Wikipedia talk:Attribution/Archive 7

Latest comment: 17 years ago by JulesH in topic Third exception
Archive 1Archive 5Archive 6Archive 7Archive 8Archive 9Archive 10

Towards concision

I wonder if it's time to workshop some parts of the proposed page for concision. I'd like to make some suggestions, but only about the language. If work on improving the language leads to people wanting to change meaning, I'll desist from this idea until the ideas themselves are settled.

Here's a first gobbet for the slab:

Gobbet 1: Questionable or self-published sources in articles about their authors.

Material from questionable or self-published sources may be used as primary source material in articles about that source's author. The material:
  • must be relevant to the person's or organization's notability;
  • should not be contentious or unduly self-serving;
  • must not be used to support claims about topics not directly related to the source or about third parties.
Articles about such sources should not — on the grounds of needing to give examples of the source's work — repeat potentially defamatory claims about third parties, unless the claim has also been published by reliable sources, which should be cited.

My main suggestion here would be to remove the last paragraph entirely, because it is covered by the third bullet point. Its last sentence adds an example which in effect cancels itself, because the reliable source can be cited instead. That last sentence in particular is the sort of elaboration I'd like to see cut from this page. Too many hedges and you end up with a maze.

Any comments? qp10qp 01:13, 15 November 2006 (UTC)

We've had POV pushers from a satirical magazine claim that their stories were "directly related" to them and that therefore they were allowed to use their stories about other people as examples of their work in articles about themselves. That sentence was added to make sure that doesn't happen again. SlimVirgin (talk) 01:52, 15 November 2006 (UTC)

I disagree. That sentence is needed for clarification on the specific situation that the self-published or questionable source is defamatory. What we are saying is: it is OK to use self-published sources in articles about the author(s) of the article, but only if are related to the author's notability and not to third parties when the intention of that author is to defame. Plain and simple. ≈ jossi ≈ t@ 00:42, 15 November 2006 (UTC)

Well, I would ask Slim and Jossi to consider whether their explanations are readable from the paragraph, which to me is very unclear and odd indeed. Yet I feel their explanations are cleanly covered by "[Material from questionable and self-published sources] must not be used to support claims about topics not directly related to the source or about third parties". Plain and simple; and it covers "defamatory" in "third parties". It seems to me a shame to follow such an unequivocal principle with a muddy elaboration that actually appears to advise those who want to include defamatory material on how better to do it.
That's fine by me. SlimVirgin (talk) 02:26, 15 November 2006 (UTC)
Explain to me this: if an article about a questionable source's author should not repeat that source's potentially damaging claims about third parties, why should it be allowed to repeat them if the claim has also been published by reliable sources? Surely the principle on Wikipedia is that we don't use questionable sources where reliable ones are available; and so in this case the reliable claim can be put in the article and the unreliable defamatory claim remain as unacceptable as it was before. I apologise if this is just me. Perhaps my powers of comprehension have suddenly gone to pieces. qp10qp 02:37, 15 November 2006 (UTC)
Because then the fact that this claim was made becomes important. Because of the large amount of attention it received, we can report (carefully) that The Sun claimed that Freddie Starr consumed a live hamster, and in such cases it would be best to use the original issue that made that claim as a primary source, just to dispel any possible idea that somebody else might have made it all up. JulesH 10:18, 15 November 2006 (UTC)
If an article about a magazine wants to describe how it broke the story about actress X's affair with actor Y, it can do so only if other reliable sources published the story too, and we must cite the other source. In this case, the magazine is not our source; it is simply being described. SlimVirgin (talk) 12:29, 15 November 2006 (UTC)

Qp10qp's right that this section is all too complicated at present. The key principle is to use reliable sources. Then, instead of a section saying what is meant by "reliable sources" and how to identify them, we have a section introducing two different concepts, "questionable sources" and "self-published sources". Each of these terms is then defined to have a meaning other than their normal English meaning. We then have a different rule from questionable sources than we do for self-published sources, three exceptions to those rules and three exceptions to those exceptions. That's why it's all confusing.

First, if the term "reliable sources" is used in the key principles, it is "reliable sources" that should be explained in the further discussion. Second, we need to make any explanation far more straightforward.

Might I suggest something along the lines of the following (to replace all the text on "using questionable or self-published sources" and to appear under a heading "reliable sources":

The key principle is that Wikipedia articles must be based on reliable sources

Articles should rely on reliable, third-party published sources with a reputation for fact-checking and accuracy. Sources should be appropriate to the claims made: exceptional claims require stronger sources.

A number of points to note:

• You need to apply common sense when interpreting the principle.
• A source can be reliable for some claims, but not others.
• A reliable source ought to be able to withstand a reasonable level of scrutiny. This does not imply that a source needs to have impeccable reliability, just enough to be reasonably sure that it is reliable for the claim being made.

When considering whether a source is reliable, ask questions such as:

• Who wrote the source?
• Why should the source know the fact it is being used to support?
• What is its reputation for reliability?
• Who has reviewed the source before publication?
• Should there be better sources for the same information, if so, can they be found?
• In particular, if the source is non-English, why can’t an English-language source more readily accessible to readers be found?
• Who published the source?
• Does another source contradict it?
• Do other sources confirm it (in which case consider citing some of the other sources too or instead)

In principle, there's no reason why we should not assess the reliability of self-published sources in the same way as we assess any other source. That they are self-published increases our scepticism about them, and we are more aware that they may be self-serving, which may render them unreliable. But being self-published, of itself, is not determinative of the source being unreliable.

jguk 21:49, 15 November 2006 (UTC)

Gobbet 2: Biographies of Living Persons

It seems to me that the policy towards living persons is unnecessarily placed twice in the proposed page. The two entries are these:

Any unsourced material may be removed, and in biographies of living persons unsourced contentious material must be removed immediately.

  • Any material that is challenged or likely to be challenged needs a source, as do quotations; the burden of evidence lies with the editor wishing to retain the material. Material lacking attribution may be removed, but use common sense: that Paris is the capital of France needs no source. This policy should never be used to cause disruption by prematurely removing material for which reliable sources could reasonably be found. The exception to this is when dealing with claims about living persons, where unsourced contentious material must be removed immediately and not be moved to talk pages.

and

Biographies of living persons

Editors must take particular care when writing biographies of living persons, which require a degree of sensitivity because they could negatively affect someone's life and have legal consequences. Remove unattributed, contentious biographical material immediately if it concerns a living person, and do not move it to the talk page. [1][2] This applies to material in biographies of living persons and to material about living persons in other articles, talk pages, and user pages.


My suggestion would be to cut the second one altogether but move the Jimbo references to the first one. The second one seems to be elaboration—I don't see any real need to split biographies from other contexts, which are all covered in the first occurrence (this is another example of breaking principles into sub-principles for no reason). The second occurrence could be moved to the FAQ.

qp10qp 00:16, 15 November 2006 (UTC)

I am not so sure. The first one deals with the removal of unsourced material in general, and the immediate removal unsourced material from BLPs. The second deals with the details. We may attempt to combine, but not so sure it will work. ≈ jossi ≈ t@ 00:35, 15 November 2006 (UTC)

I'm sure they can be merged. And remember that the details are explained on the WP:BLP page - we need give no more than a reference to this page and a guide as to when you might wish to consult it, not a detailed exposition of what that page says. I suggest deleting the second para referred to above and changing the last sentence of the first para from "The exception to this is when dealing with claims about living persons, where unsourced contentious material must be removed immediately and not be moved to talk pages." to "When dealing with claims about living persons, unsourced contentious material must be removed immediately and not be moved to talk pages - see Wikipedia:Biographies of living persons." jguk 06:30, 15 November 2006 (UTC)

Gobbet 3: The fourth key principle

I'm not sure that the fourth key principle belongs in the company of the first three, or that its accompanying paragraph sings to the principle it trumpets. Here are the four key principles in the raw:

  • Wikipedia does not publish original research.
  • Wikipedia articles must be based on reliable sources.
  • Any unsourced material may be removed, and in biographies of living persons unsourced contentious material must be removed immediately.
  • Building an encyclopedia requires the use of good editorial judgment and common sense.

The first three are strong and direct and recognizable as Wikipedia's three core content principles. The fourth is by contrast long-winded, redundant, and imprecise. More confusing to me is that the instructions following it are mainly not illustrations of the principle, such as it is, but comprise some precise procedural advice about using questionable sources. Apart from the possibility that addressing questionable sources as part of a statement of principle is self-wounding, the material would be placed better, in my opinion, under the heading we have for such advice, Using questionable or self-published sources, lower down the page. Here is the paragraph:

  • Formal rules are not a substitute for good judgment and intellectual honesty. Formal rules cannot always determine whether material is reliable, notable, or relevant to a particular topic. In rare cases, the best source for a particular article may be one that this policy would describe as "questionable." (See below.) Editors are most likely to encounter these exceptions in the areas of popular culture and fiction, where professional sources offer shallow coverage. In these cases, and where there is consensus that the sources in question are accurate, trustworthy, and reasonably free of bias, such a source may be used. If reasonable objections are raised about the use of a questionable source, it should not be used. Questionable sources must never be used to support biographical claims about living persons, or to support fringe theories or marginal positions in the areas of history, politics, current affairs, science, religion, and other academic disciplines.

So, my suggestion is that the fourth key principle be moved because it isn't a key content principle, and that the material it headed be placed in some form under Using questionable or self-published sources lower down the page. I don't think we should try to resuscitate this principle with better wording; in my opinion, this page will be better off if aspirations about common sense, bar an obligatory mention or two, are left to be covered by citing and consensus policies and by the many essays and advice pages on editing.

The paragraph is, in my opinion, the worst written on the page, and needs a going over with the hedge trimmers. But one thing at a time. qp10qp 01:09, 15 November 2006 (UTC)

Yes you are right. I do not agree with your assessment ≈ jossi ≈ t@ 01:26, 15 November 2006 (UTC)
I agree. I don't like that para and I especially don't like where it's placed. SlimVirgin (talk) 01:49, 15 November 2006 (UTC)
Jossi is agreeing with my comment that I didn't think anyone would agree with me, which I must have cut while he replied. Anyway, I'm a married man, and so quite used to not getting my own way.qp10qp 01:42, 15 November 2006 (UTC)
;D SlimVirgin (talk) 01:53, 15 November 2006 (UTC)
:) I forgot to say why I do disagree. Content policies are neither a replacement for the good judgment of editors, nor a replacement for common sense. Have you ever engaged a malicious editor that recites WP policy by the letter to push his POV? Common sense and good judgment are an integral part of this project. And whe we forget that, the project suffers, and disruption is common currency. ≈ jossi ≈ t@ 01:46, 15 November 2006 (UTC)
In that case, I still think the principle needs to be better phrased and that a good chunk of the paragraph needs moving because instead of addressing what you've just mentioned, it addresses a questionable-source issue. qp10qp 01:57, 15 November 2006 (UTC)
I'd be fine with moving all of it, or at least the part that deals with sources, although if you remove that part, the remainder will seem a bit lame as a principle. So my preference would be to move the whole thing into the questionable sources section, because that in fact is the point of the paragraph. SlimVirgin (talk) 02:27, 15 November 2006 (UTC)
Brilliant! Robert A.West (Talk) 06:19, 15 November 2006 (UTC)
I think it's good where it's currently placed. This was how an early version handled the exceptions thing. SlimVirgin (talk) 12:26, 15 November 2006 (UTC)

Whilst it would be sensible to have something along these lines in a policy explaining what Wikipedia is, it does not fit into a policy about sources and sourcing. I agree that it would be best to remove it from this page, jguk 06:26, 15 November 2006 (UTC)

I disagree with moving it. While I think it probably shouldn't discuss specifics in the detail that it does at the moment (perhaps leaving those for a shorter paragraph in the location it has been moved to), I think it is a key principle.
My suggestion for a solution would be:
  • A short "key principle" item that reads somewhat like this:
This policy is not a substitute for common sense
Formal rules cannot always determine whether material is reliable, notable, or relevant to a particular topic. Sometimes, they may contradict what needs to be done to improve the article. In these cases, and where there is consensus of editors that it is the best thing to do, you may ignore them.
  • Leave the existing text where it is now (i.e. in the "questionable sources" section)
  • Add a note to the BLP section echoing the current sentence about that subject.
What do you think? JulesH 09:54, 15 November 2006 (UTC)

Disagree. May as well not bother writing a policy if any consensus of editors on a page (e.g. a consensus of cultists) can ignore it. SlimVirgin (talk) 12:25, 15 November 2006 (UTC)

Third exception for self-published sources

There is a third exception that has been alluded to multiple times in the discussion, but not explicitly mentioned. I think it should be placed first, simply because it is a clear case.

1. When cited by a reliable source.
Self-published material that has been analyzed by a reliable secondary source may be treated as any other primary source material. Furthermore, any work that is widely recognized as reliable by professional researchers may be treated as reliable, even if self-published.

I realize we are working for concision, and this exception could, if people prefer, come under the "common sense" rubric, but so could the other two. I think it is worth stating explicitly. If people agree, I will add. Robert A.West (Talk) 06:14, 15 November 2006 (UTC)

This seems problematic to me - if CNN cites a whacko's Kennedy assasination theory in talking about assasination theories, it suddenly becomes reliable? Are we actually trying to include any specific sources and topics here that are currently being excluded, or is this [{WP:BEANS]]? Phil Sandifer 06:17, 15 November 2006 (UTC)
That's not what it says: I added a "furthermore" to make it clear that there are two cases. If CNN cites a whacko's theory, we should be able to quote from the theory itself if so doing would make the article clearer. That doesn't constitute "Widely recognized as reliable," it is just a primary source that is commented on by a reliable secondary one. The second case covers the original works in some fields of research, when professional researchers as we understand the term were scarce, and much was done by men of independent means, who paid to have their works published. I think those works currently get in either under IAR or because no one notices they were self-published. Robert A.West (Talk) 06:25, 15 November 2006 (UTC)
It introduces a complication, and we wouldn't suddenly start to use a questionable source just because CNN does. SlimVirgin (talk) 12:22, 15 November 2006 (UTC)
Your second-last sentence is covered by the questionable sources exception (where professional coverage is shallow etc). SlimVirgin (talk) 12:24, 15 November 2006 (UTC)
If CNN cites a questionable source in its coverage and that coverage passes tests of relevance, why can we not cite from that source as we would for any other primary source? Obviously, we would not want to introduce new interpretations, but we quote from/summarize primary sources all the time. Take a case in point. The article Sixteenth Amendment to the United States Constitution covers tax protester theories in a section, because they influence significant numbers of people to break the law. Several of the most important books are self-published and the sources are certainly not reliable as to the law. The article is not about the books themselves, and there are reliable sources (authoritative textbooks and court cases) that discuss why they are wrong. Yet, there is an agreement among the editors (including several tax lawyers) that it improves the article to quote from and cite the sources themselves. I would prefer not to have to appeal to IAR for something like this. Robert A.West (Talk) 13:13, 15 November 2006 (UTC)


Second issue. As phrased, we could not include relevant quotations from Benjamin Franklin's self-published scientific monographs, or even Thomas Jefferson's Notes on the State of Virginia, which was originally self-published, because there are later works covering the same territory. That strikes me as silly. Yes, only a troll would argue the point, but there are other valid cases that are less clear. I cited Henry Chapman Mercer above, whose antiquarian work was ignored by academics of his day, but has since become the basis for authoritative professional inquiry. While The Bible in Iron should be treated like any source a century old, it should not be ruled out of court. Robert A.West (Talk) 13:19, 15 November 2006 (UTC)
We can quote Benjamin Franklin's self-published work because it has been republished; we quote it from the professionally published editions. With someone as notable as Franklin, we could in any case quote his self-published work because he is known to be reliable from other publications. In any case, why would Wikipedia want to be so groundbreaking as to quote previously unpublished work by people as famous as Franklin, when we have so much published stuff to go on? qp10qp 14:22, 15 November 2006 (UTC)
Can someone please explain to me how a self-published source could possibly be considered unpublished? I am at a complete loss to understand the last sentence above. Also, my first example is still unanswered: there are no scholarly republications of self-published tax-protester materials, yet some are valuable primary sources for certain articles.
I do not believe that publication of a scholarly edition changes a questionable source into one a non-questionable one. Scholarly editions avoid making editorial changes to the contents (as opposed to form) of the text, so publishing one does not change a work without editorial oversight into one with editorial oversight. If its assertions were nonsense before, republication doesn't make them less nonsensical.
On one of my back shelves, I have an 1812 edition of a home medical guide by Benjamin Rush that gives Dr. Rush as the publisher. As a medical reference it is mostly worse than useless. It might, however, be of interest as a primary source in an article on the history of medicine. Assuming that a reliable secondary source discusses Dr. Rush's opinion of smoking, why should I be compelled to find a reprinted edition somewhere to quote him directly on the subject? Robert A.West (Talk) 15:27, 15 November 2006 (UTC)
Well, perhaps to have prevented you being at a loss to understand me, I should have added the word "reliably" before "published". I normally find your contributions here well-informed, but I have to say that your last two comments are way adrift of Wikipedia policy. Believe me, a massive amount of editorial work goes into publishing documents, even in facsimile. No, of course, nonsense doesn't become less nonsensical by being professionally published and edited, but it becomes usable on Wikipedia, that's all. The policy makes perfect sense to me. qp10qp 16:21, 15 November 2006 (UTC)
As for the book by Benjamin Rush, you should certainly reference a reliably published edition. That may be inconvenient for you, but many inconveniences exist with sourcing, which we have to live with. And, by the way, in an article about the history of medicine, it would not be up to you to find primary sources to quote from; you should use secondary sources which have researched primary sources for themselves. qp10qp 16:31, 15 November 2006 (UTC)
This is silly; the 1812 edition is almost certainly a more reliable text than any modern edition, unless Library of America has gotten around to Rush. Septentrionalis 20:44, 16 November 2006 (UTC)

(unindent) Thank you, Qp10qp, for the compliment about most of my comments. Either I am really expressing myself badly (which happens), or I've slipped a cog. I've tried to be very careful to discuss using primary sources to illustrate the points made by the secondary sources we are using. I thought it well understood that we only use primary sources for such purposes if they are so used by one or more of the secondary sources we are citing.

