Wikipedia talk:Attribution/FAQ

• Also see Wikipedia talk:Attribution

Mention reliable sources first?

Great structure here. My only suggestion would be that this start with a short section titled something like "What kind of sources are regarded as reliable?" Better perhaps to start with the positive? qp10qp 15:08, 13 October 2006 (UTC)

Plagiarism example

This entire paragraph is original research, because it expresses the editor's opinion that, given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it.

I disagree. "Given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it" is a fact, just like "given a definition which says that plagiarism is a cube, plagiarism has six sides" is a fact. It's a direct logical inference.

The paragraph is inappropriate because we have no reason to *care* that the Chicago Manual of Style's rules lead to a certain conclusion; the Chicago Manual of Style isn't an authority on this particular case, and probably not even on plagiarism in general. In other words, it's a reliable sources issue, not an original research issue.

And I'm aware that this example is from WP:NOR. It's a bad example there too. Ken Arromdee 22:20, 13 October 2006 (UTC)

No, it's an excellent example, because many editors don't get it, yet it's classic OR. You wrote that" 'Given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it' is a fact." But it's a fact that no reliable source has discussed in terms of Jones. That makes it the essence of OR. Anything introduced into an article (fact, argument, whatever) must have been published by a reliable source discussing that topic.

If we allowed Wikipedians to argue that "Given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it," I could just as easily add: "However, according to publication X, the Chicago Manual of Style's definition of plagiarism is inadequate." And we'd be off, swapping each other's OR i.e. personal opinions. SlimVirgin ^(talk) 05:22, 14 October 2006 (UTC)

I don't know why you're saying "If we allowed Wikipedians to argue...". I'm not claiming we should let Wikipedians argue that. I just think the example mischaracterizes *why* we don't let Wikipedians argue that.

'Given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it' is a fact. It isn't a fact we should use, because the Chicago Manual of Style isn't a reliable source, but that isn't original research--that's use of unreliable sources. Even if you classify that as original research (which I think is a bad idea, because then you've classified all unreliable sources as original research), it isn't original research *on the grounds of not being a fact*. It's original research because yes, it is a fact, but the fact isn't useful, and is being misapplied in a way which is original research.

And if it's a fact, we shouldn't say it isn't. Ken Arromdee 14:50, 14 October 2006 (UTC)

I may also add that the example refers to a real case, and describes it in a misleading way. See Talk:Alan Dershowitz#The debate with Norman Finkelstein. The difference between the real case and this is that in the real case, the statement that Jones (Dershowitz) didn't commit plagiarism according to the Chicago Manual of Style isn't an original invention of the Wikipedia editor at all. It's sourced to "James O. Freedman, the former president of Dartmouth College, the University of Iowa, and the American Academy of Arts and Sciences." We cannot ourselves use the Chicago Manual of Style to analyze whether Dershowitz is a plagiarist, but if we have a reliable source who does that analysis, we can legitimately quote him. Ken Arromdee 01:53, 24 October 2006 (UTC)

I think the plagiarism example is horribly confusing, and should be replaced. Synthesis is described as combining two statements, A and B, to reach a new conclusion C. As well as being a confusing and complicated case, the example never explains what sources A and B say, or even what sources A and B are (although source B is presumably the CMS). Without this information, the reader is left guessing at what the example is getting at, and I find it unsurprising that so many readers have come to different conclusions about what it is trying to say. In my view it should be replaced by a clear and straightforward example of synthesis. Enchanter 14:54, 28 January 2007 (UTC)

Well, it is synthesis. A is "the Chicago Manual of Style defines plagiarism as doing _____". B is "Dershowitz did not do _____". C is "according to the Chicago Manual of Style, Dershowitz isn't a plagiarist".

The problem with calling this original research is that this synthesis is a straightforward logical deduction, and those are exempt from being original research. Consider my example above; the Chicago Manual of Style claims that plagiarism is a cube, so we deduce that according to it, plagiarism has six sides.

Neither of those is original research. Rather, they are unreliable sources--the Chicago Manual of Style isn't an authority on the Dershowitz case, and can't be quoted on it. Rejecting them for being original research rejects them for the wrong reason, and sets a bad precedent for using straightforward deductions. Ken Arromdee 17:04, 28 January 2007 (UTC)

I'm not familiar with the original case it is based on, but the "B, Jones did not do ____" is not actually there in the example. So if this was the intended interpretation it's not properly expressed in the example. Enchanter 21:43, 28 January 2007 (UTC)

Bloat

We need to watch the bloat here, or we'll end up with a page with the problems RS has. We also need to watch that we don't contradict the policy. I found a few sentences that were arguably inconsistent with it. SlimVirgin ^(talk) 05:18, 14 October 2006 (UTC)

• The problem with WP:RS is that it was both bloated and {{guideline}}. I would suggest, if it hasn't been suggested already, that this FAQ be {{essay}} or similar. Bloated things which are non-normative are less of a problem. --EngineerScotty 05:22, 14 October 2006 (UTC)

Even so, the writing needs to be clear and crisp or people won't read it, and we do need to be very careful that everything in here is consistent with the policy. SlimVirgin ^(talk) 05:23, 14 October 2006 (UTC)

Agreed. --EngineerScotty 05:50, 14 October 2006 (UTC)

Structuring as an FAQ, though, means that you don't need to read it all for it to be helpful: most people will read the parts that are relevant to sources they are interested in, and ignore the rest, I suspect. JulesH 07:47, 14 October 2006 (UTC)

I very much agree that this will need a hawk's eye kept on it to prevent contradiction of the policy statement. This is the likeliest place for interest groups to try stitching in loopholes that suit themselves. On the other hand, the policy must take priority; the FAQ should not be considered canonical, only the policy. This shouldn't even claim to be a guideline, or we'll have to post armed guards on it 24/7.

I agree with JulesH that few people will read this page all the way through and so we need worry less about its overall shape; it is incapable of taking the form of an essay, though its status could be equivalent to Wikipedia essays. Each entry should certainly be as crisp as possible; nevertheleless, some style carbuncling is inevitable with this sort of page. qp10qp 13:38, 14 October 2006 (UTC)

Examples of web sites...

Do we need a separate section for each web site that anyone can contribute to? Or could we just lump MySpace, YouTube, Wikis, and so on in a single section. And maybe even combine them with blogs. That seems better to me, as the same problems apply to all:

• They're self-published
• The authors are usually anonymous

This means the same exceptions may apply to all of them. JulesH 07:54, 14 October 2006 (UTC)

I agree. SlimVirgin ^(talk) 07:58, 14 October 2006 (UTC)

Wikipedia:Attribution/FAQ#What do I do when reliable sources conflict?

This section in the proposal currently doesn't seem to answer the question enclosed in the section title. In the single paragraph of that section there are currently some circumstantial explanations that reliable sources can conflict (maybe amounting to something like "don't panic" as part of an answer to the initial question). But no real answer to What do I do...

May I suggest to add something in this vein as a second paragraph to that section:

When each of the differing versions are notable in their own right, are based on reliable published sources, and there is enough difference between the competing versions to be noteworthy, then proceed with what is described in Wikipedia:Neutral point of view: "(...) all significant published points of view are to be presented (...)", each with their own reference. An example that also the difference between the sources needs to be non-trivial: for instance, if competing official sources give a few percent difference in population numbers for an average city, that is maybe not a "notable" difference - just round the number and put "approx." before it, referring to the available reliable sources.

? --Francis Schonken 07:59, 23 October 2006 (UTC)

Newspapers

The FAQ refers to A New York Times account of a George Bush speech is a secondary source. This is creating the false impression that all newspaper articles are secondary sources, when actually, a large amount of newspapers accounts are primary sources. Wording needs to be found to make that distinction clearer. The reason I am rasing this, is that some editors are claiming that an article that appears in a small town newspaper (e.g. 'The Post Crescent" From Menosha, Wisconsin) is "a reliable secondary source", because of this FAQ, and claims that that small-town newspaper is more reliable than a book on the subject by a respected author. So, please, let's clarify these issues in the FAQ. ≈ jossi ≈ t • @ 06:25, 10 November 2006 (UTC)

You desribe the book that you propose as an alternative source as encyclopedic which is a tertiary source. Andries 06:33, 10 November 2006 (UTC)

And it is untrue that the book on the subject is by a respected author. The book discusses many subjects and this is just one of many subject and its treatment of the subject is not very good and thorough. Andries 06:35, 10 November 2006 (UTC)

There is no problem in using tertiary sources in WP, on the contrary. ≈ jossi ≈ t • @ 06:39, 10 November 2006 (UTC)

And still better that 'The Post Crescent" From Menosha, Wisconsin... ≈ jossi ≈ t • @ 06:41, 10 November 2006 (UTC)

I agree that a tertiary source may be fine for use in Wikipedia. Andries 08:03, 11 November 2006 (UTC)

Back to the discussion, see what WP:RS had to say about newspapers: More recently, primary sources have been put online, such as the complete run of the London Times, the New York Times and other major newspapers. Some of these are proprietary and must be accessed through libraries; others such as “Making of America”, publishing of 19th century magazines, are open to the public.; See also what our own secondary sources say: An example of a secondary source would be the biographyof a historical figure which constructed a coherent narrative out of avariety of primary source documents, such as letters, diaries,newspaper accounts, and official records.. Clearly the FAQ is misleading editors in asserting that newspaper accounts are secondary sources. ≈ jossi ≈ t • @ 06:39, 10 November 2006 (UTC)

Jossi, then why do you say that a report in a magazine is a secondary source [1]. I see little difference between this article in salon.com and a newspaper article. Andries 08:03, 11 November 2006 (UTC)

A newspaper can be both primary and secondary sources. That is what I am trying to say. Some material in newspapers can be a primary sources and other material can be secondary sources. A 1941 newspaper article thatthea reports the Perl Harbor attack is a primary source. A newspaper column in which a reputable journalist compares Perl Harbor with 9/11 attacks is a secondary source. Magazines are mostly secondary sources, for example The Economist as they not provide news, but analysis of news. ≈ jossi ≈ t • @ 16:12, 11 November 2006 (UTC)

Following your definition (that I do not trust as useful), the article 2004 Indian Ocean Earthquake violates Wikipedia policies, because it uses primary source material i.e. newspaper reports from 2004 and early 2005. Andries 16:40, 11 November 2006 (UTC)

There are thousands of articles in WP that do not conform to policy. In articles about which there is no controversy, sometimes you see situations such as 2004 Indian Ocean Earthquake that uses only primary sources. Nevertheless, the use of primary sources as the only sources for an article is highly discouraged, in particular as these tend to violate WP:NOR, when these attempt to forward a position based on new analysis/commentary/synthesis of these primary sources. ≈ jossi ≈ t • @ 16:54, 11 November 2006 (UTC)

I do not believe that the 2004 Indian Ocean Earthquake violates Wikipedia policies in this respect and I challenge you to test your opinion that the article violates the policies by leaving a message on the article or on the talk page. I do not think that the contributors there will agree with you. Andries 16:58, 11 November 2006 (UTC)

You are misinterpreting my comment. What I am saying is that there are articles that rely mostly on primary sources, and in some cases, such as 2004 Indian Ocean Earthquake that is perfectly OK. Nevertheless, the use of primary sources as the only sources for an article has been always highly discouraged, in particular as these tend to violate WP:NOR. ≈ jossi ≈ t • @ 16:50, 13 November 2006 (UTC)

This is interesting. My take on it is that articles using entirely primary sources may "tend to" violate NOR, but not so much because they use entirely primary sources. An article using entirely primary sources may violate NOR if primary sources are being used to make unsupported analytical claims, but secondary sources being used to make unsupported analytical claims are equally unacceptable. If the issue is lack of analytical context at all, that's an NPOV issue, unrelated to type or quality of sources. Meanwhile, NOR presently states that "Although most articles should rely predominantly on secondary sources, there are rare occasions when they may rely entirely on primary sources (for example, current events or Braunfeld v. Brown)". That's missing in this version -- I'm not certain why. -- Bailey^(talk) 00:01, 16 November 2006 (UTC)

(undent) Is a court ruling a primary source (on the ruling) or a secondary source (on the meaning of the law)? One can argue each way: appellate judges are professional scholars interpreting material, but the cases then are further subjected to analysis, especially for a case of first impression.

