Jump to content

Wikipedia talk:Attribution/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3Archive 5

Original research definition

The definition needs to be rewritten, because it confuses defining original research with prohibiting the publication of original research in Wikipedia. It says "Original research refers to material for which no reliable published source could be found...." I'm sure Science would be horrified to discover they don't publish any original research. --Gerry Ashton 05:40, 11 October 2006 (UTC)

We have an entire encyclopedia people can fill with words detailing the fine points of "original research". Our policy should help editors make content choices, not pretend it can substitute for clue or to substitute for encyclopedia articles such as one on "original research". WAS 4.250 05:47, 11 October 2006 (UTC)
Are you suggesting the original research key term be deleted? --Gerry Ashton 05:53, 11 October 2006 (UTC)

The original research key principle now reads:

Wikipedia does not publish original research.
    • Original research refers to material for which no reliable published source could be found; that is, it refers to material that is not attributable to a reliable source. It includes unpublished arguments, concepts, data, ideas, statements, and theories. It also includes any unpublished analysis or synthesis of published material that appears to advance a position.

This way of phrasing it perpetuates a way that many Wikipedians have of writing about original research on talk, policy, and guideline pages. Anything that cannot, after a reasonable effort, be attributed is original research. This differs from the meaning of original research outside of Wikipedia, where original research involves creating new ideas, as contrasted with regular research, which involves organizing and summarizing existing ideas.

Wikipedia is a publication. To the extent Wikipedia publishes original research, it violates its own policy, but the material is published nonetheless. Furthermore, much of the material in academic journals and newpapers is original research, yet it is published in a reliable source. To avoid a self-contradictory paragraph, I suggest this (italics are to distinguish proposed wording from my commentary):

Wikipedia does not publish original research that is not attributable to a reliable source. Previously unpublished arguments, concepts, data, ideas, statements, and theories are excluded, and so are any previously unpublished analysis or synthesis of published material that appears to advance a position.

--Gerry Ashton 06:17, 11 October 2006 (UTC)

Gerry, OR is material for which no reliable source exists, or could reasonably be found (we can never say definitively that no source exists). That is, it is material that is not attributable to a reliable source. Unattributed = unsourced. Unattributable = original research. SlimVirgin (talk) 06:22, 11 October 2006 (UTC)
Suppose I create a Wikipedia article that is a verbatim copy of a U.S. patent. It's original research; the law says it has to be novel to get a patent. It's a reliable source. It's not protected by copyright. So Wikipedia has published original research. The reason it does not violate policy is that Wikipedia is not the first publisher; the U.S. Patent Office was the first publisher. To say that something stops being original research just because Wikipedia republishes it is a concept that only exists within Wikipedia.
I think a real benefit of combining the No original research and Verifiability policies is we can stop trying to distinguish whether a Wikipedia editor performed original research while composing an article, or whether the original research was done by some other person whom the WP editor copied, or whether the material is original research at all, and just focus on whether the same ideas can be attributed to a reliable source. --Gerry Ashton 06:55, 11 October 2006 (UTC)
I don't believe the USPTO counts as a reliable source. They may claim that everything they publish is a novel invention, but the facts just don't back it up. They've published a lot of junk in the past. JulesH 08:02, 11 October 2006 (UTC)
That isn't especially relevant to the point I was making, but I'd say the USPTO is a reliable source when compared to the spectrum of sources that are accepted on Wikipedia. I wouldn't trust USPTO as much as Science, but I'd trust it as much as the typical evening TV newscast in a medium sized city. --Gerry Ashton 09:55, 11 October 2006 (UTC)
I'm not sure how long an article which is a direct copy of a U.S. patent would last on Wikipedia, and this seems a very convoluted example. Are there any other areas where this might cause a problem, because at the minute I'm happy that verbatim copies of U.S. patents would be deleted. Hiding Talk 15:18, 11 October 2006 (UTC)
This discussion is about whether, using the normal meaning of "original research", and not any jargon meaning that only applies within Wikipedia, should Wikipedia publish original research? On the "yes" side, an article or a passage in an article might be copied from a reliable source that presents original research, and we are allowed to copy from such sources as long as we follow the other policies too. On the "no" side, the process of the Wikipedia editor finding and copying the material is non-original research, so by the time it gests to Wikipedia it isn't original research anymore. --Gerry Ashton 17:34, 11 October 2006 (UTC)
I guess since this is a Wikipedia policy we'd use the Wikipedia definition. I'd also make a stab that issues like that are so rare and complex they don't really matter. If the only example we have is copying a U.S. patent I'm not sure we need to document that. Hiding Talk 18:15, 11 October 2006 (UTC)

(unindent) I'd say that something as prominent as policy, to which new editors will be referred frequently, should use as few specialized definitions, or jargon, as possible. And copying a patent is hardly the only example; anytime we copy material from a primary source, this applies. --Gerry Ashton 18:26, 11 October 2006 (UTC)

I don't think we do copy from a primary source, we quote a primary source. At least, that's my take on it. Copying would violate copyright. And material sourced from a primary source isn't necessarily original research. And I'd say policy on Wikipedia should do what Wikipedians want it to do, and be Wikipedia specific. Otherwise we'd be unable to use the term Wikipedia and Wikipedian, for starters. As long as we explain what we mean, I don't see the issue. Hiding Talk 18:39, 11 October 2006 (UTC)
Quoting and copying are nearly the same thing, although quoting has connotations of citing the source and brevitiy while copying does not have those connotations. Not all sources are covered by copyright, for example, patents, US government publications, laws, and court decisions. --Gerry Ashton 18:53, 11 October 2006 (UTC)
Quoting and copying are nearly the same thing, but as you note, aren't for important . They get put in WikiSource. Hiding Talk 19:10, 11 October 2006 (UTC)
Whenever Wikipedia copies or quotes previously published original research, it publishes original research, and doing so is perfectly fine as long as the material is properly attributed and copyright laws are obeyed. The quoting or copying is apt to be more extensive if the first publication is not copyrighted. When Wikikpedia assign a jargon definition to "original research" and then says that it is against policy to publish original research, this is an affront to the English language and I oppose abusing the English language in this way. --Gerry Ashton 19:19, 11 October 2006 (UTC)
I'm not sure there's much I can do to address that, is there? Hiding Talk 20:31, 11 October 2006 (UTC)
I'm amazed at how the term "original research" has grown from its original usage as a simple tool to expel kook ideas & other silliness from Wikipedia. The emphasis has gone from defining how to exclude information that is obviously wrong to what we can include whether it is right or wrong. I'm not sure that it is an improvement. -- llywrch 22:42, 12 October 2006 (UTC)

Some key terms section

The section titled "Some key terms" does not confine itself to defining terms, it also expresses policy (e.g. edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge). This is a problem, because people who are looking for the policy may not think to look in the key terms section. --Gerry Ashton 05:52, 11 October 2006 (UTC)

Is it better now? SlimVirgin (talk) 06:00, 11 October 2006 (UTC)
Yes, this arrangement is more elegant than anything that came to my mind. --Gerry Ashton 06:18, 11 October 2006 (UTC)

Nutshell

I have expanded the Policy in a Nutshell section to include a point of WP:V that I believe is vital to any content policy - that is, material which is not attributed to any reliable source can be challenged or summarily removed. The burden of attributing information to a reliable source must fall on the person adding or defending such material. FCYTravis 06:06, 11 October 2006 (UTC)

FCY, I've avoided the secondary/tertiary issue because people find it confusing. Also, to say that reliable sources are needed because reliable information is needed is repetitive. We should try to keep unnecessary words to a minimum. SlimVirgin (talk) 06:17, 11 October 2006 (UTC)
  • I'm not convinced that a nutshell is in fact necessary if it simply restates the first few sentences of the page. >Radiant< 15:31, 11 October 2006 (UTC)
  • Can any material which is not attributed by summarily removed? This seems like a terrible position to hold. Material which is unattributed and which one has reason to believe is not true can be summarily removed. If there's an unsourced statement that Henry VII was King of England from 1485 to 1509, we shouldn't remove that. john k 16:47, 11 October 2006 (UTC)
Anything, cited or not, can be removed for lots of reasons. I might take something out because it gives undue weight, because it isn't accurate, or because it is just bad writing: yet another thing that can be eaten on toast. It is (sadly) not false that lard and sugar can be eaten on toast, nor do we necessarily need a citation to say so. I remove it because it is item 17 on an increasingly lame list of condiments, after peanut butter, chutney, and mayonaise. I remove stuff all the time, just in the normal course of editing. Tom Harrison Talk 17:16, 11 October 2006 (UTC)
Of course, but information which one knows to be completely correct, but which is not cited, should not be removed on the basis that there is no citation. FCYTravis seems to be saying that if I come across an article saying that Henry VII reigned as King of England from 1485 to 1509, but that fact is not cited, I am perfectly within my rights in removing that item, simply on the basis that no citation is provided. That is completely absurd. john k 21:19, 11 October 2006 (UTC)
Yes, I agree. It can't be used to frivolously remove stuff, or unreasonably demand citations. Saying "may be challenged or removed" in the nutshell may be taken to suggest otherwise. Tom Harrison Talk 23:10, 11 October 2006 (UTC)

Sourcing and notability

Again, not necessarily OR. If I'm staging a play that no one has written about, I can't write about it on WP because it's not notable enough, not because my article would be OR. Why do you want to avoid mentioning notability? SlimVirgin (talk) 06:19, 11 October 2006 (UTC)
If you're staging a play nobody has written about, then it's prima facie OR. There's two concepts here. Regardless of whether something is "notable" or not, if we can't find reliable sources to talk about it, it can't be written about. But just because we can find reliable sources, doesn't necessarily mean it's notable. I think we need to clearly separate the two, because otherwise we end up with the concept that "if we can find a single reliable source that talks about it, that means it's notable and we can write about it." Which isn't necessarily the case at all. FCYTravis 06:23, 11 October 2006 (UTC)
Yes, I see your point. I'm just trying to avoid confusing people about what OR is. But maybe you're right. SlimVirgin (talk) 06:29, 11 October 2006 (UTC)
And it has to be secondary or tertiary (britannica articles only). If you are notable playwright and publish your play on your blog, it doesn't deserve an entry in Wikipedia until a secondary source publishes something about it. The abuse is that notable persons publish views that are not covered by secondary sources. It would technically be Original Research to summarize the play, but if the summary is provided by the playwright, it should still not be in Wikipedia because it's a RS issue. There is also a number of articles that publish original thought by notable people that self-publish outside their area of notability. This type of primary sourcing is particularly insidious because it has the color of authority even though it is not peer-reviewed or covered as a secondary source. --Tbeatty 06:36, 11 October 2006 (UTC)
Er, no, writing an article on a play nobody has written about is not original research, as long as the article is entirely based on the play. It is not original research to use a work of fiction as a source for a description of the plot of that work of fiction, and this is what one would be doing, presumably. The problem is the play isn't notable. Has anybody ever written a plot summary of a work of fiction (novel, play, movie, television show, short story) on wikipedia that was based on a secondary source, rather than on the work of fiction itself? john k 16:52, 11 October 2006 (UTC)
  • Yes. The Adventures of Tintin. Plot summary based on the work can stray into OR because of issues of interpretation. I would much rather source plot summary from secondary sources, where possible. I appreciate this isn't always possible. Hiding Talk 17:05, 11 October 2006 (UTC)
  • Me too. I've often written about a book from book reviews, or based on a description in the writer's biography, or from a survey of the literature of a particular cultural tradition, etc. - Jmabel | Talk 04:43, 12 October 2006 (UTC)

I changed the first sentence in "Requesting sources" to read "If an article topic has no reliable sources that can be cited, it cannot be written about in Wikipedia because this is not the place to pubish information for the first time."

I made this change because I think this is the covers the gamut of reasons why sources are necessary better than the previous wordings. One previous wording mentioned notability, but something can be notable even if reliable sources about the event do not yet exist. Another wording mentioned that information for which no reliable sources exist is by definition original research. This is not strictly true; a Wikipedia editor copying from a privately held source is not doing original research, but the material must nevertheless be excluded from Wikipedia on the grounds of lack of attribution. --Gerry Ashton 06:40, 11 October 2006 (UTC)

That's a good change, thanks. SlimVirgin (talk) 07:06, 11 October 2006 (UTC)

The "Requesting sources" section currently contains this passage:

Note the difference between unsourced material and original research:
    • Unsourced material = material not yet attributed to a reliable source.
    • Original research = material that cannot be attributed to a reliable source.
Original research should be removed immediately, but give people a chance to provide attribution for unsourced material unless you're fairly sure none exists.