You seem to be distinguishing among the following cases of primary sources. All are cited by a reliable secondary source that we use in the article.

  1. It cites a manuscript that has been published in an authoritative edition.
  2. It cites a book published by a commercial publisher, using the one and only edition.
  3. It cites a self-published book, using the original edition, which is available in some libraries and sporadically in used book stores.
  4. It cites a self-published book, using a scholarly reprint.

As I read you, we can include a relevant quote from the primary source in cases 1, 2 and 4, but not 3. You seem further to argue that we could use a different edition from that actually used by the secondary source, if a scholarly republication exists. That makes no sense to me. Robert A.West (Talk) 20:02, 15 November 2006 (UTC)

The word "scholarly" doesn't come into this; reliably published covers that for our purposes. You can use a different edition from those quoted in secondary sources, so long as you make no interpretation of it. Any critical or other comment from a secondary source will usually use the edition quoted in that secondary source, though you may use another edition without commenting on it (for example, you might reference a secondary source which says that Caesar was lucky to win at Pharsalus but quote battle details from Pharsalus from Caesar's own commentaries, which exist in many editions and translations); the first is referenced to the secondary source, the second to a published edition of the commentaries. The best thing is to use the quotations given in a secondary source, quoting the secondary source's page number but adding another published edition of the primary source in the books section at the bottom of the page, for general interest, if the one quoted in the secondary source is not available (not obligatory). You can't go quoting parts of a self-published source not already quoted in a secondary source. But there are exceptions for notability, and Franklin would qualify, I'm sure; I've not heard of Rush, and my instinct is that he wouldn't, but you might be able to make the case and win consensus to do so, if the information was not available in any other editions. qp10qp 21:52, 15 November 2006 (UTC)
You say that "There is an exception for notability." This is nowhere stated in the proposal. Its a reasonable idea, but it is not there. The only exception that comes close is for "professional researchers".
Ironically, given your expressed opinion, Dr. Benjamin Rush, Professor of Medicine and Medical Practice at the University of Pennsylvania would probably qualify as a professional researcher self-publishing in his field of expertise. His more famous contemporary, Dr. Franklin, was never paid for research in his life, so could not be considered a professional researcher.
The policy as written militates in the opposite direction from your instinct, and from current Wikipedia practice, and I do not think either your instinct nor the practice of the professional historians in our midst is bad. We mostly do the right thing as regards sources from before the era when nearly all research is done by professionals, and I predict we will continue to do so. I would just prefer that policy actually support what we do. Robert A.West (Talk) 22:50, 15 November 2006 (UTC)
No, I think the proposal as written covers all of this well. The exception for notability is covered by both the following:
1.Material from questionable or self-published sources may be used as primary source material in articles about that source's author. The material:
  • must be relevant to the person's or organization's notability;
2.Professional self-published sources:
When a well-known, professional researcher in a relevant field, or a well-known professional journalist, produces self-published material, we can rely on it if material produced by that writer would normally be regarded as a reliable source.
An exception for self-published material by Franklin could be justified by both of those. Rush, I doubt, but it's not my field. Remember that it's current notability that counts, on which score Franklin trumps Rush just as Van Gogh trumps Anton Mauve. Remember also that the word "research" in Wikipedia also means "thought", and that Franklin's thought was professionally published after his death. Personally, I think the issue of writers who self-published long ago is not the policy's main worry here; it's contemporary self-publishers who worry us. Should we attend to what few writings by Franklin may remain unpublished professionally when we have quite enough material by him to use in Wikipedia articles already? qp10qp 02:28, 16 November 2006 (UTC)
I didn't think that it was necessary to point out to anyone that I am choosing Franklin and Rush as illustrative examples, not because I have an agenda in mind. As you have demonstrated by your reasoning, getting to a reasonable conclusion from the policy as written requires considerable exegesis. Note that you rule that Rush (one of the most famous professional researchers of his day) is not a notable professional researcher, and the Franklin, who was never paid for research in his life, is a professional researcher. That is counterintuitive to say the least. (I don't buy the argument that popular notability should rule the day for figures of the past. Notability among historians of the period would be a better gauge, and there Rush is unquestionably notable.) These are not out-in-left-field cases -- these are the sort of thing that comes up a lot. We should not rephrase policy in a way that requires Talmudic reasoning. Robert A.West (Talk) 09:47, 16 November 2006 (UTC)
P.S. Obviously, I am getting nowhere with convincing anyone on this point. Reasonable editors will do the reasonable thing, as they would if the policy were reduced to the nutshell description. What unreasonable (or unduly mechanistic) editors will do with it will be revealed only when this becomes policy (assuming it does). Robert A.West (Talk) 10:11, 16 November 2006 (UTC)

Purpose of the "Key Principles" section.

At a couple of points above, I've seen suggestions of moving something out of "key principles" into a later section where other similar things are discussed. I think this negates the purpose of having a principles section, which is to provide a summary of the ideas upon which the rest of the policy is based. A reader can stop reading after the key principles section and have a good idea of the kind of thing that is in the rest of the page. This wouldn't be true if we moved the point about always removing unsourced BLP material into the BLP section, and it isn't true now because of the fact that the "use common sense" bit has been moved down into questionable sources.

Can we get agreement that having a key principles section that summarises the rest of the points is a good goal? JulesH 10:12, 15 November 2006 (UTC)

That section about questionable sources wasn't a key principle. Key principles aren't a summary of the page. SlimVirgin (talk) 12:20, 15 November 2006 (UTC)
I acknowledge that some of the text that was included there isn't a key principle, but I think there is a key principle that's now missing: that the rules aren't perfect and that that sometimes it's best to ignore them. Yes, I know that's covered in WP:IAR, but I feel it is an important enough point that it needs mentioning here, right up-front, in a prominent place. Thinking about it further, it doesn't only mean that there are questionable sources that we can use in some situations. It also means that there are otherwise-reliable sources that might not be appropriate to use in all situations. That these rules aren't (and cannot ever be) perfect is an important point, that needs to be as clear as possible. JulesH 12:28, 15 November 2006 (UTC)
I disagree. The issue of good sourcing is one of the most important things we deal with. We can't have a policy on it that advices people to ignore it. SlimVirgin (talk) 13:32, 15 November 2006 (UTC)
Jules, I think the key principle you are talking about was not clear. Was it about good editorial judgement or about common sense? They are not the same, because good editorial judgement usually requires a lot more than common sense, and common sense (don't stroke lions) may not require editorial judgement. The principles of good editorial judgement and common sense will not do any harm being mentioned on this page, but they are not so much content principles as Wikipedia principles, addressed in may other places. As content principles they are vague, lack rigour, and, unlike the first three principles, do not give active guidance to the editors. qp10qp 14:13, 15 November 2006 (UTC)
The principle I'm thinking of is basically "these rules aren't always right; if they get it wrong for one particular source, go ahead and ignore them". It's essentially a statement of IAR, which is a policy I feel doesn't get enough attention these days -- wikipedia has attracted too many "rules lawyers", and I'd like to see their attention drawn to the fact that rules don't cover every situation perfectly wherever it seems appropriate to do so. And here's a good place. IAR has consensus; it's already policy. We've discussed it on this page and on the pop-culture discussion page, and it has been almost universally acknowledged that we can't write a set of rules that will work for every situation. So let's remind people of that, right up front, in the key principles section in bold type.
I proposed the following phrasing above.
This policy is not a substitute for common sense
Formal rules cannot always determine whether material is reliable, notable, or relevant to a particular topic. Sometimes, they may contradict what needs to be done to improve the article. In these cases, and where there is consensus of editors that it is the best thing to do, you may ignore them.
SlimVirgin objected to it on the grounds that it allows consensus of editors to override policy, which she suggested was undesirable. However, I think IAR disputes that anyway. Consensus can override policy, that's the key implication of IAR. But if there's some phrasing that would better get the message across, I'm open to suggestions. JulesH 16:20, 15 November 2006 (UTC)
Consensus of editors on any given page can't override NOR, NPOV, and V, as it says clearly on those pages. SlimVirgin (talk) 21:46, 15 November 2006 (UTC)
I think Jules means to say editor's consensus can determine a specific source is reliable for certain claims. Not invalidate a policy; which can never account for every situation anyways.--Birgitte§β ʈ Talk 22:08, 15 November 2006 (UTC)
No, I think he wants to introduce an "ignore this policy when editors agree to" clause. SlimVirgin (talk) 23:54, 15 November 2006 (UTC)
...and that is unacceptable. IAR does not mean IAPWC (Ignore All Policies When Convenient) ≈ jossi ≈ t@ 02:40, 16 November 2006 (UTC)
My intent here is exactly as Birgitte suggests; simply to point out that the policy cannot be right for all situations and that there are some situations where it should be ignored. I.e., just place a prominent reminder on the page that IAR exists. I believe this is important. JulesH 08:23, 16 November 2006 (UTC)

"Consensus of editors on any given page can't override NOR, NPOV, and V" is a contradiction in the rules, since Ignore All Rules states it can override anything, even those. I've complained about the contradiction before, and nobody seems willing to let it be fixed. If consensus really *can't* override those, then we need to change WP:IAR to say so, rather than contradicting. Ken Arromdee 14:26, 16 November 2006 (UTC)

IAR is by definition a contradiction of the rules, and so there is no way to remove the contradiction between the policies except by abolishing IAR. IAR is in my opinion a wise and realistic policy, not because there is anything wrong with Wikipedia policies but because we all know that you can't legislate for every situation. I don't think this policy proposal tries to legislate for every situation (thank goodness) and so it doesn't need to flag IAR ostentatiously.
No rules benefit by announcing that they can be ignored. In my family, the children have a policy called "ignore all parents", but I'm blowed if I'm going to shoot myself in the foot by reminding them of that myself when I'm laying down the law. qp10qp 15:39, 16 November 2006 (UTC)
Yes, but I dare say in your family you don't have to deal with rules lawyers who won't budge an inch from their exact interpretation of what is and isn't allowed. JulesH 20:17, 16 November 2006 (UTC)
Well, I try. qp10qp 22:07, 16 November 2006 (UTC)

BLP section

I'd suggest removing this. It doesn't say anything useful that isn't already in the key principles section. Opinions? JulesH 10:31, 15 November 2006 (UTC)

I think it's worth retaining because the Foundation regularly stresses the importance of the need for good sources in BLPs. SlimVirgin (talk) 12:18, 15 November 2006 (UTC)
But the point is that it's covered in the key principles section. Nothing would be lost from the policy by combining the two parts that deal with the issue; but concision and focus would be gained. (IMO) qp10qp 13:58, 15 November 2006 (UTC)
I agree with Qp10qp and JulesH. And as long as we link to Wikipedia:Biographies of living persons it is clear that further guidance is available for those wanting it. jguk 21:36, 15 November 2006 (UTC)
Remove policy content that will exist in other policies when ATT replaces WP:V, WP:RS and WP:NOR. We are trying to get this down to 1000 words, right? WAS 4.250 21:42, 15 November 2006 (UTC)
No, we're not trying to get it below any particular number of words. It should be as long as it needs to be, with the emphasis on "needs." SlimVirgin (talk) 21:44, 15 November 2006 (UTC)
WAS 4.250 is right that this policy should not seek to duplicate what is in other policies. And being concise is a virtue in any rule or policy, jguk 21:52, 15 November 2006 (UTC)
To make my point clearer, there's no question of removing the living-persons clause, which is actually one of the few non-negotiable principles here; what's needed is to give the guidance in one place on this page rather than two. It's just an issue of good writing and policy-page presentation. qp10qp 00:21, 16 November 2006 (UTC)

Copy edit

I tightened the writing slightly, and removed two things: first, the prose attribution section, which isn't the sort of thing a policy should address (and we didn't give advice on when to do it and when not, so the section didn't actually say anything), and secondly, I removed the thing from the intro about sources being reliable "for the claim being sourced." This isn't great writing, but I can't see how to improve it, because it's not clear what it means — unless the word "appropriate" would do. Can someone give an example of a reliable source not being reliable "for the claim being sourced." SlimVirgin (talk) 12:16, 15 November 2006 (UTC)

All sources are reliable for some things and not reliable for other things. To put all sources into either a bin marked "reliable" or a bin marked "unreliable" is shear idiocy and the worst sort of black/white thinking. WAS 4.250 21:46, 15 November 2006 (UTC)
Harald, you restored it without addressing the point. 4.250 or Harald, can you give me an actual example of a reliable source that we would decide not to use on a particular topic because it was not reliable "for the source being claimed." SlimVirgin (talk) 22:00, 15 November 2006 (UTC)
I think this would mainly come into play with out-of-date sources which can be used to give information about within the context of history and not in general.--Birgitte§β ʈ Talk 22:11, 15 November 2006 (UTC)
The movie "Star Wars" is a reliable source for the content of the movie, but it would not be correct to quote the dialogue as an example of the correct use of the word parsec. --Gerry Ashton 22:25, 15 November 2006 (UTC)
But what that means is that Star Wars would not be a reliable source for the use of that word. To add that sources must be "reliable for the claim being sourced" is to misunderstand what we mean by reliable. Of course part of reliable is that sources be "appropriate." Historians for history, physicists for physics, and so on, wherever possible. The policy isn't recommending reliable but inappropriate sources, and to have to point that out is kind of odd looking. SlimVirgin (talk) 23:19, 15 November 2006 (UTC)
I agree with the removal of the phrase. The new, reduced phrasing, is clear enough, and, best of all, doesn't inflict that little double take where you wondered why "for the claim being sourced" had been inserted with such prominence and suspected you were failing to grasp something. It may just refer to a normal citing decency, because too many references do dissolve somewhat on being checked. But I have two other theories about the phrase: one is that it was a balance to the deceptively authoritative term "reliable sources", which doesn't actually ensure reliable references, as such; the other is that it attempts to cover the crack in the Wikipedia doorstep opened up by the word "threshold".
I've never been happy with that word "threshold"; I've always believed metaphor is best left out of instructions because it leads to mental puzzling. A threshold is a doorstep, an entry point, a boundary; in my opinion that metaphor creates a problem at the beginning of the policy page because a reference doesn't automatically qualify for inclusion in a Wikipedia article just because its source is attributable to a published source—it must then meet further criteria. So I wonder if "for the claim being sourced" was someone's instinctive attempt to fill the semantic gap that lurks beyond this figurative "threshold" of ours. qp10qp 00:09, 16 November 2006 (UTC)
The new, reduced phrasing ... doesn't inflict that little double take where you wondered why "for the claim being sourced" had been inserted with such prominence and suspected you were failing to grasp something.
Exactly. We need to get rid of any writing like that.
As for threshold, I like the metaphor: don't come in here if you're not well-sourced. For me, threshold invokes a necessary but not sufficient condition, and it emphasizes attribution not truth. I don't see it as meaning that a good source is the only thing required (although it pretty well is the only thing). SlimVirgin (talk) 00:19, 16 November 2006 (UTC)

(reset tabs)So you believe that the statement like "The yellow race probably is not inferior to the white race in intellect and fortitude; though anthropologists incline to the opinion that physically it is nearest of all races to the child and that it may represent the most primitive extant type of man." is not used as a source because it inappropriate by some other WP policy not because it is unreliable?--Birgitte§β ʈ Talk 02:18, 16 November 2006 (UTC)

