Wikipedia talk:No original research

Recent edits

If people want to make substantive changes, could they please propose them here first? SlimVirgin ^{talk|contribs} 02:16, 27 August 2009 (UTC)

What about that change makes it substantive? It's a great effort to remove a giant block of needless text that has no substantive content. Frankly, it's rather confusing, and belongs in an essay. "Here comes the original synthesis" indeed. If you don't have an objection beyond "you can't change anything until you check if someone can think of an objection at the talk page", then just don't revert.    M   02:36, 27 August 2009 (UTC)

It's an excellent example of WP:SYNTH, you'll need to find consensus to either remove it or replace it. Dreadstar ☥ 02:41, 27 August 2009 (UTC)

It would have still been easily accessible through a link at the end of the first paragraph of SYNTH in this version of WP:NOR, until it was removed from the main part of Wikipedia:No original research/Examples very recently by this edit. --Bob K31416 (talk) 22:54, 29 August 2009 (UTC)

Rich Farmbrough's points ("Remove the Smith Jones example per WP:POL because one example is enough. Also it is an example discussing sourcing which can confuse") are perfectly valid. What exactly does it illustrate that isn't already clear, and which justifies 4 extra paragraphs and 300 extra words? I can't see anything, but if you do please let us know.    M   02:51, 27 August 2009 (UTC)

These are excellent and longstanding examples. I can't imagine the policy would be better off by removing them; quite the opposite, in fact. Also, I see no consensus for making these major changes to longstanding policies. Jayjg ^(talk) 23:43, 27 August 2009 (UTC)

Agree with Jayjg. JN466 00:20, 28 August 2009 (UTC)

The Smith-Jones example is an example of bad logic. It doesn't serve to show that synthesis to prove a point is improper in Wikipedia (I think such synthesis is improper), it just shows how bad logic can be misused to "prove" something that it doesn't prove. We moderns are at a disadvantage when it comes to logic: we aren't educated in the various forms of syllogism nor in the various logical fallacies. It seems to me to be a grave error to use an example of bad logic as a claimed example of why synthesis in general is improper. If the use of logic is appropriate for a topic then I suggest that any proper synthesis that helps further understanding within that topic is not just valid but is desirable. "Proper" synthesis would be synthesis that follows valid logic and which is based on well-accepted premises. I can appreciate that some would think this can go to far. I'd welcome examples that show "going too far" (with the provision that the example not really be an example of invalid logic masquerading as valid logic.) Minasbeede (talk) 14:39, 28 August 2009 (UTC)

Actually, what makes this synthetic is the use of disperate sources (neither of which supports the conclusion) to back the logic (whether faulty or not). The correct/incorrect use of logic is a seperate issue. If the conclusion was cited to a reliable source (one that used the same sources to reach the conclusion), it would still be faulty logic, but it would no longer be a WP:SYNT vio. Blueboar (talk) 17:56, 28 August 2009 (UTC)

As Blueboar points out, the Smith/Jones example shows the use of reliable sources to synthesize an argument not present in any of the sources. It happens all the time on Wikipedia, which is why a good example like this is needed to point out why it is against policy to do this kind of synthesis. Jayjg ^(talk) 21:56, 28 August 2009 (UTC)

Perhaps someone here can give the link to a diff where they, or anyone else, have used the Smith/Jones example to clarify SYNTH for another wikipedian. Thanks. --Bob K31416 (talk) 07:14, 29 August 2009 (UTC)

Bob, I believe you've tried to remove this example earlier this year, and there was no consensus for doing so, after very lengthy discussions on the topic. Has something changed since then? Jayjg ^(talk) 07:06, 30 August 2009 (UTC)

It doesn't matter if this is a good example. Does it need to be here? Can it be stated just as well in three sentences? Could someone please state what rule it is supposed to illustrate? For example, the rule "don't vandalize" might have the example "Jones posts swears and deletes large blocks of content, so Smith temporarily blocks her". What exactly is the rule behind this section?    M   21:43, 29 August 2009 (UTC)

"No original research" is a policy that is difficult to understand, so examples are required to clarify. The policy on vandalism, a concept actually much easier to understand, is filled with examples of what is and isn't vandalism. A more comparable policy is WP:NPOV; not surprisingly, it also contains examples, see for example Wikipedia:ASF#A_simple_formulation, Wikipedia:ASF#Article_naming, WP:GEVAL. Jayjg ^(talk) 07:06, 30 August 2009 (UTC)

I would question your first sentence that implies it is a good example. I don't know of any evidence from actual use, that would indicate that it is a good example, or that any of the people here have found it good enough to use in their discussions with less knowledgeable wikipedians, to help clarify Synth. And if someone here did use it, were they successful in clarifying Synth, with the Smith/Jones example, for another wikipedian. That's why I made the request above, " Perhaps someone here can give the link to a diff where they, or anyone else, have used the Smith/Jones example to clarify SYNTH for another wikipedian. Thanks." --Bob K31416 (talk) 22:13, 29 August 2009 (UTC)

Bob, pointing to WP:SYNTH is enough, since it contains the examples. Jayjg ^(talk) 07:10, 30 August 2009 (UTC)

"What exactly is the rule behind this section?"

User:M, to answer your question directly "What exactly is the rule behind this section?", there is no agreed upon rule. Although, several attempts have been made in the past to establish a clear definition of the synthesis rule, no consensus has been reached. However, most users here have agreed on a set of examples that serve to illustrate the rule: [1].

1. Everyone agrees that:

Editors should build an article by summarizing the sources available on the topic of the article. Any information added should therefore be based on reliable sources that present this information in direct connection with the article subject.

2. However, some find that definition too restrictive and deem it necessary to add an exception:

In some cases, supplementary information from generic sources that cover a broader subject area than the specific article topic may be deemed to add value to an article, in order to clarify places, people or things mentioned in the article (WP:PCR). For example, an editor might want to add a detail from a reliable source that describes the historical context in which the subject of an article lived, even though the cited source does not mention the article subject.

3. Among those who agree with that exception (2), there are some who believe that the exception itself requires an exception:

Any information that advances a position or point of view with respect to the article topic must be based on a source that presents this information in direct connection with the article topic.

4. And among those who agree with this exception (3) to the exception (2), there are some who believe that it also requires an exception:

Any information that advances a position or point of view with respect to the article topic must be based on a source that presents this information in direct connection with the article topic, unless the information represents a mainstream view which is contrasted with a fringe or pseudo-scientific claim within the article. According to WP:PARITY, if a claim is fringe, it will likely not have been made in a reliable source; so we can reference a blog or a website that criticizes that claim; we don't necessarily have to find a reliable source that debunks it, which is sometimes hard to find. In those exceptional cases where a fringe claim is published in a reliable source without any reliable rebuttals or when a fringe claim is made in a non-reliable source without any rebuttals at all, we can rely on WP:IAR and add a rebuttal. Editors should, however, resist the temptation to add their own explanation of the discrepancy between the mainstream view and the fringe view. An unpublished synthesis or analysis should not be presented for the readers' "benefit". Let the readers draw their own conclusions after seeing the related facts in juxtaposition.