I wholeheartedly agree with the objective of the primary/secondary distinction. I think it a mistake to try to force a term from historical studies to neatly fit other areas. And, even among professional historians, the term does not have the same quasi-juridical import that it has on Wikipedia. Robert A.West (Talk) 00:10, 16 November 2006 (UTC)

Truth

If "Doesn't Wikipedia care about truth" is really the most frequently asked question about this subject, wouldn't we be doing ourselves a favour to remove the two words "not truth" from the start of Wikipedia:Verifiability? In other words, isn't any confusion about this created by the wording we ourselves have chosen in our most prominent content policy after our copyright policy? jguk 19:02, 21 November 2006 (UTC)

It draws the attention to what exactly Wikipedia is and what is not. Asserting Verifiability, not truth highlights an extremely important distinction. Let the question continued to be asked, and let reply with a considerate explanation about this seemingly unusual formulation. ≈ jossi ≈ t • @ 19:19, 21 November 2006 (UTC)

(BTW, our most prominent content policy is WP:NPOV, not WP:COPYRIGHT). ≈ jossi ≈ t • @ 19:21, 21 November 2006 (UTC)

As has been noted, Verifiability/NOR can be seen as a consequence of NPOV -- we cannot know that an article is NPOV unless it has reliable sources. Of course, NPOV gets confused with moral relativism. These sorts of things will just happen. Robert A.West (Talk) 19:26, 21 November 2006 (UTC)

Jossi, the problem is that it detracts from what Wikipedia is. If it didn't detract greatly, why is it seen as the most frequently asked question? By all means explain common misperceptions in the policy page - just do this after you have explained what the policy is. I've done enough teaching in my life to know that doing otherwise just confuses people. Greatly. jguk 19:30, 21 November 2006 (UTC)

I've inserted "merely" into the policy page. Now I would like to append a sentence to the end of the first paragraph of the FAQ section "Does this mean we have to include every crank view that can get itself published?". I propose to add: "However, editors should try to avoid inserting or leaving in material they have reason to believe to be false unless it's reported as a quote or indirect speech." --Coppertwig 13:44, 5 March 2007 (UTC)

I'm not crazy about that one, although maybe it could be refined. I might believe that one of the two sides to the global warming debate or the October surprise conspiracy is true and the other false. However, I should aspire to report all significant verifiable viewpoints in proportion to their prominence, whether or not I think those viewpoints are true. Maybe we should say something like all controversial viewpoints should be attributed to at least a general source, so that my global warming debate becomes "Global warming proponents argue X. Global warming opponents argue Y." TheronJ 14:36, 5 March 2007 (UTC)

Not crazy about that formulation either. ≈ jossi ≈ (talk) 15:25, 5 March 2007 (UTC)

Can you suggest another formulation that meets the need? What I'm trying to say in no way contradicts what TheronJ says. I have no problem with "Global warming proponents argue X. Global warming opponents argue Y." What I would like to guard against is people thinking the "not whether it is true" provision in the new version of the policy is a licence to edit a Wikipedia article to say "X." (not "Global warming proponents argue X.") even though they have reason to believe that X is false. Could you also please explain what it is you don't like about the way I worded it, so that I can try to find another wording that might work better for you?

Guessing, how about this? "If there is reason to believe that some material is false or is likely to be false or is contentious, then it must either be deleted or have a prose attribution." --Coppertwig 00:19, 18 March 2007 (UTC)

Inline Citations as a minimum

Look at the first sentence of the answer to "How do I format citations?"

At a minimum, any substantial, surprising, or controversial claim in an article should be accompanied by an inline citation . . .

I am uncomfortable with the "at minimum" formulation, especially since all articles should contain at least one substantial claim. Footnotes exist to aid the reader in locating the relevant fact, either to verification or for further research. In longer articles with many sources, this is good, but there are many articles for which footnotes would serve no purpose and would even create a distraction.

One can write a perfectly reasonable short biographical article for some members of the british peerage using three sources: the Dictionary of National Biography, Burke's Peerage and The Complete Peerage. In most cases, the entries are a page or two at most, indexed by name, and the sources simply provide a cross-check on one another. One editor even argued that s.n. (Latin, sub nomine, [look] under the name) should suffice in place of a page number. Since most of the information can be found in any of the three sources, I see no purpose in associating a particular datum with a particular source.

In a lot of mathematical articles, the oft-mentioned reason for lack of citation is an embarrassment of riches: look it up in any textbook. My proposal to solve that is a "Further reading" section, such as the Britannica used to have, which would serve both as reference and to direct interested readers. The same advantage does not apply to inline citations: the whole argument appears in any of the sources, possibly in a slightly different form. Furthermore, for GFDL reasons, the examples and diagrams are often original, so not footnotable, which would require reduplicating footnotes before and after the unsourced text. Robert A.West (Talk) 19:48, 21 November 2006 (UTC)

I guess the important thing is that no statement should be immune from a request for attribution. So many articles, especially advanced topics in mathematics, can very well be written with simply one textbook reference. This is no problem and there are very few edit wars on advanced math topics :) However, if one editor doubts the correctness of a proof or requests attribution, it should be granted. Even "self-evident" mathematical reasoning should not be immune to requests for inline citation, but only when someone really doubts the correctness, not because it is fashionable to have inline citations. --Merzul 16:32, 4 January 2007 (UTC)

However, you get people who go through random articles and say "cite or I delete" simply because they enjoy forcing people to jump through hoops, not because they sincerely doubt the correctness of something. It's essentially vandalism, and is popular because making the demand takes a few seconds while saving one's article takes much more time.

It's also popular among POV-pushers, who selectively demand citations for material they know very well is true but which is embarassing to their side.

Wikipedia's policies are needlessly encouraging such people. We really don't *want* anyone to be able to demand attribution under any circumstances. Ken Arromdee 21:29, 5 January 2007 (UTC)

"Are Wikis reliable sources?"

"Wikis, including Wikipedia and other wikis sponsored by the Wikimedia Foundation, are not reliable sources, because the content is self-published, anyone can edit them, and the current version of an article may change at any time."

That seems wrong. If you link to a diff or a particular version, it is not true that anyone can edit the version you're linking to and it's not true that the current version can change at any time.

Which means that the only objection left is that it's self-published.

Self-published sources are acceptable under some limited circumstances, so therefore it's possible for a Wiki to be a reliable source.

How about this?

"Wikis, including Wikipedia and other wikis sponsored by the Wikimedia Foundation, are self-published, so they may only be used when self-published sources are permitted. If a Wiki is used in such a manner, the source should be a specific dated version of an article, not the current version, since the current version may be edited by anyone and may change at any time." Ken Arromdee 15:03, 4 January 2007 (UTC)

See also Talk:Jimmy Wales#Question RS and Talk:Paul_McKenna#Legal section struck for examples where people want to use Wikipedia as a self-published source in an article about the subject. Ken Arromdee 15:06, 4 January 2007 (UTC)

Our policy on using self-published sources allows us to use them only in the case that the author is known and recognised as an authority (for one reason or another). In general, on a wiki, there is no single author. I suppose in the unusual case that a wiki page had only ever been edited by people who could be identified as authorities, a version of that page might be a reliable source. This strikes me as a highly unlikely situation, however, where WP:IAR would be the best solution. JulesH 21:45, 17 February 2007 (UTC)

On Wikipedia, there *is* a single author for each individual change, since you can link to a specific diff and see only the change the individual author made. The fact that the whole page was written by many different authors doesn't change this. Ken Arromdee 21:33, 20 February 2007 (UTC)

A Wikipedia article can be thought of in these two ways: as a collection of individual self-published contributions, or as a whole collaborative article. A novel printed as a paper book can similarly be seen in those two ways: as mostly self-published by the author but with some corrections of spelling errors etc. self-published by unidentified proofreaders and the choice of title self-published by the editor.

I would like to remove the statement that wikis are not reliable. It's OK to have a statement that Wikipedia is not to be used as a source for Wikipedia articles, but stating that Wikipedia is not reliable just contributes to Wikipedia's poor reputation for reliability. There's no need for that. Some Wikipedia articles have a good level of reliability. I suggest that the atomic numbers of the elements in Wikipedia is probably at a fairly high level of reliability. Some Wikipedia articles are written by professionals in their field and are closely watched. The collaborative development on Wikipedia is essentially similar to (or the same as) peer review and editing that occurs in some other media. Some Wikipedia articles have gone through Wikipedia peer review, which, while it is not the same thing as peer review in a scientific journal, can actually be more thorough in some ways. Each source needs to be judged on its merits, not just thrown out because it happens to be in wiki form. --Coppertwig 00:21, 18 March 2007 (UTC)

Not truth section

I would like to change

By demanding reliable sources for all assertions, Wikipedia depends on the research and opinions of experts.

to

By demanding reliable sources, Wikipedia depends on the published research and opinion of experts.

Since (1) WP:ATT does not "demand sources" for all assertions and (2) the changed wording emphasizes that we are looking for published sources, not hearsay from experts. CMummert · talk 13:17, 9 February 2007 (UTC)

Citation

I have added what ATT actually now says, and commented out the paragraph that would be a, quite controversial, change in policy. I have an idea on how to reduce this to questions, but at a minimum, this FAQ should represent what ATT proposes. If someone wants to smoth this over, great; if not, I should be back within 24 hours. Septentrionalis PMAnderson 03:23, 10 February 2007 (UTC)

Query

NE2, could you say exactly what the sentence you restored (in bold below) means?

MySpace: MySpace is generally not a reliable source because its material is virtually always self-published, and much of it is anonymous or pseudonymous. An exception in which it may be an acceptable primary source is when a reliable source has confirmed who it belongs to.