I'm not sure I understand the question. The source is a students' tertiary source from 1914, and therefore not reliable. I wouldn't use it as a source for anything, but I wouldn't use a students' encyclopedia from 2006 either. SlimVirgin (talk) 06:37, 16 November 2006 (UTC)
I am not going to debate that a source is unreliable because it was marketed towards students, because this is not limited to those sources. I do not think anyone can successfully argue that the Encyclopedia Britannica of 1911 is not a reliable source according to WP. Many articles are sourced from this and much of the text itself is still extant in WP articles. I hope you will concede that the EB1911 contains similar statements on race (if I must dig through the unsearchable scans to find a quote let me know). The question that remains is why do we not source these kinds of statements policy-wise? Is it because we recognize a source may reliable for some things and not for others? Or are these statements inappropriate to use as a source material for some other reason?--Birgitte§β ʈ Talk 20:48, 16 November 2006 (UTC)
I wouldn't rely on the 1911 EB as a secondary source for anything. Are you sure that you would? Isn't it the case that you'd only rely on it if you knew the material wasn't controversial i.e. if you knew you could source it elsewhere? But if that's the case, then you're not in fact relying on it. This is why I wonder about the issue of a source being reliable "for the claim being sourced." If all that's meant is that a source should be appropriate, that's fine, but when I added "appropriate," it was reverted. If what's implied is "we can use the NYT for X, but we don't like what it says about Y," then that's not acceptable. SlimVirgin (talk) 21:40, 16 November 2006 (UTC)
Including one subject for which I have used it, the industrialization of Berlin around 1900? There are better sources, but they are much harder to come by. Septentrionalis 00:20, 17 November 2006 (UTC)


In case I didn't answer your question directly, I would not use as a source for anything a publication that contained the sentence "The yellow race probably is not inferior to the white race in intellect and fortitude; though anthropologists incline to the opinion that physically it is nearest of all races to the child ..." Whether because it's an old source, or racist, or just plain nuts, I wouldn't care. I wouldn't use it because it's clear it can't be trusted. To say "but it can be trusted for some things," misses the point. Even Stormfront can be trusted for some things. But if you have a source where you keep on finding mistakes, then you know not to use it. We have a world of very trustworthy sources available to us. We don't need to use bad sources. SlimVirgin (talk) 21:45, 16 November 2006 (UTC)
No large publication by many hands can survive this standard. The NYTimes, within the last decade, published a hagiography of Samuel Chase, presented as a Style section article on Baltimore, which was (IIRC) equally provably in error; I'm sure it's done worse. No source can be trusted unconditionally. The 1911 Britannica's views on race are held by a vanishingly small set of modern anthropologists; the solution to them is to exclude them as such, per undue weight. Septentrionalis 00:20, 17 November 2006 (UTC)
I took "reliable for the claim cited" to cover things like not using a newswpaper for details of a scientific discovery. The problem is that the assertion is outside the domain of reliability for the sources. So also the physicist who holds forth on evolution, which is outside his domain of expertise. His reliability as a physicist is not compromised. The "reliable for the claim cited" phrasing seems more natural and direct than "appropriate." Robert A.West (Talk) 22:05, 16 November 2006 (UTC)
A good newspaper is a reliable source for a scientific discovery. It's possibly not the best source, but it might be. If you use the scientists' paper, they're likely to talk it up, whereas if you use a good newspaper, you might also get someone saying it's nonsense. It's obvious that there are good, better, and best sources, and that everything depends on context. All this policy seeks to do is lay down a minimum standard. The phrase "reliable for the claim being sourced" is not good writing, and is either trivially true or raises more questions than it answers. SlimVirgin (talk) 22:14, 16 November 2006 (UTC)
In response to never using such a source. I think that a source containing that statement or a similar one could be used in Race (historical definitions). I don't care about the exact phrase "reliable for the claim being sourced". But the fact is sources which are reliable in some areas can be unreliable in others areas. This is completely obvious in to me and think statements to the contrary are simply wishfull thinking not how things really work. --Birgitte§β ʈ Talk 23:44, 16 November 2006 (UTC)
It could be used in Race (historical definitions) as a primary source.
I think your point is already covered by the second key principle: "Reliable sources are credible published materials with a reliable publication process; their authors are generally regarded as trustworthy or authoritative in relation to the subject at hand. How reliable a source is depends on context." I don't see any benefit in the confusing "reliable for the claim being sourced." SlimVirgin (talk) 08:03, 17 November 2006 (UTC)

Response to RFC

A few thoughts on this suggestion from my perspective. I'm probably quite stringent since I teach research methods to young managers and consultants:

  • It's useful that the article has been dragged down to a digestible size, it's now about the maximum length for retaining attention in a meaningful way, so I think anything else now has to be done using supporting pages. With that in mind the corresponding FAQ is too long, that should probably be cut down to a number of more focussed pages dealing with different aspects of what the article is talking about. I realise that this may appear to be instruction creep, but the main issue is readability and retaining attention.
  • A lot of the content has now been so thoroughly discussed it doesn't actually say anything useful to people looking for advice, or guidance.
  • The general overview is reasonable but a bit weak, if this is intended as one of two key policies then the opening needs to be punchy and direct.
  • The section on Original Research leads me to two questions. What is Reliable and what is Published, I note that the next section then conflates these two issues into a pretty vacuous mush of ideas. Both of these issues need to be underpinned by something meaningful. Personally I prefer to provide principles of the subject rather than specific examples. Schoolchildren seek specifics, adults will perform better with more latitude to think for themselves. The subsequent amplification on OR doesn't really answer this since it's all a bit vague and lacks punch, the same as the opening header.
  • I'd suggest that the issue about Biography probably needs to be drawn out more distinctively given the potential legal implications.
  • Questionable sources leads me to more questions that answers. What is meant by the various statements? Widely acknowledged, by whom? Extremist, in whose opinion? Promotional. that would take out most of the discussion of organisations since marketing strategies include an element of uncontrolled media, any discussion of a commercial outfit is promotional. How do you define Vanity Press in a manner that isn't challengable?
  • Formal rules are not a substitute for good judgment. is a really important point and it is lost in the text. It needs drawn out, early in the article to make clear that this is about being grown up about the editing process.
  • Citing yourself is a platitude. Personally I'd lean towards the not unless absolutely required school of thought and that should be made clear. If this is intended to be a policy give it some backbone.
  • The references are both the personal opinion of the founder, which are technically inadmissible under the guidance. There is plenty of material which can be referenced about how to conduct research and to assure the integrity of the material underpinning that research. As much as possible this article should be an exemplar of its subject so should be rigorous in it's foundations.

All in all it's an admirable effort although it's clearly suffering the fate of the camel, a horse designed by a committee. I'd suggest that one individual needs to take responsibility for crafting the policy so that it doesn't become too weak to be useful. I do recognise that is an argument for giving my views a stiff ignoring to since there are already too many people sticking their oars in. An interesting challenge though, making it both useful and acceptable to the broader community.

ALR 22:23, 15 November 2006 (UTC)

Thank you for your comments. Some of the points you make are well made, but I find peculiar your last statement. Wikipedia articles, and policy, are written via a collaborative effort. No one individual can supersede that. In my experience, with time and patience, the prose becomes more fluid, as agreement on the principles is reached. ≈ jossi ≈ t@ 23:01, 15 November 2006 (UTC)
Ah yes, the tyranny of imagined consensus. In my experience there is a critical mass where consensus actually means factionalisation and alienation, whilst leading to a risk averse culture and bland ineffective strategic direction. That critical mass is about 12 to 15 people in my experience. In seeking consensus in larger groups one must resource adequate time to work in small groups then synthesise the outcomes from those small groups using representatives in their own groups of 12 to 15.
I am aware of the Wikipedia ideal, I'm just not convinced that it actually exists, consensus is used as shorthand for majority and **** anyone who dissents. Mediocrity is the preferred default position as a result.
ALR 23:16, 15 November 2006 (UTC)
If that is your assessment, may I ask why you edit Wikipedia? My viewpoint is that all these theories about critical mass, and "small groups" have bot matured yet to include social networks such as Wikipedia. Thus, I would take all these theories cum grano salis until such time these new environments are incorporated into that body of work. ≈ jossi ≈ t@ 23:22, 15 November 2006 (UTC)
On an unrelated note, I tend to share your preference for principles over examples, but believe it's most beneficial for Wikipedia policy to employ both redundantly. You say, Schoolchildren seek specifics, adults will perform better with more latitude to think for themselves. Perhaps true, but we shouldn't forget that schoolchildren actually do edit Wikipedia, and as long as we're a wiki, probably always will. -- Bailey(talk) 23:24, 15 November 2006 (UTC)
I have no problem with that position. At the moment this article is neither, for the reasons outlined above. My preference would be for an illustration of principles illustrated with examples.ALR 08:40, 16 November 2006 (UTC)
I agree with many of the points ALR makes, but they come over as rather patronising; John R has been taking a similar tone at Wikipedia talk: Reliable sources. This nose-wrinkling approach tends to come from those who see the Wikipedia experiment as a bit of an amateur mess, with too many cooks spoiling the broth. All I can say is that perfectionism is doomed here; the process is more like building sandcastles on the tideline. I would also say that this page is only about a month old and is only now starting the process of trying to turn itself from a mass think-in to a precise page that fully knows and shows what it is about. In many places it is still groping for the right words and terms; as it finds them, one by one, it will gradually improve. Despite my reservations, ALR's comments are valuable, and his work on redrafting Wikipedia:Reliable Sources in line with the straw poll there is much appreciated and I'm sure will influence this page over time.qp10qp 01:31, 16 November 2006 (UTC)
It's not a question of perfectionism, although in taking the position that it is you usefully illustrate my point about factionalisation. There are clear issues with the form and structure of Wikipedia with the function being a worthwhile effort. Organic or incremental change will only take a system so far before there is a requirement for revolutionary change as size and complexity become greater than existing control mechanisms can cope with. Unless the requirement for that change is recognised there is a tendency to try to achieve it with a proliferation of rules which ends up influencing the change process adversely.
At the moment much of the content is an amateurish mess, the process of improving that is supported by a taut kernel of guidance on how to improve the content. In this sense form leads function and if form lacks discipline then function suffers as a result.
ALR 08:40, 16 November 2006 (UTC)
Why do I edit? Well a number of reasons:
  • I'm a specialist in knowledge management and business change, in that sense it's an interesting social experiment and it has the potential to become useful as it evolves. We need to take account of the fact that the delivery mechanism is discrete from the knowledge creation process but the form of the delivery mechanism does influence the efficiency of the process. There is a great deal of work on technically facilitated communities of practice some of which includes the nature of Wikis as a delivery mechanism. The general direction of that work is that CoPs gain a great deal from having identified facilitators to support the process and to mitigate for the absence of audio-visual indicators in discussion which exacerbates the impersonal nature of text based interaction. As WP evolves these catalysts will emerge, to an extent they already are in different forms but have little effective penetration as yet.
  • In some areas a level of consensus can be achieved because the numbers of editors involved is quite low. The quality of content is not homogenous, there are patches of decent quality in a sea of mediocrity. Some of that mediocrity has a high rate of change, some of it just isn't interesting enough for people to bother about. Over time as facilitators, mediators etc emerge then it is likely that the signal to noise ratio will improve.
  • I enjoy an interesting discussion/ argument as much as the next man.
ALR 08:40, 16 November 2006 (UTC)
I think people looking for advice and guidance should be looking at Help pages not policy pages. This should be a firm statement of consensus; not a tutorial. --Birgitte§β ʈ Talk 04:57, 16 November 2006 (UTC)
Whilst in principle I can see merit in that argument there are a number of areas which require more authority than the Help pages. Attribution as a policy will need to be enforcable. Many admins aren't in a position to actually understand this subject themselves so need something to support them, that isn't their fault, they just haven't finished being educated yet.ALR 08:40, 16 November 2006 (UTC)
Ignoring any issues over how meaningful conensus is (which is not what we're here to discuss), a few comments:
The section on Original Research leads me to two questions. What is Reliable and what is Published, I note that the next section then conflates these two issues into a pretty vacuous mush of ideas.
Well, I think defining "reliable" is one of the things we're here to do. If we've failed, perhaps we need to rephrase the "key principles" section that attempts to do that. As for "published", we should perhaps insert a definition; it has certainly been a frequently asked question on WP:V's talk page. How about:
A source is published if it has been distributed in such a way that it is available to general members of the public, including the case where it is deposited at a public library.
Next point:
*Questionable sources leads me to more questions that answers. What is meant by the various statements? Widely acknowledged, by whom? Extremist, in whose opinion? Promotional. that would take out most of the discussion of organisations since marketing strategies include an element of uncontrolled media, any discussion of a commercial outfit is promotional. How do you define Vanity Press in a manner that isn't challengable?
I'd say leaving these words open for interpretation is actually a good thing -- it allows the arbitration committee some latitude in deciding how to enforce the rules. Realistically, if there's any danger of source being labelled as one of these, it should probably not be used in most circumstances.
*Formal rules are not a substitute for good judgment. is a really important point and it is lost in the text. It needs drawn out, early in the article to make clear that this is about being grown up about the editing process.
I definitely agree with this; it's a point I raised above.
*Citing yourself is a platitude. Personally I'd lean towards the not unless absolutely required school of thought and that should be made clear. If this is intended to be a policy give it some backbone.
Wikipedia is currently trying to persuade more subject experts to contribute. Part of this is being seen as expert-friendly, and denying an expert the chance to cite themselves could be seen as unfriendly. I'd leave it as it is for this reason.
*The references are both the personal opinion of the founder, which are technically inadmissible under the guidance.
That is a problem, yes. These references should certainly be acceptable according to these rules -- I think the discussions (above) about clarifying that primary sources on people's opinions can be used when relevant will fix it, though.
There is plenty of material which can be referenced about how to conduct research and to assure the integrity of the material underpinning that research. As much as possible this article should be an exemplar of its subject so should be rigorous in it's foundations.
Good idea. Since you're clearly familiar with the subject, would you like to provide some references? JulesH 08:17, 16 November 2006 (UTC)
I've added a new version of the "formal rules are not a substitute for good judgment" key principle; hopefully nobody will interpret this one as a loophole that needs to be closed, as apparently happened with my last suggestion. JulesH 08:35, 16 November 2006 (UTC)
With respect to Reliable and Published, I think all the posturing about WP:RS has rather got in the way of defining the first. Personally I'm ambivalent about where it is but think that WP does need something firm for editors and administrators, I've put the bones of some material on Reliability as a sub-page to that article already, it's not easily captured in a couple of paragraphs. Publication is similar, I'd disagree that it is as simple as being available to the public, there is a lot of material which has been published and isn't available to the general public, equally there is a lot which is reliable and representative which hasn't been published. In a couple of articles I've reverted to using my course books, privately printed by Cranfield University and vetted by Kings College, London, which talk about things like Gunfire support and similar. They're accurate and reliable, but to access them you either need to be a senior military officer, specially selected civilian candidate or a senior police officer.
Does Arbcom have a view on whether they should be left with latitude or not? Essentially what you're advocating is a common law approach rather than statute law approach to governing the editorial process. From my perspective Arbcom should be dealing with exception reports, rather than routine noise. If this is really seen as a viable alternative to extant policy and guidance then it should pin things down a lot better than they currently are. Inform that from Arbcom decisions, but if it's just a licence for ongoing arbitrary tyranny then why bother.
There is a problem with hoping that experts will contribute when the extant policy and the proposed policy don't allow them any latitude to actually make use of their expertise in a sensible manner. There are much broader issues to address than citing yourself. If the proposed policy can be made adequately robust then I don't have an issue, but at the moment the wording is wooly and won't suit anyone. It needs to swing one way or the other.
Three is a difference between citing Wales' view on the sourcing strategy and citing credible foundations for the execution of that strategy. Personally I rather wish the community could move away from the cult of personality, which I don't think is useful for the further development of the encyclopedia anyway. I've already identified my preferred textbook on research in the RS discussion, I teach from this which covers a lot of the issues. It's pitched at post-grad level for Masters and Doctoral candidates in the business field, which clearly has a different context from the social sciences which seem to predominate in the discussion here.
ALR 20:48, 20 November 2006 (UTC)

Questionable sources exception

When and why did this get turned back into a pop culture exception? I just don't get it. All some people seem to care about is having their loophole in place. ..."the best source for a particular article may be one that this policy would describe as "questionable"" is a f**king awful phrase. Marskell 08:08, 16 November 2006 (UTC)
I agree it's not ideal. What was the last version you preferred, Marskell? SlimVirgin (talk) 08:11, 16 November 2006 (UTC)
I would've left previous which stood for five or so days: Formal rules are not a substitute for good judgment and intellectual honesty. Formal rules cannot always determine whether material is reliable, notable, or relevant to a particular topic. An editor who thinks that a source is reliable or unreliable contrary to the formal rules should convincingly justify his or her judgement, and should not use a source if reasonable objections are raised.
I don't think we need Questionable sources must never be used to support biographical claims about living persons, or to support fringe theories or marginal positions in the areas of history, politics, current affairs, science, religion, and other academic disciplines. Marskell 08:18, 16 November 2006 (UTC)

(edit conflict)

I looked at the last version you edited to see what it said about that area, and it had an even more problematic paragraph, in key principles, the "formal rules are not a substitute" thing:

An editor who thinks that a source is reliable or unreliable contrary to the formal rules should convincingly justify his or her judgement, and should not use a source if reasonable objections are raised. Questionable sources must never be used to support biographical claims about living persons etc.