I hope this is helpful. Cheers, Phenylalanine (talk) 03:12, 30 August 2009 (UTC)

Keeping in mind, of course, that 3. is basic policy (not an exception to an exception) while 4. is a clear and obvious violation of WP:NOR, and there are no "special exceptions" for WP:V just because we really, really, want to rebut a fringe argument. As has been said many times before, if David Icke claims that George W. Bush and Queen Elizabeth II are reptilian humanoids, we don't bother to counter by saying "However, Bush and Elizabeth are, in fact, fully mammalian." Jayjg ^(talk) 07:06, 30 August 2009 (UTC)

Indeed... and if we did add such a counter statement, two minutes after adding it the line would look like this:

• "However, Bush and Elizabeth are, in fact, fully mammalian. <citation needed>"

(and good luck finding a reliable source that actually supports that). :>) Blueboar (talk) 20:46, 30 August 2009 (UTC)

Why not replace the examples with this?    M   20:10, 31 August 2009 (UTC)

Because this is too simple an example, and suffers from WP:V issues as well. Jayjg ^(talk) 01:30, 4 September 2009 (UTC)

Just a general comment on the use of the phrase "too simple an example". Please be careful not to exclude the interests of beginning editors who aren't familiar with the concept of Original SYN. --Bob K31416 (talk) 02:13, 4 September 2009 (UTC)

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 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 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.

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 [2] 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)

WP:SYNTH - why does this section exist?

The edit war over the WP:SYNTH examples reminds me that I’ve always been confused by WP:SYNTH and why it exists as part of policy. The section starts out with “Do not combine material from multiple sources to reach a conclusion not explicitly stated by any of the sources.” I'm thinking “Why isn't this reduced to a more general rule such as 'Do not reach a conclusion not explicitly stated by any of the sources?'” What's special about synthesis of multiple sources that's not covered under the more general rule?

My confusion deepens when I see the the first example. It's a single sentence about the UN and 160 wars that's entirely out of context and seems to bear no relationship at all to “RS A, RS B, synthesis C” that was used in the section lead.

The second example has paragraph with two sentences that presumably are based on one, maybe two RS. It just says “The first paragraph was properly sourced.” It’s a better example in that it shows sourced material followed by a WP:OR conclusion. However, I don’t see this as a direct example and clarification of “RS A, RS B, synthesis C.” It is a good example of “RS A, synthesis or conclusion B” but not “RS A and RS B synthesized into C.”

I'm guessing the WP:SYN section got added as someone must have wikilawyered their way around the more general "Even with well-sourced material, however, if you use it out of context or to advance a position that is not directly and explicitly supported by the source used, you as an editor are engaging in original research; see below." (the below links to WP:SYN). --Marc Kupper|talk 07:56, 3 September 2009 (UTC)

Actually, no... the SYN section was developed quite a while before the line you quote was added. The section was created to tell editors not to include what they see as logical conclusions... conclusions based upon reliable sources, but not actually made by any of the sources they cite. The examples are there to show that this is not always easy to detect (in part because the conclusions often are logical). We probably should have called the section "No original logic" instead of "Synthesis". Blueboar (talk) 17:24, 3 September 2009 (UTC)

Even if Mark Kupper's guess was off the mark, his point is valid. Renaming it to "No original logic" sounds sensible, but doesn't quite address his point. The general 'Do not reach a conclusion not explicitly stated by any of the sources' would cover "No original logic" too. -- Fullstop (talk) 19:24, 3 September 2009 (UTC)

I think we could easily remove the A+B=C sentences. That's sometimes what SYN is, but not invariably. SlimVirgin ^{talk|contribs} 23:06, 3 September 2009 (UTC)

Thank you, that helped a lot. While it seems redundant I now see that it's necessary to have a section like WP:SYN as there are some who will argue that logical conclusions are not "original" research.

The use of the word "synthesis" still bugs me a little but I found it is correct. It turns out there's a lesser used definition of synthesis. It's in the World Book Dictionary as the very last entry and is "according to Thomas Hobbes, Isaac Newton, and others, deductive reasoning." Thus "synthesis" can cover an editor derived conclusion or deduction, even if the article is based on one RS. I can't say I'm comfortable with needing to use a definition that only shows up in the largest dictionaries but I can't think of a better word at the moment. I'm going to follow this with the current lead paragraph, save the edit, and then immediately edit the lead to a proposed version. Use the history/diff to see the changes.

Do not combine material from one or more 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, regardless of how logical it may seem. 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 conclusion in relation to the topic of the article.

The tweaks are minor but would then cover synthesis from one source. I changed that last "argument" as while it's the correct word it seems to link better with the earlier use of "conclusion" in the paragraph. --Marc Kupper|talk 08:23, 4 September 2009 (UTC)

What do you think of getting rid of A+B=C? SlimVirgin ^{talk|contribs} 08:21, 4 September 2009 (UTC)

It's late for me and I can't think of a clean way to replace it. I like how it's showing that we have a reliable source(s), synthesized conclusion, and then restating that it should not be done. I'll sleep on it and will see if a wording comes to mind that has less emphasis on synthesizing multiple sources vs. deductions or conclusions from a single source. --Marc Kupper|talk 08:29, 4 September 2009 (UTC)

I think I see the problem and hopefully the solution or the road down which the solution lies. The focus in the present language is wrong. We don't care whether material is from one source or multiple sources for purposes of original research, but we do care that items of information are being combined to form a conclusion regardless of whether the data (and never the datum) is taken from a single source or multiple sources. The information is always going to be multiple items thereof and that's where the plural comes in. The restatement above goes some way to clear up up but still keeps the focus in the latter half on the plurality of sources rather than on the information in the sources. I think this resolves the problem (among other tweaks, note the changes from by sources to in sources)

Do not combine items of information (facts, opinion, etc.) from a source or sources to reach a conclusion not explicitly reached in the source or sources. Editors should not make the mistake of thinking that if A is published in a reliable source, and B is published in 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.[7] "A and B, therefore C" is acceptable only if one or more reliable sources have published the same deductive chain and result in relation to the topic of the article.

--Fuhghettaboutit (talk) 12:02, 4 September 2009 (UTC)

(Here's my 2 cents.) Perhaps the simplest way is to just modify the first sentence of the first paragraph in WP:SYNTH:

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

Do not combine material from multiple sources to reach a conclusion not explicitly stated by a ny of the source s. 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.

In this way, we remove the implication that this applies only to the case where the two facts are in different sources. --Bob K31416 (talk) 18:34, 4 September 2009 (UTC)

Here's a version that does not use A, B, C. I also moved the concluding paragraph up and reformulated the first example so that it shows the source sentence first and then the synthesis.