I belive this usage of "original research" is Wikipedia jargon. Material that cannot be attributed to a reliable source might be original research, or it might be a blatant non-libelous lie. The material should be immediately removed either way, but I don't think that making up blatant lies constitutes original research. We could say "unatributable material" or "unsourceable material" rather than "original research"; any other suggestions? --Gerry Ashton 09:14, 11 October 2006 (UTC)

This is just our own internal definition of original research. It means "stuff that Wikipedians have made up." This could be a lie, or some other type of fabrication or creative input. The important point is that there exists no reliable source for it. It's not just that we can't find one; but also that we're fairly certain none exists. SlimVirgin (talk) 09:30, 11 October 2006 (UTC)
Indeed; that's what I meant by the word "jargon". If we are going to use "original research" that way in a policy, the first time we use it we should say something like "for purposes of this policy, original research means any material that cannot be attributed to a reliable source with reasonable effort". But if we can think of a phrase that means what we want without having to make up our own definition, that would be better. --Gerry Ashton 09:40, 11 October 2006 (UTC)
It's such a famous phrase in Wikipedia, I'm not sure we could get rid of it. SlimVirgin (talk) 12:19, 11 October 2006 (UTC)
Yes, it is jargon. But it would be over-ambitious to replace it with a different phrase, I think, given its currency within wikipedia. john k 16:52, 11 October 2006 (UTC)

Secondary and Tertiary sources

I think it's important to mention that articles have at least one secondary source. Tertiary sources were mentioned to include the old Britannica articles that aren't covered by copyright. There are articles with only primary sources that should be deleted and we might as well close this door before it's opened if we are amending policy. I can think of no case where a primary source only article is a valid entity on Wikipedia. But the circular reasoning is if Person A is notable (through whatever means), his blog entry can be cited as an article for Theory X even though there is no secondary source coverage of Theory X. It's a loophole and it is being abused and is not the intent of wikipedia. --Tbeatty 06:28, 11 October 2006 (UTC)

It's a hard loophole to close without ruling out blogs entirely, which is something that annoys a lot of good editors, who want to be allowed to use them in measured ways. SlimVirgin (talk) 06:31, 11 October 2006 (UTC)
I don't want to eliminate blogs entirely, just as the sole source. Or more specifically, if an article's sole sources are from a "walled garden" of blogs and self-published sources, what good is it? If it's covered by the NY Times, and a blog is used to back it up that's great. But if Blog A supports Blog B and that supports Joes Web Page that points to Blog A, it's ugly. --Tbeatty 06:39, 11 October 2006 (UTC)
I agree. Some editors get very exercised about the idea of insisting on at least one secondary source per article, but I think it's reasonable myself. SlimVirgin (talk) 07:05, 11 October 2006 (UTC)
A straightforward summary of a notable work of fiction can be useful, and this would be an example of an article that might only have one primary source. Some uses of such an article might be to understand allusions to the work in other sources or to help parents select books for a child. Of course, the use of secondary sources would improve such articles, but I don't know that they fall below the minimum standard to be included in WP. --Gerry Ashton 07:07, 11 October 2006 (UTC)
Yes, although wouldn't you always need a secondary source to show notability e.g. this novel was nominated for a prize, or The Guardian said it was wonderful, or whatever? SlimVirgin (talk) 07:09, 11 October 2006 (UTC)
In some cases, a book could be its own source for notability, e.g. my copy of Catch-22 states on the cover that it had sold over 250,000 copies by the time that edition was printed. Regarding using blogs as sources, see the section on self-published sources I'm about to add. JulesH 07:17, 11 October 2006 (UTC)
If anyone questions the notability of the work of fiction, an independent source would have to be provided. Usually the independent source would be secondary, but if a sufficiently prominent primary source mentions the work of fiction, that might do. For example, if the official website of President X quotes her as saying Prime Minister Y is about as trustworthy as character C from movie M, and the two heads-of-state both lead nuclear powers, I think that makes movie M notable.
Getting away from works of fiction, another use for primary sources is summarizing or paraphrasing expensive works so people can get at the information free. This could apply to technical standard, for example. The notability of the standard might be so obvious that it never occurs to the original editor to provide a secondary source about notability, and it might never occur to the readers to ask about notability. --Gerry Ashton 07:29, 11 October 2006 (UTC)

I can think of many examples, but here is a massive one: the U.S. Census is technically a primary source, so all of the articles that were automatically generated from census data used only a primary source. There are probably still 10,000 or more of these that have had no further enhancement or sourcing. - Jmabel | Talk 04:50, 12 October 2006 (UTC)

But one should be able (in most cases) to find secondary sources for those articles. The proposal makes a distinction between articles that can be made policy-conforming and those that cannot. The former should be fixed or tagged, the latter deleted.
In addition, this excellent proposal still leaves is with the classification problem: what is the dividing line between a primary and a secondary source? For example, I would judge that an article in a peer-reviewed journal is a secondary source, even though it proposes a novel theory and is written by the scientists who did the fundamental experiment, whereas the raw data, lab notebooks and so on are the primary sources. Other Wikipedians hold the view that such a journal article is a primary source, because it first presents a new theory. Similarly, is Plutarch a secondary source when he quotes a work no longer extant? I would judge not. In fact, I would regard practically all ancient and medieval sources as primary, because of the huge cultural and linguistic abyss that separates us, necessitating expert intepretation. Other editors disagree in good faith. Robert A.West (Talk) 00:10, 13 October 2006 (UTC)

Self-published sources section

I still think the self-published source section needs some work. My primary concerns are:

  • It's still biased towards academic subjects. We can use a "professional researcher"'s self-published work, but we can't use a subject expert's work. Consider, for instance, articles on Microsoft software, e.g. Microsoft Exchange Server. There are a lot of experts on this software out there, many of them accredited by Microsoft themselves (e.g. Microsoft MVPs) whose opinions are notable and worth including in articles on the software, yet currently they are excluded by the wording here (as they are by WP:V).
  • It allows us to use a self-published work in an article on the author of that work, but not on an article about the author's theories. If interpreted literally, it would leave us substantially short of sources for articles like Time Cube and Scientology and others, because most of the body of work that those articles are about was self-published.
  • A self-published work should always be usable as a source for the opinion of its author. If the opinion of the author is deemed relevant, it should be possible to include it on any page, so long as it is phrased as an attributed opinion. I think there would have to be a separate guideline about when it's useful to include opinions, but this is probably needed already anyway. For instance, the self-published opinions of John Hemming MP in his blog on gas supply issues should be citable on Energy policy in the United Kingdom, if anything he says is relevant to that page.

See also my unfinished essay on this subject at User:JulesH/Self Published Sources. JulesH 07:36, 11 October 2006 (UTC)

Another thought: I think that for a person's opinions, a primary source should be prefered. Secondary sources frequently misstate opinions, and it's rare for a primary source to do so. JulesH 07:51, 11 October 2006 (UTC)

Another possibility not mentioned in the Self-published sources section is self-published sources that become reliable after publication because independent reliable sources recommend them. --Gerry Ashton 07:53, 11 October 2006 (UTC)
I have added a fourth exception to cover self-published sources that are recommended by independent reliable sources. As an example, consider the Internet Resource Links page at the National Institute of Standards and Technology. Two of these links, Metrication Matters and US Metric Association each appear to be under the control of single individuals, and one might argue they are self-published. But since they are recommended by NIST, there is really no point in having such an argument; they're probably reliable. --Gerry Ashton 08:31, 11 October 2006 (UTC)
Let me add a slightly more clear-cut example. Explanatory Supplement to the Astronomical Almanac, edited by P. Kenneth Seidelmann and published by University Science Books in 1991, p. 574. contains this reference: "Herget, P. (1948). Computation of Orbits Published privately by author." --Gerry Ashton 08:42, 11 October 2006 (UTC)
That doesn't mean P. Herget's book becomes a reliable source. Someone's citing it isn't the same as it being recommended. And indeed, Seidelmann may have cited the only sentence it in he's prepared to stand by, for all we know. (I'm not commenting on Herget's work here; just using it as an example.) SlimVirgin (talk) 08:48, 11 October 2006 (UTC)
It's true that mere presence in a reference or citation list does not make a work reliable; the independent reliable sources would have to give a favorable evaluation of the self-published work. My real point is that self-published works do get mentioned in reliable sources; it isn't just an abstract situation that never comes up in practice. --Gerry Ashton 08:56, 11 October 2006 (UTC)

section heading

The section heading What does this policy exclude? seems a bit problematic. The way I read it is 'what does this policy not cover' rather than the intended 'what does this policy disallow'. Maybe this should be changed?-Localzuk(talk) 08:36, 11 October 2006 (UTC)

Good point. I've changed it. SlimVirgin (talk) 08:38, 11 October 2006 (UTC)

But brevity

In my opinion this is far too long. It succumbs to the same structural fault of the policies it combines by tailing off into details and dog-show definitions. My preference would be for either this new policy or each of the existing two to be reduced to a maximum of, say, a thousand words (we kid ourselves if we think most people will plough through more than that). In my opinion, the policy or policies should state their principles only, using a minimum of examples and those only where unavoidable; each policy, however, should be connected by links or summary style to a guideline, the repository for all the niggling little ifs and buts that accrete on a policy like fluff on the tail of a rat. qp10qp 13:28, 11 October 2006 (UTC)

I disagree. That's what went wrong with the other policies. WP:V started filling up with material that people felt was too detailed for NOR. Then V became policy, and people wanted to streamline it too, so material was moved to RS. RS was only a guideline so lots of people piled in to edit it, and it ended up a mess. Let's not go down that road again.
In my view, we need one policy on sources that is comprehensive on the important, basic points, but succinct, and no forking to more detailed guidelines that end up contradicting it. SlimVirgin (talk) 13:37, 11 October 2006 (UTC)
This is currently 2,400 words, replacing NOR and V, which are over 4,800 between them, and making RS redundant (apart from the advice on where to look for sources), which is over 6,000.
2,400 words instead of nearly 11,000 isn't bad. If you cut it down too much, you risk losing the essentials. SlimVirgin (talk) 13:44, 11 October 2006 (UTC)
It's not forking, it's the difference between front-of-house and the back cupboard, in my opinion. Guidelines on this subject can't help but disintegrate into detail, which is where, as you say, they become a mess. That's inevitable, I think, so lets push that mess as far away from the statement of principles as possible, where the pokers, pedants, and cruft-shufflers can nitpick and spraint to their heart's content without fouling the policy pages in the process. That's my opinion, anyway; but I'm impressed by what you're trying to do; certainly we need a lot fewer words than in the combined originals. qp10qp 13:52, 11 October 2006 (UTC)
I completely agree with you on front-of-house and back cupboard, and RS was meant to fulfill the function of providing a space for detailed discussion of issues that would only serve to confuse the policy. But then links to RS were provided on V and NOR — in sentences like "verifiability depends on reliable sources — and people started getting the impression that V depended on RS; and RS was a mess, therefore V was fatally flawed. Then of course you have to make sure nothing in RS contradicts V, but whenever I tried to make copy edits to get rid of creeping contradictions, I'd get reverted because I wasn't one of the editors who were on the talk page every day. So it all becomes very maintenance heavy. SlimVirgin (talk) 14:06, 11 October 2006 (UTC)
Ah, I see the point. qp10qp 14:31, 11 October 2006 (UTC)

Scientific controversies

This proposal has major implications for scientific controversies.

The "truth" of the Theory of Evolution would immediately be downgraded from a "fact" (needing no attribution) to a "mainstream view" (requiring attribution to a source).

  • We'd have to write According to biologists natural selection et al. produces new species rather than Natural sselection et al. produces new species.

The controversy over anthropogenic (human-caused) Global Warming would likewise need recasting. The proportion of "consensus" vs. disgreement would no longer be something any Wikipedia contributor could simply assert as a "fact". Instead, the contributor would have to write:

  • According to an essay (or editorial) in Science, scientific journals have never published a peer-reviewed article denying the anthropogenic global warming theory.