First, the writing: "convincingly justify his judgment." :-( It doesn't say who is to be convinced, but implies that any group of editors on a page may decide to use any source they want. I realize the current version implies the same, but at least by mentioning popular culture, we give people an idea of the areas we're talking about. The version above also introduces the idea of "questionable sources" before they've been defined.
Why do you think the current version would allow more than the above? I see it as tighter. Or do you dislike both equally? SlimVirgin (talk) 08:19, 16 November 2006 (UTC)
If you think we don't need "Questionable sources must never be used to support" etc, then you've done a complete volte- face. Your version without that would mean that any consensus of editors on any page for any subject could ignore this policy and use whatever sources they wanted. SlimVirgin (talk) 08:22, 16 November 2006 (UTC)
Hold up: I don't like last because it can be interpreted to mean questionable sources can be used for subjects not listed (i.e., I'm concerned about an inference based on omission, which I don't think is a volte-face!) "Your version without that would mean that any consensus of editors on any page for any subject could ignore this policy and use whatever sources they wanted". Maybe you're right, but that's totally not my intention, which leads to "do you dislike both equally"—yes, I'd be perfectly happy if none of this were on the page.
I did prefer previous for "reliable or unreliable". It suggests that in some cases we need to raise the bar, and I like how that balances out the point. And I also prefer not mentioning pop culture for all of the reasons already suggested to you (how do we demarcate the subject area, why should we have dual standards, etc.)
"Can be convincingly justified"...ya, far from perfect. Perhaps we can try an altogether new wording. If there is consensus to remove it entirely, I'd be happy, but I don't know that there is. Marskell 08:36, 16 November 2006 (UTC)
I've tweaked it a little, and removed pop culture and fiction. Does that help? SlimVirgin (talk) 08:43, 16 November 2006 (UTC)
Yes, that's better, but I'm obviously hard to please on this topic :). Below is something I've been toying with. I'm contradicting myself slightly by re-introducing cultural examples (it's also a little long), but I think we need to think about aspects of topics and not topics themselves. That is, you don't have carte blanche to do up your video game as you please, but allowances can be made for certain areas of the topic.
"In certain cases, professional or academic sources offer shallow coverage of aspects of a topic. This is particularly true for obscure subjects (e.g., folk receipes or crafts) or for "in-universe" aspects of certain new media and culture articles (e.g., song lyrics or a video game plot summary). In these cases, if there is consensus that a source the policy would not normally recommend can be trusted and there are no reasonable objections, it may be used for these limited aspects. For all other regular information [need better wording], including, but not limited to, statistics, quotes, and judgements of quality, only reliable sources as defined by this policy may be used. If sources cannot be found, such information should not be added. Such sources must never be used in the areas of history, politics, current affairs, science, religion, or other academic disciplines; to support fringe theories or marginal positions in any area; or to support claims about living persons." Marskell 09:30, 16 November 2006 (UTC)
I'd go along with that. SlimVirgin (talk) 20:42, 16 November 2006 (UTC)
Looks good to me. Jayjg (talk) 00:48, 17 November 2006 (UTC)
I like that much, much better. I was going to propose that the wording be changed along these lines -- "new media" works much better than "fiction" (which seems to suggest that professional and acedemic sources are lacking for all of literature) and I'm happier with the new acknowledgement that "fringe theories" can occur in any subject area, academic or not. -- Bailey(talk) 14:57, 17 November 2006 (UTC)

I still see some problems:

  • wordiness - the solution fails our goal of making WP:ATT a simple, unbloated, straightforward policy. Example: "if there is consensus that a source the policy would not normally recommend can be trusted [...]", or "[...] "in-universe" aspects of certain new media and culture articles [...]" - there must be ways to say this in a less convoluted fashion;
  • folk receipes example:
    • typo - shouldn't this be "folk recipes"?
    • I wouldn't use this as an example at all, as there is a long standing tradition that recipes (which of course includes folk recipes) should be moved to WikiBooks/Cookbook.
  • current affairs: "Such sources must never be used in the areas of [...] current affairs, [...]": this seems to contradict a recent ArbCom ruling: "In the case of articles which chronicle a developing current event it is not a violation of Wikipedia policy to temporarily include links to blogs which contain contemporary opinion and observations about the event." (my bolding, see the "remedy" proposed in Wikipedia:Requests for arbitration/Israel-Lebanon#Use of blogs) - unless WP:ATT is an attempt to change current positions of WP:V/WP:NOR/WP:RS on this matter, so that if the Israel-Lebanon ArbCom case were redone, the outcome would be different. At least "current affairs" should be removed from this "no-exception" list.
  • The current wording leans too much towards unrefrained inclusion of trivia imho. Note that for example Wikipedia:Trivia ultimately relies on a strict definition of the type of sources we accept. I don't think we should (for example) move towards a more lenient posture regarding "in-universe" and other amateur websites on all kinds of pop culture trivia ("how many dogs are owned by popular film actress XYZ?" "does she like cats notwithstanding the fact she only owns dogs?" etc - compare also this WikiEN-l posting). On the other hand, the exception should apply to, for example, developments in open-source software, where "facts" are often established long before they appear in print, e.g. Linus Torvalds' determination to keep with GPLv2 for Linux, before the next version of GPL is even published. --Francis Schonken 11:03, 18 November 2006 (UTC)
Regarding folk recipes, I can see that they are useful sources for statements in articles about food, particularly on discussions about unusual variations of preparation method. An example would be the (currently unsourced) statement in fudge that reads "some recipes call for making fudge using marshmallows". An example of a recipe that did this would be an excellent source for this statement. JulesH 14:39, 18 November 2006 (UTC)

Primary and secondary sources

We really need to include something about primary and secondary sources in this policy, as is found in WP:NOR. It's a fairly critical distinction when dealing with original research, because primary sources often require interpretation to understand properly, and that interpretation needs to come from secondary sources. We need to avoid the situation where someone inserts into the Jesus article "Jesus advocated eye-removal (Matthew 18:9, Mark 9:47) and castration (Matthew 19:12) for his followers", and insists it be kept because it is properly attributed. Jayjg (talk) 23:37, 16 November 2006 (UTC)

Secondary sources just as often require interpretation, thus see my counter proposal above (point 3 takes care of your issue). And with that I definitely leave this discussion, good luck! Harald88 23:51, 16 November 2006 (UTC)
I agree that we need something. WP:NOR makes the distinction, and I see editors rely on it all the time, myself included. SlimVirgin (talk) 00:00, 17 November 2006 (UTC)
Here's what we had before, under key principles. I think we need something like that again, whether in key principles or elsewhere. SlimVirgin (talk) 00:14, 17 November 2006 (UTC)

I agree we need something of the sort. I would prefer to make it "sources requiring interpretation" or some such phrase; because "primary source", "source close to the event", and "source requiring interpretation" are not exactly the same thing. To pick only one of the eight possible cases: The Anatomy of Melancholy is not close to the events that it describes; and it is explicitly and almost entirely secondary; but it does require much interpretation. Septentrionalis 00:30, 17 November 2006 (UTC)

What is a "source requiring interpretation"? It seems less clear to me than "primary source". Jayjg (talk) 00:38, 17 November 2006 (UTC)
That's a primary source, Sept. It may have been a secondary source in its day, but it's a primary source now. We can add more details about issues like that (our relationship to the sources) in WP:ATTFAQ, and link to that discussion from the proposal page. SlimVirgin (talk) 00:42, 17 November 2006 (UTC)


I've added something. It's not perfect or complete, but then no policy about this will ever be, because judging which sources are appropriate for anything is a difficult business. This is just to give editors a bare minimum standard. SlimVirgin (talk) 00:40, 17 November 2006 (UTC)
Jay, I added your Bible examples, because they illustrate the problem very well. SlimVirgin (talk) 00:51, 17 November 2006 (UTC)
Please explain how Anatomy of Melancholy is a primary source and newspapers reporting on current events are not? --Birgitte§β ʈ Talk 01:15, 17 November 2006 (UTC)
The issue of how to use sources is very complex, and no policy page on Wikipedia can cover it. People who want to learn more will have to read academic books. Some editors seem to be saying that, because discussions about primary/secondary sources can't reflect the complexity of the topic, we shouldn't mention them at all, but all I can say is I disagree. Editors do use the distinction frequently, so I see no reason not to have some light guidance, so long as it doesn't include any absolutes. Anyone who wants to add more complexity could write it up for WP:ATTFAQ.
Anatomy of Melancholy is a primary source for the topic of medical textbooks in those days, what was thought about melancholia then, what Burton said about it. It's a secondary source for material about other authors such as Aristotle.
And I was thinking about Aristotle or the Lamia, because that's the sort of topic AM's normally read for. Whether it's primary or secondary on humours depends on which period you're writing about. But primary or secondary, it must be used with care, because it's not a modern scientific or scholarly work; which is one reason why sources require interpretation. Burton is not trying to do what we're doing; even when he's not dated, we should not use him in Aristotle. (This is also why we should avoid using White House press releases; they're not doing what we do.) Septentrionalis 03:52, 17 November 2006 (UTC)
An eyewitness's statement about a traffic accident is a primary source; a newspaper's story about that accident, which might use the eyewitnesses's statement as a source, is a secondary source. That same newspaper story in 100 years time might be a primary source about the state of the roads in 2006. SlimVirgin (talk) 02:18, 17 November 2006 (UTC)
A newspaper report of an accident is still a primary source. An analysis and commentary of the accident, on the other hand, is a secondary source. ≈ jossi ≈ t@ 02:25, 17 November 2006 (UTC)
It depends what you're using it for. For most purposes, a current newspaper article is a secondary source. Historians might regard them as primary sources for general information about the state of (whatever) in 2006. But an historian writing about the war in Iraq would not want to use a newspaper story as a primary source about what George Bush intended, although would use it as a primary source on how the media covered the war.
These are complexities we don't need to get into. SlimVirgin (talk) 02:57, 17 November 2006 (UTC)
I don't think anyone is against talking about how to properly uses sources. The problem is bolding Primary and Secondary Source as if it was a defining term that is clear to anyone who reads it. --Birgitte§β ʈ Talk 02:30, 17 November 2006 (UTC)
It's mostly clear. Talking about "sources that need to be intepreted" is meaningless. But pretty well everyone with a university degree or who has worked in certain professions will know (roughly) what primary and secondary sources are. SlimVirgin (talk) 02:53, 17 November 2006 (UTC)
Pretending two overlapping terms are mutually exclusive is a very bad idea. Primary and secondary sources are how you use them not intrinisic qualties that can be used to label them. I think we should talk about how to use something as a primary source, but using "secondary source" as a definition is not going to improve things. --Birgitte§β ʈ Talk 03:46, 17 November 2006 (UTC)
I'm definitely open to suggestions about how to improve the wording. SlimVirgin (talk) 04:00, 17 November 2006 (UTC)

I think the wording starts off well and then goes to pieces. As usual, the term "primary sources" is being used to cover a whole range of different things and will cause confusion in many places. The biblical example is controversial; yet the majority of primary-source use in history, literature, and art articles that I read or edit is uncontroversial. A large part of the time, the primary sources don't need any interpretation at all because they are self-evident.

"Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge." Many edits that rely on primary sources don't make any claims at all (I've just quoted from a Chekhov story on the Chekhov page merely to illustrate a bit of the story; no claim made). So does this mean edits that interpret primary sources?

No, it means you may say: "Chekhov wrote that X is good" (a descriptive claim about the publication that anyone can check). We may not say "X is good" and use Chekhov as a source. We may not say "Chekhov wrote that X is good and here is what he meant ..." SlimVirgin (talk) 04:03, 17 November 2006 (UTC)

On the other hand, a primary source may say something useful that people without specialist knowledge can't check. For example, Geoffrey Malaterra says that Erembourge, wife of Roger I of Sicily, was the daughter of William, Count of Mortain (Eremburga filia Gulielmi comitis Mortonensis). He's the only source for that and so he needs quoting, especially as he was Roger's chaplain. Even though I have an MA in history, specialising in that period, I cannot verify (in terms of truth) what Geoffrey said, and nor can anyone else.

But we never do that. SlimVirgin (talk) 04:04, 17 November 2006 (UTC)
Exactly. So, we must say what Malaterra said and leave it there.
Then I don't follow your point, sorry. SlimVirgin (talk) 04:21, 17 November 2006 (UTC)
The point is that the following is meaningless as worded: "Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge." If we assume that primary sources make "claims", how is someone without (or even with, in this case) specialist knowledge going to check whether they are telling the truth or not? But I don't consider that Malaterra claimed anything, just that he said it; and anyone can check that he said it by looking up the reference. "Descriptive claims" is just unhelpful wording, I feel.
Okay, fair point. As I said, I'm open to finding ways to improve the wording. SlimVirgin (talk) 07:58, 17 November 2006 (UTC)

What it has to do with anything is that wording framed for totally different types of issue sounds false, and potentially confusing, when applied to the run-of-the-mill work with primary sources. For my money, a published primary source such as Malaterra, particularly one edited by scholars, as Malaterra is, is as good as a secondary one in most cases.

"Wikipedia articles should rely on reliable, published secondary sources wherever possible." Why, though? I usually find that the primary sources are better. For example, James Wood recounts in a book an anecdote that "an actor" said about Chekhov. On chasing it down, I found that Stanislavski was the actor who said it; and since he said it in "My Life in Art", it is better to reference "My Life in Art", a primary source, than Wood, surely, since Wood either didn't know Stanislavski was the actor in question or chose not to mention it for fear of complicating the argument he was illustrating (and since the anecdote tells itself and does not need interpretation, no original research took place, only the referencing of a sound source). A good use of letters, notebooks, diaries etc.(published, of course) — brings biographical articles to life; I notice that the featured-article biographies rely on primary sources almost as bone structure for their information, which for me is excellent and logical practice.

They do not rely on Wikipedian's interpretations of primary sources, or at least they ought not to (I've seen pop culture ones that do, of course). SlimVirgin (talk) 04:12, 17 November 2006 (UTC)

I am not against useful advice about primary sources appearing in this proposal, so long as the different types are distinguished. I know quite well why the confusing stuff has been introduced: it's because of all the POV that comes into Wikipedia in mainly contemporary issues. I am sorry to say that if the advice stands as presently worded, I will simply ignore it when editing run-of-the-mill articles. I just hope this doesn't cause needless arguments and misunderstandings, like this at Talk:Sviatoslav I of Kiev#Primary sources, prompted by the wording at Reliable Sources. qp10qp 03:49, 17 November 2006 (UTC)

I think you're getting all sorts of issues confused. If you want to say "A said X," where A is primary source, that's fine; it's a descriptive claim that anyone with access to a library can check. If you want to say that Jesus said: "If thine eye offend thee, pluck it out": again, a descriptive claim, which is fine to use so long as you don't imply that you know what he meant. If you want to use that quote to state or imply that Jesus advised people to pluck their eyes out, no, because people differ as to what he meant. I can't see what's difficult about this. SlimVirgin (talk) 04:09, 17 November 2006 (UTC)
On the other hand, take Plutarch. His anecdotes are better told than they are in the secondary sources too. But the secondary sources will warn you when not to believe him; as Perseus says: "Plutarch relates more information about Aspasia than any other ancient author. Unfortunately, Plutarch's Lives are full of distortions and historical inaccuracies. His purpose in the Lives was to exemplify the virtues and vices of great men, not to write history."Septentrionalis 04:07, 17 November 2006 (UTC)
Exactly. SlimVirgin (talk) 08:05, 17 November 2006 (UTC)

Trying to sort out the confusion

Perhaps what's confusing people is this. We say only use primary sources to make descriptive claims that anyone without specialist knowledge can check. That is, with "A said X," anyone can go to a library, and there it is on the page.