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

Care must be taken to not form arguments or conclusions that are not explicitly stated by a source. A synthesis of published material that advances a new position constitutes original research. Stating an argument or conclusion in an article is acceptable only if a reliable source has published the same position 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 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.

example commented out per feedback.--Marc Kupper|talk 10:03, 5 September 2009 (UTC)

I did not include the second example here. If we use it, then does not need rewording as it's an example of a deliberate synthesis where the first example could be accidental but it also shows why citing the sources is valuable. --Marc Kupper|talk 09:05, 5 September 2009 (UTC)

I think we need to retain the point about combining multiple sources, which is the essence of SYN. Also, you changed the example. Could we please not involve the examples in discussing the way the introduction is written?

As the intro to that section is written at the moment, what do you see as the problem with it? SlimVirgin ^{talk|contribs} 09:09, 5 September 2009 (UTC)

I know it's not perfect talk page etiquette but I've commented the modified example out from my previous post. Using strikeout on the large block was too distracting. We can include plural sources by replacing "a source" in the first sentence with "sources." Would that work? One thing I don't like about my own rewording is that it's does not have a strong statement about "logical conclusions are OR." That got weakened as I was trying to work in both synthesized arguments and conclusions. --Marc Kupper|talk 10:03, 5 September 2009 (UTC)

I like your latest proposal. I actually complained about the current wording of the policy over a year ago; but no luck changing it back them. Asking that "A and B, therefore C" come exactly from one publication/source is unreasonable. In fact, this particular requirement is ignored by large swaths of our Mathematics articles for a good reason. By observing the current policy you'd have to commit a copyright violation (wrt. derivative work) by copying entire lines of reasoning "as is" since you're technically allowed to make only superficial wording changes. Your proposal is much more reasonable, in that A, B, and C may come from different sources as long as all are germane to the article's topic. I also like that you put it in words, instead of syllogisms. Pcap ping 15:41, 5 September 2009 (UTC)

More problematic, is that some make a synthesis from different sources that use the same terminology with different meanings. Something like A=>B (source 1), B=>C (source 2), therefore A=>C (no citation), when the two sources don't mean the same thing by B. I've not seen this often, but I've seen some occurrences. Might be an example worth adding in the examples part, but it might be overkill as well given that it's rare. Your proposed phasing still prohibits this practice. Pcap ping 15:41, 5 September 2009 (UTC)

I think an example is important as some people have difficulty understanding the description of synthesis. Chillum 15:47, 5 September 2009 (UTC)

I agree that one, or even more examples are needed, but the defitintion of "syntehsis" here should not be given in syllogism, since you can't cover all aspects that way. Pcap ping 15:51, 5 September 2009 (UTC)

Comparison

Current version

Marc's proposal

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 ... (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.

Care must be taken to not form arguments or conclusions that are not explicitly stated by a source. A synthesis of published material that advances a new position constitutes original research. Stating an argument or conclusion in an article is acceptable only if a reliable source has published the same position 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 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. (Followed by examples)

I personally found the A, B, C wording to make the concept very clear. While the new wording does cover the basic concept I don't think it explains it as well as the old wording. The idea of synthesis has been very difficult for some editors to understand in my history of dealing with people. Perhaps it will not be an issue with examples. Chillum 16:06, 5 September 2009 (UTC)

The old wording is confusing because the A-B-C part actually imposes an additional restiction not found in the 1st sentence of the current policy. "Do not combine material from multiple sources to reach a conclusion not explicitly stated by any of the sources." only implies that you need a citation for each of A, B and C; it does not imply that you need a citation for "A and B implies C". The difference is significant. Pcap ping 16:14, 5 September 2009 (UTC)

By the way, the new wording (or just the 1st sentence from the old), does allow syllogisms like: "The sky is blue" (cite 1), "the Sun is hot" (cite 2), therefore "an elephant is a mammal" (cite3); you can easily "fix" this (as a syllogism, not as a meaningful article) by replacing therefore with and. But I don't think the "SYNT" part of the policy was meant to prohibit Wikipedia:NONSENSE. That's a different issue to me than "improper synthesis". Pcap ping 16:33, 5 September 2009 (UTC)

My point here is that even the current WP:SYNT doesn't prohibit this article: "The sky is blue (cite 1). The Sun is hot (cite 2). An elephant is a mammal (cite3)." because no attempt to present one statement as a conclusion of others is made. All statements are properly sourced, so it's not WP:OR. Still this "article" is an WP:INDISCRIMINATE collection of information, which we don't want here. But the purpose of this policy isn't to prohibit such nonsensical article in here; there's another policy for that, which I've just linked to. So, the proposed change in wording separates the concerns, i.e. makes the policies more orthogonal. Pcap ping 17:20, 5 September 2009 (UTC)

Huh... perhaps it is to early in the morning for me. I will re-read this section later. Chillum 16:47, 5 September 2009 (UTC)

Looking at the comparison, I do find the original version much easier to understand. Is it perfect, no... but I think it does a better job of explaining what the issue is than the proposal. Blueboar (talk) 17:18, 5 September 2009 (UTC)

I'm in agreement with Blueboar, the current (original) version is better at explaining the issue and is easier to understand. Dreadstar ☥ 22:02, 5 September 2009 (UTC)

I think I prefer the original too. The proposal doesn't clearly address the idea that each part is sourced correctly, but that the combination is OR. Marc: what is it about the current introduction to SYN that you find problematic? SlimVirgin ^{talk|contribs} 00:33, 6 September 2009 (UTC)

SlimVirgin asks what I find problematic with the current introduction. I see that it needs extremely careful reading, pretty much at the wikilawyer level to understand it.

I believe it is beneficial to have a section of NOR that reinforces that the editor must not introduce unpublished arguments, conclusions, ideas, etc. in an article. It does not matter if those arguments/conclusions are constructed via analysis (or analysis), synthesis, fallacy, insanity, or some other method. I believe the first WP:SYN example is a good one as it shows how a factual statement from a source was modified into an argument.