SlimVirgin ^(talk) 13:07, 16 February 2007 (UTC)

If a reliable source says "famous person's myspace is myspace.com/famousperson", then we know whose it is. --NE2 13:54, 16 February 2007 (UTC)

Reworded that statement to make it less ambiguous. ≈ jossi ≈ (talk) 16:46, 16 February 2007 (UTC)

I don't agree with Jossi's rewording: "An exception to this would be those MySpace pages about which there are secondary published sources that attest to the identity of the person featured in that MySpace page." First, the edit summary claims it is to make the FAQ compliant with the Attribution policy, but I can't find where the policy says that. Second, the statement confuses reliable with secondary. We are not talking about interpretation or analysis, just a simple fact: is the purported author or publisher of a MySpace page really the author or publisher? Reliable primary sources are just as acceptable for establishing straightforward facts as secondary sources, and we should never use the phrase "secondary source" as if it were a synonym for "reliable source". --Gerry Ashton 17:01, 16 February 2007 (UTC)

I see your point. I have reworded it to make it compatible with the wording in the policy. ≈ jossi ≈ (talk) 19:44, 16 February 2007 (UTC)

I still think that the current wording is unclear. There are two problems with MySpace: first, it is generally anonymous/pseudonymous. Second, it is self-published. Establishing identity using reliable sources answers the first objection, but not the second. This should be made clearer. Robert A.West (Talk) 20:05, 16 February 2007 (UTC)

The self-published aspect is already covered in the section about self-published sources in articles about themselves. ≈ jossi ≈ (talk) 22:26, 16 February 2007 (UTC)

Yes, but so is the bit about anonymous or pseudonymous sources. This is a FAQ, not a policy page: we are explaining how policy is applied in certain common cases that come up over and over again. The audience is someone who has read the policy and doesn't quite get it, not you or me or Slim Virgin. Robert A.West (Talk) 15:29, 17 February 2007 (UTC)

Major rewrite of not truth section.

For nearly six months, the WP:NOTTRUTH section has read:

Doesn't Wikipedia care about truth?

This is a common misunderstanding. Wikipedians care greatly about producing a high-quality encyclopedia. By demanding reliable sources, Wikipedia depends on the published research and opinions of experts. For many subjects, this will be the truth, pure and simple. For others, it will be as close as we will ever get to the truth. For still others, it is merely the best information currently available. For controversial subjects and matters of faith or opinion, the truth Wikipedia strives for is to accurately describe all sides of the debate, including their strengths and weaknesses. Of course, experts are sometimes wrong, and theories are frequently incomplete. Because Wikipedia is continuously updated, it can improve as human knowledge improves, depending on reliable sources.

This has now been changed to a phrasing that I think is less helpful.

Wikipedians want to produce a high-quality encyclopedia. By insisting on the use of reliable sources, Wikipedia depends on the published research and opinions of informed commentators. Editors ensure that all majority and significant-minority opinion is included in articles — in rough proportion to the representation of those views in reliable published sources — but without otherwise indicating which of the views, if any, may be true.

We simply present reliably attributed statements, views, and arguments, then allow our readers to judge for themselves. That is, the threshold for inclusion in Wikipedia is whether material can be reliably attributed, not whether any individual editor holds it to be true.

It seems to me that the revised version is oriented more towards a person who already understands NOR and NPOV and less towards the audience of this FAQ. It misses two points that I think are essential, and somewhat conflates WP:ATT with NPOV. First, it omits that the policy acually improves the truth-value of the encyclopedia. If we discuss who believes Foo and who believes Bar, and do so accurately, we are speaking truth. Second, it omits the relation between continual improvement and this policy: as experts change their opinions, Wikipedia expects to keep up.

The discussion about significant minority opinions and letting readers decide for themselves belongs in an NPOV FAQ. We properly exclude a significant amount of attributable material from qualified researchers because the view is obsolete, or is too new to have gained significant support, or has been rejected by the vast majority of researchers. The material passes WP:ATT, but fails other policies. Robert A.West (Talk) 16:21, 17 February 2007 (UTC)

Hi Robert, I take your point, thank you. The reason I changed it is, first, that the passage had a didactic, almost patronizing, tone to it that I feel is best avoided. Secondly, it was wordy, and I tried to tighten the writing throughout the FAQ. Third, and most importantly, it implies that we do care about truth. But we really don't! We have to put the "truth" out of our minds when we're writing and stick to what authoritative sources in that field are saying, even if it's nonsense. Examples of problem sentences:

• "For many subjects, [what we publish] ... will be the truth, pure and simple."

• Will it? How do we know?

• "For still others, it is merely the best information currently available."

• The word "information" implies it's true: there's no such thing as false information.

• "For controversial subjects and matters of faith or opinion, the truth Wikipedia strives for is to accurately describe all sides of the debate, including their strengths and weaknesses."

• No, we do this for everything, not just controversial subjects, not just matters of faith or opinion. And there's no such thing as "the truth Wikipedia strives for": truth is truth.

• "Of course, experts are sometimes wrong, and theories are frequently incomplete."

• Experts are often wrong, and theories pretty well always incomplete.

• "Because Wikipedia is continuously updated, it can improve as human knowledge improves, depending on reliable sources."

• Waffle. How could it be otherwise? SlimVirgin ^(talk) 16:55, 17 February 2007 (UTC)

The shorter version is more clear, but might be too terse for the intended reader. I gave it a very slight rephrasing that I think makes it easier to see what's going on. CMummert · talk 19:36, 17 February 2007 (UTC)

I agree with Robert, and consequently I'm restoring much of the earlier version. WP is not Truth-atheistic, at most we're truth-agnostic. Where there's no reasonable debate in the sources, there's also no problem in proclaiming that the consensus is true and accurate, or at least as accurate as we can make it.

• "For many subjects, [what we publish] ... will be the truth, pure and simple."

• Will it? How do we know?

If there is no serious discord in the sources, we can pretty well take it as written. The speed of light in a vacuum is in fact 1,079,252,848.8 km/h, George Washington died on 14 December 1799, and, from the Main Page today, a fluidized bed reactor (pictured) can be used in the creation of fuel, rubber, vinyl chloride, polyethylene, and styrenes. All those facts have to be sourced, but they are indeed facts. I would include in this category facts that have been more accurately ascertained since earlier sources were published. Sources that were reliable in 1980 may not be reliable now, at least as to some subjects. Editors have to make a judgment call on what to cull, and that judgment should be informed by how likely the source is to be accurate. Old theories will sometimes be worth noting at some point in the article, in language such as "Pluto was formerly considered to be a planet," but the point remains that there are data that we can be fairly convinced are true.

• "For controversial subjects and matters of faith or opinion, the truth Wikipedia strives for is to accurately describe all sides of the debate, including their strengths and weaknesses."

• No, we do this for everything, not just controversial subjects, not just matters of faith or opinion. And there's no such thing as "the truth Wikipedia strives for": truth is truth.

Unless of course there aren't any sides in the debate. Controversial subjects, by definition, are the subjects in which it is likely that there will be a dispute among and between reliable sources. Your stylistic argument is correct, and I've fixed that in my edit.

• "Because Wikipedia is continuously updated, it can improve as human knowledge improves, depending on reliable sources."

• Waffle. How could it be otherwise?

Easy, if we give the impression that we don't particularly care about providing accurate information, which is the perception this FAQ is supposed to address. DCB4W 20:04, 18 February 2007 (UTC)

There are cases where the reliable source provides false information and the source that provides the correct information doesn't meet Wikipedia standards. In this case, the policy demands that Wikipedia include false information.

The covered bridge example is well known; the sources claim the bridge is closed; the guy who visited it saw that it's not. There's also a recent example where a vandal added a spurious entry saying someone is Iranian, and this is now being reported in reliable sources as a true fact--most likely someone just copied the vandalized information from Wikipedia, but we have no way to prove it. Ken Arromdee 22:00, 20 February 2007 (UTC)

WE are entitled, indeed, requited, to judge our sources; that is one reason "reliable source" has not been defined, because it would mean putting down in words the criteria of that judgment. We should probably avoid doing this as much as possible, but we cannot escape it.

• For the covered bridge; the effort is laudable; but the user's photos are original research, and unless he posts them, how do we know they exist? (Taking thje claim out of the article would be a reasonable comporomise.
• Circular sourcing, in which "reliable sources" say something becuase WP does, is a growing problem, and we should have a guideline on it. If it can be shown that the sources that say "Iranian" are all after that version of the article, and none of them claim primary sourcing, it may be due weight to mention that concern, or simply decide the other sources are more reliable. Septentrionalis PMAnderson 22:59, 20 February 2007 (UTC)

For the covered bridge, the point is that the reliable sources/verifiability policy require adding (or at least leaving in) false information. If you try to deny that verifiability and truth are ever at odds with each other, you're not accurately describing Wikipedia. In some cases, "verifiability, not truth" really *means* "not truth".

In the Iranian vandalism case (I found it: Talk:Sacha_Baron_Cohen#include Persian descent, the key is that none of the sources *say* they cribbed from Wikipedia. The idea that they did is purely conjecture. Going from what we actually know about the sources, they are verifiable, reliable, sources. Ken Arromdee 03:50, 21 February 2007 (UTC)

Re the covered bridge example: If the editor who saw the bridge first-hand convinces the other editors that the bridge is not covered, then they should delete from the article the statement about the bridge being covered. If there is disagreement, then there is a content dispute, and they need to discuss and resolve it; one possible result would be to replace "The bridge is covered." with "Source X states that the bridge is covered." There is nothing wrong with an editor expressing on a talk page opinions, personal knowledge or original research in an effort to convince other editors; they could email photos of the bridge to other editors, for example, to support their argument about deleting the statement, even if the photos would not be acceptable on a Wikipedia article page for one reason or another. In any case, no one should knowingly advocate leaving a false statement in Wikipedia. "Source X states that the bridge is covered" is generally not false even if "The bridge is covered" is false. --Coppertwig 19:10, 4 March 2007 (UTC)

Re requiring false information: I disagree. Wikipedia policy does not (or should not) ever require adding or leaving in false information. Also, the "not truth" wording needs to be changed to make this clear. If whether the bridge is covered or not is a relatively minor detail, editors can simply leave out this (false) detail along with many other details which are left out because they are unimportant, irrelevant to the article, etc. If whether it is covered or not is presented in the "Reliable Sources" as an important topic, (e.g. extensively discussed), then perhaps due weight requires that it be mentioned, but nothing in Wikipedia policy prevents the editors from transforming it into a true statement by inserting "Source X states that ...". --Coppertwig 19:19, 4 March 2007 (UTC)

Paradox

It's a paradox; not caring about the Truth is our best hope of approaching it. This is a guideline; we can afford, here, to reassure those who fear we are Uncyclopedia, and make up what we like. I think it is the case that the effect of WP:ATT will be to make us as accurate as we can hope to be; why not say this? Septentrionalis PMAnderson 23:17, 20 February 2007 (UTC)

When I read "As human knowledge improves, and we include more of it, Wikipedia will become more truthful" my first thought was: can we contrast the last word with "truthy"? CMummert · talk 02:11, 21 February 2007 (UTC)

I'm still not sure we should be hardwiring epistemological skepticism into the guideline. Truth isn't always unknowable, and we shouldn't imply that ignorance of the truth is our normal condition. Honestly, if we can't provide accurate information on a variety of subjects, why do we bother writing this? If we don't think our sources provide accurate information, why cite to them? DCB4W 05:22, 21 February 2007 (UTC)

Truth isn't always unknowable; Euclidean geometry is one of the fields in which the present consensus is overwhelmingly "likely to be the truth, or a special case of the truth." But all encyclopedias have had to rewrite their geology articles in the last 50 years. Septentrionalis PMAnderson 00:12, 22 February 2007 (UTC)