Some editors object to this (if I understand them) on the grounds that we're not allowed to interpret any sources; we are only allowed to publish what sources have said, primary or secondary. So why single out primary sources?

But the purely descriptive claim thing is supposed to stop primary sources being misused in a way that secondary sources aren't misused. It's meant to stop an editor adding something from the Talmud (quoting it correctly) to show Jewish attitudes to non-Jews; or to stop Jesus's words being used to show that he actually wanted us to pluck our eyes out.

It is easy to misuse primary sources in this way. They are often misused in this way. It is not so easy to misuse secondary sources in this particular way. SlimVirgin (talk) 04:40, 17 November 2006 (UTC)

Slim, since the aim here is to improve wording, may I isolate two things you said in reply above, because they may help towards clarification of the wording:
"They do not rely on Wikipedian's interpretations of primary sources..." This nails what needs to be deprecated. Primary sources should not be deprecated against secondary sources, but interpretation of primary sources should. Articles should and do rely on primary sources; without them, we'd have to exclusively rely on secondary sources telling us what Van Gogh said in his letters rather than reading what he said for ourselves.
You also used the word "claim". I think that word is confusing as applied to primary sources, since most don't make claims (they might do in controversial contemporary articles, but not in most articles). You may be using the word claim to mean "say" or "assert" or "prove", but it's not the first word that comes to my mind about primary sources. "Claims" applies more, if anything, to the interpretations of primary sources. So the main point, again, is that Wikipedia editors shouldn't interpret primary sources. We're all agreeed on that. It's the use of the primary source that's the problem, not the primary source itself, if it's been properly published. qp10qp 04:45, 17 November 2006 (UTC)
What would you say to an editor who, in the article on Jew, in a section about Jewish attitudes to non-Jews, quoted a relevant passage from the Talmud, without commentary, and quoted it correctly, to illustrate Jewish attitudes? SlimVirgin (talk) 07:55, 17 November 2006 (UTC)
There needs to be a distinction between that sort of use and the use of primary sources generally, which is usually benign. I was happier with the wording before we singled out primary sources; but if we are going to talk about primary sources, lets not do it in such a way as to cause thousands of unnecessary arguments about standard uses of primary sources. I am sad that we have imported some of the wording from Reliable Sources that caused such arguments in the past.
Your example above is covered by "The Bible cannot be used as a source for the claim that Jesus advocated eye removal (Matthew 18:9, Mark 9:47) and castration (Matthew 19:12) for his followers, because theologians differ as to how these passages should be interpreted", which I don't object to. Most uses of primary sources are nothing like that at all and needn't be deprecated en masse. qp10qp 11:48, 17 November 2006 (UTC)
That wording is not from Reliable sources. It's from NOR and it's policy (taken from there to RS). I don't know how you can know that most uses of primary sources aren't like those examples; most of the uses I see of it are precisely that, and not only in that area.
What would you say to the Talmud editor I referred to? With the NOR policy as it currently is, we have something we can say to him (find a secondary source). The way you want to fashion this proposal, we'd have nothing to say. So please, if faced with him, what reason would you give for removing his edit? I'm not asking to put you on the spot, by the way. I'm just curious to know how you'd deal with it without referring to the NOR primary/secondary provision. SlimVirgin (talk) 11:56, 17 November 2006 (UTC)
The same thing you do if the editor uses the exact Talmud quote to support a claim that President Bush is a Martian. You say "the point you're making isn't in the quote. I can read the quote, it doesn't say that Jews in general believe that". The fact that the editor only implied that his point was in the quote rather than stating it outright, doesn't change this. Ken Arromdee 14:47, 17 November 2006 (UTC)
No, my point was supposing he quoted it correctly and without commentary. SlimVirgin (talk) 16:28, 17 November 2006 (UTC)
SV, we are talking at cross purposes now because I am happy for you to make that sort of point in the proposal; it's just that any sign that we are generally disapproving primary sources will create all sorts of needless arguing among editors. Perhaps most of the uses of primary sources you see are problematic because you are an administrator. I don't necessarily want to take anything away from what you said in the policy text yesterday but only to improve the wording so that primary sources aren't generally deprecated. I feel that you have definitions of certain words in your head, for example of "claims" and "threshold", which don't allow that people might read them a different way. You said yesterday that pretty much all that really matters is that you come to Wikipedia armed with reliable sources; but for me that's just the starting point. The consequence of that view of the word "threshold" is that you get problems of the sort in your Talmud or Bible examples above, because people may feel they are entitled to use such quotations without fitting them into the context. In fact, any reference, primary or otherwise, may be challenged for a number of reasons, whether it passes the attributable threshold or not, and the reliability of a source will depend on the context and will be challenged on the context.
I remember you mentioning a case where someone wanted to prove that vegetarianism was supported by the Bhagavad Gita (or something), and I can see that you need a clause to deal with that sort of thing. Yes, lets cover that. But most primary sources are not being used to make points that way in areas such as literature or biography, where primary quotes or references to primary sources are repeatedly needed just to uncontroversially represent the works or the life. Why appear to tar all use of primary sources as dodgy? qp10qp 15:35, 17 November 2006 (UTC)
But which words in this proposal would prevent primary sources being used uncontroversially to represent the works or the life of someone? SlimVirgin (talk) 16:30, 17 November 2006 (UTC)

SlimVirgin wrote on this talk page that an eyewitness account of a car crash would be a primary source, but a newspaper account that quotes the eyewitness would be a secondary source. The proposed policy refers to an eyewitness account as a primary source, with no mention of the newspaper. The typical reader of the proposed policy will see that we don't use unpublished sources, and presume that the the eyewitness account means an account contained in a newspaper article, and the newspaper article is a primary source. After all, we already said we wouldn't use the account if it wasn't published, and where else would an account of a car crash be published?

In a courtroom, in a book/article written by the witness, or in a newspaper story as an eyewitness account. The section did mention newspaper articles when I last looked. SlimVirgin (talk) 20:01, 17 November 2006 (UTC)

Also, the primary/secondary sources refers both to publications and people as sources. It should be reworded to refer only to published sources; Wikipedia does not do interviews. --Gerry Ashton 19:10, 17 November 2006 (UTC)

No, sources are people or the documents they write. We say a professional researcher's blog is usable because the person is the reliable source; his reliability goes with him, and isn't dependent on where he's published. SlimVirgin (talk) 20:01, 17 November 2006 (UTC)
Nothing prevents it, SV, but the wording seems to discourage it. I need to repeat myself to answer you, but I suppose that's the way it has to be here (I can't be bothered to repeat my examples because I'm getting tired, having fitted in a day's work among comments since I started trying to make my point):
"Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge". I'm not quite sure what edits are meant here; are the claims intrinsic to the source or deduced from the source? Many claims are difficult to check, for example whether Erembourge was the daughter of William of Mortain or not, even by specialists. All we can check is whether the source is referenced properly, not the validity of the claim itself. (If your answer again is "but we never do that", I would guess that our cross-purposes result from different readings of the unhelpful word "claims".)
We don't check whether claims are true. The wording can be changed if you want, but most editors know that "check" doesn't mean "check whether it's true." The proposal makes that clear. SlimVirgin (talk) 20:10, 17 November 2006 (UTC)
"...but only with care because it's easy to misuse a primary source". The same care should be used for all sources; in fact, inaccuracy is more difficult to detect in secondary sources than primary ones because the latter may impose a layer of interpretation. Why should we be more careful with primary sources than with secondary ones? I think, for example, that it is easier to misuse the POV comments of Cornish historians about the Anglo-Saxon conquest of Cornwall than it is to misuse the Anglo-Saxon chronicle and the Annales Cambriae, which are best quoted raw on the matter. (I helped improve a paragraph on this by reducing the secondary commentary.)
We're not talking about accuracy. SlimVirgin (talk) 20:10, 17 November 2006 (UTC)
"Wikipedia articles should rely on reliable published secondary sources wherever possible." I'd cut "secondary" (which by implication deprecates primary), and cut "wherever possible". This would then make sense to me as "Wikipedia articles should rely on reliable published sources".
They should rely on secondary sources as a rule. SlimVirgin (talk) 20:10, 17 November 2006 (UTC)
Note here that my comments do nothing to oppose the clauses which help you fight misuses of sources of the type you have mentioned in this matter.qp10qp 19:55, 17 November 2006 (UTC)
I'm sorry you're having to repeat yourself, but so am I, and I'm genuinely confused about what you don't get here. Forget truth. It has nothing to do with truth, checking truth, checking claims in that sense. Checking a descriptive claim means: "Did Jesus actually say: 'If thine eye offends thee, pluck it out'." As opposed to: "Did Jesus say that people should pluck their eyes out?" which can't be checked, because there are disparate views.
I think the bottom line is that, as you say, there is nothing in this proposal that stops people from using primary sources in the way you use them, so is there really a problem here? Perhaps you should show me an article you've written, or substantially contributed to, where you relied heavily on primary sources, and we can check whether we're talking about the same kind of use. SlimVirgin (talk) 20:10, 17 November 2006 (UTC)
Night (book) is an example of an article I wrote that relies heavily on the primary source in the lead and in section two, when I tell Elie Wiesel's story. Anything in those sections that smacks of analysis — anything not obvious to a sane, adult, native English speaker reading the book — is referenced to a secondary source. Almost everything in sections three and four is from secondary sources. Are you talking about using primary sources any differently from this? SlimVirgin (talk) 20:23, 17 November 2006 (UTC)
I'm about two thirds through working on Anton Chekhov; I've provided all the references and sources there. You may conclude that we are talking about different things, but the point is how someone else might react to our wording: is it pellucid? The word "claim" will cause misunderstanding, I am sure. I do, of course, understand that we aren't talking about the truth; my point is that the word "claims", in the way you use it, overlaps with the notion of what may be a claim in the source. Perhaps it's me, but we didn't use the expression "descriptive claims" at my university, and if it's a well-known term, forgive me. You seem to indicate that it means that the reference checks out. qp10qp 20:29, 17 November 2006 (UTC)
Actually, having read your excellent article, I think we are broadly using sources on the same principles. Chekhov's not controversial, though. He had a short, relatively uneventful life, but he wrote not only hundreds of stories, plays, and a notebook but letters which are thought some of the best ever, and so it seemed to me appropriate to lean heavily on primary sources. Your section on Buchenwald does that in particular, and quite right too. But if you'd stuck to the principle "Wikipedia articles should rely on reliable published secondary sources wherever possible" you might have replaced those primary quotes and references with secondary ones which used the primary source in retailing the same events described in the book or broke off into analysis and historical perspective. It is likewise possible to replace the primary references in the Chekhov article with secondary ones that summarise Chekhov's own words. Why do that, though? It's more vivid the simple way. And "...but only with care because it's easy to misuse a primary source"? Well, care has been taken, but primary sources are largely allowed to tell their own story in the Buchenwald section and other places; far more likely for those extracts and references to be misused or blurred by secondary sources, surely. I just think the wording worries unnecessarily about primary sources. Anyway, I've had my say for a while. qp10qp 21:43, 17 November 2006 (UTC)
Your Anton Chekhov is wonderful. :-) It's great to see that kind of quality on Wikipedia. Please don't think you're saying too much about this (at least, not from my point of view, although you're probably fed up with it); what you're saying is very helpful. I think we actually agree; we just disagree on how the thing should be worded. I'll try to think of another phrase for descriptive claims that any non-specialist can check. I took it from NOR, [1] and I can't remember whether I originally wrote it. The point is that you wouldn't write Anton Chekhov using only primary sources, so that's one point we're trying to get across. It boils down to editorial judgment when to use primary sources, and we can't come up with an algorithm for that. Anyway, thank you for all these insights. SlimVirgin (talk) 05:17, 18 November 2006 (UTC)

Possible rewording

If I'm following this discussion properly, everyone agrees that it's ok for articles to say "Primary source A says X", as long as X is a neutrally worded description of A's content. But any interpretive, evaluative, or analytic claim about A must be attributable to a reputable secondary source. If that's right, could this change work:

Edits that rely on primary sources should only make descriptive statements that can be checked by anyone without specialist knowledge. --Akhilleus (talk) 06:02, 18 November 2006 (UTC)
That's definitely better, Akhilleus, thanks, but I "descriptive" anything might be difficult, and "checked," if I understood the objections correctly. SlimVirgin (talk) 23:09, 18 November 2006 (UTC)
An approach down these lines would be a disaster. I understand your suggestion, and when an article is targeted towards the world in general, it is good advice. However, we have very many articles that are written with a specialised audience in mind. Non-specialists aren't barred from trying to read them, but the style, content and assumed knowledge is all pitched at specialists. For these articles, it is perfectly reasonable to make a deduction from a primary source that any specialist in that area would make. For example, we have many articles covering mathematics that only professional mathematicians are likely to understand. What's wrong with them referencing to a peer reviewed theses that proves a theorem for the first time (and so is a primary source) and making deductions that anyone who is a specialist in that field of mathematics would follow? jguk 16:15, 18 November 2006 (UTC)

Hm, all I did was change the word "claims" to "statements"!

Is it really necessary for a mathematics article to make deductions that haven't been published elsewhere? If you're talking about the need to create examples, this is already covered in the "No Original Research" section, where it says "Editors may make straightforward mathematical calculations or logical deductions based on fully attributed data that neither change the significance of the data nor require additional assumptions beyond what is in the source. It should be possible for any reader without specialist knowledge to understand the deductions." If, on the other hand, the deductions involve original thought, then they should probably be published somewhere else before they appear in Wikipedia. I'm not sure there's a problem here. --Akhilleus (talk) 17:16, 18 November 2006 (UTC)

My comments are, of course, equally applicable to the suggested text on the project page! :)
To give a simpler example, the Laws of Cricket are difficult to understand if you know nothing at all about cricket, but are used as sources for all the articles on the laws. It's not a problem, there are more than enough cricket fans on Wikipedia to check that those articles properly reflect the source. However, if we were to follow the suggested wording (or rewording:) ), all those references would need to be deleted! On the plus side, you might get rid of some of those bloody Pokémon articles;) jguk 17:32, 18 November 2006 (UTC)

I think what you're objecting to is the phrase "without specialist knowledge"? What I was trying to do in the proposed rewording was deal with the problematic word "claims", which I changed to "statements". Does that seem better/worse?

As far as "specialist knowledge", I'm not convinced there's a problem here. In cricket and laws of cricket there's an external link to official rulesets, which are primary sources. I don't know a thing about cricket, but I'm confident that I could read the rules and verify that the articles properly reflect the primary sources, if I really felt like it. Furthermore, given cricket's popularity, if there are situations in which interpretation of the rules are unclear, there are probably secondary sources discussing the issue (see tuck rule for an example from American football). --Akhilleus (talk) 18:04, 18 November 2006 (UTC)

I think you've hit the nail on the head. The problem is the "without specialist knowledge" bit. Removing that would leave the advice as "Edits that rely on primary sources should only make descriptive statements". I'm tempted to suggest omitting the whole sentence rather than leaving the remainder in. jguk 18:20, 18 November 2006 (UTC)

Impact of revisions at WP:EL

(also posted at WP:RS) Please note that WP:EL has been revised... there is one section that may impact this guideline or the various rewrites to WP:RS being drafted:

Links normally to be avoided
Except for a link to a page that is the subject of the article or is an official page of the subject of the article, one should avoid:
2 Any site that misleads the reader by use of factually inaccurate material or unverifiable research. See Reliable sources.

This is the first time that I can think of where factually inaccurate material has been discussed in such a explicitly negative way in a guideline. WP:RS certainly doesn't (nor the draft rewrites that might replace it). Given this, Perhaps the time has come for us to address the issue of sources that contain factually inaccurate material as well.