I also believe it would be beneficial to remind editors to not include they see as logical conclusions... conclusions based upon reliable sources, but not actually made by any of the sources they cite. (this is from Blueboar 17:24, 3 September 2009 above)

I'm now starting to see that we should not use words like synthesis, analysis, fallacy, etc. in core policy. Policy should be accessible to a wide audience and needs to be as readable as possible. --Marc Kupper|talk 05:43, 6 September 2009 (UTC)

May I add to Marc's important remark about readability, when policy is discussed extensively here on this talk page, the terminology and concepts become familiar to the editors here. We should recognize that the typical reader of WP:NOR doesn't have that advantage. --Bob K31416 (talk) 12:20, 6 September 2009 (UTC)

Prefer the original version, per Blueboar. --JN466 12:27, 6 September 2009 (UTC)

May I also add, in the original version the explanation of the concept of synth is split up. There's most of the explanation, then the examples, then the short remaining part of the explanation at the end, where it can get lost. Marc's version puts all of the explanation at the beginning before the examples. Even if you disagree with other aspects of Marc's proposal, surely you can at least agree with this simple improvement, which is the Alternate proposal in a previous section above. --Bob K31416 (talk) 12:49, 6 September 2009 (UTC)

I also prefer the current version. Given the choices, it is clearer and more to the point. When running into WP:SYN situations, which is all too common, I find it useful to be able to point to the basic "if A and B are individually well sourced, it doesn't mean you can juxtapose them to advance a conclusion C from them, unless C is also well sourced, and all sources must refer to the specific subject." I would also add that conclusion C needs specific sourcing whether it is directly stated or just implied by juxtaposition. Crum375 (talk) 15:29, 6 September 2009 (UTC)

Crum375, What do you think about the part of Marc's proposal that has the examples following the explanation of SYNTH? Would that simple change be acceptable to you, i.e. the change mentioned in my previous message and suggested in the Alternate proposal in a previous section above? It seems better than the current version because a small important part of the explanation doesn't get lost by being alone at the end of the WP:SYNTH section. --Bob K31416 (talk) 15:41, 6 September 2009 (UTC)

I think the current version, with the point about "summarizing existing sources is not OR" at the end is fine, but I don't see that order as crucial. I think the current UN example is too confusing as presented. Crum375 (talk) 15:55, 6 September 2009 (UTC)

Crum375, I personally had the experience awhile back of not noticing the "summarizing existing sources is not OR" paragraph because of its position alone after the complex example. So it can go unnoticed in its current position. The way it's positioned, it looks like part of the explanation of the second example and can make the reading less smooth too. In what I experienced, I was checking WP:SYNTH for some info. I looked at the first paragraph and since the example was next, I figured only the example remained in the section and I stopped reading, since I wasn't interested in the example at that time. --Bob K31416 (talk) 16:10, 6 September 2009 (UTC)

I can live with it being higher up, though perhaps it could be shortened. This entire section is very wordy and confusing as it stands. Crum375 (talk) 16:20, 6 September 2009 (UTC)

Re "This entire section is very wordy and confusing as it stands." - Ain't that the truth. I think it's due to the examples. --Bob K31416 (talk) 17:17, 6 September 2009 (UTC)

The original version is superior, precisely because it provides these critical examples. Otherwise, as has been pointed out, people invariably say "but it's all sourced from reliable sources". I can't tell you how many times I've run into this problem; reliably sourced material being used to advance a position not found in any of the sources, or to make arguments on topics not actually addressed in the sources. If anything, we need more examples in the policy, not fewer. Jayjg ^(talk) 17:38, 6 September 2009 (UTC)

I think that this section is probably the most difficult piece of Wiki-policy for the average editor to understand... not because we do a poor job of trying to explain it, but because the underlying concept is very complex. The line between source based research and sythesis is often blurred and hard for the average editor to figure out. I think the examples help. Blueboar (talk) 21:34, 6 September 2009 (UTC)

Aye, you've said a true word there. JN466 21:58, 6 September 2009 (UTC)

I was playing with some new wording, read WP:SYN again, and now it makes sense to me... The wording I was working on was "Any arguments, conclusions, ideas, etc. in an article must be explicitly stated in reliable sources. Developing a new position in an article, even if it is based on reliable sources, is original research. ¶Carefully summarizing or rephrasing source material without changing its meaning 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." That will also cover OR ideas developed by methods other than synthesis such as what's used in the first example. --Marc Kupper|talk 07:58, 7 September 2009 (UTC)

Best practice?

As long as we are discussing the wording... I would like us to take a look at the "Best practice" line... Experienced editors know what the line means, but to a novice editor it may not be at all clear where the line between "best practice editing" and "synthesis" lies. The problem is that it is quite possible to take "material from different reliable sources on the topic, and put those claims on the page in your own words, with each claim attributable to a source that explicitly makes that claim" and still form an improper synthesis.

Thinking about this... I think the problem might be resolved if we swap paragraphs... put the one about good editing and best practice first... and then warn editors about forming a synthetic statement. Using the current language (with a minimum of tweeking) it would look like this:

Original research is often introduced to articles through synthesis. 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.

Do not, however, 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.[7] "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.

Then give the examples. I think this clarifies the distinction. Blueboar (talk) 13:51, 7 September 2009 (UTC)

Yep. -- Fullstop (talk) 18:35, 7 September 2009 (UTC)

I am concerned that this can still be misinterpreted. Remember the (real-life) example presented some weeks ago (included as the first example in WP:ORIGINALSYN) about the activist who had called for a boycott of a company. An editor then added material to the effect that the sales figures and share price of the company dropped after the activist's boycott call. Crucially, neither of these sources mentioned the activist, or linked the drop in sales and share price to the activist's actions.

I am concerned that the above wording does not explain clearly enough what was wrong about what the editor did, as its leading statement seems to describe just what he did. We have heard editors defending SYN in this manner before, using the same phrasing: "But I took material from different reliable sources on the topic and put those claims on the page in my own words, with each claim attributable to a source that explicitly makes that claim, just like you say. And I did not combine material from multiple sources to reach a conclusion not explicitly stated by any of the sources. Because I did not explicitly say it was the activist's actions that caused the drop in sales, I just presented the attested facts and left it to the reader to draw their own conclusions."

That "not guilty" argument is precisely wrong, and I fear this reversal of the wording may encourage editors to use it. --JN466 23:49, 7 September 2009 (UTC)

Hmmm... I suggested that we reverse the paragraph order in part because I saw the current order encouraging exactly the same sort of error ... Perhaps the problem isn't the order... but that there is a flaw in our statement about best practice? Blueboar (talk) 02:01, 8 September 2009 (UTC)

Probably so. We need a way to define the point at which a juxtaposition of unrelated material becomes, or suggests, a novel conclusion. Basically, editors combine such material because they have come to a conclusion in their own minds. The editor in the boycott example felt, at the very least, that it was "interesting" that the company's sales and share price dropped after the boycott call. But the point is no source had said it was "interesting". The boycott call had no significance in secondary sources commenting about sales that year. Even if it were true that the activist's boycott call had something to do with the company's poor performance, Wikipedia should not be the first place where people can read about this.

In the past, we had addressed this by requiring that editors only cite sources that "directly refer to" or "directly mention" the article topic. That would have taken care of the boycott call example. We then changed this to sources that "are directly related to" the article topic. I remember Jayjg pointing out at the time that this opened a huge loophole – because in the editor's mind, there was a "direct relationship" between the boycott call and the drop in sales, even though the sources documenting the drop in sales did not mention the boycott call.