Articles could also include statements from AGW sceptics like scientist Richard Lindzen and novelist Michael Crichton, M.D., who claim that numerous peer-reviewed articles have been published rebutting various aspets of AGW.

I like the new Attribution policy, because it will take Wikipedia out of the business of "getting to the bottom" of scientific controversies and put it back to where it was originally intended to be. It will not assert that "X is a fact" but rather that "A says X is true" and also "B says X is not true". This will allow readers to make up their own minds. --Uncle Ed 13:31, 11 October 2006 (UTC)

If that would be the result of the proposed policy, it's not something I could support. Tom Harrison Talk 13:45, 11 October 2006 (UTC)
This is my concern as well. --EngineerScotty 15:27, 11 October 2006 (UTC)
This proposal seems to mainly streamline and codify existing policies - I doubt that it woudl have the radical effects Ed wants it to. The Land 13:52, 11 October 2006 (UTC)
  • The only good thing I see in this is the synthesis description, which is well done. I worry that the changed semantics will end up tossing out a lot of the useful semantics tied to WP:NOR and WP:V, both of which, as I understand, are core principles of Wikipedia. I don't see a strong enough argument to change them, and think doing so may cause, as Ed notes, shifts in how we handle situations that move us away from being academic/encyclopedic. --Improv 14:22, 11 October 2006 (UTC)
I have the impression that Wikipedia:Neutral point of view is the only fundamental (foundation) policy of the three best-known (mentioned by Jimbo Wales in User:Jimbo Wales/Statement of principles and in m:Foundation issues). The other two aren't among the Wikipedia:Five Pillars and, by their nature, must adapt to changing circumstances to survive.qp10qp 14:58, 11 October 2006 (UTC)
  • Just to address the concerns above, actually it wuld be great if we could say that according to Nature of date foo no scientific journal has never published a peer-reviewed article denying the anthropogenic global warming theory; we can then say that within the scientific community it is generally regarded as accepted fact and that there is no significant informed dissent from this view in independent scientific thought. That is a very strong statement. So strong that the theory could justly be stated as fact (or as generally accepted) in the lead and the evidence stated a subsequent paragraph. It would be foolish to deny that there are climate change deniers, just as there are holocaust deniers and evolution deniers, but if we can make a directly attributable statement which unequivocally demonstrates the complete lack of acceptance of these denials by mainstream thought, I really see no problem. I am prepared to eb persuaded otherwise, of course. Guy 15:23, 11 October 2006 (UTC)
I would like to note, any encyclopedia that hedges claims about global warming or evolution with a "Group X believes" is a shitty encyclopedia. Phil Sandifer 19:31, 11 October 2006 (UTC)
Exactly. This proposal reflects the narrow interests of giving undue weight to fringe viewpoints than to a factual portrayal of the topics. FeloniousMonk 21:53, 11 October 2006 (UTC)
FM, which parts of the proposal imply that, would you say, because it certainly wasn't the intent; nothing in this proposal should adversely affect the undue weight provision of NPOV. SlimVirgin (talk) 04:54, 12 October 2006 (UTC)

wherever possible? what about when not possible?

Wikipedia articles should rely on reliable, published secondary sources wherever possible

Great. But we fail to address the issue of what to do when that is not possible. As it stands now this is a gaping hole for adding unreliable material to articles. ≈ jossi ≈ t@ 15:33, 11 October 2006 (UTC)

I don't understand the "wherever possible" at all. What is an example of when material not published by a reliable, secondary source would be acceptable? Sandy 17:44, 11 October 2006 (UTC)
The "wherever possible" refers to the reliance on secondary sources: we should rely on reliable secondary sources wherever possible; although sometimes primary sources will be appropropriate. SlimVirgin (talk) 23:29, 11 October 2006 (UTC)

This is another gaping hole: "2. Self-published sources that have been reviewed and recommended by independent reliable sources." Example? Sandy 17:47, 11 October 2006 (UTC)

See the self-published sources section of the talk page, near the bottom of that discussion, for some examples. I acknowledge that the examples I was able to come up with in a short time are not ideal examples, but I'm sure we could find better examples if we looked. The point is, reliable source do sometimes use self-published sources. If we can show that authorities in a field consider a certain self-published source to be reliable, why should we reject it for use in Wikipedia. --Gerry Ashton 17:56, 11 October 2006 (UTC)
Well, with blogs one can certainly find examples of this. Mainstream media sources certainly cite self-published blogs on occasion. And blogs that are not self-published very frequently cite self-published blogs. john k 12:45, 12 October 2006 (UTC)
I think this "reviewed and recommended" exception is problematic. First, we're an encyclopedia, not a newspaper or magazine, and we don't use sources in the same way that they do; so just because a reliable source — say, a newspaper — recommends a blog, that doesn't mean we can suddenly regard it as a reliable source. Secondly, it's usually particular material from a blog or other self-published source that's reviewed and referred to, not the whole thing. Just because the blog was good on Tuesday when the newspaper reviewed it doesn't mean it's good at any other time. SlimVirgin (talk) 13:00, 12 October 2006 (UTC)
I wonder about older books, which may have been self-published before publishing houses play the role they do today. I have never worked with this type of material, but I'm concerned about excluding anything important. Furthermore, the self-published definition has been edited to include organizations as self-publishers, so the "reviewed and recommended" exception may be necessary to use material published by organizations, with no author listed. The current wording requires us to decide if the organization is "well-known and reliable". It may be more practical to see if the material has been recommended in reliable sources. Also, the reliability of any source can always be discussed on the talk page, and the discussion might be more productive if discussing specific recommendations versus whether the organization is "well-known and reliable". --Gerry Ashton 18:02, 12 October 2006 (UTC)

I would think that with any archaic work (which I'll somewhat arbitrarily define as anything written before 1900--feel free to substitute a more appropriate guideline), such works fall into one of several classes:

  • Those which have been analyzed and commented on over the years by more recent scholars. These I have no issue with using (subject to the opinions of the scholars of their worth)--the lens of history can serve as a replacement for publication.
  • Those which have not. In some cases, an archaic work may have been lying dormant for years in an archive, recently discovered; in other cases, it wasn't considered worth commentary. In either case, the lack of historical perspective on these should disallow their use on Wikipedia; at least until modern scholars outside Wikipedia comment on and analyze them.

--EngineerScotty 22:19, 12 October 2006 (UTC)

WP:NOT

I'm not sure who came up with the idea of the "three core content policies" but I wish to point out that WP:NOT is just as much a "core content policy" as V/NPOV/NOR, and in fact predates two of the three. One could also plausibly argue for the Copyright Policy to be a core content policy, and possibly WP:OWN as well. >Radiant< 15:37, 11 October 2006 (UTC)

Fair point. I maintain that we should aim for the canonical policies to be very short. Guy 15:41, 11 October 2006 (UTC)
I would argue for Wikipedia:Lawful to be a core policy that subsumes copyright, libel, laws on privacy at least. Since the wiki foundation's charter specifies "educational", it might even subsume most of NOT. WAS 4.250 16:39, 11 October 2006 (UTC)
Maybe we could call NPOV and Attribution (or whatever we end up with) core editorial policies, rather than content. I wanted to get across the idea that there are these two policies that determine what we're allow to write, and that they work in harmony and shouldn't be intepreted in isolation. SlimVirgin (talk) 23:31, 11 October 2006 (UTC)

Clarification

In reading the comments above, I am unclear about what is the purpose of this. Can anyone clarify it for me? Is it:

  1. a replacement for WP:V, WP:NOR and WP:RS?
  2. a clarification of these?
  3. a new version of Wikipedia:Policy_trifecta, Wikipedia:Five_pillars, Wikipedia:Simplified_Ruleset?

Thanks ≈ jossi ≈ t@ 15:45, 11 October 2006 (UTC)

A more sensibly written and simpler replacement for WP:V, WP:NOR and WP:RS. Jayjg (talk) 15:54, 11 October 2006 (UTC)
Good luck to us, then. We will surely need it.... ≈ jossi ≈ t@ 17:36, 11 October 2006 (UTC)
Jossi, the proposal is to combine V and NOR into one policy; and then to turn RS into strictly a page of advice about where to look for good sources, minus anything that belongs in the policy. SlimVirgin (talk) 23:34, 11 October 2006 (UTC)

Why are we doing this? Surely this rewrite is valuable only if this rewrite clarifies what is meant by NPOV. That is, we have no independent reason to want Verifiability, ReliableSource, Attribution, . . . . except as they are required to implement WP:NPOV. For example, it would be very useful to have a few very clear pages to which we could send newcomers to explain what it takes to produce the NPOV quality of Wikipedia pages, is that not true?

Accordingly, this combination of Verifiability and ReliableSource into one page makes no improvement. I see nothing wrong with combining those two pages into this one. But the combination at WP:ATT does not address our real problem--which is how to resolve the controversies on pages.

I suggest that our goal in all of these policy pages is to tease out the details of this NPOV process of "representing significant views fairly and without bias." So we are wasting our time on these policy pages unless we are explaining to newcomers what we mean by 1) "representing," 2) "significant," 3) "fairly," and 4) "without bias." The current texts of WP:V and WP:RS are murky and self-contradictory in explaining how to resolve controversies over what is meant by 1) representing, 2) significant, 3) fairly, and 4) without bias. And the proposed WP:ATT is no better.

I give you an example of what we should be working here. We should be spending our time explaining here clearly what we mean by fairly when we say we should be "representing significant views fairly and without bias." Specifically, our policy text should state that it is "unfair" to paraphrase published ad hominem fallacies as a criticism of any political, economic, or scientific point-of-view. That is, it is "unfair" to paraphrase on the Pythagorean theorem page any WP:V and WP:RS publication that states that Pythagoras's writing "calls for the use of deception, misinformation and divisiveness to [conquer his opponents]." Even if some WP:V and WP:RS scholar actually said that Pythagoras's writing "calls for the use of deception, misinformation and divisiveness to [conquer his opponents]" that would not have relevance to any proof or disproof of the Pythagorean theorem.

So, I repeat with others above: What is the purpose of this combination page? Our grand mission here should be "representing significant views fairly and without bias." The current texts of WP:V, WP:RS, and WP:ATT mischaracterize that grand mission. For all of the above reasons, we should organize the sub-pages of WP:NPOV around explaining in clear and self-consistent text the elements of 1) "representing," 2) "significant," 3) "fairly," and 4) "without bias." --Rednblu 19:22, 11 October 2006 (UTC)


Fact vs opinion

My main concern with where this seems to be going, it that it blurs the distinction between claims of fact (statements of the form "P") and claims of opinion (statements of the form "according to source S, P")--where P is a proposition (which may be true, may be false, may be disputed, or may be entirely personal opinion) and S is a source for that proposition (which may be authoritative or relevent, or not). (Assume that in both cases, S is cited in the notes).

Current policy doesn't really address this well; WP:NPOV comes closest to dealing with it.

Obviously, a claim of fact should only apply when:

  • The fact is manifest and undisputed ("the sun rises in the East")
  • The fact is a matter of established historical record, and undisputed ("Babe Ruth hit 714 home runs in his major league career")
  • The fact is a matter of established scholarship or science, and is undisputed ("The bite of the King Cobra is toxic to humans").

Here, "undisputed" excludes cranks and the like--the fact that a few crackpots think they can square the circle with only compass and straightedge (though none have shown how, and it's been long proven otherwise) doesn't denigrate the claim that "it is impossible to square the circle using only compass and straightedge" from the status of fact.

As an encyclopedia, we should be presenting facts as facts.

Claims of opinion--which should always be attributed in prose--should apply when:

  • The claim is widely regarded as false or nonsense ("The Earth is flat", vs. "According to the Flat Earth Society, the Earth is flat.")
  • The claim itself is an opinion ("Ishtar is a lousy movie." vs "According to Roger Ebert, Ishtar is a lousy movie").
  • The claim involves an active controversy ("The War in Iraq is a failure" vs "According to many critics of George W. Bush, the War in Iraq is a failure").
  • The claim involves an area of scholarship or science which is significantly disputed, or presents a significant minority opinion ("According to most scientists, human activities have caused global warming"; "according to proponents of chiropractic, many maladies can be cured with spinal adjustments")
  • The source is of questionable reliablity. ("According to Jane Fonda, plutonium is the most toxic substance on Earth"--I heard that one from an old chemistry prof many years ago). In this case, of course, better sources that Fonda are available on this topic, and she ought not be quoted at all--but there are some propositions that are significant enough to merit mention but not terribly well-sourced
  • The source is a statistcal aggregate ("According to a Gallup poll, 62% of Americans believe that..."). Note that two sources--the researcher (Gallup), and the group being sampled ("62% of Americans"), are needed for this sort of claim.