I know that we have to be careful not to contradict the priciples expressed in WP:NPOV... but should we state something similar to what is on WP:EL. I.E. any site that misleads the reader by use of factually inaccurate material or unverifiable research should not be considered reliable. Any thoughts? Blueboar 19:18, 17 November 2006 (UTC)

Wow that is unfortunate change. "misleads the reader by use of factually inaccurate material or unverifiable research" Now editors have to agree A) whether material is true or not B) whether research is verifiable C) if either A or B are false then they must agree if it is used in a misleading way. Could you not have excluded the external links you where aiming for with less problamatic wording? I will just say I would hate to see such wording enter into this policy regarding sources.--Birgitte§β ʈ Talk 20:09, 17 November 2006 (UTC)
The Attribution proposal limits the use of questionable sources, which has roughly the same effect as normally avoiding "any site that misleads the reader by use of factually inaccurate material." The difference is the Attribution proposal leans towards examining the fact-checking reputation of the source, while the External links wording leans toward checking the truth of the facts themselves, with no admonition against using original research to check the facts.
I have more of an issue with the word "unverifiable". If the Washington Post bases a story on anonymous sources, the fact taht we can't telephone the sources and ask for verification does not mean we should delete a link to the Washington Post story. --Gerry Ashton 20:25, 17 November 2006 (UTC)
As I think of this; maybe some of this is not so bad for external links as websites should be able to correct any inaccuracies without much diffculty if they wish to be reliable. However it would be a bad idea to say a book source is not reliable if it contains any inaccurate material. I really don't get the misleading bit. How people are supposed to decide if inaccurate material is being used to purposefully mislead or not is beyond me. If you drop the intentions; how can inaccurate material on it's own not mislead? --Birgitte§β ʈ Talk 20:41, 17 November 2006 (UTC)

"[...] should we state something similar to what is on WP:EL. I.E. any site that misleads the reader by use of factually inaccurate material or unverifiable research should not be considered reliable. Any thoughts?" I think the sites that the phrasing at WP:EL is intended to catch are those that are described as "questionable" here. I think the phrasing they've chosen is poor, and substantially too broad for what they were trying to achieve. JulesH 15:08, 18 November 2006 (UTC)

This idea opens a can of worms. On contentious issues, each side often claims the other is making factual mistakes or being misleading. As a result, saying "this source makes mistakes, so it shouldn't be used" ends up being the same as "I personally can argue against what this source is saying, so it should not be used". I could point out that the ACLU has stated mistaken things about the Patriot Act (and in turn, other people could claim that those mistakes are immaterial to the gist of what the ACLU is saying). Should I remove the ACLU as a reliable source?
See also Wikipedia:Requests_for_arbitration/Bowling_for_Columbine/Evidence. Ken Arromdee 16:08, 20 November 2006 (UTC)

More about sources and original research

Sources

The advice about sources is still somewhat confused. First sources are mentioned in bold in the bold proper with a reference to "reliable" sources. Instead of a straight discussion or explanation of "reliable", we have a rule about "primary" and "secondary" sources, both of which need defining. We then have an alternative rule introducing concepts of "questionable" and "self-published" sources, both of which are defined away from their normal English meanings. This alternative rule is subject to three detailed exceptions, some of which are subject to further exceptions.

I'm sure this can be simplified. Maybe we could take the current wording in WP:V and work on improving that? jguk 15:34, 18 November 2006 (UTC)

Original research

The proposal also appears to overemphasise original research. WP:NOR started off as a device to prevent people from typing in their own wacky physics theories, or what have you. It was only much later in Wikipedia's development that WP:V, and an emphasis on providing sources when questioned, developed. Largely the concept of WP:V (the requirement to provide a reliable source) means that we don't need a separate rule of WP:NOR (don't use material that cannot be sourced). NOR may be important, but it's a subset of V. This can be seen in the definitions of "unsourced material" (material without a reliable source) and "original research" (material which cannot have a reliable source). Everything fitting within the second definition necessarily fits within the first. Had V been about in Wikipedia's first days, it is likely that NOR would always have been expressed as a subset of V.

Despite this, the special case of V (NOR) is mentioned first. V gets one paragraph, and then the rest of the discussion is centred around NOR. This seems to be a mistake. We should mention the general rule, V, first, and note what V means (which is largely absent from WP:ATT at the moment). Only then should we describe NOR as a particularly important special case of V. This approach would also make it easier for newbies and those reading the page because they want to learn from it to understand what they should be doing when making new edits. jguk 15:34, 18 November 2006 (UTC)

The policy should be rewritten from scatch starting with the justification that because anyone can edit we must establish credibilty, followed by the method which is to properly source claims with sources that are reliable for those claims, and then provide general advice for adults concening the principles of good sourcing. The associated guideline should contain advice for the young or inexperienced (this is where all the "the world is black and white" stuff goes), and examples. WAS 4.250 21:00, 18 November 2006 (UTC)
I do not see any feasibility in a proposal to re-write policy from scratch, WAS. We are having a hard enough time attempting to clarify current policy through attempts such as WP:ATT. ≈ jossi ≈ t@ 03:39, 19 November 2006 (UTC)
I guess I wasn't clear. ATT is already rewording existing policy - poorly. I am merely suggesting it be done right. Don't create a strawman, if you please. By "rewrite" I don't mean any more than the "clarify" you mention - done right. WAS 4.250 05:07, 19 November 2006 (UTC)
I agree with WAS. The approach he suggests is far more logical and, if adopted, would make the whole policy much easier to read - for newbies and for experienced editors. Maybe if he has time he could mock up what this approach would mean in practice. jguk 08:28, 19 November 2006 (UTC)
I'm willing to put in some time if others are willing to help. Any volunteers? WAS 4.250 16:42, 19 November 2006 (UTC)
Extensive work by many editors over more than six weeks has resulted in the current formulation of ATT. Why the dismissal of all that hard work? Why do you thing that a new round will produce anything better? ≈ jossi ≈ t@ 17:08, 20 November 2006 (UTC)
Any rewrite will be resisted. Over 60 editors have worked on this and it's looking not bad. We have to work the FAQ up to standard, and wait for the "tweaking" (or in the case of one editor, the disruption) of this proposal to end, and then we can talk about whether it can be presented. SlimVirgin (talk) 19:55, 21 November 2006 (UTC)


  • I don't see why one would have to restart from scratch to address the problems indicated by Jguk and WAS. A paragraph on justification wouldn't hurt and could be added to the top. If the page focuses to much on NOR, that could be trimmed down, or the section that is not on NOR expanded. Don't throw the baby out with the bathwater. (Radiant) 16:01, 21 November 2006 (UTC)
Maybe if we added a justification at the top, as WAS suggests, reincorporate the key points of WP:V that have been lost from WP:ATT at some stage, and trim down some of the NOR section so that we have a project page that is succinct (other contributors have suggested 1000 words max as a guide) it would look much better. Shall I have a bash, or does someone else want to have a go? jguk 18:56, 21 November 2006 (UTC)
What aspects have been lost from WP:V? I do not see the need to reduce the size of the text relating to NOR. ≈ jossi ≈ t@ 19:03, 21 November 2006 (UTC)
The emphasis on the three parts of the WP:V policy (the bits in the policy box). If we are to retain a succinct policy (even using 1000 words max as a guide rather than a diktat), and we are already at around 1000 words, and we add more on WP:V and a justification, which is a worthy suggestion from WAS, it follows that some of what is already there needs to be reduced in size. jguk 19:06, 21 November 2006 (UTC)
I'd remove the "policy in a nutshell" box and replace it with something along the following (the first three points are taken verbatim from WP:V):

The policy:

  1. Articles should contain only material that has been published by reliable sources.
  2. Editors adding new material to an article should cite a reliable source, or it may be challenged or removed by any editor.
  3. The obligation to provide a reliable source lies with the editors wishing to include the material, not on those seeking to challenge and/or remove it.
  4. It follows from the above that editors should not add material that constitutes original research.
jguk 19:10, 21 November 2006 (UTC)
I don't follow. The current nutshell is succinct and captures very well the current formulation. All these points are well covered in the article. As for the reduction in size, I do ot see the need to abide by a 1,000 word formulation. It either works as is or it does not. A random decision about size is unnecessary. ≈ jossi ≈ t@ 19:15, 21 November 2006 (UTC)
I'll reformat the wording in the "nutshell" box into a WP:V policy box - maybe it'll look better that way. jguk 19:19, 21 November 2006 (UTC)
Actually, I'm surprised by how much I now like the proposal after my reformatting. The second half (the discussion on sources) still needs work, and I still think WAS's justification up front idea should be worked upon. But it looks better. I hope my tweaks (I haven't deleted anything, just reformatted) sticks. jguk 19:26, 21 November 2006 (UTC)
No, no, no, Jguk. You cannot do that. We are discussing this, there are objections, and you still proceed? What is going on? ≈ jossi ≈ t@ 19:37, 21 November 2006 (UTC)
jguk, I'm not at all surprised by how much you like your proposal, but I can't see why anyone else would. Let's try to collaboratively work towards changes that actually improve the policy, instead of what you are doing. Jayjg (talk) 19:47, 21 November 2006 (UTC)

We'll discuss it below. jguk 20:31, 21 November 2006 (UTC)

Some random suggestions

A few thoughts:

  1. Generally, this is looking pretty good. Congratulations to everyone who put in the hard work to get it this far.
  2. Reviewing the article at this point, I think it gets hung up on the term "published source." Are "published source" and "self-published source" defined anywhere? Do they need to be?
  3. If I were king of the universe, or at least the Wikipedia, I would be inclined to say that while combining WP:OR and WP:V into a single policy appears achievable, WP:RS may still need to be its own page. This page would be a lot simpler if it just said that sources had to be reliable, had a single paragraph about what makes sources reliable, and then left self-published sources, exceptions, partisan sources, and all the rest for the RS page.

Sorry if all of this has been resolved before, TheronJ 21:05, 20 November 2006 (UTC)

"This page would be a lot simpler if it just said that sources had to be reliable, had a single paragraph about what makes sources reliable, and then left self-published sources, exceptions, partisan sources, and all the rest for the RS page." Exactly. Will you volunteer to help write such a paragraph as I proposed above? WAS 4.250 21:49, 20 November 2006 (UTC)
TheronJ, you may have missed WP:ATT/FAQ. Please check if the FAQ takes care of your concerns. ≈ jossi ≈ t@ 22:57, 20 November 2006 (UTC)
Also note that there is an excellent explanation of "reliable" and "published" in the article itself under the Key Principles heading, subheading "Wikipedia articles must be based on reliable sources" (copied below) ≈ jossi ≈ t@ 23:00, 20 November 2006 (UTC)

Reliable sources are credible published materials with a reliable publication process; their authors are generally regarded as trustworthy or authoritative in relation to the subject at hand. How reliable a source is depends on context. In general, the most reliable sources are books and journals published by universities; mainstream newspapers; and magazines and journals that are published by known publishing houses. What these have in common is process and approval between document creation and publication. As a rule of thumb,the more people engaged in checking facts, analyzing legal issues, and scrutinizing the writing, the more reliable the publication. Material that is self-published is generally not regarded as reliable, but seebelow for exceptions. Any unsourced material may be removed, and in biographies of living persons unsourced contentious material must be removed immediately.

  • The FAQ answers most of my comments very well. I still see some possible ambiguity on "published," however. This is a perennial problem with the existing guidelines - is a free weekly like The Stranger a "self-published" source? Is an online only publication like Media Matters or Snopes self-published, and how do we tell? Does it make sense to draft a section of the faq explaining the difference between "published" and "self-published?" TheronJ 23:21, 20 November 2006 (UTC)
Those nuances may be better left to be addressed on a case-by-case basis. The examples you provided can be good sources depending on the context and the specific situation, regardless if they are self-published or not. But sure, give it a go and add some material to the FAQ if you think it is necessary. ≈ jossi ≈ t@ 15:25, 21 November 2006 (UTC)
Thanks, I'll sleep on it and give it a shot. (On reflection, the current discussion of "self-published" does offer a lot of guidance on this question). TheronJ 15:31, 21 November 2006 (UTC)

Not OR if false

I removed the sentence saying false statements aren't OR, because I couldn't see what it was getting at. If I add 2 plus 2 equals 5 to an article, and can't provide a source, it's OR. OR is anything for which no source can be found, and for which it's assumed no source exists (not attributable), because it's the editor's own opinion. Whether it's true or false doesn't come into it. SlimVirgin (talk) 19:52, 21 November 2006 (UTC)

  • My point was that while everything OR is unsourceable, not everything that is unsourceable is OR. Someone who says "2 + 2 = 5" is likely not using his own logic to create support for a position, but simply making a mistake or repeating something he heard somewhere. Indeed, such a statement is neither original nor is it research. (Radiant) 09:29, 22 November 2006 (UTC)
In certain science fields, the OR policy is used to prevent fringe or "crank" theories from being published on wikipedia before they appear in peer-reviewed publications. It is often impossible to determine whether these theories are true or false because they are so poorly presented. The editors of these articles are skiled at wikilawyering to try to keep their inappropriate articles from being deleted. The burden of determing that something is not false before it can be called OR would make it very difficult to use the OR policy to delete these articles. The general definition of OR that is used to delete these articles is "if it hasn't appeared in a peer-reviewed publication, it's OR, whether or not it is correct". CMummert 12:15, 22 November 2006 (UTC)
  • It is irrelevant if we are unsure whether something is false or original research - because in both cases we don't want the information on Wikipedia. The point is, then, that we don't want unsourceable information regardless of whether it's "original" or "research" (or, indeed, "true"). (Radiant) 12:26, 22 November 2006 (UTC)
Yes, we don't want it. As SlimVirgin said, "original research" has always meant "unable to be sourced" which made it easy to delete. If you change the definition of original research to really be about original research, it would become impossible to verify that anything is actually original research, because you would have to prove that it is original. That's the motivation for the strange meaning of the words "original research" we use. CMummert 12:43, 22 November 2006 (UTC)
  • So basically what is being said here is that "original research" is a misnomer, since something on Wikipedia can be OR while being neither original nor research. If we're going to go by that somewhat misleading definition, this needs a footnote to clarify it to newcomers. I personally would never have considered "2 + 2 = 5" OR, but I would have removed it anyway. (Radiant) 12:52, 22 November 2006 (UTC)
Yes, that's it. As WP:OR says "Original research is a term used in Wikipedia to refer to material that has not been published by a reliable source". See note 1 in WP:OR for an explanation. Maybe that note could be copied as a note into this article. CMummert 13:35, 22 November 2006 (UTC)
But that's not what you did; instead, you changed the definition. It has nothing to do with what kind of information it is. All that matters is that it is not attributable. SlimVirgin (talk) 18:02, 22 November 2006 (UTC)
  • Well, then please reword it. If "original research" does not, or not always, mean what it intuitively does (i.e. something that is "original" and "research") then this needs to be pointed out. Otherwise novice users will get confused. (Radiant) 09:08, 23 November 2006 (UTC)
    • Oddly, I have never found "original research" to be counterintuitive nor inappropriate (although it is often a euphemism for "utter crap"). Wikipedia should never be the publisher of first instance for any idea. Isn't it as simple as that? Robert A.West (Talk) 19:52, 24 November 2006 (UTC)

Reformatting

I'd like to suggest reformatting the start of the proposal as follows (note this is a reformatting, nothing is deleted):

The policy:

  1. All material in Wikipedia must be attributable to a reliable, published source.
  2. Unsourced or poorly sourced material may be challenged or removed.

Wikipedia:Attribution is one of Wikipedia's core content policies. It complements Wikipedia:Neutral point of view. Jointly, these policies determine the type and quality of material that is acceptable in articles; that is, content on Wikipedia must be both attributable and written from a neutral point of view. Because the policies are complementary, they should not be interpreted in isolation from one another. For examples and explanations that illustrate key aspects of this policy, see Wikipedia:Attribution/FAQ.

==Key principles==

Wikipedia articles must be based on reliable sources
Reliable sources are credible published materials with a reliable publication process; their authors are generally regarded as trustworthy or authoritative in relation to the subject at hand. How reliable a source is depends on context. In general, the most reliable sources are books and journals published by universities; mainstream newspapers; and magazines and journals that are published by known publishing houses. What these have in common is process and approval between document creation and publication. As a rule of thumb, the more people engaged in checking facts, analysing legal issues, and scrutinizing the writing, the more reliable the publication. Material that is self-published is generally not regarded as reliable, but see below for exceptions. Any unsourced material may be removed, and in biographies of living persons unsourced contentious material must be removed immediately.
Wikipedia does not publish original research
Wikipedia is an encyclopedia, not a publisher of original thought. The threshold for inclusion in Wikipedia is whether material is attributable to a published source, not whether it is true. Wikipedia is not the place to publish your own opinions, experiences or arguments.
Although everything in Wikipedia must be attributable, in practice not all material is attributed. Editors should provide attribution for any material that is challenged or likely to be challenged. The burden of evidence lies with the editor wishing to add or retain the material.
Original research refers to material that is not attributable to a reliable, published source. This includes unpublished facts, arguments, ideas, statements, and neologisms; and any unpublished analysis or synthesis of published material that appears to advance a position. Material added to articles must be directly and explicitly supported by the cited sources.

end of extract

This organisation has the following benefits:

  • aesthetically, it looks better
  • it emphasises the key aspects of the policy better
  • as for any article, it starts with the name in bold followed by a brief description
  • it accentuates the positive aspects (ie what you should do) before noting the negative aspects (ie what you should not do) whilst retaining the meaning of the latter
  • it reads better as it is in a more logical order, jguk 20:30, 21 November 2006 (UTC)
(copied from above) No, no, no, Jguk. You cannot do that. We are discussing this, there are objections, and you still proceed? What is going on? ≈ jossi ≈ t@ 19:37, 21 November 2006 (UTC)
jguk, I'm not at all surprised by how much you like your proposal, but I can't see why anyone else would. Let's try to collaboratively work towards changes that actually improve the policy, instead of what you are doing. Jayjg (talk) 19:47, 21 November 2006 (UTC)
Same as usual. SlimVirgin (talk) 20:44, 21 November 2006 (UTC)

Not sure this cosmetic change helps. I would prefer using my time to work on the FAQ, than to keep massaging the current formulation. ≈ jossi ≈ (talk) 20:58, 21 November 2006 (UTC)

I tried to go through RS the other day to find anything worth keeping in there to add to the FAQ here. In particular, I was looking at the area-specific sections. It's a bit of a mess, so I'll need more time to do it than I have at the moment. SlimVirgin (talk) 22:27, 21 November 2006 (UTC)

Regarding Jguk's suggestion: I think it works better placing the "reliable sources" principle above the "original research" one. It seems more balanced that way. The other changes don't seem to add anything useful to me, though. Too much useful text was removed from the lead section, and changing the 'nutshell' box to one that suggests that those two sentences are the entire policy is just misleading. Perhaps you could argue that the policy is contained within the "key principles" section and the rest is just explanation (although not so much now there's no longer a "use judgement" principle which could be used to justify the exceptions) but suggesting the policy could be simplified to two sentences is taking it too far, IMO. JulesH 08:48, 22 November 2006 (UTC)

Jules, thank you for your positive comments regarding reordering.
As far as policy boxes go, a policy box has worked really well on WP:Verifiability for around nine months now. It allows people to get to the real nub of what we mean without reading the discussion below (if they so choose). I agree with your comment that the current "nutshell" on this page is too short for this. Maybe we should copy what is already done at WP:V and paste it in here - after all, as it's worked well at WP:V, it will work well here. jguk 11:17, 25 November 2006 (UTC)

Provision of WP:V that is not here

One sentence that I think is important was lost in the transition from WP:V.