Perhaps a way to address this is that in all cases involving explicit or implicit conclusions that might be in any way contentious, the sources cited must directly mention the article topic, to ensure that Wikipedia is not the first place where a significant novel conclusion is published. JN466 12:06, 8 September 2009 (UTC)

Isn't this concern covered by a saying not to state a "conclusion not explicitly stated by any of the sources"? Blueboar (talk) 12:24, 8 September 2009 (UTC)

It isn't, because the editor never stated the conclusion that the boycott call had anything to do with the poor sales. They just said there was a boycott call, and then said that the company's sales were poor that year, citing a source that did not mention the boycott call. JN466 12:27, 8 September 2009 (UTC)

Another example I discussed with SlimVirgin a while back was where a person widely accused of having lied about his achievements claimed to have commandeered a particular ship during their military career. An editor reported the claim in an article and added a sentence reporting the "interesting fact" that there was no ship of that name listed in the Dictionary of American Naval Fighting Ships. Obviously, the Dictionary of American Naval Fighting Ships did not mention the article subject. I felt adding this sentence was OR, because no RS could be found that had reported the discrepancy before Wikipedia, but SlimVirgin felt it was alright. It's these sorts of issues that are at stake here.

If we allow sources that do not mention the article topic, we have to find some way of clarifying which types of "interesting facts" are okay to mention and which ones are not. JN466 12:35, 8 September 2009 (UTC)

Blueboar, I like your suggested wording. I think it makes things clearer.

Jayen, I agree with you in principle, but I wouldn't want to see SYN tightened even more. Although the example you gave above was arguably OR, it was also an interesting and relevant point that was added. The danger with tightening things even more is that we risk strangling an editor's ability to communicate his understanding of a topic. We do that a little anyway, and we have to do it somewhat, but we also need to allow some leeway. SlimVirgin ^{talk|contribs} 13:41, 8 September 2009 (UTC)

I thought the fundamental principle of Wikipedia was that we should communicate the understandings expressed in reliable sources, rather than the understanding of our editors. I'm wary that editors with limited subject matter expertise often come across facts that strike them as interesting and relevant, where an expert might readily recognise a fallacy, or have more detailed background knowledge that puts things in a different light. Surely we have all learned, if we are old enough, that some of the things which appeared quite logical to us and which we passionately believed in 10 or 25 years ago were based on partial information.

Taking the boycott case as an example, an expert might know that other retail companies had a bad pre-Christmas season as well that year -- there was a global economic crisis and share prices generally tumbled. That would explain why no sources linked the company's poor performance to the boycott call: it was unremarkable, conforming to the general market trend at the time. I think editors should resist the temptation to communicate their own hunches, findings and understandings, especially where these are based on "inspired" cross-connections drawn from extraneous sources, as in the boycott case, and concentrate instead on reviewing the best published sources that are focused directly on the specific article topic. On average, this will raise rather than lower the quality level of Wikipedia articles.

Looking at Blueboar's proposed wording, how about saying, "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. Make sure that you do not combine material from multiple sources to reach or suggest a conclusion that is not explicitly stated ..." etc. JN466 16:34, 8 September 2009 (UTC)

Current version	Blueboar's proposal	Jayen's proposal
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 ... (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.	Original research is often introduced to articles through synthesis. 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. Do not, however, 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. (Examples) ...	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. Make sure that you do not combine material from multiple sources to reach or suggest a conclusion that is 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.

Jayen, I'm not seeing a substantive difference in your proposal. For example, Blueboar says, "Do not, however, combine material from multiple sources to reach a conclusion not explicitly stated by any of the sources." And you say, "Make sure that you do not combine material from multiple sources to reach or suggest a conclusion that is not explicitly stated by any of the sources."

Is there a real difference between "reach a conclusion" and "reach or suggest a conclusion"? I can see "suggest" is maybe a little weaker, but what would a real example of that difference be, in a WP context? SlimVirgin ^{talk|contribs} 18:03, 9 September 2009 (UTC)

(This is my first, and maybe only, post in this thread.) I think Jayen's proposal is very nice. It fixes something that I have always thought was unfortunate about the present language here, which does not say that best practice actually includes doing any broad research of sources or even reading significant parts of them. Instead, the present language suggests cherry-picking facts out of numerous sources without regard for the topic as a whole. Jayen's version restores the proper order of events: first, do some general reading on the topic, taking note of what you see. Second, write the article, making sure it agrees with the things you read. — Carl (CBM · talk) 19:23, 9 September 2009 (UTC)

Carl, what do you see as the difference between Jayen's and Blueboar's? SlimVirgin ^{talk|contribs} 19:24, 9 September 2009 (UTC)

The phrase "researching the most reliable published sources focusing on the article topic"; that activity is not mentioned in Blueboar's version. If you asked me what the best practice is when writing articles, I would say: find the best sources on the topic, read them carefully to see what they say about it, and then write an article to accurately summarize what they say. This is true both in Wikipedia and in general life. — Carl (CBM · talk) 19:36, 9 September 2009 (UTC)

I agree with that. SlimVirgin ^{talk|contribs} 19:43, 9 September 2009 (UTC)

What if we combined my proposal with Jayen's Blueboar (talk) 19:52, 9 September 2009 (UTC)

Combined version

Original research is often introduced to articles through synthesis. 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 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. Do not, however, combine material from multiple sources to reach or suggest 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. (Examples) ...

Yes, it's good, though I'm a little worried about "suggest." I'd like to hear how broad the scope of that is intended to be. SlimVirgin ^{talk|contribs} 19:54, 9 September 2009 (UTC)

Good point... we are talking about implied conclusions here (ie stating A and B in such a way that conclusion C is clearly implied, even if it is not actually stated)... So let's use the word we mean... "reach or imply a conclusion" would be a better way to phrase it (or "reach or clearly imply a conclusion" if we want some leeway). Blueboar (talk) 20:42, 9 September 2009 (UTC)

I like the combined version and that it use imply rather than suggest. Clearly imply has merit but but I can't see how it would help prevent or cause future battles over if something is implied in an article any more or less than just imply. --Marc Kupper|talk 22:36, 9 September 2009 (UTC)

"Imply" is fine with me. Also, personally I prefer it without the bold. SlimVirgin ^{talk|contribs} 22:44, 9 September 2009 (UTC)

Seems like you folks are performing organ transplants when the patient only needs minor surgery, and you're harming the patient. Yikes! You are using a term "synthesis" in the first paragraph without explaining what it is until the second paragraph and you didn't address the issue of synthesis of material from the same source. (Didn't SlimVirgin bring up the "same source" issue originally?) The following version modifies just the first sentence, addressing the "same source" issue, and moves the last paragraph to after the first paragraph. Note that we still have "synthesis" described before the term is used, which is the way of this and the current version.

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

Do not combine material to reach a conclusion not explicitly stated by any source. 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 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.

(Examples)

--Bob K31416 (talk) 23:12, 9 September 2009 (UTC)

Original images and license

The section makes reference the "GFDL or another free license". Perhaps the reader should be directed to Wikipedia:Image use policy. I also think the Creative Commons should be explicitly mentioned. –droll [chat] 21:30, 5 September 2009 (UTC)

Wikipedia does not publish original research or original thought.