Of importance to this instant topic is the following observation: Claims of fact require stronger evidence. WP:RS, for its faults, attempted to note that some sources are stronger than others. This current proposal seems to gloss over or ignore the difference. It would be a massive step backward were Wikipedia to be reduced to Uncle Ed's carciture above.

--EngineerScotty 15:49, 11 October 2006 (UTC)

You would be astonished at how many things people assert are "undisputed fact", and how often they disagree on just what those "undisputed facts" are. Jayjg (talk) 15:57, 11 October 2006 (UTC)
Not at all.  :) Generally, nothing is undisputed--for any proposition P, I'm sure someone can be found who will claim ¬P (whether they believe it or are just trolling is another matter). In some cases, all claims of ¬P can be dismissed as original research; in other cases, such claims will remain. The general policy is that if ¬P remains a fringe viewpoint, it can still be ignored--there are sources advocating Phrenology, but we still treat it as a discredited theory as opposed to a minority opinion. --EngineerScotty 16:04, 11 October 2006 (UTC)
Not that I have any idea how to put this into policy, but in science having at least one journal article disputing a claim, can be seen as an indication that the claim is not totally crankish. On the real crankish claims, no scientist would waste his efforts critizing them. That's the reasons it's hard to find good critical sources for Heim theory or New Chronology (Fomenko). --Pjacobi 16:22, 11 October 2006 (UTC)

Bible examples

Are the examples currently given regarding the Bible (it being a good source for the story of the Garden of Eden or details of the Ten Commandments) accurate? Religions differ significantly regarding exactly what the Bible says about both. Jayjg (talk) 16:00, 11 October 2006 (UTC)

Interpretations will differ, but the best source for the claim 'Genesis contains a narrative about Adam and Eve in the garden' is surely the text of Genesis itself. That the ten commandments include an injunction against theft, is likewise best verified from Exodus.--Doc 16:24, 11 October 2006 (UTC)
Theft is a good example. Could one say that the Ten Commandments contains an injunction against theft? The traditional Jewish interpretation, for example, is that that particular injunction forbids kidnapping (i.e. theft of people). Jayjg (talk) 16:40, 11 October 2006 (UTC)
It may not be the only citable source, but are you actually saying one could not cite the Bible as a source for this? I cannot imagine an article on either of these topics that did not cite the Bible. - Jmabel | Talk 16:12, 11 October 2006 (UTC)
See the theft example above. How would you break up the verses, translate them, etc.? These are all points of contention. Jayjg (talk) 16:40, 11 October 2006 (UTC)
I was under the impression that this new proposal will be tight and not too verbose. All these examples, can be moved to a separate article in which we give examples from different domains of knowledge. ≈ jossi ≈ t@ 16:14, 11 October 2006 (UTC)
IMHO any amount of Bible exegesis is original research and should strictly be avoided. Too many people claim but it's said in the Bible and too many peoply disagree about what's acually said. Scholars of theology have to be cited for exegesis. --Pjacobi 16:44, 11 October 2006 (UTC)
To a point yes. but statements like 'the Bible contains many references to Jericho/flowers/trees/donkeys' whilst strictly exegesis, are probably best referenced by citing some of the Biblical references. Likewise statements like 'according the Genesis Adam and Eve had a son called Cain', or 'According to John's gospel, Jesus claimed to be the 'true vine etc. To draw a distinction between a 'plain reading' and a 'hermeneutic' is in theory vety difficult, but in practice many readings will be uncontested. It is only when someone points out that 'hey, some schollars read that differently' that we need to insist on secondary sources (for both the traditional and alternative reading), it is simply a matter of common sense here (as with any primary source). Let's not over-complicate or prescribe until a problem emerges.--Doc 13:44, 12 October 2006 (UTC)

Foreign-language sources

My pet issue, I'll admit: as a person who reads about half a dozen languages decently, I find the strong preference for English language sources annoying. Try doing research on Romania, or on Spanish-language literature, with only English-language sources. Especially, try doing online research on these topics with only English-language sources.

Right now we say "English-language sources should be used whenever possible…" I would like to change that to something like "There is a strong preference for including English-language sources…". The fact that it is possible to find English-language sources on these topics (in a university library in the English-speaking world, for example) should not be an argument against using readily available sources in the relevant languages. A Romanian in Bucharest writing on Romanian historical figures should not be told that he or she must hunt down an obscure book in a library 2000 kilometres away rather than using equally good materials readily at hand. To be honest, even I, living in Seattle, shouldn't be told that: for many topics, my personal library contains useful books in other languages, the public library is useless, and nothing else is readily available to me unless it happens to be in the one university library I have access to.

It's great if someone wants to add English-language references to such articles, but it is ridiculous to say that, because they could, in principle, be found, other equally reliable and more readily available sources are not acceptable. - Jmabel | Talk 16:25, 11 October 2006 (UTC)

The problem is that it becomes really difficult to verify non-English sources. If someone claims that a seminal work on a particular subject has been written in Basque, and you find the claims made extremely dubious, what can you do? Jayjg (talk) 16:41, 11 October 2006 (UTC)
There is generally no problem in verifying material written in languages that are widespread among Wikipedia editors, such as French, German, and Dutch. Andries 17:04, 11 October 2006 (UTC)
Or Romanian. At least five administrators read it well (four natively) and are about as politically diverse etc. as anyone could ever ask for. - Jmabel | Talk 17:29, 11 October 2006 (UTC)
Yes, there can be a problem: differences in translation can be a source of heated POV discussions. If English sources are available, since this is the English wiki, they should take precedence. I agree with the policy of only using non-English language sources when nothing is available in English, and emphasizing the importance of trying to locate a reliable source in the language of the Wiki. Sandy 17:39, 11 October 2006 (UTC)
Y'know: on most topics this is no problem at all. There are perhaps 2% of pages where things are really contentious, and where people start worrying about the subtle connotations of sources. But that happens just as readily with English-language sources. And, frankly, where the primary materials are in another language and the area is controversial, it is as likely as not that a professional translator has an axe to grind, too. - Jmabel | Talk 18:09, 11 October 2006 (UTC)
Good point. SlimVirgin (talk) 04:56, 12 October 2006 (UTC)

do not delete ... "historical" it

I think that deleting / redirecting WP:V would be a bad idea. If it does become obsolete to this new one, it should be tagged with a "historical" tag to show how the policy progressed. - Che Nuevara 16:29, 11 October 2006 (UTC)

Concur strongly. - Jmabel | Talk 18:10, 11 October 2006 (UTC)
Note that if it is redirected (which I would prefer), you can still access its history by clicking on the notification you were redirected, and then choosing 'history'. JulesH 19:15, 11 October 2006 (UTC)
It needs to become historical, not deleted. Lincher 21:24, 11 October 2006 (UTC)

Why the extra rules for deceased people?

I thought the extra strict rules were only meant for living people, but this article writes that the same standards apply for deceased people. This is new and incomprehensible for me. Andries 16:35, 11 October 2006 (UTC)

That's a good point. If we are going to make a change here, I think it would be good to apply the rules to people who died recently, if false disparaging information is likely to harm living people or organizations that the dead person was closely associated with. --Gerry Ashton 17:16, 11 October 2006 (UTC)
AFAIK, the only reason the rule exists is to prevent legal issues. This ceases to be an issue the moment the subject dies. JulesH 19:19, 11 October 2006 (UTC)
That is not true. Jimbo has repeatedly stressed that it is also "the right thing to do." Put another way, we're not just concerned about libel because we might get sued. We're concerned about libel because saying untrue things that are hurtful is wrong. Phil Sandifer 19:29, 11 October 2006 (UTC)
Following the same way of reasoning, saying untrue things that are hurtful would not be wrong only for living or deceased people, but also about organizations, religions, countries, soccer clubs, companies, etc. To summarize for every subject with with some people (partially) identify with. Andries 22:18, 13 October 2006 (UTC)

Interesting

I found this highly interesting in a particular context: If you are published outside of Wikipedia, you may cite your own publications just as you'd cite anyone else's work. Such citations must fulfill all requirements of this policy; make sure your material is relevant and that you're regarded as a reliable source for the purposes of Wikipedia. And, the context is that a number of our highly productive and prolific editors have chosen to remain shrouded inside a thick veil of anonymity! I do not expect that they shall disclose their identities as they have every right to protect their identities. However, such editors can always cite their publications if otherwise in conformity with the "redefined" policies and enjoy the publication of their work on Wikipedia in absentia. I do not presume that it is happening currently! I was just talking of a possibility with due respect to all the mighty editors of Wikipedia. However, I do not find anything unusual with this point of the policy. --Bhadani 18:35, 11 October 2006 (UTC)

Thanks EngineerScotty. I will carefully read the page very shortly. --Bhadani 18:56, 11 October 2006 (UTC)
I notice the ArbCom case contains the following principle: "Editors should avoid contributing to articles about themselves or subjects in which they are personally involved, as it is difficult to maintain NPOV while doing so." That is so outrageous. Lets make sure we don't let anyone who actually understand a subject contribute to an article on the subject. --Gerry Ashton 19:12, 11 October 2006 (UTC)
I think you've got the wrong end of the stick. People "who actually understand a subject contribute to an article on the subject" aren't necessarily equal to people "contributing to articles about themselves or subjects in which they are personally involved". There's a difference. Hiding Talk 19:27, 11 October 2006 (UTC)
The problem with the "principle" from the ArbCom is it says it is difficult to maintain NPOV while contributing to articles about themselves (very true), or subjects in which they are personally involved (often true), so these experts should just not contribute to these article. It isn't good enough to be extra careful, or to confine oneself to factual issues and avoid matters of opinion; it isn't good enough to confining oneself to mentioning material published in reliable sources, these experts should just get the hell out, we don't want them around here. --Gerry Ashton 19:36, 11 October 2006 (UTC)
See, I don't see that interpretation. I see it as saying that if you can do it from a NPOV feel free to do so, but be advised that we generally ask people to avoid for these reasons. I don't see the part where we tell people to get the hell out. Perhaps you didn't quote that bit? Hiding Talk 20:43, 11 October 2006 (UTC)

What's the improvement?