I also removed (just now) a rhetorical flourish. ("Not attempting to proclaim the truth gives us the best prospect of accuracy.") I think it's stylistically superfluous, and logically incorrect. (If anything, including divergent views makes us less accurate, although it makes it more probable that the truth will be included in the various positions included in the article.) Following my own logic, I also rewrote the flourish at the end of the subsection. DCB4W 23:58, 21 February 2007 (UTC)

See the rest of the third paragraph, for which it was written. We are more likely to attain accuracy, in the long run, by following the debate of experts, than by trying, with our limited judgment, to declare which one is right. In the short run, we can be accurate in saying that "both X and Y are now supported by scholarship"; saying simply "X" or simply "Y" is guesswork. Septentrionalis PMAnderson 00:12, 22 February 2007 (UTC)

You're missing my point about the wording. Comprehensiveness is the enemy of accuracy. If we have eleven incorrect theories and one correct one, the article is horribly inaccurate (eleven errors!) but it does have the answer contained therein. Our policy is the way it is not because it leads to accuracy, but because we are not always in a position to discern truth from falsehood. I think we're disagreeing over wording rather than policy. DCB4W 00:23, 22 February 2007 (UTC)

No, that's an argument that comprehensiveness is the enemy of precision, which I suppose it is. You should also consult Robert West's post above the section break: saying "these are the twelve theories" can be done with perfect accuracy. Septentrionalis PMAnderson 00:27, 22 February 2007 (UTC)

Not necessarily. For one thing, the words "precise" and "accurate" are usually used interchangeably. To the extent that they're used distinctly, "accurate" means "free of error" whereas "precise" means "exact." Look at Accuracy and precision for the gunshot metaphor. The comprehensive approach is the shotgun: it's neither precise nor accurate but the target is extremely likely to be covered. The Wikipedia approach is the shotgun; aiming may imply POV or other editor bias. The statement "these are the twelve theories" may be accurate, but an article that includes 12 theories, 11 of which must logically be incorrect, is inherently inaccurate. It's an inaccuracy we accept as the price of doing business, because we're willing to sacrifice accuracy for comprehensiveness. Again, it's not clear to me that we're arguing over anything more than semantics at this point. DCB4W 00:47, 22 February 2007 (UTC)

Emphasis

Actually, I think DCB4W is arguing over emphasis; he has a different set of concerns than I do. He seems to be chiefly concerned with the pseudo-scientists and their websites; I don't think WP:ATT can, or is intended to, deal with them. WP:NPOV#undue weight and several guidelines exist for them. Perhaps some additions to the section, outlining distinct concerns, are in order. Septentrionalis PMAnderson 20:08, 23 February 2007 (UTC)

Faq and WP:RS

What does WP:RS have that this page or WP:ATT doesn't? Hiding Talk 21:32, 27 February 2007 (UTC)

Good question. Answer: Nothing that I can spot. ≈ jossi ≈ (talk) 21:55, 27 February 2007 (UTC)

Leads me to believe we could think about opening up the deprecation of WP:RS in favour of this. Personally I'd rather see WP:RS redirected to where WP:V currently is. Where to point WP:V is a problem though, maybe WP:ATT#Key principles? Hiding Talk 14:50, 28 February 2007 (UTC)

I would agree to WP:RS >> WP:ATT#Key principles ≈ jossi ≈ (talk) 15:59, 28 February 2007 (UTC)

Wikis as reliable sources

Wikis, including Wikipedia and other wikis sponsored by the Wikimedia Foundation, are self-published, so they may only be used when other self-published sources are considered reliable. Even as self-published sources, Wikis should only be used when they are not subject to arbitrary editing; for instance, the current version of a Wikipedia article may be edited at any time, so a better source would be a version from a specific date, which does not change. +

[2]

I did discuss it, or tried to. I also pointed out places where people wanted to use Wikipedia as a self-published source under circumstances where our current policy allows self-published sources. The objections to Wikis as sources are that

• They are self-published
• They may change, and you can't tell which user wrote something or even be sure that it'll still be around
• The real-life identity of a user may be uncertain

The second objection just *isn't true*, since Wikipedia allows you to link to a specific dated version, or even to a specific diff, which is less likely to change than most other web-based sources. The first and third objections are easily taken care of: use them only when self-published sources are allowed and only when there's no reasonable doubt about who the author is. It's true that the latter is a judgment call, but WP:RS *itself* says that self-published sources are allowed under certain circumstances "so long as there is no reasonable doubt who wrote the material".

It may not have been worded that exact way before, but the paragraph in fact is consistent with policy, and the flat-out statement that they are never, ever, allowed, is not. Ken Arromdee 14:33, 28 February 2007 (UTC)

• If you're basing this on Talk:Paul McKenna#Legal section struck then no way would I allow Wikipedia diffs as source there. Wikipedia attempts to avoid self reference, for starters, and that's before the legal implications. Let the lawyers do what they are paid to do. Hiding Talk 14:56, 28 February 2007 (UTC)

No, I'm basing it on policy, I only thought of the McKenna issue afterwards.

Other examples include:

Talk:Jimmy Wales#Question RS

Talk:Eric S. Raymond/Archive 2#Response by the target of the article, specifically the point where Eric says that the person who accused him of making pro-war entries to the Jargon File apologized for it.

And remember what self-reference is. Self-reference, according to WP:SELF, is implying that the article is a Wikipedia article. Using Wikipedia *as a source* doesn't violate self-reference.

I'd also claim the McKenna reference should be allowed, since we have an article about a person, quoting a self-published source to show what the person has said. This fits Wikipedia's rules governing use of self-published sources. (Yeah, I commented on that page while forgetting this; I goofed.) Ken Arromdee 18:52, 28 February 2007 (UTC)

Then I see no point in taking this further, since I believe it is counter-productive to do something people have been asked not to do by the foundation's legal counsel. I would suggest that these examples violate WP:POINT. I would suggest they are counter-productive, and that they are harmful to the project. I would make the point that in the McKenna instance you haven't met the third point, you haven't found McKenna self-publishing anything. the same at the Eric Raymond page, you have no qualative evidence that is the subject. Posts to Wikipedia talk pages are posts to message boards, and we don't use those as sources. What you have is original research, wiki-lawyering and counter-productive editing. If someone wants to use one of Brad Patrick's statements from the edit history on why we can't edit a page, what do you believe will be the result of that? Brad patrick will simply lock articles, leaving no comment and lots of people will become upset. We have a saying in my house, "don't upset the apple cart". Jimmy Wales' article is the only example that works, because you have a reliable source backing up the point. But like I say, I'm done here. I can't see such diametrically opposed points of view reaching consensus. I support the policies as they stand. Hiding Talk 14:12, 1 March 2007 (UTC)

A post to a Wiki *is* a self-published source, by definition. I don't understand why you think it isn't.

The reason we don't use posts to message boards as sources is that, as stated in the FAQ, we don't know who actually wrote them. If, in fact, there is no reasonable doubt they were written by the subject (as stated in the rules about using self-published sources), then they no longer fall under the prohibitions on using posts to message boards.

And if we use a self-published source in an article about the author, and the author comes by and starts locking articles, and it isn't an office action or some other situation where the author's following Wikipedia rules by locking the article, then that author is abusing Wikipedia. We should not have rules which say "don't do this because someone might abuse Wikipedia to get back at you". Ken Arromdee 14:27, 1 March 2007 (UTC)

Ken's second objection to wikis as a source, "They may change, and you can't tell which user wrote something or even be sure that it'll still be around" is more nearly true than his comments suggest. If I happen to stumble upon the edit that introduces a passage into an article, then it's not true. But usually, I read the passage of interest, see that the article has a long edit history, and I have considerable difficulty in discovering which edit introduced the passage. There is no function, that I know of, to automatically determinte which edit introduced a certain passage. --Gerry Ashton 19:31, 28 February 2007 (UTC)

While that is true, the proper conclusion is not "and therefore a Wiki can never be a source". Rather, the proper conclusion is "and therefore, a Wiki can never be a source, unless you manage to find the particular edit that introduced the passage". It may, in fact, be hard to find the proper edit; but if someone manages to do it anyway, then they've overcome this particular objection.

And anyway, it's usually easy to find edits. Do a binary search; if an article has 200 edits, see if the edit is in version 100. If not, look in version 150; if so, look in version 50, etc. It doesn't take that long. It's even easier in most cases where you really want to use a Wiki as a source because you already know the name of the person; that person may have not made many edits to the whole article. — Preceding unsigned comment added by Ken Arromdee (talk • contribs)

What kinds of sources are generally regarded as reliable?

I started a new subsection. It needs expanding. ≈ jossi ≈ (talk) 19:15, 4 March 2007 (UTC)

I have added material that was originally in RS. But I am concerned that some of that text is outdated or confusing. I have cleaned up the obvious ones, but it needs further attention. ≈ jossi ≈ (talk) 18:41, 10 March 2007 (UTC)

Thanks!  : ) I think it might be best to rank them in order of importance, and state that they are ranked in that way. However, that would require finding a consensus as to their relative importance. — Armed Blowfish (talk|mail) 23:47, 14 March 2007 (UTC)

Proposed or official policy?

The Parent article was promoted to official policy on Feb 16. What about the FAQ page?? --Otheus 21:19, 4 March 2007 (UTC)

We are working on it to eventually upgrade it to a guideline. ≈ jossi ≈ (talk) 21:30, 4 March 2007 (UTC)

I'd support calling it a guideline (rule of thumb that clarifies issues around policy) Night Gyr (talk/Oy) 19:10, 14 March 2007 (UTC)

I would definitely not support making this a full policy. I do not know if I would like it to be a guideline or not, but I am happy with some of ≈ jossi ≈'s recent changes. — Armed Blowfish (talk|mail) 19:46, 14 March 2007 (UTC)

The discussion has always been about eventually upgrading this to a guideline, not to policy. The tone and verbosity of a FAQ is not good for a policy. WP:ATT is the policy, this is ab FAQ. ≈ jossi ≈ (talk) 19:53, 14 March 2007 (UTC)

Something wrong with this page?

What happened to all the new sections on this page? I can see them when I go back in the history [3], but not here. Please excuse me if I've made a stupid mistake. Sparkzilla 06:45, 20 March 2007 (UTC)

Seems to be better now. Very strange :( Sparkzilla 06:46, 20 March 2007 (UTC)

Do we really need a separate section for every kind of source that is nearly always self-published?

See Wikipedia:Attribution/FAQ#Questions_about_the_reliability_of_specific_sources [4]

Almost everything listed there is a type of source that is nearly always self-published. Do we really need a separate section for each one? Why not combine them into one big section that discusses the problems of self-published sources in general, and lists them as examples?

— Armed Blowfish (talk|mail) 23:44, 14 March 2007 (UTC)

Because it is a FAQ, that's why. People find it much easier to look for their question and see it answered, rather than look for the answers to questions that are not been presented clearly. ≈ jossi ≈ (talk) 00:10, 15 March 2007 (UTC)

Maybe it is helpful for other people. I personally find it confusing since it seems to imply that not all self-published sources should be handled the same way. A different explanation is given for each kind of self-published source. Perhaps you could combine it into one big section that says something like, "Are message boards, mailing lists, USENET postings, blogs, wikis, IRC, Myspace, and YouTube reliable sources?" or, "What kinds of sources are usually self-published, and what should I do about it?"