If an article topic has no reliable, third-party sources, Wikipedia should not have an article on it.

This proviso is needed to fully justify deletion of articles that have only unreliable sources, and/or are based entirely on self-published claims. Some editors are claiming that WP:V explicitly allows use of self-published works to establish notability, whereas (as I have always understood) the exception is that one can use self-published works to the extent that they comment on the thing for which the publisher is notable, as established though third-party sources. I cannot conceive that we allow notability to be established via self-published sources.

I would like to add this phrasing back to where it logically belongs, under the exceptions in "Questionable or self-published sources." I boldface the additions for clarity on the talk page only -- there is no need for emphasis on the project page.

* * *

1. Questionable or self-published sources in articles about their authors
Material from questionable or self-published sources may be used as primary source material in articles about that source's author. The material:
  • must be relevant to the person's or organization's notability, as established using independent, third-party sources;
  • should not be contentious or unduly self-serving;
  • must not be used to support claims about topics not directly related to the source or about third parties.
  • must not be the sole basis for the article. If a topic has no third-party, independent sources, Wikipedia should not have an article on that topic.

* * *

I think that this will address this misunderstanding, which I have actually seen. Personally, I would extend the prohibition to articles where the only source is questionable, but I am not confident that will meet with approval. I cannot think of a valid reason to have an article where the only possible sources are self-published. Even if we use a notable person's or a municipality's website to write a basic, uncontroversial article, we are confident that there are other sources available, so the article should eventually be improved. Robert A.West (Talk) 12:20, 25 November 2006 (UTC)

Indeed that one sentence, "If an article topic has no reliable, third-party sources, Wikipedia should not have an article on it" is a nice, concise alternative way of saying all that's in those paragraphs about no original research. jguk 12:33, 25 November 2006 (UTC)

You may want to take a look at the proposal at WP:ATT where V, NOR and RS have been combined into a single article. It addresses these concerns quite well. ≈ jossi ≈ (talk) 18:59, 26 November 2006 (UTC)

:<confused look>. This is the talk page for WP:ATT, and I have spend considerable time working on this page. I am commenting on the existing text of this project, which omits the key phrase above and, IMO, does not address this particular misunderstanding so well as the current WP:V text. Robert A.West (Talk) 19:06, 26 November 2006 (UTC)

SOrry for the confusion above. That specific issue is already covered on the first bullet point: must be relevant to the person's or organization's notability. If the subject is notable, third-party sources will be plentiful. ≈ jossi ≈ (talk) 05:13, 27 November 2006 (UTC)

  • Well, not necessarily. Lots of things on the Internet can be well known without having any actual verifiable information about them. In some cases, we don't even know if the story is true, because there are no third-party reliable sources, but the story is famous. And, of course, notability is not actually a policy, so consensus can agree to keep a non-notable article, which could then be validly sourced solely from self-published crap as the current version stands. I don't see why you object to retaining this phrase which is part of current policy and has served Wikipedia well. Robert A.West (Talk) 05:28, 27 November 2006 (UTC)
I do not oppose that phrase, just that it seems to be already covered on the first sentence. If you feel strongly for its inclusion, you may try consolidating these two phrases into one. ≈ jossi ≈ (talk) 05:37, 27 November 2006 (UTC)
I have been suffering through arguments that notability can be established via self-published sources. [2] I'll accept your invitation. See what you think. Robert A.West (Talk) 05:56, 27 November 2006 (UTC)
I'm not entirely sure that the meaning of "The material ... must be relevant to the person's or organization's notability, as established using independent, third-party sources" is clear. Perhaps changing "as" to "which must be", or adding a final requirement to the clause that the importance of a subject must be established by third-party sources before self-published material can be used, or similar. JulesH 12:36, 27 November 2006 (UTC)
I see your point. Unless I misunderstand, "Notability" here means not the fact that the person or organization is notable, which is a separate question, but the subject matter area for which he/she/it is notable. Thus, an NFL quarterback's opinion on league policy is relevant. His self-published opinion on foreign policy is probably not. The fact that he is an NFL quarterback needs independent confirmation. I'll try to rephrase more clearly. Robert A.West (Talk) 12:53, 27 November 2006 (UTC)

Questionable sources redux

If this were analogous to the current WP:RS, the end of the sentence should have the bolded-underlined text added: “…except in articles about themselves or their activities.” Is this deletion intentional, and if so, why? Thank you. -- Avi 01:54, 27 November 2006 (UTC)

I don't believe so; I think the text came from the version at WP:V which, as far as I can tell, has always been missing this clause. I think it may have been added to WP:RS at a later stage and not taken across to WP:V -- exactly the kind of problem that this proposal is intended to prevent. JulesH 08:46, 29 November 2006 (UTC)

Begging the question

"Material from questionable or self-published sources… must be relevant to the person's or organization's notability, as established using third-party, independent sources." In short, you may only use questionable or self-published sources when you have better sources? - Jmabel | Talk 08:21, 29 November 2006 (UTC)

On second thought, I find this this is even stranger than I did at first. One of the most likely uses of self-published sources is to fill in general biographical information, precisely the sort of human interest material that is not relevant to the person's notability: marriage, number of children, where they went to school, hobbies, etc. - Jmabel | Talk 08:24, 29 November 2006 (UTC)

It's an ambiguous sentence. Does "as established using third-party, independent sources" pertain to "relevant" or to "notability"? In other words, are we asking for better sources to verify the relevance of bad sources (silly) or asserting that, once something's notability is established, we can use unreliable sources (just plain wrong)? Maybe the original thrust of the sentence was that unreliable or self-published sources, although they may be used in some connections, can't be used to establish notability. I doubt the editor who wrote it meant to imply that, for matters not related to estabishing notability, it's a free-for-all. -GTBacchus(talk) 08:29, 29 November 2006 (UTC)

OK, what I think the sentence means, and what it should probably be replaced with is something along the lines of:

Material from questionable or self-published sources must be relevant to the article's subject, whose notability must be established with information from third-party sources, and must be in a field in which the author is considered authoritative -- for example, it is usually acceptable to use biographical details from a self-published source.

A bit wordy, perhaps. And perhaps I've misunderstood the intent of the original sentence; I've never been entirely happy with it, even over at WP:V where it seems to have originated. JulesH 08:42, 29 November 2006 (UTC)

That is what I have always understood it to mean. Once a person is notable for Q, you can, with care, use his self-published remarks on subject Q, and possibly on closely-related subject P, but certainly not on unrelated subject R. Robert A.West (Talk) 22:31, 29 November 2006 (UTC)
How about this? "Must be relevant to a topic for which the person or organization is notable according to reliable, third-party, independent sources." Does that avoid confusion yet preserve brevity? Robert A.West (Talk) 22:35, 29 November 2006 (UTC)
That is contradictory. The original formulation of "must be relevant to the person's or organization's notability" is sufficient to explain this point: If you use self-published sources, the material must be relevant to the person or organization notability. That is, you cannot base a claim about a person or organization, based on a self-published source, if that material has nothing to do with the person or organization's notability. ≈ jossi ≈ (talk) 03:26, 1 December 2006 (UTC)
I have added JulesH wording as it makes it quite clear what we are referring to. ≈ jossi ≈ (talk) 03:32, 1 December 2006 (UTC)
I like the new wording, it limits to exceptions, rather than seeming to give permission. Atom 12:51, 1 December 2006 (UTC)
The current version says, "Must be relevant to the person's or organization." This is either an incomplete thought, with the word, "Notability" missing, or a stray possessive. Frankly, I hope that the intention is to get rid of the ambiguous word, "notability," since I disagree with (and am astonished by) Jossi's (apparent) assertion above that it is acceptable to use self-published sources to establish notability. Robert A.West (Talk) 17:46, 1 December 2006 (UTC)
Considering that "notability" is nothing more than a topic's having been covered non-trivially in multiple independent sources, it would be quite the contradiction to establish it using self-published material. -GTBacchus(talk) 18:56, 1 December 2006 (UTC)

Usenet posts?

Sorry if I'm jumping in late here, but does this proposed policy change our stance on using usenet posts or non-notable blogs as sources for disputed statements of controvercial fact? Hipocrite - «Talk» 14:07, 11 December 2006 (UTC)

I don't see how they could be. On usenet you never know for certain who a poster is. If you look at S. M. Stirling, where a usenet post Stevie boy made many years ago is being cited as proof of Islamophobia, it's far from provable that the poster in question was SMS, even though it's very, very likely. Unless there's proof from some independent, fact-checked source, I wouldn't accept a usenet post as proof of anything. Not even the existence of Archimedes Plutonium ("there is no assurance that all posts attributed to Plutonium on Usenet were by him"; an understatement to be sure). Angus McLellan (Talk) 16:04, 11 December 2006 (UTC)
I think the issue is different for blogs than it is for usenet posts. There will be blogs out there that we can reliably determine to be by the stated author. However, because real blogs (as opposed to corporate PR puff blogs) are supposed have top of the head, what the author thinks today content - they are not fact checked to the same extent that the author's website or written publications are. So, using the same evaluation process I'd apply to a personal web page, I'd come down to concluding that the blog just isn't a reliable source for much. As for the question of the author's attitude, it is a primary source, and I personally would refrain from viewing an isolated comment or two as an adequate basis for asserting a general way of thinking. Analysis of the context and frequency of such contents is best left to biographers, rather than encyclopedia editors. GRBerry 18:23, 11 December 2006 (UTC)


I think the issue is with this statement: "Personal websites and messages on USENET and Internet bulletin boards are considered self-published." That's giving USENET messages the same weight as a personal website, which they shouldn't have. Note that a self-published source would be quite valuable for even controversial claims about the subject: "I, Joe Schmoe, am gay." However that is for messages for their personal site, verifiably theirs, not for any Usenet or forum post by someone claiming to be Joe Schmoe. I'd make a separate section for them, something like this:

With rare exceptions, messages on USENET and Internet bulletin boards can not be verified to be by the person they are claiming to be from. Unless the person has admitted authorship of these messages in more reliable sources, Usenet and forum messages are to be considered anonymous, and should not be used to verify anything. AnonEMouse (squeak) 18:33, 11 December 2006 (UTC)
Unless there is reason to doubt the identity of the author, I don't see any particular problem with usenet/message board posts. Yes, they can be faked, so using them to back up controversial claims about the apparent poster is not usually a good idea. But in most cases where a usenet post is used on wikipedia, that isn't the case. I would suggest the following addition to the end of the "self-published sources in articles about their authors" section:
  • should not normally be used if there is any doubt about the authenticity of the source (e.g., sources in semi-anonymous media like USENET and Internet message boards), particularly if the claim made by the source is controversial.
Comments? JulesH 10:07, 12 December 2006 (UTC)
I would add something to cover the situation where people link to their usenet posts from their blogs or their website. That links the usenet post back to a more verifiable source that claims the usenet post as being from that person. Similar to blogs that link to Wikipedia talk page comments by the blogger. Carcharoth 10:11, 12 December 2006 (UTC)
I don't think that's necessary. That's clearly not a case where there is "doubt about the authenticity of the source", so the rule wouldn't apply. JulesH 10:34, 12 December 2006 (UTC)
And the remaining rules about self-published sources apply with full force. Robert A.West (Talk) 11:25, 12 December 2006 (UTC)

Phrasing of limitation of third exception

I understand User:AnonEMouse's concern that the third exception not seem to contradict the first. I understand User:JulesH's concern that we not appear to create a redundant exception. I have rephrased to, I hope, address both concerns. I have also corrected the punctuation: a semicolon should separate clauses, not phrases within a clause. Robert A.West (Talk) 11:27, 12 December 2006 (UTC)

I'm not sure I at all like the resulting wording. Despite Robert's edit, the the third section seems to directly contradict the first, and then some. "This exception never applies in the areas of history, politics, current affairs…" Everything is arguably either history or current affairs. Perhaps these are meant in a different sense than I am taking them? Often, an organization's web site is the only decent source for organizational history. Does this mean that we should not trust, say, the site of CAMERA for the identity of its founder?
As for politics, it would seem absolutely clear that (for example) a party's or candidate's own materials are usually the best source for their official views. - Jmabel | Talk 16:56, 15 December 2006 (UTC)

What is not original research - two comments

It says: "For example, if a published source gives the numbers of votes cast in an election, it is not original research to calculate a percentage, so long as it is a simple calculation and the original numbers the source offers accompany it."
--- I don't see why the last part is necessary. It should be enough that the source contains the original numbers. The underlying principle is that readers should be able to go to the source and check without too much trouble that the information attributed to it is correct. An example where this issue is more clear: suppose a source has a table giving the population of a city divided into age-group and sex, but omits the total. We should be able to add the numbers up and just give the total in wikipedia, without needing to copy the whole table into wikipedia. I propose replacing "and the original numbers the source offers accompany it" by "and it is clear from the source which numbers should be used in the calculation". --Zerotalk 09:28, 21 December 2006 (UTC)

Maybe it is sufficiently clear already, I'm not sure, but such elementary deductions should be able to span sources. If source 1 says "he was born in 1823" and source 2 says "he died in 1883", we should be able to write "he lived for about 60 years [cite1][cite2]". Can that example go in the FAQ? Another example: Source 1 says "Joe died in 1923", source 2 says "Alex was born in 1939", we report "Joe died before Alex was born[cite1][cite2]." --Zerotalk 09:28, 21 December 2006 (UTC)

I agree with both comments, except where it says "Maybe it is sufficiently clear already," -- we wouldn't be having this discussion if it was clear already.... would we? --Michael C. Price talk 09:32, 21 December 2006 (UTC)

What is and isn't OR: a dozen comments

Hi. I got pointed here from a discussion on Wikipedia_talk:No_original_research about whether a "synthesis" is permissible in some situations. I posted a list of possible guidelines ( I posted an earlier version of this list to talk:NOR ), do these make sense? Squidfryerchef 03:07, 22 December 2006 (UTC)

1. Just because the facts "A is a B" and "B is a C" are each asserted in reliable sources, does not by itself allow us to claim that "A is a C" ( this is a paraphrase of the "synthesis" paragraph from NOR ).

2. If a reliable souce can be found that makes that same inference, then the editor can simply cite the source and skip the rest of the guidelines.

3. Logical inferences that are made to advance a thesis are not allowed. For example, a literature term paper is OR just as a science experiment or a journalistic interview is OR. It is acceptable, however, to compare and contrast several points of view provided the other guidelines are met.

4. Sources used in logical inferences are held to a higher standard than simply being reliable. They must also be an authority on the subject. This doesn't mean they have to be an authority on the main subject of the article, but to use a source for "A is a B", it must be an authoritative source on A or B.

5. If there are several conflicting authoritative sources on the subject, the editor should not cherry-pick one that best supports an inference. If there is a standard reference book which everybody uses for the given subject, its OK to use the one. If there's an "oligopoly" of standard references, it's OK to compare a couple of those. If there is no one authoritative source but several minority views, the article should summarize several of those.