Olin, the two pages and the two ideas (material that can't be sourced (OR) versus material that simply isn't sourced) have caused confusion; the aim is to merge and streamline, and thereby hopefully avoid that confusion. SlimVirgin (talk) 23:52, 11 October 2006 (UTC)
  • What Slim said, and it's a good idea to reduce the amount of policy pages to reduce confusion. >Radiant< 16:13, 12 October 2006 (UTC)
    • I respect the sentiment; the only fear I have is that by cutting out the policies too much, you might create more ambiguities than already exist. If you take out examples and clarifying remarks in the name of reductionism, you might create confusion by making the policy interpretable.Olin 20:35, 12 October 2006 (UTC)

Primary sources and articles in scientific journals

I've been wondering for some time whether an article in a scientific journal (a typical article, not a review article) is a "primary source" in the Wikipedia meaning of that term. I think that, according to the standard definitions in science, articles are primary sources (see e.g. [1] and [2]). However, in my experience articles are used extensively in current practice at Wikipedia, contravening "Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge" if articles are indeed primary sources. -- Jitse Niesen (talk) 07:23, 11 October 2006 (UTC)

I guess science is out then, except for those popular media commentaries. ;)
Or not... Science publications may be primary sources, however, the synthesis that needs to happen to make up the publication means that if you are commenting on the underlying issue, the science publication acts like a secondary source, which is possibly more appropriate under this proposal. Ansell 07:28, 11 October 2006 (UTC)
Ansell wrote "the synthesis that needs to happen to make up the publication means that if you are commenting on the underlying issue, the science publication acts like a secondary source." That may happen in some cases, but many scientific journals consist of a collection of unrelated articles with no attempt at synthesis.
The problem I have with "Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge" is the phrase "specialist knowledge". Suppose I spot something in an existing article that does not seem right. I try to trace the questionable material back to the source, and when I do, I find that (1) the original source is a primary source, and (2) the WP article does indeed need correction. So far, no problem. But what if the WP article is written using different units of measure, and I have to convert the units in the primary source into the units in the WP article? Is that specialist knowledge? Where do we draw the line?
In the current policies and guidelines, the part about specialist knowledge only appears in a passage about articles based predominantly on primary sources. By expanding the scope in this way, it makes the use of primary sources for narrow fact-checking of articles that are mostly based on secondary sources a bit more difficult. --Gerry Ashton 07:46, 11 October 2006 (UTC)
I was trying to get at the concept of synthesis of ideas within scientific articles. I realise that any two articles in a journal are not likely to be similar to each other. Ansell 12:15, 11 October 2006 (UTC)
IMHO when and where to rely on (non-review) articles depends on the topic. In articles about established and broad topics like entropy, using or referring to a single journal article would be the exception (the history sections will refer to important first publications like Caratheodory, C., Math. Ann. 67, 355 (1909)).
The more exotic the topic, the fewer the textbooks and review articles, which would be preferred as sources, so Discrete Lorentzian quantum gravity will definitively be based on journal articles.
Pjacobi 14:05, 11 October 2006 (UTC)
IMO, Textbooks are lower than Peer Reviewed journal articles in terms of reliability. Textbooks generally have editors who are general purpose employees of publishing companies, whereas, the editors of Peer Reviewed Journals are intimiately involved with the subject area. This often results in Textbooks having false statements in my experience. In my current research, I have not used a single "textbook" on the area, preferring to stick to articles, as the textbook commentaries are not useful to me. I do not see Wikipedia's role as being a parrot for textbook companies, who have a distinct profit motive for their work. As you say though, it depends on the maturity of the area in question in the end. Ansell 00:27, 12 October 2006 (UTC)

Primary sources

I've added this to the section about secondary sources being preferred in general: "Articles about pop culture and works of fiction may sometimes rely on editors' interpretations of primary source material, because of the lack of secondary sources. In these cases, editors should use good editorial judgment and common sense."

I did this because editors who write about pop culture often complain that the current versions of V and NOR (and RS) would prevent them from writing their articles if they stuck to them. This is a point Phil Sandifer has made many times, pointing to the featured article Spoo as an example of an article that relies on primary sources. Therefore, it seems reasonable to add an exception for pop culture. SlimVirgin (talk) 08:59, 11 October 2006 (UTC)

I think that could be very helpful, yes. JulesH 09:45, 11 October 2006 (UTC)
I think the cautioning needs to be much stronger. Frankly, many of the pop culture articles should be removed under V, NOR and RS. I've seen way too many pop culture articles that are original research and deserve deletion. This particular discussion, I believe, should be separated from the discussion on merging the policy documents.
This also contradicts WP:NOT (which is echoed at WP:FICT) "Plot summaries. Wikipedia articles on works of fiction should contain real-world context and sourced analysis, offering detail on a work's achievements, impact or historical significance, not solely a summary of that work's plot. A plot summary may be appropriate as an aspect of a larger article, or as part of a series of articles per Wikipedia:Article series." The key is that the plot summary needs to be sourced, and we should not create an exception allowing sourcing only to the original work of fiction.
This also contradicts Wikipedia:Manual of Style (writing about fiction)#Sources for articles on fiction, which never suggests original research by a wikipedia editor as a source. GRBerry 16:02, 11 October 2006 (UTC)
  • For better or worse, stripping Wikipedia of articles on pop culture is not likely to happen any time soon. One could argue that a book of fiction is a reliable source as to what happens in that fictional world. Is it original research to state that Thorin was buried with the Arkenstone? >Radiant< 16:30, 11 October 2006 (UTC)
  • I think if there's analysis that's not in the original source material, then that would be original research (plot summaries, though heavily discouraged by themselves, fall under this). I do have another question though. Popular works of fiction like Star Wars have sourcebooks written about them. So does Anime like Dragon Ball and Naruto, but the difference is that many anime sourcebooks will be written by the creator, but Star Wars, where there's the creator (Lucas) and other people who have been licensed to write extra material. Are these all the same in terms of being "primary"? To me it is because these tend to be either written by the creator or paid by the creator to write. (If we're going to open this can of worms, might as well let them all out.) ColourBurst 06:20, 12 October 2006 (UTC)
    • I'd say you are correct in that most factbooks and the like about any fictional setting are written by the author of that setting, or someone in his pay. Actual analytical works of books of fiction exist but are rare; Tolkien is one of the few fantasy authors I can think of that has third-party books about his work. However, Wikipedia has zillions of articles on fiction. We could then (1) make an exception in this policy with respect to fiction; or (2) accept the fact that this policy doesn't apply to all articles; or (3) embark on a massive deletion of material related to fiction. I'm sure we're agreed that (3) is not such a good idea and will get a lot of editors angry? >Radiant< 16:04, 12 October 2006 (UTC)

Material for which no reliable published source could be found

Firstly, I very much welcome this initiative in re-framing policy.

It has been widely discussed, and, I think, agreed, that some types of material may be added for which no already-published source may be available. A good example is converting a length previously published in yards so it can also be reported in metres. Conversion factors have obviously "already been published", but not the specific quantities involved.

I would like it to be explicit that this is not only permitted, but encouraged.

Another example that was contentious at the time. I wished to report that Western Avenue in London is 11 to 12 miles long. Nobody actually believes this statement is false but I have never been able to find a source reporting it. Reference to a map was rejected on grounds that OR was involved in computing the length. Reference to Multimap travel directions from Old Oak Common Lane Acton W3 to Denham Roundabout Denham UB9 was rejected by some on grounds of OR and by others on grounds of verifiabliity (that the length had not already been published). The article Western Avenue has never been challenged over this but many editors disliked Longest streets in London and wanted to avoid its continued existence through references to systems such as MultiMap.

Under these sort of circumstances I have found comforting the statement in WP:OR "All articles in Wikipedia should be based on information collected from published primary and secondary sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia". Should this be resurrected?

I think the policy should say it is disruptive to remove material on grounds of lack of source unless one believes the material to be false. Thincat 12:44, 11 October 2006 (UTC)

  • The idea that reference to the map constitutes original research is wrong, unless you drew the map. I'm staggered that that argument won the debate. How do people think maps are created, and for what purpose? Do people not realise what the scale on a map does? People can go away and verify it and state you got your measurement wrong and amend the info, but that's the same as text which is sourced, people can misrepresent those. We really need to tweak our policies if people are pushing those sorts of fallacies. Hiding Talk 13:08, 11 October 2006 (UTC)
We can't cover every evantuality in a written policy, without it becoming ridiculously involved, or so vauge that even POV pusher or tinfoilhat merchant can drive a coach and horses through it. At the margins we need to simply use common sense. Maps look like one such case. --Doc 13:20, 11 October 2006 (UTC)
  • Oh, I agree entirely. I like the page and haven't commented because I like it doesn't really help. But it's been a long time coming if you ask me. But the idea that looking at a map constitutes original research is just... it just doesn't sit right. I'm not convinced that an article on the Longest streets in London has merit though. Hiding Talk 13:34, 11 October 2006 (UTC)

Ed, I reverted your edits because I feel they introduced confusing tangents. The aim of this page is to keep things as clean and streamlined as possible. SlimVirgin (talk) 13:22, 11 October 2006 (UTC)

The immediately previous comment by SlimVirgin seem to relate to the "Scientific controversies" section below, rather than to the current section. Thincat 14:23, 11 October 2006 (UTC)

WP:NOR

While I accept that RS needs to be removed and V made simpler and stronger, as discussed in the EN mailining list, I would opposse any attempt to dimminish the policy of WP:NOR by merging it. ≈ jossi ≈ t@ 14:48, 11 October 2006 (UTC)

I'd oppose dimminishing it strongly too. But your statement contains a non-sequitur.--Doc 15:04, 11 October 2006 (UTC)
Concur with Doc. I can't see how this diminishes it. Hiding Talk 15:08, 11 October 2006 (UTC)
Indeed -- replacing NOR is a major point to the proposal -- if you object to that it would probably be best to discuss it here first rather than evicerate the proposal. olderwiser 15:16, 11 October 2006 (UTC)

The lenghty discussion in Wiki-EN did not incude replacing NOR. As I said, I am open to be convinced that our work here will not be a waste if time. I really doubt at this point that we can change or replace long standing policies. Tens of thousands of articles and discussions refer to these policies and guidelines. The momentum is way to string for such overraching change. Let's attempt to demove the need for RS by tightening V, first. If we are successful, we could take the next step. As for my "evisceration" of this proposal. Feel free to revert. My point stand, though. ≈ jossi ≈ t@ 15:26, 11 October 2006 (UTC)

The fact that NOR is embedded is a bad argument. NOR could simply be redirected here - as it is in essence contained here. You talk of 'momentum' - but in truth this seems to simply be an argument against change because change is bad and difficult. What is your substantive objection? I'm unclear.--Doc 15:30, 11 October 2006 (UTC)
Doc, my issue is one of process. I simply do not believe that we can change long-standing policies (3 years!) with one wave of the hand. Thousands of editors have applied it and hundreds of thousands of articles have been developed on the basis o
Here's where I come from on this: I would like to see all Wikipedia policies reduced to a very small statement of the policy. What I see here is the potential to reduce the formally agreed canonical policy back to its basis: Wikipedia is not a publisher of original thought. With enough work we might be able to get all the canonical policies onto a single page. That would be fantastic, because it's really really important to have as few absolute rules as possible, and for them to be as unambiguous as they can be. If we then replace all the clarifications and hedging on V and NOR with this, and make this a guideline - a descriptive guide to how the policies are appplied in practice - then we have the opportunity to replace RS. We can also chip away at EL, VAIN, conflict of interest and others. Guidelines grow to prevent "that stupid thing Bob did three years ago" and should be ruthlessly pruned periodically, but if we can get the formal policies right down then it will be much easier to argue from policy not from guidelines, because in the end that is what we should be doing. Foo is notable is subjective; Foo has (or has not) generated significant verifiable coverage in independent sources of known and acceptable authority to allow a verifiably neutral article is far less likely to be a cause of dispute.
So: I see a positive benefit to retaining the policies, stated in plain language and without ornamentation, whether singly or in a single article. That way the policies will remain above the hurly-burly of debate which inevitably affects guidelines, and the policies will become much more important and the guidelines much less, to the extent that the guidelines can start being drastically shortened, merged or quietly dropped. Policy -> Application guideline -> Subject specific examples. And preferably nothing below that level; although some user essays are extremely useful in some circumstances (WP:NFT for example), who can find these things without being pointed to them? Guy 15:37, 11 October 2006 (UTC)
(replying to comments some ways up after edit conflict) From the the very first post that SlimVirgin made regarding this proposal to the EN list, it was clear that the proposal involved merging NOR and V. As I understood it, that is one of the main points. If it is to merely be a re-write of WP:V, then why bother with renaming it? ASFAIC, one of the main weaknesses of NOR is the discrepency between the idiosyncratic usage within Wikipedia and the more common understanding of original research. Seems there was significant volumes of tedious discussions attempting to address this. I think this is a bold new direction for addressing that (and other) shortcomings in the policies. I'm not entirely sure what I think about it overall just yet, but I do think the proposal should be given a chance. olderwiser 15:43, 11 October 2006 (UTC)
One of the main purposes of this proposed policy is to replace NOR and V with a unitary policy. Jayjg (talk) 15:55, 11 October 2006 (UTC)
Yes, the key idea is to combine V and NOR into one policy. SlimVirgin (talk) 23:32, 11 October 2006 (UTC)

What is not original research

I have added a section on "What is not original research", drawing on Wikipedia:These are not Original Research. I realize this may be slightly controversial, but I believe we need something like this, because the omission of this in the previous policy statement has been the source of some of the most ridiculous wikilawyering I've ever seen. - Jmabel | Talk 16:15, 11 October 2006 (UTC)

Without, as far as I can tell, engaging in discussion here on the talk page, User:Marskell removed this, commenting only "rm, for now. there are no good examples here". Seems to me that these are all highly relevant examples: each illustrates a point over which I've seen major fights, that is why I added the section.
To avoid an edit war, I am adding the material here on the talk page; comment would be appreciated as to whether this (a) belongs in the policy (b) belongs in a guideline or (c) does not belong at all. - Jmabel | Talk 16:01, 12 October 2006 (UTC)

[begin cut passage]

What is not original research

Conversely, the prohibition on original research should not be carried to absurd extremes. If a published source gives the numbers of votes cast in an election, it is not original research to calculate percentages. Similarly, if a source says that someone was born in 1954 in Baltimore, Maryland, it is not original research to extrapolate that she was born in the United States.