No matter how many sources or citations one includes in adding to content, the converse of the subject/headline is always true. (That is "Wikipedia does publish original research or original thought.") At some point there was original research and original thought. It is humanly impossible to publish otherwise. So the entire premise of "no original research" is a lie. And non-truth is currently supported by the verifiability policy. Then, neutral point of view is automatically violated as policy, due to human error in writing an outright lie and allowing its support. Is there any concensus here to correct the error of policy? Edward Palamar (talk) 12:23, 7 September 2009 (UTC)

"Original" applies only to the person adding text to Wikipedia; as long as the originality of thought is removed one step from WP (and of course considered reliable) then there's no violation. --MASEM (t) 12:28, 7 September 2009 (UTC)

One step, two steps - the actual publishing contains original thought. Why are the two uses of the word "original" not either placed in quotes and redefined, or have citations of their own, to make the sentence as printed a non-lie? Edward Palamar (talk) 12:39, 7 September 2009 (UTC)

To clarify... What we mean is that information appearing in Wikipedia should not be the original thought of the editor who adds it. To put it another way: Wikipedia should not be the original venue of publication for information (this last statement, or something similar, used to be in the policy, we should probably re-add it). If information is previously published elsewhere, it did not originate with the Wikipedian who added it; Wikipedia is not the original venue of publication. Blueboar (talk) 13:24, 7 September 2009 (UTC)

Your second point is correct, but not necessarily the first. We do allow an editor to publish his original thought on Wikipedia if it has already been published in a reliable source elsewhere. So the point of originality, as you say, is that Wikipedia should not be the original publisher of the material. Crum375 (talk) 13:30, 7 September 2009 (UTC)

OK... I stand corrected. Information appearing in Wikipedia should not be the original thought of the edtior who adds it UNLESS it has been previously published elsewhere (in a reliable source). The rest stands.

What do people think about adding: Wikipedia should not be the original venue of publication for information? Blueboar (talk) 14:01, 7 September 2009 (UTC)

The problem is that we do allow some types of original information, such as self-made drawings, diagrams, photos, audio or video recordings, etc. So it's not quite that strict. Crum375 (talk) 14:23, 7 September 2009 (UTC)

That's true. But let's face it. The term "original research" is a misnomer. What is meant is unpublished research or facts.

It's interesting that if it was properly named in the beginning, we would just as easily be talking about UR instead of OR, and there would be less confusion. --Bob K31416 (talk) 14:31, 7 September 2009 (UTC)

Re: self-made drawings, diagrams, photos etc. Not quite... read the section on user-created images again... if the information presented in the drawing constitutes OR, we don't allow it. What we do allow is a user-created image that visually depicts information that has been reliably published elsewhere. It is a visual equivalent of an editor summarizing information from a reliable source into their own words. The image may be original... but the information is not. Blueboar (talk) 15:58, 7 September 2009 (UTC)

That may sound good in theory, but is not what happens in practice. Photos depict tons of new information not in any published text, and diagrams and drawings, or Wikipedian-made "computer renderings", have lots of unsourced details. You could start arguing about major issues, but the more minor ones are not worth the debating time and energy. As bottom line, all those allowed things introduce new material. Crum375 (talk) 16:31, 7 September 2009 (UTC)

It's definitely a problem. We've discussed maps of genetic distribution, where an editor may cherrypick from published source from different times, even using the sources incorrectly, and it's really hard to disentangle such maps. Dougweller (talk) 16:36, 7 September 2009 (UTC)

Sure, identifying OR can sometimes be a problem. But the difficulty in identifying OR is not (and should not be) a license to violate NOR. -- Fullstop (talk) 18:47, 7 September 2009 (UTC)

Aside from the fact that this flaw in Wikipedia policy, that is the stated lie "Wikipedia does not publish original research or original thought." is conducive to violations of which I have been accused falsely (i.e. - what goes around, comes around), the actuality of what the rest of you seem to imply is - "An editor can only copy and paste another's writing." This statement uses 52 characters compared to 65 found in the lie. With all due respect to Wikipedia's efforts to have a comprehensive encyclopedia inclusive to contributions by anyone, there should be a similar practice of ethics in valueing anyone who may read it. None of you have seriously looked at the statement to the point of seeing how damaging it already is, else you wouldn't be trying to defend it. Edward Palamar (talk) 16:45, 8 September 2009 (UTC)

Edward, I think you fail to see that while the wording of "No original research" may not reach your standards, editors are aware of the practical implications of the rule itself. You are defining "original research" to mean "research that originated somewhere," which yes, the wording can be taken that way. But your interpretation makes "original research" redundant; if something is "unoriginal research," what does that mean? That it is research copied from other research? However, the vast majority of editors see "original research" to be unpublished or otherwise unverified research. And that is the consensus, and thus is the rule for WP. Wiki-lawyering the wording of a specific rule of guideline isn't going to turn over a long line of history of editting under the guideline. WP is not like legislation; just because a word or phrase can be taken in two different directions does not overturn or change a rule.

In sum, the consensus definition for original research/thought is different than yours. While you may dislike the wording, the importance is the spirit of the guideline and the consensus behind it. Angryapathy (talk) 17:03, 8 September 2009 (UTC)

"Original research" is a confusing term that we're stuck with. As I mentioned, it should have been called "Unpublished research".

As to Edward's other point, the last paragraph of WP:SYNTH helps avoid the misapplication of NOR, but unfortunately, it is located in a place where it can easily be overlooked. I had a hard time finding it when I needed to refer to it, and perhaps some of you had similar experiences. There is a proposal to improve its location. Please see Alternate proposal in a previous section above and add your comments there. Thanks. --Bob K31416 (talk) 18:14, 8 September 2009 (UTC)

Analysis:

Wikipedia has written a sentence including the statement that it does not publish original thought, that is, what one would normally assume to be mental activity within the mind of a man.

In order for any words formulated from the mental activity to be verbalized, orally or in writing, there needs to be Grace, a source of strength, to overcome the inability of man to control his tongue. Man, without Grace, in the basic condition of a being doomed to die, is incapable of survival on his own, hence he has lost the ability to control his tongue, including thought. The enlightenment of the mind with Grace permits both the ability to think and to verbalize those thoughts. Any editor inside or outside of Wikipedia was under these bounds, the enlightenment due to Christ's Redemption being the release from original sin (inclusive to said bounds). If we take the use of 'original' as in 'original thought' to be thought before being enlightened, we would have to be alive before the time of Christ's Redemption, or we would have to reject Him willfully. In terms of free thought, to reject Him willfully would be a violation of neutral point of view. And as we who are alive are not alive before the time of Christ, the use of 'original' must refer to something else after Christ's Salvific work. As it is humanly impossible to verbalize without thought, 'original thought' can only refer to thought originating in a specific person or persons after Christ's coming. The thought is necessary to verbalize, so in this it remains intact as a consistent whole to the end of an auditory (oral) verbalization or of a written (published) message. For whatever reason one chooses to advocate the cutting off of the necessary thought to the end of publication does not change the fact that the thought is still present in the publication. Hence, it is not an issue of the wording of the sentence, it is the content and its implications.