It is also confusing to me in the sense that it focuses more on the source being a blog, message board, wiki or whatever, rather than focusing on the complete lack of editorial oversight. For example, if I found a InterUniversityWiki that was a joint project for inter-university collaboration, and only trusted representative professors were able to get accounts, would you still consider that self-published?

— Armed Blowfish (talk|mail) 00:49, 15 March 2007 (UTC)

How about a statement under the general heading: "Most of the following, except encyclopedias, are usually self-published; in addition, they may have other problems>,"? Septentrionalis PMAnderson 02:12, 15 March 2007 (UTC)

As I understand it, the problems of self-published sources are basically all the same. You mostly get random Joe Bobs whose opinions don't count for anything on Wikipedia. The random Joe Bobs do not represent expert or majority or even large minority opinions. In the rare case you get a Subject Expert rather than a Joe Bob, there is rarely any guarantee that the Subject Expert is indeed the Subject Expert and not a fake. No one has reviewed Joe Bob's in any way, with rare exception. In addition, even if you could say, "This page is reliable at this point in time," many self-published sources, like wikis and personal websites, can change suddenly at any time. Standard exceptions apply to all, but are rare. Did I miss anything? — Armed Blowfish (talk|mail) 03:07, 15 March 2007 (UTC)

Mute inglorious genius

Having just spent a day arguing with an advocate of a book, much of which is criticized by general consensus, I added a new section addressed to the editor who wants to include the Truth as revealed by their pet crank.

I'm sure we've all had this argument. I tried to summarize my experience with it. I do think it belings here; one reason for guidelines is to have a central home for arguments we've had again and again.

But please tighten and refine; if you want to remove it again, please discuss here. Septentrionalis PMAnderson 02:37, 15 March 2007 (UTC)

That text is more appropriate for an essay than for a proposed guideline. I have removed it ≈ jossi ≈ (talk) 02:40, 15 March 2007 (UTC)

The text in question is:

From time to time an editor declares that we must include the views of a some author who has not received favorable reviews, with the explanation that if the Genius received his due, we would include his views in full. He is only ignored, because he opposes patriarchy, or the liberal/conservative conspiracy (pick one), or the fiendish plots of the Foolanders; and we cannot deny the Genius's accuracy without seeing the full file, which is unpublished.

Other editors often see this as the flip side of the section above; but it may be that we sometimes miss the next Einstein, the next Galileo. This is an unavoidable cost of the way Wikipedia functions.

Wikipedia is not a crystal ball, to see what presently unsuspected knowledge will be the commonplace of the future. We are not wise enough to decide what unconventional theory we should proclaim, despite the rejection of the experts. We will be fortunate if we are no worse, no more in error than humanity at large; it is not our business to proclaim the Truth, but to record the error which the present humanity possesses.

This addresses a question frequently asked. I hope someone can express it better. Septentrionalis PMAnderson 03:17, 15 March 2007 (UTC)

WP:NPOV#Undue_weight says, "Minority views can receive attention on pages specifically devoted to them — Wikipedia is not paper. But on such pages, though a view may be spelled out in great detail, it should not be represented as the truth." WP:FRINGE suggests there should be another, adequately reliable source to prove the subject is significant enough to even have an article. I don't know if that helps. — Armed Blowfish (talk|mail) 03:18, 15 March 2007 (UTC)

It doesn't. This is a case where there is no question that the book concerned is notable; and the sources which prove its notability agree that 40% of it is tosh. The lone believer insists that attributed statements to thiw effect not be included.

I'm not asking for help; I think the lone editor recognizes that he is outnumbered. I'm just tired of having the same conversation over again. I suppose the argument strictly belongs in WP:NPOVFAQ, though. Septentrionalis PMAnderson 03:39, 15 March 2007 (UTC)

The anti-blog stuff

I want to echo some embedded comments above. It's, well, basically kind of silly to declare that blogs are unreliable. It's like saying "sources printed on A4 paper are unreliable". A web log is simply a medium, and the content conveyed by that medium has to be evaluated on its own merits. Jakob Nielsen's blog is a damned-well reliable source, as he is arguably the most emminent writer in the field of computer-human interface usability (and absolutely in the top five even if you have another favorite.) Notably (no pun intended) he cites sources (often to usability studies performed by his own company, but I don't see anyone, anywhere challenging them, and their methods are sound within the bounds of social science). The blog of Jimmy Smith, high school sophomore in Norman, Oklahoma, is not likely to be reliable for much of anything other than what girls Jimmy thinks are cute. But the same would almost certainly be true of a book somehow published by him, or a Jimmy's World TV show hosted by him. It's not the medium, it's the reliability of the message. The problem with blogs (and wikis) generally is not really a problem with blogs, etc., it's a problem with lack of editorial control, and that problem is something that some blogs, etc., actually escape. — SMcCandlish [talk] [contrib] ツ 00:32, 18 March 2007 (UTC)

"Synthesis" does not mean what some editors seem to think it means

Okay, I am amazed, nay stupified at how people cannot seem to grasp this concept. Thusly, I am going to go over it again. Slowly.

This is the dicdef of synthesis:

the formation of something complex or coherent by combining simpler things

Thus, the word "synthesis" is interchangable with "collection" or "combination". This is not debatable. That's the definition as far as Wikipedia is concerned.

Now let's look at the FAQ text everyone seems so intent on keeping:

Original writing is desirable; it is original ideas or new synthesis of ideas and viewpoints that violate policy.

That might initially seem acceptable, but let's replace "synthesis" with the above words which mean the exact same thing.

Original writing is desirable; it is original ideas or new combinations of ideas and viewpoints that violate policy.

Original writing is desirable; it is original ideas or new collections of ideas and viewpoints that violate policy.

Neither of those are in line with the Wikipedia policy on OR, because they effectively prohibit any article from drawing on multiple "ideas". Don't believe me? Here's an example, taken from a real article:

Research suggests that ADHD arises from a combination of various genes, many of which affect dopamine transporters. Suspect genes include the 10-repeat allele of the DAT1 gene, the 7-repeat allele of the DRD4 gene, and the dopamine beta hydroxylase gene (DBH TaqI).

Seems fine and dandy, right? But wait, look at that second sentence. It clearly uses not one, not two, but three different "ideas", or a collection/combination of ideas, if you will. And since I can't find any other source which used that exact collection/combination of ideas, that makes it a "new" collection/combination of ideas. This means it's original research, and must be removed from the article immediately.

Obviously, the above reasoning is inconsistent with the actual policy, but if someone were to edit an article based on this FAQ that would be the conclusion they would have to come to. Hence, the word "synthesis" should not appear in that sentence.

Now as to Jossi's assertion that "there's a new definition of synthesis according to policy": No. You are wrong. Try again. The word already has a well established definition in the english language. If I decided to re-define the word "pederast" to mean "experienced Wikipedia editor" and then wrote a policy FAQ using that "new definition", exactly how many seconds would it be before the FAQ was deleted and I was banned? (And yes, this is an extreme example, but I'm trying to prove a point here) The word already has a definition that almost all english speakers understand, and trying to re-define it will only confuse people.

I am amazed that this is such a big issue. This is a mind-bogglingly simple concept to grasp: the word you are intent on using has a definition that makes the sentence inconsistent with policy. I realize it's a nice-sounding word (most latin-derived words are) but no amount of shoehorning here will fit it into the FAQ properly. Someone made a bad word choice a while ago and nobody caught it until now. It's a single word; find another one to use. -- Y|yukichigai (_ramble argue ^check) 10:22, 19 March 2007 (UTC)

I'm not following your objections. Can you make them more succinctly, please? SlimVirgin ^(talk) 10:48, 19 March 2007 (UTC)

You cannot be serious. But in case you are: the word "synthesis" means "collection". That's it. That's all it means. What policy sets out to prohibit is a collection of ideas that implies or leads to a separate, new idea. The term "synthesis" doesn't cover that; it covers any collection, meaning anything made up of multiple elements. This includes things like the above sentences from the ADHD article, which are obviously not OR. The word "synthesis" is not the right word to express that particular concept of the OR policy, unless it is specifically mentioned that said synthesis must lead to a new idea or conclusion. -- Y|yukichigai (_ramble argue ^check) 21:44, 19 March 2007 (UTC)

I am not following either. And BTW, my comment was not that "synthesis" has been defined in policy as a term. That is a straw-man argument. What I said is that synthesis has been used in this context in WP policy, and it is a long-standing formulation as it applies to no original research. ≈ jossi ≈ (talk) 14:36, 19 March 2007 (UTC)

Then let me rephrase my example: if I wrote a policy where I used "pederast" in the context of "experienced Wikipedia editor", and nobody removed it for, say, a year, would it be okay to leave it in (and also insist that the same use appear in all other policy documents) because it was "a long-standing formulation"? No, it would not. The word clearly has a well-established meaning that is different than what we want to talk about. There is no excuse for using the wrong word. -- Y|yukichigai (_ramble argue ^check) 21:44, 19 March 2007 (UTC)

Your example on ADHD, upon which you base your case, is not original research as it is simply providing information that is well sourced and attributed without attempting to draw a conclusion or assert a synthesis of these viewpoints. ≈ jossi ≈ (talk) 14:40, 19 March 2007 (UTC)

YES! EXACTLY! THAT WAS THE POINT!!! That sentence is not original research, but the langauge in the FAQ says it is while the current "synthesis" wording remains. And again, you used that word "synthesis"; from the context of your sentence, it seems like you think it means something like "a new creation based on assumptions from multiple, lesser elements." Am I right? That's the problem; that isn't what "synthesis" means. The word "synthesis" is interchangable with the word "collection". I don't know how I can make this any clearer. -- Y|yukichigai (_ramble argue ^check) 21:31, 19 March 2007 (UTC)

I think synthesis can cover a range of writing: simple collection of facts from various sources at one extreme, to combining earlier ideas to form a new idea at the other extreme. I think editors could write some passages that could be called synthesis and yet be acceptable in Wikipedia articles. So I tend to agree with Yukichigai, although his use of synthesis does not seem to include the generation of new ideas. --Gerry Ashton 15:00, 19 March 2007 (UTC)

I'm with Yukichigai on this, completely, but I think this dicussion needs to happen at Wikipedia talk:Attribution since this is really a dispute about the language of the policy; the language in the FAQ is simply derived from that. The word "synthesis" is clearly being misused here, and even if someone can dig up an alternative dicdef somewhere that supports the way the term is being used here, that won't solve the problem, as ambiguity about the meaning will remain, and the policy will be open to too much interpretation. — SMcCandlish [talk] [contrib] ツ 15:55, 19 March 2007 (UTC)

Unstated or implied synthesis

I have raised this before, but this is a good time to raise it again... I often come across articles that imply a synthesis but don't actually state it. ie, where the various parts of the synthesis are laid out in such a way as to strongly lead the reader to a given conclusion, but where the conclusion itself is not actually stated. In our model it would be saying: A + B (= C). When I have raised the issue at RfCs, I am often told it is fine, as long as the article just states the facts without actually reaching a conclusion. The reader can reach their own conclusions. The problem is that given the facts stated (often cherry picked) and the way they are laid out, the reader is led directly to the unstated conclusion the writer wants them to reach. Thoughts? Blueboar 15:06, 19 March 2007 (UTC)

That is why, Blueboar, there is no policy or guideline that trumps editors' common sense and good judgment. If a collection of facts is presented in such a way as to lead the reader in a specific direction, that would not be acceptable, would it? ≈ jossi ≈ (talk) 16:46, 19 March 2007 (UTC)

This is what I am trying to determine... when is it acceptable and when is it not? Where do we draw the line between "present the evidence and let the reader decide" (which, in theory, I agree with) vs. "present a synthesis with an implied conclusion" (which, in practice, I do not agree with). I am wondering if we need to say something about unstated/implied conclusions in the FAQ, to further explain the idea behind WP:SYNT. Blueboar 19:48, 19 March 2007 (UTC)

You are welcome to try and add some examples, but I think that these nuances are better addressed in article's talk by involved editors, and if there are disputes about a specific case, they can take it to WP:DR. ≈ jossi ≈ (talk) 20:38, 19 March 2007 (UTC)

Example(s) of unpublished synthesis

I've removed the example of synthesis from Wikipedia:Attribution on the grounds of it being ambiguous and hard to follow. I propose that a more straightforward example is given on this page to illustrate what Wikipedia means by an unpublished synthesis of ideas. One possible example could be:

Example:

"In 1950, average rainfall in Northland was 10 centimeters per year_{Reliable source 1}, whereas in 2005 the average rainfall was 9 centimeters per year_{Reliable source 2}. Therefore the climate in Northland is changing, and getting dryer over time_{No source}."