If there is an obvious typographical error in a source, it is not original research to take the sentence for its obvious meaning (clearly, an author is not claiming that England had a king named "Hen3ry VII"). Similarly, if a source article refers to "the painter Dalí", it is not original research to presume that they mean Salvador Dalí, unless the context suggests otherwise.

Jmabel, I think these are useful points to make; we often have editors on the NOR talk page asking questions like that. I'd say that the first paragraph would be enough though. SlimVirgin (talk) 18:33, 12 October 2006 (UTC)
I have restored that paragraph. - Jmabel | Talk 21:35, 12 October 2006 (UTC)
I notice that Joe did not include the third case from this essay -- where a cited authority is obviously contradicted by the facts. Am I the only one who does not think it falls under the ban on original research? -- llywrch 22:31, 12 October 2006 (UTC)
I would agree. I suspect others might not (I know that Slim has argued for an interpretation of NOR that would not allow us to point out possible mistakes in a source unless we can find a reliable source that does the same.) There certainly are any number of articles on wikipedia that do this without coming under serious scrutiny - articles on historical errors in films, for instance, tend to merely cite sources about the real history, not sources that point out the particular error in the particular film. Of the others, I think that "obvious deductions" needs to be laid out clearly, in particular, as some of the stronger presentations of the NOR policy have tended to suggest, at least, that obvious deductions are forbidden as OR. john k 17:49, 13 October 2006 (UTC)
There is a controversy as to whether Alexander Hamilton was born in 1755 or 1757. This clearly applies to that issue. But do we really need to wait for a RS to warrant our ignoring the biography which says, as one does, that he was born in 1957? Septentrionalis 19:54, 13 October 2006 (UTC)

There should be some sort of advice to be careful, however; for example, published election results often omit third parties or spoiled ballots without notice. Septentrionalis 19:37, 13 October 2006 (UTC)

"Could be attributed" + NOR

"Not all material must actually be attributed to a source, but it must be capable of being attributed." Why should this be emphasized? "Ya I could source, but it's so obvious I don't need to" is a problem—this will only encourage it.

I'm also uneasy about merging NOR. It's the hardest content policy to grasp IMO, and deserves a distinct page. Marskell 16:55, 11 October 2006 (UTC)

If you tell potential editors that they must source every statement, they are apt to think that they must spend all there time writing citations for thinks like "Bill Clinton was a U.S. president" and decide they have better things to do than write Wikipedia articles. Alternatively, they will observe that hardly any articles provide sources for obvious statements, so will consider this policy to be one of those rules that exists but no one obeys, and just ignore the entire policy. --Gerry Ashton 17:21, 11 October 2006 (UTC)
My suggestion is that we tell them nothing with regard to what is obvious. Commonsensically, we do not need to source "Bill Clinton was a U.S. president" every time it's written, but if this were to be one of two content policies I'd hate to see that emphasized off the top. Marskell 17:36, 11 October 2006 (UTC)
Agree. I have removed that statement. ≈ jossi ≈ t@ 18:00, 11 October 2006 (UTC)
An interesting question, although perhaps it shouldn't be codified here, would be whether or not "Bill Clinton was a U.S. president" is well sourced or not. There are plenty of sources on Clinton's article that confirm it, and we've just linked directly there. JulesH 19:22, 11 October 2006 (UTC)
Regarding the sentence: "Not all material must actually be attributed to a source, but it must be capable of being attributed" — this is an important sentence. It goes right to the heart of the issue. Material that cannot be attributed is original research. Just because something doesn't have a source doesn't mean that it's OR i.e. that no source could be found for it. And just because something is in Wikipedia doesn't mean it necessarily has to be sourced. The sentence you're objecting to sums that all up: material must be attributable, but need not actually be attributed. And it then goes on to say when we do need a source: for all material that is challenged or likely to be challenged. SlimVirgin (talk) 23:40, 11 October 2006 (UTC)
This is an absolutely vital distinction. Phil Sandifer 02:11, 12 October 2006 (UTC)
I agree with Slim, et al. There are way too many people who say something is "OR" simply because there is not an attribution. That is not what OR is, and that needs to be made clear. A citation only needs to be found if someone challenges a piece of information, and people should only challenge if they think something is inaccurate, or might be. john k 17:52, 13 October 2006 (UTC)

Acceptable sources

Being WP:BOLD, I added in material indicating the mininum threshold of acceptability for a source. Not that being acceptable doesn't imply reliability or authority; but the items listed paint a clear picture of what is obviously not appropriate. --EngineerScotty 16:56, 11 October 2006 (UTC)

Isn't this turning into another bloated policy page? Wasn't that one of the things it was trying to avoid? Jayjg (talk) 17:36, 11 October 2006 (UTC)
I agree. Less is more. Examples and details can be moved to separate guidelines/essays, etc. ≈ jossi ≈ t@ 17:38, 11 October 2006 (UTC)
Examples should be moved, agreed. The core of the policy, however, is not bloat. --EngineerScotty
How about creation of Wikipedia:Attribution/Examples, which would be a place for examples to illustrate the policy? Unlike this page, which would ideally have policy status; the example page might be an essay or guideline at best; and be non-normative--in other words, a pedagogical aid to illustrate the policy, but not binding. --EngineerScotty 18:06, 11 October 2006 (UTC)
That's a good idea, although some examples on the page are good too, for basic illustrative purposes. The one problem with another page, of course, is that you have to constantly watch that it doesn't end up contradicting the policy. But so long as we made clear that there had to be consistency, it would be fine. SlimVirgin (talk) 23:43, 11 October 2006 (UTC)
The section on pop culture and fiction seems to change the intent of WP:V, see the current Wikipedia:Articles for deletion/List of Dragon Ball special abilities, where the article would be deleted in its current state (without #4 of the self-published section), but not so if Attribution passes. I see a whole lot of DRVs in the future if it does. (oops, I forgot to sign. Addendum ColourBurst 05:58, 12 October 2006 (UTC))

Policy vs. guideline

This is getting bloated. We need to separarate policy vs. guideline.

Perhaps the correct approach to this all is:

  • This becomes a relatively terse policy page. Very few examples are provided and those are intended to be illustrative rather than exhaustive.
  • WP:V and WP:POV (or equivalents) continue to exist, but are demoted to guidelines illustrating how to conform with the policy.

- Jmabel | Talk 18:13, 11 October 2006 (UTC)

I'd like to try to have it all on one page but without the bloat. It may not be possible, but the problem with having (a) the policy page and then (b) a page on how to comply with the policy is that you end up with contradictions. This has been part of the problem with NOR, V, and RS: trying to keep them meaningful and consistent. It was heavy maintenance and not too successful. SlimVirgin (talk) 23:45, 11 October 2006 (UTC)

How about the possibility of rules here, with no more than one example each, but a guideline page (or pages) elsewhere that gives numerous illustrations what the rule does and does not mean? - Jmabel | Talk 04:55, 12 October 2006 (UTC)

That sounds good. One good example illustrating the point, where necessary, but otherwise examples go on an FAQ page or similar. SlimVirgin (talk) 04:58, 12 October 2006 (UTC)

Primary vs secondary sources

As mentioned in Wikipedia:Reliable sources/Flaws, there is a lot of confusion on this topic; I think this policy would do well to avoid the "primary/secondary/tertiary" taxonomy where possible--instead focusing on other issues like publication, editorial oversight, etc.

In particular, many in science use "primary source" to refer to research papers. WP:RS largely uses the definition from historiography, wherein primary sources are artifacts such as public records, interviews with persons of interest, physical objects of import, and other things of a non-descriptive, non-synthetic, non-analytical nature. Papers, books, and other forms of scholarship, to a historian, are always secondary sources. Were science to adapt history's terminology, primary sources would be things like fossil records, laboratory notebooks, the output (and source code) of computer simulations, samples of chemicals, samples of biological tissue, biological specimens, medical records, etc.

Some primary sources (under the definition of historiography) are published--court records and other records of official acts, published literature (including textual works, recorded music, film). Many of these lack analyses, though some (decisions of a court) contain an analytical component.

IMHO, the properties of a source (analytical or not; published or not; is it an idea or is it a thing) are more relevant to the distinction between "primary" and "secondary"; we ought to focus on properties and not the terminology used in historiography (and abused in science).  :)

--EngineerScotty 18:24, 11 October 2006 (UTC)

I'm increasingly thinking the primary/secondary distinction isn't helpful, but I know that other editors rely on it. It is useful to be able to say to a newbie that they can't base an article directly on their view of the New Testament because it's a primary source. It cuts through all the explanations about not using source material properly, which is the main point, because secondary sources can be misused too. But it requires a lot more explanation to get that across without mentioning the primary/secondary distinction. It's also useful to talk about the need for secondary sources to establish that something is notable enough to be written about. SlimVirgin (talk) 23:49, 11 October 2006 (UTC)
Any source can be misused if the editor is not careful to avoid "editorialising". I think that the distinction is not accurately portrayed between history and science "primary sources". In reality, the primary sources in science are the actual data that is collected during experiments. Articles on the topic will contain summarised statistics based on this data, but they are also required to contain analysis of the data, which is IMO, a second "layer" of knowledge. A "tertiary source" in this sense will be the heavily summarised version of these articles which are included in textbooks. Whether the restriction in content that the textbook publisher puts on the final version affects the "accuracy" of what can be said about the original articles is past what Wikipedia should be using or not, well, I don't know about that. Ansell 00:33, 12 October 2006 (UTC)
I agree that any kind of source can be misused, and ideally I'd like not to have to even mention the primary/secondary/tertiary distinction. But it can be useful; for example, an editor arrives at Animal rights and adds that the Sikh religion supports animal rights, using the Bhagavad Gita as his source. The easiest way to explain the problem with that kind of editing is to say it's his own interpretation of primary source material, and that he needs to produce a secondary source that supports what he wants to add. How else would you get that across, succinctly, without relying on the primary/secondary distinction? SlimVirgin (talk) 05:06, 12 October 2006 (UTC)
Perhaps say "I note that you have used Bhagavad Gita as your source for saying that the Sikh religion supports animal rights. Whether this source supports that conclusion or not is debateable. Wikipedia requires sourcing that directly supports what you say and not indirectly. If you supplied a reliable source that itself drew that conclusion through evidence such as Bhagavad Gita, then we can add in the article that source X says the Sikh religion supports animal rights based on Bhagavad Gita." WAS 4.250 16:31, 12 October 2006 (UTC)
Yes, you're right; it can be done without reference to primary/secondary. SlimVirgin (talk) 06:36, 14 October 2006 (UTC)

original research section

I've removed this, my understanding is that this will be what the new WP:OR page will look like. Hiding Talk 19:18, 11 October 2006 (UTC)

I'm not sure where that understanding comes from, but it's explicitly stated at the top of this discussion that this page is intended to replace WP:OR. In fact, I think this probably is the new WP:OR page you're referring to. JulesH 19:30, 11 October 2006 (UTC)
I think at the top it says that examples of original research will be on the original research page. That's why I removed it. But I might have it wrong. It did also make the page shorter, which was nice. Hiding Talk 20:45, 11 October 2006 (UTC)
The intention of the proposal is that this will be the main NOR policy page, Hiding; we might want to retain the other page for examples, but we might not. Any essential examples for the purposes of illuminating the policy should be on this page. SlimVirgin (talk) 23:53, 11 October 2006 (UTC)

Websites

I think this proposal will set off a huge firestorm if it is enacted and then fully acted upon, but I'm willing to humor it for the nonce.