There is not enough data to make the statement concerning consensus definition, Angryapathy. There are billions of people, with only eight showing. Also, I haven't failed in seeing anything as regards the sentence as quoted. Even if one only copies and pastes from another's writing, one has to think while doing it, original thought. Anything with more effort requires more thought, more original thought. When one selects "Save Page" one is publishing, even more original thought. And one has had to do original research to "Save Page". And unless one becomes brain dead while saving the page, there still is original research being put into publishing. But most importantly, there is no spirit in a lie. "Wikipedia should not be the original venue for publication." is much better as it allows, if necessary, some positive group interaction. Edward Palamar (talk) 13:34, 9 September 2009 (UTC)

Edward, we don't need to consult with billions, we have eight here, and now nine when I press [Save Page]. The consensus reached is among the nine of us. Hundreds may be lurking but WP:SILENCE applies. Also, consensus can change at any time. You have commented on the issues surrounding the words "no original research." What words do you recommend be used instead of "no original research?" --Marc Kupper|talk 23:06, 9 September 2009 (UTC)

In place of "Wikipedia does not publish original research or original thought." "Wikipedia should not be the original venue for publication." This would allow for a prophet/prophetess to prove him/herself rather than striking the prophet/prophetess with a lie. Edward Palamar (talk) 02:29, 10 September 2009 (UTC)

I have no problem with that but as it's very close to the previous wording I took a look to see why it was added and then removed. Here's the previous wording at the time it was added 13 September 2007 (talk page) and removed 17 Dec 2007 (talk page). The talk page links are the section in the talk page related to the change at the instant the change was made.

• Wikipedia is not a venue for publishing, publicizing or promoting original research in any way. no original research, or NOR, is a corollary to two other policies:
• Wikipedia is not a venue for publishing, publicizing or promoting original research. No original research, or NOR, is a corollary to two other policies:

It looks like both the addition and removal were for clarity rather than to deal with specific issues. I prefer the 2007 wording (final version) over the new proposal as it seems easier to understand. It'll be the very first sentence in NOR. --Marc Kupper|talk 05:38, 10 September 2009 (UTC)

In terms of clarity, it is best not to repeat the word(s) which one is trying to define in the definition. "Wikipedia should not be the original venue for publication." is not close to the lie as exposed in this : the putting of a sword into man between mind and work in "original thought" carries over as the same for "original research". And in both occurences, "original thought" and "original research", it constitutes assault. It is better to define "No original research" as a policy without the absolute, i.e., the use of "should", to allow for editors to place material they feel is correct and of benefit to everyone else, not in defiance, but to improve Wikipedia. There is always the discussion page to iron out things. Edward Palamar (talk) 14:46, 10 September 2009 (UTC)

The boundaries of OR?

If I claim that 777 × 286 = 222222 without citing a source, is that "OR"? Michael Hardy (talk) 19:53, 7 September 2009 (UTC)

Of course it isn't, if presented merely as a claim about arithmetic. We have a relatively well-established practice regarding the use of calculations such as that, and more complex ones as well. Did some more specific situation come up? — Carl (CBM · talk) 20:22, 7 September 2009 (UTC)

OK, what if it's merely an algebraic identity that anyone can easily check in a few seconds, such as the following?

${\begin{aligned}\left(a^{2}+b^{2}\right)\left(c^{2}+d^{2}\right)&{}=\left(ac-bd\right)^{2}+\left(ad+bc\right)^{2}\ \qquad \qquad (1)\\&{}=\left(ac+bd\right)^{2}+\left(ad-bc\right)^{2}.\qquad \qquad (2)\end{aligned}}$

Or a trigonmetric identity, such as this?:

$\sec(\theta _{1}+\cdots +\theta _{n})={\frac {\sec \theta _{1}\cdots \sec \theta _{n}}{e_{0}-e_{2}+e_{4}-\cdots }}$

where e_k is the kth-degree elementary symmetric polynomial in the n variables x_i = tan θ_i, i = 1, ..., n, and the number of terms in the denominator depends on n.

That one might take more than a minute, but it's just a secondary-school exercise. Michael Hardy (talk) 23:16, 7 September 2009 (UTC)

I think OR using BASIC (grade school level) math... simple multiplication, division, addition and subtraction of numbers is OK. But I would draw the line there. If you have to use algebraic or trigonometric calculation (even if fairly simple) you need a source. Blueboar (talk) 02:07, 8 September 2009 (UTC)

No and yes. See WP:NOR#Routine_calculations. A Quest For Knowledge (talk) 02:49, 8 September 2009 (UTC)

The algebraic identity shown above is just a routine calculation of which there are too many to count in Wikipedia; we do not have any problem with such calculations in practice, and they do not violate the OR policy. The secant example is more questionable, especially without any context to help. I would be more concerned about its accessibility and clarity than its originality.

In the end, the spirit of the OR policy is to keep crackpot theories off Wikipedia, not to stop people from writing clear explanations of standard material. The more obvious it is that something is just a neutral explanation of standard material, the less problem anyone can have with it. For example, if multiple textbooks make the same sort of calculation, including it on WP does not violate any policy, even if the form used here is not identical to the form used elsewhere. — Carl (CBM · talk) 04:10, 8 September 2009 (UTC)

If something is just a routine textbook calculation, then it should be easy to cite a source, and citing a source is better for article stability. Problems arise if an editor says, "But this is the same as so-and-so," and they can't cite a reliable published source that says so.

