inadvertantly added Original Research to the article. That said, I agree... the best articles are often the result of a collaboration between "experts" and those who know little about the topic. The experts know the material and what the best sources are... but often can not see the forest for the trees when it comes to figuring out what to put in the article. They also often assume that others have underlying knowledge and this may not be the case. The editor who knows little about the topic can point out where the article gets bogged down in confusing details, uses confusing jargon, assumes the reader has underlying information, or has other problems that would make the article less than useful to the average reader. Blueboar (talk) 18:32, 8 November 2009 (UTC)

I think that this is a different (and to me at least more productive point) than the discussion about what kinds of deduction count as OR, which is arguing about the definition of a fuzzy concept. Editors necessarily make choices as to which are reliable sources. Further they need to undertake 'scholarship' if not 'research' to discover what are reliable sources. These editorial choices can alter the 'general view' of a topic as follows. Secondary and tertiary sources copy extensively from each other (this is as true outside science as in it). If an early source makes a mistake, then that mistake will be replicated in many subsequent sources. An editor can then happily repeat the mistake in WP, citing many sources. Now consider the following two scenarios.

• The editor discovers from a primary source that a mistake has been made. My reading of WP:OR is that if the editor says that the sources are mistaken, quoting the primary source, and other editors object, this will fall foul of OR.
• The editor discovers from a neglected (but reliably published, etc.) secondary source that a mistake has been made. Now the editor can, I think, say without being challenged something like "Many sources (list of refs) state X; however Y has stated (ref) that this is incorrect because ..." (But what about "Y has established (ref) that this is incorrect"? There is also the issue of balance if Y's statement is expanded and the other not.)

The science/history distinction comes up in this way. If I were editing a science article and I found that many sources repeated some piece of physics which I knew was wrong and could find a reliable source to say was wrong, then I would not be happy to write that many sources say A but that X says not-A. I would want to say that although many sources say A, this is not correct (ref to X). Would I be right? Does it matter whether it is science or history?

Concrete example: in working on an article on botanical gardens, I discovered that many sources (as a Google search quickly verifies) say that Theophrastus inherited Aristotle's garden. Only one source that I can discover – an academic article published in the proceedings of a relatively obscure journal – says that this is incorrect and has been repeated due to a mistake in a very early secondary source. Personally I am utterly convinced that this is correct. Being wary of OR, I used the "Many sources say A; X says not-A" format in the article. But this is actually weasel wording from my POV. I don't believe what I've written; I just wanted to avoid claims of OR. Am I right? Peter coxhead (talk) 09:50, 20 November 2009 (UTC)

I don't think what you are saying is OR (as both stories can be sourced) ... it may or may not be WP:Undue weight.... a lot depends on who X is. For example, if X is a noted clasicists, his translation of ancient texts may well be more reliable than the accepted wisdom of thousands of botinists. But if X has a poor reputation, then it would not be. The best way to deal with this to raise the issue on the article talk page and discuss the situation with other well informed editors. Blueboar (talk) 14:45, 20 November 2009 (UTC)

I'm not sure that you've replied to quite the point I was trying to make. If I write "Many sources say A; X says not-A", this is not OR given that everything is sourced (and undue weight doesn't apply). But what if I write "Although many sources say A, this is wrong because X says not-A"? Isn't this "synthesis of published material that advances a position" and therefore not allowed?

A quite separate point is that counting sources can be very unhelpful. It's patently obvious when you research some topics that previous encyclopedists have just copied from each other. Thus there may appear to be a long list of sources asserting A, and only one asserting not-A. But in reality, the assertion A comes from one original secondary source. Peter coxhead (talk) 15:20, 20 November 2009 (UTC)

The problem is, we can not determine if it is OR without the specifics. This isn't a question that can be answered in the abstract. We would need to know exactly what X said about theory A and how he phrased it. and we would need to know exactly what the various other sources say. In other words... to answer your question we need to know what article you are talking about, and what statement in the article you find OR. Then we would need to examine the sources.

That said, in the abstract, I do find the statement "... this is wrong because X says not-A"... to be problematic... but my issue with that wording is based more on WP:NPOV than WP:OR... that wording does not present the dispute in a neutral tone. However, we again need to examine the specifics... there are cases where an authorative source is correcting a commonly repeated error, and it is not POV to state this. Blueboar (talk) 16:05, 20 November 2009 (UTC)