What with the emphasis on publication and reliability of sources, it seems to me that this proposal will outlaw the use of a wide variety of websites. Specifically, it seems to me that it will exclude all but the following:

  • Official websites as a primary source about the sponsoring entity
  • Archives on official sites of publications which would be acceptable as print sources

It would specifically appear to outlaw use of websites as secondary sources for documents, because one cannot be sure that the website is presenting an accurate copy thereof. (I ran across a case of this just today.) Perhaps more problematically it will outlaw the kind of reference that is common in articles about "things" (toaster, apple, etc.): websites that do not relate vetted research, but might by a reasonable person be considered reliable. There are plenty of articles of this kind in which a strict application of the proposal would remove all but the trivia, because only the trivia is readily amenable to such documentation.

Only the second bullet is addressed in the document as a legitimate source. -- Mangoe 19:47, 11 October 2006 (UTC)

The current wording actually makes a very clear distinction on these issues and fully addresses your both points. Official websites are OK as primary sources on articles about such sponsoring organizations, and archives as well. Please re-read. ≈ jossi ≈ t@ 19:57, 11 October 2006 (UTC)
Modulo official websites, it is clear as to what it includes. I deduce that it excludes this other sort of site I described, but I suspect that a lot of editors-- maybe the overwhelming majority-- might not catch onto the implied exclusion. Or worse, there would be radically conflicting views as to what was excluded, because levels of mistrust do vary wildly. It seems to me that the kind of sites it specifically excludes are obviously excluded by a reasonable person anyway; but it also seems to me that perhaps a majority of articles rely on some grey-area source which implied to be unacceptable but not so stated. Mangoe 20:34, 11 October 2006 (UTC)
I would be interested to know which passage(s) in the proposed policy outlaw the websites Mangoe is concerned about. The only issue I could come up with is the paragraph on "Traceability to author(s)". If one interpreted insisted on traceability to individual human authors, it would be a problem, but there is no problem if you also accept corporate or organizational authors. --Gerry Ashton 21:26, 11 October 2006 (UTC)
A quick check shows that a lot of the sections under what is now "reliable sources" have changed sufficiently to probably ameliorate the issue. Mangoe 01:04, 12 October 2006 (UTC)

Some issues with this page

Some of these are minor things I could probably fix myself, but I'd rather get input on the changes first.

  • We should add a parenthetical "not necessarily in print" after the first use of the word "published"
  • Let's call it popular culture instead of pop culture.
  • "this policy should never be used to cause disruption by removing material for which reliable sources could reasonably be found" - Change to "never be used to remove material for which reliable sources could reasonably be found." Nothing should be used to cause disruption, so that's redundant. The point is that this shouldn't be used for POV warring. (Which could become a problem, on articles from Scientology to Evolution very, very quickly.)
  • I'd like to keep the reference to disruption. What we're basically saying is "Don't be a dick," but more politely. I can imagine people removing material for which sources could be found but doing it non-disruptively and sensibly; it's people making nuisances of themselves that that sentence was aimed at. SlimVirgin (talk) 00:03, 12 October 2006 (UTC)
  • I'm a little wary of "unpublished facts" as a phrase. It seems to me to go against the ways in which we can appropriately use primary sources. It's not clear to me that, for instance, a statement like "Kant says 'X'" is a published fact. Obviously we don't want all our quotes from Kant to go to secondary sources, which this ends up implying we do. Perhaps "data, experimental results, and accounts" would cover the things we want a little more carefully?
The problem happens if X is not a word-for-word quote of Kant, but rather is a summary of his position. You'd need a secondary source to actually state that Kant's words were equivalent to X. Ken Arromdee 20:19, 12 October 2006 (UTC)
  • Self-publication about one's self needs to be broadened a bit to include "or about something one was heavily involved in." Going back to my usual example with such things, Spoo is not an article about J. Michael Straczynski, but his self-published material is its main source. It is an article about Babylon 5, which is a show he created, wrote most of, produced, and is the sole determinant of what is and is not canon for. This is neither quite a professional in his area of expertise nor directly about him. Really, what we want is a master category of "things the subject is unquestionably qualified to talk about."
Obviously this specific example comes under the heading of "pop culture and fiction", so the sources described there are perfectly adequate under the policy. But I get what you're talking about... it's the same as the second point I raised in Wikipedia_talk:Attribution#Self-published_sources, only I didn't phrase it quite so well. JulesH 20:34, 11 October 2006 (UTC)
  • The pop culture and fiction section is, I think, a bit off-target because the anti-blog, message board, Usenet, and Wiki tone has largely been removed, making it an extraneous clarification. (All of those now seem to fall under self-publication, with the issues of authenticity dealing with identity verification). So there's nothing new in this section to deal with things like the fact that a lot of comics news comes off of creator's message boards, or the importance of Usenet to B5. (Which is good - that was only ever one example. Scientology articles also NEED Usenet as a primary source) There's still something to be said here so that we can deal with something like, say, The Lurker's Guide to Babylon 5 for B5, or Outpost Gallifrey for Doctor Who. Perhaps a note about how popular culture topics often have excellent resources that are self-published by fans, and that these sites are perfectly reasonable sources, and a note about deferring to the consensus of editors on those topics.
Seems reasonable to me. JulesH 20:34, 11 October 2006 (UTC)
  • A note about self-published sources as primary sources about the self-publication is probably important here. I'm thinking here of webcomics and blogs, which can both be taken as reliable sources about themselves. Also, again, Scientology, where the Usenet postings are valuable primary sources for the large amount of stuff that happened on Usenet.
This is a point I've debated before. Yes, I think such things should be possible. JulesH 20:34, 11 October 2006 (UTC)
  • How much of "What is not original research" violates WP:BEANS and how much of this has ever happened?
I think it's all happened. Just like people who go through articles and tag every sentence with {{fact}} if it doesn't have a footnote right there. JulesH 20:34, 11 October 2006 (UTC)
  • Drop the "Areas of Special Concern" header, and go straight to BLP. Its current format is an invitation for instruction creep of other areas of alleged special concern.

That's what I've got for now. Possibly more later. I'll probably put a number, though not necessarily all of these into the page over the next 24 hours, especially if nobody strongly objects to them. Phil Sandifer 19:55, 11 October 2006 (UTC)

As far as I understand it, this attempt does not increase the reliability of USENET, blogs or personal web pages. These should not be accepted as reliable sources for anything, with the exception of popular culture and non-controversial subjects. Let's not re-write well established principles. My understanding is that this is an attempt to merge several policies into one, and *not* an attempt to re-write policy. ≈ jossi ≈ t@ 20:02, 11 October 2006 (UTC)
It was, I believe, a merge and revision. Certainly there's nothing in there that inherently rules out blogs, Usenet, etc on the grounds that they are blogs, Usenet, etc. They're ruled out because the specific source fails a test within - it's not that they're Usenet posts, it's that they're not easily attributed, or because they're not a reputable scholar, etc. In other words, enshrine the reasons we were banning Usenet posts before, and then any exceptions to the rule get taken care of automatically. Phil Sandifer 20:24, 11 October 2006 (UTC)
Well, here we disagree. I believe that a blog, a USENET post, or a personal homepage are inherently unreliable sources, unless very specific qualifications/conditions apply. Subtle, but important difference with yoiur formulation. ≈ jossi ≈ t@ 20:28, 11 October 2006 (UTC)
Well, I guess I disagree too. Sources have to be evaluated in context. We can't create rules which say x is acceptable but y isn't, because they issue is transformed by context. Hiding Talk 20:47, 11 October 2006 (UTC)
This page will need to be debated to determine if it has concensus or not. It's hard to get consensus together for small changes to policies, because not many people are interested in such small changes, so it is difficult to overcome resistance. Now is therefore probably the best time to debate any such changes.
I don't see any objection to using a Usenet posting (etc.) as a primary source, as long as it is clear that it means what it is claimed to mean (e.g. if the identity of the poster is an issue, there may be concerns that it is a forgery). Obviously the usual caveats of prefering secondary sources apply.
I think the suggestion that such sources are more welcome on pop culture articles is very welcome. Specifically, it's hard to write about (say) Usenet culture without refering to Usenet sources. This should make the task easier. JulesH 20:34, 11 October 2006 (UTC)
All sources are unreliable, unless they meet one of the criteria in the policy. In the case of a blog/USENET post or a personal homepage, it would have to be one of the 4 exceptions in the "self published sources" category, because the only thing distinctive about all of these is that they are self-published. JulesH 20:36, 11 October 2006 (UTC)
yes, I can see that this is already addressed on the section "Self-published sources". ≈ jossi ≈ t@ 20:43, 11 October 2006 (UTC)
Right. To smash blogs separate from self-publishing seems needless. The flaw in blogs and Usenet are not special flaws. They're flaws shared by other self-published material, as are teh exceptions. Phil Sandifer 22:09, 11 October 2006 (UTC)

Bloat danger

It's good that lots of people are editing this, but please bear in mind that the aim is to streamline: to produce a policy with no fluff to replace the most important bits of V and NOR, but without inheriting their problems. It's gained 800 words since I last checked, and some editors were already saying it was too long. The key to getting people on board with this is to keep it as simple and light as possible, but without leaving out anything essential. SlimVirgin (talk) 21:25, 11 October 2006 (UTC)

yes... the pendulum is swinging in that direction. Hopefully this half period is over, and we can now reduce the text to the essentials. ≈ jossi ≈ t@ 22:01, 11 October 2006 (UTC)
Thesis, antithesis, synthesis? :-)
Sentences like this, for example: "Whilst particular caution must be exercised with websites, the websites of major media outlets, such as [blah, blah] and so on, are considered reliable. Indeed, referencing to articles on their websites is often useful, as they are much more accessible to the average Wikipedia reader than a hardcopy document such as a newspaper."
All that is self-evident and doesn't need to be spelled out, especially the second sentence.
I also agree with Phil that the thing that's distinctive about blogs, Usenet posts etc is that they're self-published. The problem is not that they're online or that it's Usenet per se, but the self-publication, so it's the latter we should concentrate on as being often (but not always) problematic. The occasions that it's okay to use self-published material apply equally to paper and online material. SlimVirgin (talk) 22:40, 11 October 2006 (UTC)

The "Self-published sources" section begins "A self-published source is a published source, online or on paper...." We already have an explaination of what publication means, and it's long. I suggest that elsewhere in the policy we only say "published source" or "publication", and avoid spelling out all the possible publication media. Otherwise we will have bloat and contradictions. (This would not apply to a section that addresses unique charactristics of on medium.) --Gerry Ashton 23:58, 11 October 2006 (UTC)

Starting with the content of WP:V was a bad idea, IMHO. It hops from topic to topic, it's not clear where the policy begins and ends, and every statement threatens to explode with further rambling explanations and peripheral info. I would rewrite it from scratch. Well, not scratch, exactly, but pretend the WP:V page doesn't exist at all. 1. Express your policy in precise terms, as a bulleted list of single sentences, with no explanations. This is the only part that is going to be normative (and it should expressly say so). If you find yourself writing "which means that" after any sentence, then start over. 2. Precede the list with an executive summary. I wouldn't try to restate all the points in the list; instead, make it say something more to the effect that the list exists, comprises the policy, and addresses topics toward a general goal. 3. Follow the list with an expansion of each item in the list: the item's motivations (e.g. "Wikipedia is / is not a ..."), examples and verbose explanations of every little nuance, as needed. —mjb 23:49, 11 October 2006 (UTC)