6 There needs to be an exemption for "general knowledge" inferences. Otherwise, ( here's a tip of the pen ) saying "city A is west of city B" would be original research when the source says "city B is east of city A". The only facts required to make that inference would be a lexicon that west is the opposite of east.

7. I was told once that general knowledge, in a research paper, encompasses anything you might find in an encyclopedia. Obviously this definition wouldn't do for us. General knowledge for our purposes might be material included in a children's dictionary, encyclopedia, atlas, etc.

8. It is permissible, but by no means necessary, to include a reference for a general-knowledge inference.

9. The flip side of the "general knowledge" guideline is that if a source asserts that "A is a B", but uses an unconventional definition of what a B is, that is, one that contradicts the popular definition of what a B entails, then the editor is encouraged to compare and explain this usage of B.

10. Some editors may feel that closely-placed citations suggest an inference to the reader. For example, if "Y accused X of P because X did Q" is followed by "Note that authoritative reference R does not include Q in its definition of P", some editors will say "NOR" even though the inference "X did not commit P" does not appear in the text. While on one hand we can't sneak inferences into WP like that, on the other hand we can't have NOR being invoked to tell people they can't make citations.

11. What constitutes "advancing a thesis" depends in part on whether opinion is involved, if there is a chain of inferences involved, or simply on the length of the synthesized material. For example, a clause pointing out a Shakespeare reference in an episode summary of a cartoon is not a "thesis", because the editor does not have an opinion vested in there being a Shakespeare reference, but it aids in summarizing the episode. However, if it went on for several paragraphs it would be a thesis.

12. Adding items to categories, "See also" lists of internal links, and deciding which sources to cite should not count as "advancing a thesis". Otherwise we could get into meta-meta situations where you need a source just to cite a source. This does not mean to enourage "listcruft"; what belongs in a list is subject to consensus, and some things do and don't belong. However, WP:NOR is not the right policy to settle such discussions.

Squidfryerchef 03:07, 22 December 2006 (UTC)

Policy tag

Okay, I've been bold and added the policy tag. It's been stable for two weeks now and the page is getting linked to across Wikipedia, which shows it is becoming regarded as a policy. I'm extremely open to reversion, it's a bold move, but I think the time is right and that at worst it generates discussion as to what more it needs. I'd ask that anyone who feels like reverting ask themselves this one question: Is there any reason that any changes you'd like to make to the page could not be made whilst the page was policy? The policy tag, to my mind, shouldn't set in stone a specific wording or version of the page, just the basic idea. I'm of the opinion that we're all now of the opinion that this page at some point will become policy, and if that is the case I'd ask people not to revert the tag but to discuss improving the page. However, that said, please feel free to revert the tag back to proposal, noting reasons why here so it can generate improvement of the page. I haven't removed the policy tags at WP:NOR and WP:V as yet, I reckon that there needs to be some sort of discussion as to how that process works, whether we redirect those pages or tag them as historical, superceded by this page, so as to allow links to then to convey historical context. Also, if the move is reverted, I'd like to ask that no-one makes my boldness the first shot in a tag revert war. Discussion is the key, I think we all agree. Anyway, season's greetings. Hiding Talk 16:47, 27 December 2006 (UTC)

Does a regular user have the prerogative to add this tag? I have one reason why it's not ready: exception three is still a BS loophole that shouldn't become policy. Marskell 16:35, 28 December 2006 (UTC)
I think it's good that we have this page. What are we going to do with the three pages it was intended to supersede? >Radiant< 16:56, 28 December 2006 (UTC)
That all needs to be worked out still. The thing got stymied a bit for various reasons I won't go into. :-) Perhaps we could wait until Marskell has more time (in a few days, I believe) before discussing it in more depth? SlimVirgin (talk) 17:18, 28 December 2006 (UTC)
I'm pleased. (I fear this will be reverted, though, by people unversed in the lengthy discussions.) In my opinion the other two policies should stay policy for the time being: let the strongest page(s) win. Apart from anything else, it will take some more time for article editors to become aware of this one. We can legitimately fend off reversion here if we respond thoughtfully to accusations of contradictions, not that any spring to mind. qp10qp 17:16, 28 December 2006 (UTC)
I'd be worried about trying to keep three policies (NOR, V, and ATT) running in parallel, and we need to discuss what to do with the controversial section that I believe Marskell removed. I'd like to wait until he's less busy in real life so he can take part in the discussion. SlimVirgin (talk) 17:20, 28 December 2006 (UTC)
Oh, OK. I haven't any idea how these things are done. At least the idea hasn't been dropped. qp10qp 17:22, 28 December 2006 (UTC)
  • Regular user? {{Speechless}}. I forget my last edit to the thing. I think you're wrong not to discuss it with the tag in situ, to be honest, but that's just me. I think you'll never get the tag on if you wait until everyone is happy. Still, I'm not going to revert, I'm not even going to touch the page again until it's tagged as policy again. I am out of here. Hiding Talk 17:26, 28 December 2006 (UTC)
I didn't mean regular user badly. We're all regular users AFAIK (one of the nice things about Wiki). I was under the impression that serious policy changes required a specific nod from Jimbo. In any case, this has received very few edits in December. Perhaps we should determine first if people do want to continue to work on it? Marskell 00:15, 29 December 2006 (UTC)
  • That impression is incorrect; the community can certainly make policy without involving Jimbo's opinion, and has done so in the past. If you look through CAT:P you'll notice that many of our policies didn't have his involvement. >Radiant< 09:34, 29 December 2006 (UTC)

I agree that making page into policy doesn't require Jimbo's confirmation. However, it does require a broad consensus, not just between editors of this policy. I don't see a link to poll on it, ant it would better be held first. CP/M comm |Wikipedia Neutrality Project| 10:08, 29 December 2006 (UTC)

  • We don't vote on policy either (note, again, that most of CAT:P wasn't ever voted upon) and consensus is not determined by polling or voting. This page is widely published and edited by a lot of people. Many objections have been raised and addressed. All that is very solid consensus-building. And note that nobody's going to protect the page, so we can still improve it. The question is whether there's substantial disagreement, and if so what that is based upon. We should not get hung up on "procedure" since Wikipedia is not a bureaucracy. >Radiant< 10:32, 29 December 2006 (UTC)
Hm, well WP:HCP notes that serious policy changes have generally been done under Jimbo's auspices. The list there shows PROD to be the only really major innovation he wasn't involved in. Note we don't just need consensus for this, but also for the scrapping of V and NOR. I don't think BOLD is quite enough for that, and I hate the idea of this proceeding (particularly enshrining exception three) without much broader input. Marskell 18:28, 29 December 2006 (UTC)
I think we're never going to get there if we take the view that anything is enshrined simply by adding a policy tag. Consider Wikipedia:Verifiability's state when the policy tag was added, [3]. Note that it allowed a far greater degree of flexibility than is being allowed now. To be honest, I think that page works a lot better than the current one does. Note the guidance on dubious sources: one must assess whether the source is verifiable. In the case of a source of facts: is the source a noted expert in the area? Does the source write blatant errors? Has the source followed journalistic or academic standards of ethical investigation? In the case of a source of opinion: is the source notable? Does it stand for a large group of people? Would that we could be so wise now. As to those listings at WP:HCP and Jimmy's involvement, Jimmy never actually had input into Wikipedia:Verifiability either. Hiding Talk 18:44, 29 December 2006 (UTC)
I would comment that while this has been widely advertised, I don't believe that it has been widely advertised when in the current form. I also haven't noticed anyone citing it in discussions yet. I know that I have not considered it stable enough to refer to.
My personal prefernces would be to take a three step process. First get this in use regularly. Second use that as a basis for declaring it policy. Third, after that is done, get the agreement to downgrade the old pages to "historical policy" or something similar. The essay Wikipedia:There is no deadline applies as much to changes of policy as it does to article editing, although written in the latter context. GRBerry 18:48, 29 December 2006 (UTC)
Citations at Talk:Assassination and Wikipedia:Articles for deletion/List of dictators to pick two. Still, it's all horses for courses. I take your point, but I can't see that people are going to cite this over WP:V when WP:V has a policy tag and this doesn't, because of the rules lawyering that would generate. Hiding Talk 19:08, 29 December 2006 (UTC)

Third exception

I am concerned about the third exception to the usage of questionable or self-published sources, which permits them in areas where professional coverage is shallow. My understanding of how an encyclopedia works is that if professional coverage of a certain topic shallow, so should be that of an encyclopedia. On the other hand, we must bear in mind that this policy should be applied together with WP:NPOV, which in some circumstances cannot be followed to the letter if questionable sources are ignored. One example that comes to mind is a case when an otherwise questionable source has stirred a sufficiently notable controversy, which deserves being mentioned in the article on the subject of the controversy. Some examples: The Thirteenth Tribe, From Time Immemorial, Icebreaker (Suvorov). Beit Or 20:03, 28 December 2006 (UTC)

One example given was a guide to Buffalo nickels which is considered the definitive work on the subject by coin collectors, but which is self-published by someone who doesn't have a degree in Buffalo nickels or a job related to them.
A bigger example is popular culture. Most popular culture topics aren't covered in professional sources except for the occasional mass-market book which isn't really peer-reviewed or fact-checked. Ken Arromdee 21:12, 28 December 2006 (UTC)
Then the intent of the exception is as clear as so much mud; numismatics is as much an academic field as those listed. Septentrionalis PMAnderson 21:25, 28 December 2006 (UTC)
Popular culture is not exactly the subject where professional coverage is shallow. Beit Or 21:28, 28 December 2006 (UTC)
It was pop culture we had in mind with that exception; things like comics. SlimVirgin (talk) 23:30, 28 December 2006 (UTC)
Citing Pierre de Fermat in the article on diophantine equations would violate policy, because Fermat was a lawyer, not a professional mathematician, much of his work was self-published, and the field is a scholarly one. Forget the fact that he practically invented the field as we know it. In contrast, random fanboy speculation about the significance of Dragonball Z would seem to be admissible on the assumption that there are no scholarly writings. Please forgive me for failing to see how this special distinction for comics improves the encyclopedia. Robert A.West (Talk) 00:02, 29 December 2006 (UTC)
Might I slyly suggest that the half dozen posts above (re)foreground the lack of consensus on this... Yes, we can discuss it while it's up there. Or we could just spike the thing, agree to move forward on the policy tag, and then if someone wants it or something like it they'll have to pass the usual high consensus bar as is done now on policy talk. Marskell 00:22, 29 December 2006 (UTC)
Basically, we're back to the old dispute on Wikipedia talk:Reliable sources: which source is reliable and which one is not. I humbly reiterate my old point: there is no substitute for common sense. Beit Or 12:13, 29 December 2006 (UTC)
Which comes back to my old point: common sense should be mentioned here, up at the top in the principles section.JulesH 14:10, 29 December 2006 (UTC)

It's not just comics

Contradicting myself, I know, but the exception wasn't built with comics in mind as much as with any pop culture. My expertise, such as it is and not strong enough to merit my thoughts being usable, happens to be in comics. The problem is with the second exception, and the words professional, researcher and journalist.

I've argued time and time again that sources simply be weighed against the material being sourced and that strong claims require strong sources, but people feel differently. The issue comes about because whilst USENET postings would be unacceptable in an article on some scientific theory, they are actually useful in documenting certain pop culture phenomena. This is because different topics use different forums for discussing and researching ideas. The problem tends to be that tradition held that diaries and letters were private and as such couldn't widely be cited.

Time has moved on and a lot of people make their diaries and letters public through email archives, USENET postings and blogs, making their citation and verification easier. Now these sources shouldn't be inviolate, they should simply be measured for their worth, and summarised in a NPOV. If the information contained within a source is some person declaring he has devised the theory of everything, then USENET is a bad place to build from, because in this instance the source has a bias. However, USENET is being cited for instances of new words being coined by the OED, since the source can support such a fact neutrally.

Opinion can be sourced from USENET, as can the fact that something has been published on USENET. An example, and yes I'm using comics here, it's a field I'm versed in, would be both Rich Johnston and Michael Doran, both of whom started by posting news and gossip to USENET, and both of whom have advanced in the field to become reliable sources. The issue is not one of using unreliable sources, it is in using sources which are reliable within context, but which we can't actually nail down with a definition of context.

Take Eddie Campbell; according to the guidance here his blog is off limits because he is not a "well-known, professional researcher in a relevant field, or a well-known professional journalist". His failing is in being a well known, professional artist in a relevant field. And I tell you, I'd go all the way to arb-com to cite anything relevant he posts there, because not doing so is inane. It would be like disregarding Einstein's blog or Hitchcock's; again, not actually professional researchers or journalists, but practitioners and theorists. I mean, look at those words there. Researcher and journalist mean very specific things; they don't cover commentators, they don't cover practitioners, they don't cover observers. Sometimes this stuff is useful, mostly in articles on art, culture and society. Sometimes this stuff is detrimental, typically in articles on science, but these are judgement calls which require constant revisiting and revising. Luckily, a wiki is the best way to enable such revisiting and revising, and luckily that's what Wikipedia is. Don't throw the baby out with the bath water.

So how would I do it? I'd remove the bits about journalist or researcher. Whatever door you're trying to bolt there is already double locked by the rest of the second exception, which notes that "we can rely on it if material produced by that writer would normally be regarded as a reliable source. However, exercise caution: there may be a good reason no reliable third party has published the material. If there is reasonable doubt about the reliability of a self-published source, or the relevance of the material to the subject matter, don't use it." That covers it. If the person publishing the stuff would normally be a reliable source, that already covers well-known, professional researchers in a relevant field, or well-known professional journalists. And I don't get why well known journalists can pronounce on any field. Are Martin Bell's self published thoughts on flower arranging of more relevance than George Dyson's?

But look. You don't need to suit me. I'll just carry on editing as I always have. When my contributions are judged as harmful to the encyclopedia, I'll have worked out that what I'm editing isn't really worth my contributions anyway. I think whilst this debate rages, there are hundreds, maybe thousands of articles where breaches of the more fundamental point, that of WP:NPOV, are being committed and allowed to fester. Whatever you write here isn't going to stop that. Hiding Talk 11:17, 29 December 2006 (UTC)

An earlier version of this proposal had 'expert' in place of reasearcher. What happened to that? JulesH 14:10, 29 December 2006 (UTC)
Someone said define expert, at a guess. Most likely there was a point made about how anyone can call themselves an expert and so it was defined as such. Like I say, why not simply state that Where a person who would normally be regarded as a reliable source self publishes material we can rely on it to a certain extent. However, the material should concern the field in which the writer is held to be a reliable source. Further, it should be considered that there may be a good reason no reliable third party has published the material. If there is reasonable doubt about the reliability of a self-published source, or the relevance of the material to the subject matter, don't use it. Hiding Talk 14:24, 29 December 2006 (UTC)
That leaves the question of when a person "would normally be regarded as a reliable source" -- how do we make this determination? The current solution is to let reputable publishers make it for us. JulesH 14:35, 2 January 2007 (UTC)
When a person has multiple works published by reputable publishers, and also self-publishes material that concerns the same field, we may be able to rely on the self-published material to a limited extent. Nevertheless, there may be a good reason that no reliable third party has published the material. If there is reasonable doubt about the reliability of a self-published source, or the relevance of the material to the subject matter, don't use it. Robert A.West (Talk) 17:46, 2 January 2007 (UTC)
I'm not sure I like that solution. The current policy doesn't require the author of a self-published source to be previously published, and I'd hesitate to introduce any new restrictions like that which could potentially invalidated sources in existing articles. JulesH 23:12, 2 January 2007 (UTC)

Publishing houses

The text says: "In general, the most reliable sources are books and journals published by universities; mainstream newspapers; and magazines and journals that are published by known publishing houses." This preference for books published by univerisites over those published by "known publishing houses" is unwarranted. There is simply no evidence that the editorial oversight is better at university presses. In the field of history, I'd prefer a book printed by Routledge over a book that came out of Columbia University Press or University of California Press (both have churned out more than their fair share of propaganda or widely discredited nonsense), and I'd rank Brill Academic Publishers way above any univeristy press. I'm open to arguments, but at the moment I can see no basis for relying so much on university publications. Beit Or 12:41, 29 December 2006 (UTC)

I've never been happy with this phrasing either. Perhaps something like:
In general, the most reliable sources are books and journals published by professional specialist publishers with a good reputation in the field, followed by mainstream publications of well-known publishing houses.
I'd also suggest adding something about publications of academic interest groups - for instance on computer science related topics the best sources are often published by groups like the IEEE or ACM, who aren't really publishers because that isn't their primary business. JulesH 14:45, 29 December 2006 (UTC)