To be honest, I don't care if it can be shown on the talk page that mathematically some A equates to B. If no source has bothered to point that out before, then we'd be the first to do so, and that is original research, whether it's correct or not. That's not our business; our business is summarising what reliable sources have said about a topic, presenting viewpoints and statements in proportion to their prevalence in reliable sources. JN466 11:44, 8 September 2009 (UTC)

We don't equate originality of presentation with originality of content. Of course, whenever there are serious objections to something, we remove it. But "it wasn't written exactly that way in any book" is not an objection about original research. Unless we are going to make articles consisting of entirely direct quotes, paraphrasing and original writing will always be necessary. In the context of mathematics, that may include rephrasing algebraic deductions as well as rephrasing prose, because algebraic deductions are part of the language that is used to communicate mathematics. I am speaking here with a great deal of experience with both good and bad mathematics articles on WP, experience cleaning up mathematical OR and experience with articles that have original content that is not OR. There is actually a relatively clear line between unacceptable and acceptable content in mathematics articles, but that line will not be easily found by exploring the text of the page WP:NOR as it is currently written, because the present language is written with a very different set of articles in mind. — Carl (CBM · talk) 13:16, 8 September 2009 (UTC)

The discussion above seems to be losing the point of having OR in the first place. We're here to build an encyclopedia. Obviously we want it to be of high quality, and to do so we require contents to be supported by reliable sources with the hope that the quality of the sources translates to that of Wikipedia. In other words, OR is only "the means to the end". If some math examples didn't appear in exactly the same way in any source, it could be, strictly speaking, considered "synthetic". But what is the harm? What is wrong with OR isn't because it is "original" but because OR materials tend to be unverifiable or are incredible claims. Simply saying "synthesis" is wrong is very wrong: it has to take this context into account. Otherwise we get a silly situation that many math examples are judged synthetic and are subject to deletion. -- Taku (talk) 12:48, 8 September 2009 (UTC)

I don't think anyone objects to an editor making up an example, as long as equivalent examples (same structure, but different figures for example) are used in reliable sources. JN466 13:03, 8 September 2009 (UTC)

There are usually issues other than originality that are more compelling. Examples that are written in a way that does not make them obviously right are probably not written well, and probably will not help the intended audience of the article. Once the examples are clarified and simplified to be appropriate for the average reader, it is usually very clear which ones are "standard" and which ones are not. — Carl (CBM · talk) 13:16, 8 September 2009 (UTC)

Ugh. This was a problem I tried to point to & start a conversation about with my draft of Wikipedia:These are not original research, specifically:

Simple logical deductions [are not original research]. 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, again, not be included under this case because they require skills that not all readers possess, and involve a large number of steps that introduce the possibility of errors.

My intent was to close off points of contention whose only purpose was to wikilawyer over statements otherwise perfectly acceptable to an average reader -- tendentious editing. For example, no one needs to provide a source for the statement "X is a town in the country of Y" if all of the population statistics is taken from the government statistical office of Y -- an intelligent reader will be able to deduce that fact. Responding directly to Michael Hardy's example above, we assume readers can perform basic mathematical calculations -- or have a calculator at hand. However, we do not expect them to perform integral calculus because most Wikipedia users do not know how to do this. (Where we should draw the line in this discipline -- & others -- has never been discussed to my knowledge.) -- llywrch (talk) 16:29, 10 September 2009 (UTC)

What are the basic rules here?

Hi Everyone. I've been lurking here for a while, trying to make sense of what the OR rule actually says. In the above discussions, I've noticed the disagreement over how restrictive or loose this rule should be, and over the best way to state the rule. I'd like to make a suggestion. (Although I know expertise holds no sway in Wikiland, I will mention that I am a professor who studies the law and behavior, specializing in how people deal with rules.)

I wondered how the law (at least in the US) would deal with the OR problem. After thinking about it a bit, I think I am able to boil this whole thing down into two simple rules (and one exception):

RULE 1: Any assertion of fact must be verifiable by a reliable source.

This rule basically overlaps with WP:V (as does much of the existing OR rule), and simply says that all fact are to be verifiable. No exceptions. (The law would also take the time to define what a "fact" is, which, in our case, would be a statement that explicitly or implicitly presents "true" information.)

RULE 2: Any inferences derived from facts must be attributable to a reliable source.

This second rule basically means that any commentary, explanation, insight, or other inference must not be that of the editor, but rather of some other entity that can be verified. The rule also implies the inference should be attributed to it's source, and not stated plainly so as to suggest that the inference is a fact (e.g., According to _____....).

EXCEPTION: Inferences derived through formal logic (e.g., mathematical transformation) are acceptable.

This exception is likely to cause some worry among some of you. In my impression of what the OR rule is supposed to do (as someone else put it "to keep crackpot theories off of WP"), some information may be perfectly acceptable, but would be excluded by the above rules. I propose that the inference "Alice Walton is 5th on the Forbes list of wealthiest Americans, and if numbers 1-4 are all males, and if Alice Walton is a female, then Alice Walton is the wealthiest female in America" is acceptable. In cases where formal logic leads to one and only one possible conclusion (which, for the most part involves math equations), I suggest that WP allow the info to be stated. (Perhaps with a footnote.)

Anyway, I thought I'd just throw out my version for your consideration. -Nick^talk 04:26, 10 September 2009 (UTC)

I'd trim rule 2 to "Any inferences must be attributable to a reliable source." That removes wiggle room for secondary inferences and may be a good nutshell summary of the NOR page. --Marc Kupper|talk 05:59, 10 September 2009 (UTC)

Unfortunately, rule 3 ("exception") is not right. The original and fundamental goal of the NOR policy is to prevent people from adding unpublished/crank/crackpot theories to Wikipedia (for example, new theories of physics and mathematics). It is too time-consuming to read the often-poor exposition of these theories to find the errors, and the most persistent authors will not accept the flaws in their reasoning if those flaws are pointed out. So we adopted a different policy: it makes no difference if a new theory is correct or incorrect, if it has not appeared in the literature.

So if we added an exception that all "logically-correct" deductions are acceptable, it would nullify the purpose of the NOR policy. What is more accurate is that deductions that do not go beyond the spirit of the published literature are OK. Your example about wealthy women is fine, for example, as long as you attribute it to Forbes ("Jane Doe is the wealthiest woman on Forbes' list [3]"). Deductions that do not materially extend or change the literature, but merely explain what it says, are common and accepted. — Carl (CBM · talk) 10:29, 10 September 2009 (UTC)

I disagree that the only reason for NOR is to keep crackpot ideas out... it is to keep all original deduction, analisis and conclusions out... even the ones that are sane and perfectly logical. The excpetion for numerical computation was included to allow for BASIC things like converting kilometers into miles. The exception is not based on the fact that such calculations are logical, but because they are BASIC and will be helpful for our readers. Blueboar (talk) 16:15, 10 September 2009 (UTC)

I didn't say "only" and "crackpot"; I said "original and fundamental" and "unpublished/crank/crackpot",. See e.g. the version from December 21, 2003, which relies on the text on "primary research" that was in WP:NOT about the time. The reason that "whether it's true or not" is in bold there is because the point of the then-new policy was to avoid the need to disprove crackpot theories in order to remove them (disproving them is hard if they are written in a poor way and the creator is tendentious). The mailing list post from which Jimbo's famous quote was taken is part of a longer thread about minority views in the article on special relativity.

Since the page was created, the actual text of WP:NOR has drifted quite a bit, to the point that by 2005 WP:NOR was already a rephrasing of WP:V in more complicated language [3]. But we still have little difficulty, in practice, with "original" deductions that don't materially extend the published literature nor advocate a controversial position. The difficulty is entirely with things that either attempt to materially extend the literature (e.g. by introducing materially new theories) or advocate controversial positions for which there is not support in the literature. — Carl (CBM · talk) 16:59, 10 September 2009 (UTC)