Under the attribution policy, the conclusion that the climate in Northland is getting dryer is not permitted, where it is not backed up by an appropriate source. This is because coming to a conclusion like this requires careful consideration of the source data, which Wikipedia is not in a position to verify. A climate researcher would consider how the measurements were made, and analyse the available data and relevant theories, before coming to a conclusion. Wikipedia is not in a position to verify this background research itself. Instead, a conclusion from combining other ideas must be attributed to a reliable published source, unless it is straightforward and uncontroversial.

Lots of alternative examples of what Wikipedia means by unpublished synthesis would also be possible. Any comments on whether the example is appropriate, how it could be improved, or alternative suggestions? Enchanter 22:09, 19 March 2007 (UTC)

I like the explanation but not sure I like the example all that much. It is very succinct and gets to the point, but it might mislead some into thinking they can't draw more obvious conclusions, e.g. "In 1950, Northland had 10 centimeters of rainfall,^{[reliable source 1]} but was considerably drier in 2005,^{[no source per se]} with only 2 cm.^{[reliable source2]}" Or am I being paranoid? — SMcCandlish [talk] [contrib] ツ 22:30, 19 March 2007 (UTC)

PS: Something for the example: It is also possible that the data from 2005 only reflect a particularly dry year, and that 2004 and 2006 both saw 12 cm of rainfall. — SMcCandlish [talk] [contrib] ツ 22:34, 19 March 2007 (UTC)

I agree with both points. One way of addressing the first point might be to add your comment into the example, by including something like "On the other hand, a straightforward observation which draws no novel conclusions, such as 'In 1950, Northland had 10 centimeters of rainfall^{reliable source 1}, but was considerably drier in 2005^{no source per se}, with only 2 cm^{reliable source2}', may be acceptable.". Alternatively, we could look for another example altogether. Enchanter 22:43, 19 March 2007 (UTC)

I kind of like just explaining it. Draws a line (possibly a fuzzy one, but still a line nevertheless). I think the This Is OR, This Alternate Is Not OR example format will be helpful and reduce strife and confusion. — SMcCandlish [talk] [contrib] ツ 22:53, 19 March 2007 (UTC)

NB: Please do see the thread above about misuse of the word "synthesis" and related dicussion about rewording the entire OR section at the main talk page. I don't think either conversation has a whole lot to do with this example's merits, but it's better to have a sense of the entire meta-thread, I feel. — SMcCandlish [talk] [contrib] ツ 23:06, 19 March 2007 (UTC)

Putting the above together, and adding some introductory wording, I propose inserting:

==How do I avoid doing original research?==

Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position.

===Example of an unpublished synthesis of published material===

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

Example:

"In 1950, average rainfall in Northland was 10 centimeters per year_{Reliable source 1}, whereas in 2005 the average rainfall was 9 centimeters per year_{Reliable source 2}. Therefore the climate in Northland is changing, and getting dryer over time_{No source}."

Under the attribution policy, the conclusion that the climate in Northland is getting dryer is not permitted, where it is not backed up by an appropriate source. This is because coming to a conclusion like this requires careful consideration of the source data, which Wikipedia is not in a position to verify. A climate researcher would consider how the measurements were made, and analyse the available data and relevant theories, before coming to a conclusion. Wikipedia is not in a position to verify this background research itself. Instead, a conclusion from combining other ideas must be attributed to a reliable published source, unless it is straightforward and uncontroversial.

On the other hand, a straightforward observation which draws no novel or controversial conclusions, such as "In 1950, Northland had 10 centimeters of rainfall^{reliable source 1}, but was considerably drier in 2005^{no source per se}, with only 2 cm^{reliable source2}", may be acceptable.

Any suggestions for improvement, objections, or alternative ideas before putting it into the FAQ? Enchanter 01:07, 21 March 2007 (UTC)

Proposed new Q/A, on "experts"

This would address a situation I've run into more than once:

• Q: I'm an expert in this field, but my edits keep getting reverted; why do I have to cite a source when I am a source?
• A: No Wikipedia reader can know that you are an expert in the field that the article is about, nor judge the particular information you have edited or added (or even removed) on the basis of the expertise of its original author or later editors. Who you are and what qualifications you have are completely transparent to users of the encyclopedia. This is why sources must be cited; they are the only means that users have for ascertaining whether the article has a reliable basis. Telling other Wikipedia editors what your qualifications are and providing proof of them does nothing to get around the core issue. While Wikipedia welcomes and is enriched by editors with expertise in any field, the fact that you are an expert does not mean that you are exempt from Wikipedia policies about the attributability and reliable-source verifiability of information. If you believe that sourced information in an article is incorrect and/or that its sources are unreliable feel free to remove the questionable material (or correct it, if your correction still agrees with the cited sources, or you provide a new source). If other editors object, take up the matter on the article's talk page rather than get involved in an edit war over it, and demonstrate, don't just assert, that the information is faulty.

Could surely be made more concise, but that's the gist I'm trying to convey. See the largest thread at Talk:Albinism for a good example of why this is needed.

— SMcCandlish [talk] [contrib] ツ 22:51, 19 March 2007 (UTC)

Alternate version:

• Q: I'm an expert in this field, but my edits keep getting reverted; why do I have to cite a source when I am a source?
• A: No Wikipedia reader can know that you are an expert in the field that the article is about, nor judge the particular information you have edited or added (or even removed) on the basis of the expertise of its original author or later editors. Who you are and what qualifications you have are completely transparent to users of the encyclopedia. This is why sources must be cited; they are the only means that users have for ascertaining whether the article has a reliable basis. Telling other Wikipedia editors what your qualifications are and providing proof of them does nothing to get around the core issue. While Wikipedia welcomes and is enriched by editors with expertise in any field, the fact that you are an expert does not mean that you are exempt from Wikipedia policies about the attributability and reliable-source verifiability of information.

• Q: But what if I know something in the article is wrong, and that the source was misinterpreted, was never actually reliable, or is now outdated?
• A: If you believe that sourced information in an article is incorrect and/or that its sources are unreliable feel free to remove the questionable material (or correct it, if your correction still agrees with the cited sources, or you provide a new source). If other editors object, take up the matter on the article's talk page rather than get involved in an edit war over it, and demonstrate, don't just assert, that the information is faulty.

— SMcCandlish [talk] [contrib] ツ 22:57, 19 March 2007 (UTC)

I don't believe the word transparent is being used correctly. --Gerry Ashton 17:31, 20 March 2007 (UTC)

"Invisible" then? — SMcCandlish [talk] [contrib] ツ 00:22, 21 March 2007 (UTC)

Anonymous sources

I would like to ask for some clarification on anonymous sources.

An anonymous source is an unnamed person or a work created by an unnamed author. Anonymous sources are not acceptable in Wikipedia, because we can't attribute the viewpoint to its author. Anonymous sources whose material is published by reliable secondary sources, such as Deep Throat in The Washington Post, are acceptable, because Wikipedia's source in this case would be the newspaper, not the anonymous source.

In the article National Union of General Workers I have been discussing the inclusion of an anti-union editorial/commentary by what appears to be an anonymous teacher that was published in Metropolis magazine Killing the Golden Goose. The source magazine is a reliable publication, having been published for 13 years or so. There is a long history of editorials in the magazine, though they are usually not anonymous. In this case, I suspect the writer was concerned about reprisals from the union. However, the person highlights some very important points about the Union's activities that I feel should at least be added as claims.

My proposed text is:

The union has been criticized for its aggressive militancy in pursuing the Shakai Hoken (National Health Insurance) issue. An editorial in Metropolis argued that the union's actions would actually harm the majority of teachers and force the wider community of foreigners working in Japan to pay high premiums with less coverage than alternative private schemes. The article argued that allowing foreigners to opt out of the insurance scheme was a benefit to all but a small number of long-term employees.

There already has been a long, long discussion about this on the Union talk page, that ended in the source being removed as being unreliable. However, the guideline above (in bold) says that anonymous sources can be used when in reliable publications. Should this particular source be included or not? Thank you for your comments. Sparkzilla 01:06, 20 March 2007 (UTC)

I see no reason why not. But you will have to discuss the details at talk page and find common ground with other involved editors. If you cannot find such common ground, you should pursue dispute resolution whose first step would be to ask for third opinions via an request for comment. ≈ jossi ≈ (talk) 03:36, 20 March 2007 (UTC)

Just note that op-eds cannot be used to assert a fact, although these can be used to assert an opinion, if you attribute such opinions to those that hold them. ≈ jossi ≈ (talk) 03:38, 20 March 2007 (UTC)

Unfortunately, I don't think there is much common ground. I had hoped that by placing the question here that I would be able to find opinions regarding the issue (use of anonymous source) that were not infected by the bias of the participants. I appreciate your input and if I could get a few more comments from other editors I would very grateful. Sparkzilla 06:17, 20 March 2007 (UTC)

I concur with Jossi on this, word-for-word. — SMcCandlish [talk] [contrib] ツ 15:25, 20 March 2007 (UTC)

While text from an anonymous source that was published by a reliable publisher can be used, it shouldn't be taken out of context so it seems to have more credibility in Wikipedia than it had in the context of the reliable publication. This would apply to op-eds, advertisements, or quotations that the reliable publication expresses doubts about. --Gerry Ashton 17:36, 20 March 2007 (UTC)

A request for comments has been filed about the use of anonymous sources in reliable publications. The RFC can be found by the article's name in this list, and the actual discussion can be found on Talk:National_Union_of_General_Workers#Request_for_Comment_-_Use_of_anonymous_sources_in_reliable_publications in case you wish to participate. Thank you for your contributions. Sparkzilla 06:50, 25 March 2007 (UTC)

Is it okay to use verifiable sources which support one side of a controversy?

The thread that used to be here was moved to User_talk:Ed Poor, since User:Blueboar argued this was not the right area for it. If anyone wants to continue the discussion, they can do so over at Uncle Ed's talk page. EdJohnston 02:40, 21 March 2007 (UTC)

Synthesis, part deux

Yes, I'm back. Yes, it's about synthesis again. No, I'm not going to stop until the FAQ accurately reflects policy. Now that we have that out of the way....

So, after the little "incident" on the WP:A talk page (which I'm really, really sorry about) I was finally able to get a clear statement on what y'all (or at least SlimVirgin) think "synthesis" means.

Synthesis is a putting together.

— SlimVirgin (talk · contribs)

So if I can take this one step further, we can swap the terms "synthesis" and "a putting together" from now on. Good, on to step two.

Step two is to examine the sentence I've changed. (repeatedly) Prior to my change it read:

Original writing is desirable; it is original ideas or new synthesis of ideas and viewpoints that violate policy.

So if we do that little word swap I described above (plus some grammar fixing) that sentence means:

Original writing is desirable; it is original ideas or a new putting together of ideas and viewpoints that violate policy.

Read that again. "A new putting together of ideas and viewpoints violates policy?" I think not. It does if that "new putting together" advances a point, but the sentence doesn't say that. It doesn't even say "can violate policy" or "may violate policy," it says that all "new putting together of ideas and viewpoints" violates policy. That is incorrect; hence, the sentence needs to be fixed.

Now, of course, comes the part where someone will inevitably say something like, "synthesis is an important part of policy, it needs to be in the FAQ." I agree, it does. Just not in that sentence. Certainly it should go in that section, and probably should have its own section of the FAQ. The problem is that the "synthesis" part of WP:NOR really cannot be adequately explained in a single sentence, certainly not in a manner that is as simple and easy to understand as is really necessary for a FAQ. After all, a FAQ is supposed to be clear, well-spoken, and so easy to understand that there is virtually no chance for people to come to incorrect conclusions regarding what it says.

Finally, I'd like you to just evaluate the new sentence as it stands after my changes:

Original writing is desirable; it is original ideas and viewpoints that violate policy.

Is that sentence not true? No. Does it in any way distort or confuse the reader about policy? No. In fact, I'd argue that because it is so simple and to the point it is vastly improved; as it stands now it is a short, simple statement that is easy to understand and remember, and goes to the very core of the WP:NOR policy: "original ideas = bad for Wikipedia." Hell, it even covers the "synthesis" issue, because a "synthesis" violation must "advance a point" to be a violation since that "point is" (say it with me now) an original idea. -- Y|yukichigai (_ramble argue ^check) 21:43, 20 March 2007 (UTC)

Yep. — SMcCandlish [talk] [contrib] ツ 00:24, 21 March 2007 (UTC)

WP:RS -- Archives of Usenet postings as a historical citation

The ability to cite references to archived Usenet articles from such services as Google Groups as a historical record of events such as software release announcements is an invaluable resource, and should certainly be an exception to the rule against citing Usenet posts. While it is agreed that a Usenet posting in itself is not an authoritative source of verifiable information, Google Groups can certainly be used as a verification that a particular Usenet article was posted on a particular date, and is quite useful as a historical record. --Thoric 03:10, 22 March 2007 (UTC)

Yes, very strong support from my quarter. While the content of a Usenet post may be quesionable under WP:RS (wherever that goes by the time I finish typing this), the timestamps are sometimes of inestimable value. — SMcCandlish [talk] [contrib] ツ 05:04, 22 March 2007 (UTC)

I question why we would need to demonstrate that a particular usenet posting was made on a give date and time. Perhaps this would fall under the "article about itself" clause ... Such time stamped postings can be used in articles about the Usenet itself, but not in other articles. Blueboar 12:43, 22 March 2007 (UTC)

I explained why above -- as a citation of a historical event that wasn't historical enough to be cited in an official publication... for example the announcement of a release of a piece of software. Sometimes such archived, timestamped Usenet postings are the only solid record of a software release date that exists. --Thoric 15:14, 22 March 2007 (UTC)

That raises two questions in my mind ... 1) If the release date of a piece of software is not notable enough to be cited in a more reliable publication, is it notable enough to be mentioned here? and 2) Since he main objection to usenet postings is that we have no way to verify what is being posted or who is posting it... how do we know that a post saying a piece of software was released is a valid record for the release date? Blueboar 15:23, 22 March 2007 (UTC)

To answer #1, I would say that the release of well over 90% of GPL software gets released without any publication in a "notable" publication. For example, you would never find a newspaper article boasting the merits of the next version of GCC, yet archives of gnu.gcc.announce would certainly be a reliable source for release history. As for #2, I would argue that Usenet deserves more credit than you are attributing to it -- especially in the case of archived posts from the days when the World Wide Web was young (or nonexistent). Archived Usenet articles are a valuable historical Internet record. --Thoric 18:01, 22 March 2007 (UTC)

Absolutely. When I was at EFF in the early days, virtually all of our organizational communication to our constituency was through comp.org.eff.talk and comp.org.eff.news other than direct e-mail of our newsletters to subscribers. What I or Mike Godwin said on comp.org.eff.talk was literally EFF speaking, in many cases (if they were just random conversations, not, but when we were speaking as EFF spokespeople on EFF policy matters, they are in fact reliably sourced statements about EFF policywork at that point in time.) — SMcCandlish [talk] [contrib] ツ 20:05, 22 March 2007 (UTC)

Reliability of wikis

This FAQ seems to indulge in the fallacy that all wikis are attempts to build encyclopedias. Now, I can think of several reasons why citing such a page would not be a terrific idea, but remember, in general a wiki is just a way to build a web site. Depending on who is allowed to edit and the policies in place, a wiki may have in principle as much reliability as any other web page. (This may, of course, be damning with faint praise.) What about, for example, software documentation which is made available in wiki form? If the documentation derives from the software developers themselves, this seems just as good a source as any other company website. As long as the citation links to the specific version of the page, what would be wrong with this?

Just an idea for consideration. Anville 18:59, 26 March 2007 (UTC)

Using Wikipedia as a source conflicting with WP:SUMMARY?

This FAQ currently states, "Wikis, including Wikipedia and other wikis sponsored by the Wikimedia Foundation are not regarded as reliable sources. However, wikis are excellent places to locate primary and secondary sources. Many of them license content under the GFDL, which might be worth importing into Wikipedia, but once imported, the material is subject to Wikipedia:Attribution and Wikipedia:Neutral point of view."

I wholeheartedly agree that Wikipedia should not be allowed to be used as source or reference for some fact. This will result in circular reasoning ("We should write X in Wikipedia because Wikipedia writes X here"). This type of logic is fallacious and only begs the question. (This also reiterated in [[[WP:OR]], where it states only primary and secondary sources should be used to cite statements, and tertiary sources, such as Wikipedia, are to be avoided.) However, we currently policy that states Wikipedia article can, in fact, be used as sources, per WP:SUMMARY. The policy states we can summarize the article, without giving references, so as to ensure article do not overly long. Of course, very long articles are burdensome for readers, and summaries of the main points are necessary to condense the general ideas. However, simply using a different Wikipedia article seems to violate the WP:OR policy and the one stated in this FAQ. Thoughts? ~ UBeR 03:25, 27 March 2007 (UTC)

I think that WP:SS does not contradict WP:ATT in letter or in spirit; if the material is sourced in the subarticle as WP:SS requests then it is certainly attributable and so meets the requests of WP:ATT. Of course there are issues where a citation ought to be included in the summary - for example BLP issues, surprising facts, and direct quotes - but common sense says that if the reader is interested in the summarized material then she/he will go to the main article. The summary paragraph should be a short, high-level summary; if it is long enough that it starts to meed many references, it is too long anyway. The whole point of summary style is not to get into much detail. CMummert · talk 03:55, 27 March 2007 (UTC)

Right. It's not WP sourcing itself, its simply cross-referencing to where the sources are. That said, I don't see any problem whatsoever with SS articles having source citations in them for what they do assert; anyone who removed sourced from an SS article should be WP:TROUTed.  :-) — SMcCandlish [talk] [contrib] ツ 04:55, 27 March 2007 (UTC)

Okay, what about intros/leads, however? There are people who are asserting putting two brackets around a phrase so that links to another Wikipedia adequately counts as a citation/reference. They use WP:SS to justify this assertion, but I honestly do not think it is applicable here. Thoughts? ~ UBeR 19:15, 27 March 2007 (UTC)

Are wikis reliable sources? part 2

Sanjiv swarup 07:00, 27 March 2007 (UTC) Are wikis reliable sources? Wikis, including Wikipedia and other wikis sponsored by the Wikimedia Foundation are not regarded as reliable sources. However, wikis are excellent places to locate primary and secondary sources. Many of them license content under the GFDL, which might be worth importing into Wikipedia, but once imported, the material is subject to Wikipedia:Attribution and Wikipedia:Neutral point of view.

My take is that the above statement on the article may have been correct three years ago. IMHO, wikipedia IS a reliable source today. Wikipedia has matured enough to be considered a reliable source. The millions of users everyday definitely think so too. Comments requested.

Wikis (including Wikipedia) are absolutely not reliable. They may be a great resource, but they are inherantly unreliable by their nature. An article you cite to may change dramatically after you cite it (and could in fact say the exact opposite of what you were trying to support). The article you read may reflect a vandalized version and thus not be reliable. Edits are made anonimously with no way to attribute who said what. Need I go on? Blueboar 14:00, 27 March 2007 (UTC)

Not all wikis necessarily need be anonymous. Suppose I found a wiki, InterUniversityBiologyWiki, which did not allow edits without an account, and only admins (who were all heads of biology departments at universities) were allowed to create accounts. Furthermore, these biology department heads only created accounts for their top biology professors and researchers, under these peoples' real names. Granted, I've never seen a wiki like that, but disallowing anonymous edits and making account creation a privelge is possible, and is done on some wikis.

— Armed Blowfish (talk|mail) 14:05, 27 March 2007 (UTC)

One more thing. Wikis do have a permalink feature. — Armed Blowfish (talk|mail) 14:39, 27 March 2007 (UTC)

As I mentioned two sections ago, another example might be the user manual for a piece of software, text which is hosted on a wiki but written by the software developers themselves. (See, e.g., Instiki.) And, of course, if Citizendium ever gets off the ground. . . . Anville 14:51, 27 March 2007 (UTC)

I tried to change this. It got reverted.

My own concern was with using statements by people on talk pages, etc., which should count as self-published sources usable in articles about the author. What discussion there is is still on this page under Wikipedia_talk:Attribution/FAQ#.22Are_Wikis_reliable_sources.3F.22 and Wikipedia_talk:Attribution/FAQ#Wikis as reliable sources

I find the objections to this to be specious. Some of them amounted to "it's difficult to do this". Fine, only allow it when someone has done it despite the difficulty. The biggest objection, that a Wiki can change at any time, is answered by using a permalink. Ken Arromdee 15:08, 27 March 2007 (UTC)

Verifiability: language of sources

This page (and the new Attribution page) have not included the following section of Wikipedia:Verifiability#Sources in languages other than English. Why? Shouldn't it? Tazmaniacs 18:19, 2 April 2007 (UTC)

General positive comment

Although I have problems with WP:ATT in general, this FAQ is a useful policy, I particulary like the extensive 'what is reliable' and 'what is not' section. Good job, and this should become a policy.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  06:38, 8 April 2007 (UTC)

Where should i post questions about the Attribution policy (proposal)?

Where should i post questions about the Attribution policy (proposal)? Itayb 08:45, 10 April 2007 (UTC)

Try Wikipedia talk:Attribution. Blueboar 12:11, 10 April 2007 (UTC)

OK, thanks. Do i understand correctly that the questions that find themselves into this FAQ are those questions, which are frequently asked on the Wikipedia talk:Attribution page? Itayb 12:25, 10 April 2007 (UTC)

Sort of... WP:ATT (the common abbreviation for the Attribution Policy) is fairly new, and is a merger of several older Policies. Most of these questions are a summarized version of many years worth of questions that have come up on the talk pages for the various policy pages that were merged into ATT. Others are questions that no one has actually asked but which were invented, either because someone might ask it, or to help expand on the intent and meaning of the policy... this is a common practice on most FAQ type pages. 13:12, 10 April 2007 (UTC)

Wikipedia_talk:Reliable_sources#Update_examples

Please see my post there about using parts of this policy to update WP:RS.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  20:26, 15 April 2007 (UTC)

Historical or Proposed?

Radiant tagged this as Historical... but I don't think it is. As this is a sub-page of WP:ATT... which has been tagged as "Proposed"... I would think this page should be tagged the same way? Blueboar 15:08, 10 May 2007 (UTC)

Is a court opinion a primary or a secondary source?

The second paragraph here in the FAQ says that: historical documents (of) ... trials are an example of a primary source. But, typically, when a court documents its' opinion, the court includes discussion of their reasoning which typically involves the court citing prior history and prior court precedence, which on its face seems it may be secondary. Under the guidelines of WP:ATT, when courts issue written opinions of court cases, should those documents be given weight as a primary source (as the FAQ seems to say) or given weight as a secondary source? SaltyBoatr 20:18, 15 May 2007 (UTC)

I'd claim that court decisions are usually primary. It might happen that you need to create an article about an important topic, and you need to refer to a court decision. Under the presssure of necessity you might use it, but generally a secondary source would be preferable. (Can you imagine an important court decision that was never discussed in the press?). EdJohnston 21:13, 15 May 2007 (UTC)

I should probably have pointed to this specific example which got me thinking about this question, (see the last two paragraphs). Here, the quotation is from a court document which contains an analysis of historic primary documents. The court is functioning, in a way, like a secondary researcher. Yet, their opinion is wrapped in primary documentation of a modern court. Other issues; I am not sure that I see that court opinions have peer review or fact checking, which is desirable per WP:ATT. I guess that sometimes court opinions are subject to review on appeal. (Yet, in the case here, the Supreme Court refused to review.) Another issue, a selective quote from a court document might easily be argued as requiring 'specialized knowledge' for a reader to confirm, because legal jargon is found in court documents. SaltyBoatr 21:38, 15 May 2007 (UTC)

Dear editors: Editor SaltyBoatr has raised some interesting points, hitting in part at something I have previously discussed in Wikipedia.

First, we can distinguish between a primary source as that term is used in Wikipedia, and Primary authority as that term is used by legal scholars (at least in the United States). A print or reprint of the actual, verbatim text of a court opinion is primary authority in the legal sense, as is the actual verbatim text of a statute, an administrative regulation, a treaty, etc. In the Wikipedia sense of a primary source, a court opinion would be a "primary source" on the issue of what the court said (whereas a law professor's treatise discussing the court opinion would be a secondary source). Similarly, in the legal sense, the professor's treatise would be Secondary authority.

At least two other points should be made: First, if for some reason you are looking to the court opinion for a pronouncement on non-legal matters such as history, then I guess you might argue that the court opinion is a secondary source -- in the Wikipedia sense -- for that limited purpose.

Second, getting back to the point about Primary authority (in the legal sense), court opinions are different from statutes, regulations, and so on. Only a small part of the actual text of a court opinion is what we call the holding (or holdings) of the case, the Precedent or "ruling" (if you will) for which that case stands. See also Stare decisis. Most of the actual verbatim text of a court opinion consists of recitation of the facts, a description of the procedural history, statement of the issues, and statements about the law which, although generally "correct," constitute non-binding material that legal scholars refer to as obiter dicta, or just "dicta." Parsing a case to separate the holding or holdings from the non-binding dicta is indeed difficult for non-lawyers. In some texts the holding or holdings is not clearly stated and must be "distilled," if you will. The talent of analysis of case law in an English common law system (including the United States) cannot be adequately learned by studying one or ten or even a hundred cases.

I have previously raised the question in Wikipedia regarding how to interpret the "no original research" rule and the problem of how an editor can reliably distill the holdings of court opinions without the specialized expertise of a lawyer. An overly narrow interpretation of the rule against original research in this context could lead, in some cases, to exclusion of important court decisions -- for the simple reason that there simply will not be a "secondary source" (in the Wikipedia sense) interpreting every significant court opinion. The volume of case law is simply too great.

Happily, this problem has been obviated to large degree in the Wikipedia legal articles (where I spend most of my time) by virtue of two simple, fortuitous facts: First, there are a lot of lawyers here in Wikipedia. The quality of some legal articles (like articles in many Wikipedia categories) is weak, but many legal articles are surprisingly well researched and presented.

Second, many non-lawyer Wikipedians do make significant, reliable contributions to legal articles. Yours, Famspear 01:18, 16 May 2007 (UTC)

Famspear, Regarding your point that dicta is non-binding; does selectively quoting from the dicta carry a risk of falsely implying, or misrepresenting, authority of the court to the idea being quoted? If yes, then I would see that the WP:ATT warning against quoting from primary sources where specialized knowledge is required would apply in this specific example. SaltyBoatr 01:42, 16 May 2007 (UTC)

Dear SaltyBoatr: Yes, selectively quoting from the dicta carries a risk of misrepresenting what the court actually ruled. But that's not a risk unique to reporting on court opinions. That's similar to a pervasive kind of risk in quoting (or paraphrasing) from many documents in many fields. Let me explain.

In court opinions, the courts themselves often quote from (or paraphrase) dicta found in prior court cases -- generally without even mentioning that the quoted or paraphrased material from the prior case was dicta in the prior case. So, the courts themselves don't even bother to point out that what they are citing really wasn't part of the ruling in the prior case. One reason this practice in the courts is not that much of a problem is that court opinions are written by lawyers to be read by -- other lawyers, of course.

Another reason this habit of courts (of referring to a dictum in a prior case without clearly pointing out that it is a dictum) is not that much of a problem is that the vast majority of statements in the form of dicta are correct statements of law -- it's just that they're not part of the holding of the particular case in which they're found. So the main risk -- as long as you're accurately quoting or paraphrasing, as long as you're not taking anything out of context, as long as you're not making false inferences, etc. -- is that you might say that the court ruled such and such, when in reality the court only said such and such. Either way, the chances are pretty high that the statement will still be a correct statement of what the law is.

Now, I want to compare that to something else - something I find certain people trying to do when editing the tax protester articles in Wikipedia, where I edit quite a bit. Tax protesters (i.e., people who make frivolous, nonsensical arguments against the validity of the Federal income tax laws) are fond of quoting dicta from court opinions and then taking that dicta out of context. For example, there is a statement in many U.S. Supreme Court opinions to the effect that the Sixteenth Amendment (which deals with income tax in a fairly technical way we won't go into here) "created no new power of taxation" (or words to that effect). Now, that particular statement is a correct statement of the law -- it's just that it's not a ruling in most of the cases where you find that language. It's usually found in non-binding dicta.

What tax protesters do is to infer from that statement that the Congress has no power to impose an income tax. They hang on the word "no" and ignore the word "new," and they also argue that Congress had no power to tax incomes before the Amendment. (Thus, they argue, if Congress had no such power before the amendment, and the amendment granted no new power, then Congress must not have the power to impose an income tax.)

But of course, that's not what the quoted statement says at all. Worse, the protesters ignore the explanatory language in the court opinion that typically accompanies that statement. The explanatory language says that the Congress didn't need the Sixteenth Amendment to "create" a "new" power to tax incomes, since under the original text of the Constitution the Congress has always had that power.

The key point is that this kind of error really has little to do with whether the statement is non-binding dicta or, alternatively, part of the holding (the binding ruling) in the court opinion in which it is found. And, in some sense, it also has little to do with the fact that the law is so technical or specialized.

What the tax protesters are doing is simply falsely and deliberately stretching to infer a meaning that is not there. This kind of error can occur in straining to interpret any text the way someone desperately wants to interpret it, whether it's a court opinion, or whether it's a historical document, or whether it's a treatise on Einstein's theory of relativity, or whether it's a description of how to perform brain surgery. It's not really a function of the specialized legal knowledge needed to understand the difference between dicta and holding.

In effect, making a false inference from a text is itself a kind of original research -- a pernicious, incorrect original research. The risk of doing this in connection with a text on Riemannian geometry or elementary particle physics is arguably just as great as doing this in connection with a court opinion -- any time someone is pushing a non-neutral point of view.

By contrast, merely quoting or paraphrasing from dicta in a court opinion, or from a holding, or even distilling a holding in a court opinion, is not really "original research" as that term is used in Wikipedia. What is required here as everywhere else in Wikipedia is that the editor be true to the source. The source should actually stand for the statement for which the editor claims it stands. If the editor is true to the source, then the editor has satisfied the rule on attribution (both as to the requirement of verifiability and as to the rule against original research).

I know I haven't answered your question directly. I just want you to consider these ideas in reaching your own conclusion. Yours, Famspear 03:46, 16 May 2007 (UTC)

Questions about the reliability of specific sources

This is a very very useful subsection that should definietly be part of existing policies. Please note that there exists a similar page - Wikipedia:Reliable sources/examples; some form of merger, one way or another, should be considered. Also, per discussion here, I suggest adding a note on reliability of newspapers.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  17:35, 25 May 2007 (UTC)

Deletion of page

This page has just been deleted and replaced with a redirect. There's no explanation here and I'm not sure if it was deliberate or not (the change was marked as a minor edit).

Even if there are good reasons for deleting the material, incoming links need to be deleted before it is redirected so that we don't end up with meaningless links. I have reverted for now for this reason. Enchanter 22:01, 25 June 2007 (UTC)

By all means fix links (not sure what you mean by "incoming"), but there's no need for this page because the ATT proposal didn't fly, and this page was always a mess anyway. I've therefore redirected it to the ATT essay. SlimVirgin ^(talk) 22:04, 25 June 2007 (UTC)

By incoming links I mean links to this FAQ - there is one at Wikipedia:Attribution for example. Ideally you should make sure that links to pages are removed before a page is redirected or deleted, to avoid there being misleading links. I'm surprised at the redirect - is this as a result of a community discussion elsewhere, or a bold edit on your part? Either way, I would encourage you to discuss and explain edits on the talk page and include more descriptive edit comments. Enchanter 22:15, 25 June 2007 (UTC)

Historical?

Wikipedia:Attribution is NOT marked as historical, and it DOES link to this page for further clarification. Therefore this page should not be marked as historical, not until there is a consensus WP:ATT is historical.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  13:07, 6 July 2007 (UTC)