I like the idea of keeping bloat to an absolute minimum, but there were good sections in V and NOR that did work. I wouldn't want to see us ditch the tried-and-tested in the interests of a completely fresh start. SlimVirgin (talk) 00:07, 12 October 2006 (UTC)
Bloat is inevitable, I think, unless examples, quibble-fodder etc. are shifted to a species of FAQ subpage where sheep may safely graze. I'd like this proposed policy (or the existing policies) to be restricted to hard policy. And though sub-guidelines might become ragbags, no-one would want to read them whole anyway, so that wouldn't matter so much as it does on a policy page. The policies could point to the subpages ('see examples here', or whatever) but be constructed never to depend on them, so ensuring that the subs could never, as RS is at the moment, be mistaken for policy themselves. I'm serious: unless a decision is quickly taken to limit the detail level of this policy proposal page, it's doomed, in my opinion, to, as SV fears, inherit the windiness of the existing policies and guidelines. If we don't make this a slim version, so to speak, I'm out (for what then really would be the point?). qp10qp 00:11, 12 October 2006 (UTC)
I'm going to go in later and try to remove bloat and make sure essential examples are there, but no others. SlimVirgin (talk) 00:23, 12 October 2006 (UTC)
An FAQ page is good, but some key examples are needed on this page too. You say that subpages don't matter so much, but it was in part the bloat on RS that brought V into disrepute (not only that, but it contributed mightily). SlimVirgin (talk) 00:20, 12 October 2006 (UTC)
How do we start that FAQ page? Should we make a subpage such as Attribution/FAQ or Attribution/Examples or a different article alltogether? ≈ jossi ≈ t@ 00:22, 12 October 2006 (UTC)
I like Wikipedia:Attribution/FAQ. SlimVirgin (talk) 00:23, 12 October 2006 (UTC)
I'd certainly like to see a big hierarchical gap between a policy page and the next step down, because "guideline" was starting to sound too important a page type, as if it was "almost policy". The Wikipedia way, in my opinion, is that "ignore all rules" caps intricate what-is-a-published-source guidelining, which is why, for example, I advised Plange to go ahead and use original documents housed in official collections, if she so wished, rather than worrying too much about what it said on RS. Let people object at the editing level, if they want: we don't need mandatory rules at the microlevel, surely; but nuances can appear that way if they squat on a policy page or a jumped-up guideline page. qp10qp 00:36, 12 October 2006 (UTC)
Right. I wouldn't want to see this page plus associated guideline explaining it; then we end up with another V and RS situation. SlimVirgin (talk) 00:40, 12 October 2006 (UTC)

Acceptable vs. Reliable

I am not sure about the need for such distinction. Either we are asking editors to use reliable sources or we are not. Authority and trustworthiness need to be added to the other qualifications, and we nee drop that distinction calling sources: "Reliable sources." This wil eliminate also the whole caveat about "minimum standards". ≈ jossi ≈ t@ 22:11, 11 October 2006 (UTC)

Agreed. We need to use one word to describe the sources we regard as acceptable for use in Wikipedia. Over the two years I've been editing, we've used reputable, credible, acceptable, and reliable, and reliable always seemed to be the one that stuck with most people, so I suggest we go with it. SlimVirgin (talk) 22:50, 11 October 2006 (UTC)
Now that we have merged these two distinctions, we need to remove any circular references from the text. We also need to establish what a reputable publisher is, IMO. ≈ jossi ≈ t@ 00:00, 12 October 2006 (UTC)
Is a "reputable publisher" a concept which relies on the topic in question? And hence it would not be wise to define it in general? Ansell 00:36, 12 October 2006 (UTC)
I'm not sure we should make those distinctions. We have the idea of some document or some person or some publisher being a reliable enough source for Wikipedia. Do we need any further distinction (e.g. reputable publisher). SlimVirgin (talk) 00:42, 12 October 2006 (UTC)
Not sure we need a definition of a "reputable publisher" any longer. The criteria is getting quite tight as it is. That's good. ≈ jossi ≈ t@ 00:45, 12 October 2006 (UTC)
  • We might want to point out somewhere that a "credible" source on Wikipedia really is the same thing as a "reliable" source. It reduces confusion and keeps away the vexlits. >Radiant< 16:15, 12 October 2006 (UTC)

This is an unnecessary and harmful change

The two policies which this seeks to replace both have clearer names and are more readily understood. There is no need to try to reduce the number of basic policies to two. Calsicol 00:41, 12 October 2006 (UTC)

There's a conceptual overlap between V and NOR, Calsicol, so that they naturally suggest themselves for a merge. SlimVirgin (talk) 00:43, 12 October 2006 (UTC)
This concern/objection will resurface again and again, I am afraid. I guess that only when this page is in good shape and offers a real and better alternative to RS, V and NOR, we will be able to convince of its value. ≈ jossi ≈ t@ 00:47, 12 October 2006 (UTC)
Yes, I think that's right. It's going to have to be an obvious improvement before people will trust it. SlimVirgin (talk) 05:08, 12 October 2006 (UTC)

Take something like recodings of Danny Boy. That is a useful list, largely built up bit by bit by readers who have heard various people sing the song. Very few of the entries are attributed, and the suggestion that a secondary source is needed for each point, rather than what was probably largely a verifiable (but not verified) combination of "original research" and "primary sources", would destroy one of Wikipedia's strengths. --Audiovideo 00:59, 12 October 2006 (UTC)

Yes, indeed. But I think pointing out that sources are needed for material that is "challenged or likely to be challenged" gets round that; and also that people should not use requests for sources to cause disruption. SlimVirgin (talk) 05:10, 12 October 2006 (UTC)
It isn't original research by the definition here, which is that no source could ever be found. Perhaps we need something along the lines of "it's OK to add material without a source, but be prepared to back it up with a source if it's challenged." Or maybe that's something that doesn't need saying, because people will do that anyway. JulesH 06:58, 12 October 2006 (UTC)

Another key principle

Per some discussions we had at the September 2006 meetup in Seattle, I'd like to suggest adding one more "key principle", something along the lines of:

  • At times, building an encyclopedia this requires use of judgment about sources, not just application of formal rules. Formal rules are not a substitute for intellectual honesty. Formal rules can not generally determine whether an item of information is notable, or whether it is relevant to a particular topic.

- Jmabel | Talk 05:16, 12 October 2006 (UTC)

Looks good to me. BTW, I've added a bit about out-of-date sources; I tried to remove a bit of fluff as well. --EngineerScotty 05:24, 12 October 2006 (UTC)
Exactly. Our goal should be "representing significant views fairly and without bias." So where do we talk about the "judgment of sources" that will actually bring us to "representing significant views fairly and without bias"? While it is true that we would like ambitious newcomers to attribute what they write to reliable sources, where do we give guidance on the "judgment of sources"? --not in WP:NPOV and not in the formal rules of Wikipedia:Attribution--where? --Rednblu 06:21, 12 October 2006 (UTC)
WP:RS is probably the best place for anything in depth along those lines. JulesH 06:59, 12 October 2006 (UTC)

Judging sources for journal articles in science may include checking citation counts, including the often sobering test, whether the paper is only ever cited by his author. Also the citing papers should get a check, whether to actually refute the original paper. --Pjacobi 07:48, 12 October 2006 (UTC)

Separating the basic original research

No original research made it quite clear that the writer should not be making new theories, concepts, or personal analysis. Verifiability is about proper sourcing. They are tied, but then, they're not. The overlap is moreso in "It introduces an argument, without citing a reputable source for that argument, that purports to refute or support another idea, theory, argument, or position." I think that OR needs to be more streamlined insted.. It talks too much about primary and secondary sources that should be better covered in Verifiability. The section "Citing oneself" in OR I think is esspecially importantant, considering we ought help clear up the difference between original research, and someone adding stuff their own wealth of knoledge to wikipedia. Kevin_b_er 07:42, 12 October 2006 (UTC)

Well, let me put another one. I don't think the policy hammers home enough that its an encyclopedia, not for writing your personal opinions or attempting to establish a new viewpoint on something through wikipedia. This then ties to NPOV and Conflict of interest. (You notice many of the core policies tie together?) The 1st pillars expresses this well by stating "Wikipedia is not the place to insert your own opinions, experiences, or arguments." That is what needs to be rexpressed here. Kevin_b_er 08:02, 12 October 2006 (UTC)
Does that help? JulesH 08:32, 12 October 2006 (UTC)


There's a lot of stuff on WP:NOR that just isn't covered here. I'm not sure how much of it is important, so I'll summarise here and let people decide.

  • Description of primary & secondary sources.
  • "Research that creates primary sources is not allowed" and "Articles which draw predominantly on primary sources are generally discouraged, in favor of articles based predominantly on secondary sources."
  • "An article or section of an article that relies on primary source should (1) only make descriptive claims the accuracy of which is easily verifiable by any reasonable adult without specialist knowledge, and (2) make no analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on entirely primary sources should be exceptionally careful to comply with both conditions."
  • "An edit counts as original research if it proposes ideas or arguments. "
  • Use of neologisms without a source as OR
  • The 'Why OR is excluded' section
  • The example of synthesis of an argument (although we have a general description of what this would mean)
  • The 'explaining theories' section (which reads more like a 'howto' than a policy)
  • Original images (I think this is probably fairly important here)
  • I don't think we need to spell it out to that degree; although I do want to restore the synthesis of an argument example. But if you want to spell it out, we only have to restore the "What does this policy disallow" section, which someone removed. SlimVirgin (talk) 09:32, 12 October 2006 (UTC)

So what of this needs adding back in? JulesH 08:56, 12 October 2006 (UTC)

Reliable sources section

I'm not keen on this section. It's long; too detailed; and it mixes up key concepts. For example, it says anonymous sources aren't allowed unless they've been interviewed by someone trustworthy e.g. Deep Throat. But Deep Throat wouldn't have been Wikipedia's source; he was Bob Woodward's source, and Bob Woodward might have been ours via the Washington Post.

I think we should remove this section entirely. Any thoughts? SlimVirgin (talk) 09:29, 12 October 2006 (UTC)

I've gone ahead and removed it, because it seemed too problematic. SlimVirgin (talk) 10:23, 12 October 2006 (UTC)

I've added a section on popular culture to give it two explicit exceptions to the reliable sources rule: (1) that pop culture articles may rely on pseudonymous posts on bulletin boards and blogs; and (2) that in some areas, self-published non-professional secondary sources may be used, such as fansites.

Some articles about popular culture rely on self-published posts on bulletin boards, blogs, and Usenet, because few other sources exist for them. In such cases, the material used must have been posted by named, or well-known pseudonymous, individuals with a known expertise in the area, although the individual need not be a professional in a relevant field. Anonymous posts should never be used.

Some areas of popular culture also rely on self-published secondary sources, such as well-known fansites. If in doubt about how to use a source in this area, consult the relevant WikiProjects for advice.

This exception applies only to articles about popular culture.

This is a significant departure from the previous policies. The reason I added it is twofold: first, the policies are supposed to be descriptive as well as prescriptive, and it's a fact that editors working on pop culture do use these types of sources. The second reason is that a number of good editors have repeatedly criticized V and NOR as too restrictive for pop culture articles; and in fact someone recently wrote to the mailing list saying he had decided to ignore V and NOR because otherwise he'd get no work done. There's clearly a problem when good editors are saying they can't work within the policies.

At the same time, we don't want to open the floodgates to Usenet posts and bulletin boards. The wording of this section is therefore important, so please everyone take a look and say whether it needs to be tightened and made more specific. Editors working in that area should say whether it becomes too specific. SlimVirgin (talk) 10:23, 12 October 2006 (UTC)

My only concern is what the definition of "popular culture" would include. For instance, if I were writing an article on the history of a particular USENET group, would this section allow me to use posts from that group as sources? It certainly seems like the sort of thing that should be allowed, but it's not currently certain that it would be. JulesH 12:02, 12 October 2006 (UTC)
Yes, it would. That's the kind of thing it's intended to cover. But I agree it might be worthwhile to try to define popular culture to make sure the sweep's not too broad. SlimVirgin (talk) 12:11, 12 October 2006 (UTC)
Slim, you misspoke when you said "This is a significant departure from the previous policies" because it is only a significant departure from the previous wording of policies. The fact that all that popular culture stuff exists and is being expanded shows we know it meets actual policy even if the current wording of policy is insufficiently clear about that. Please see Jimbo's talk page for further reading about the map is not the territory. Let's not confuse people into thinking new policy is being written here. We are only rewording. OK? WAS 4.250 16:47, 12 October 2006 (UTC)
I was told in WP_talk:V that there was no distinction between the wording and the policy. The policy is what's written. the fact that all that popular culture stuff exists and is being expanded shows we know it meets actual policy — not unless you can say that those articles are being edited by people who are aware of and fully understand the policies, which is doubtful, expecially in the pop culture articles. All it shows is that to date, there has been more momentum behind building those articles up than behind stripping them down for lack of reliable sources. —mjb 01:18, 13 October 2006 (UTC)
WP:IAR is part of the wording of policy too. Trying to get literal serves no good purpose when the words are logically self contradictory. WAS 4.250 01:31, 13 October 2006 (UTC)

I don't like this section. Who gets to decide what is a "reliable", self-published fan site and what is a non-reliable, self-published fansite? A concensus of the editors working on the article? They're most likely all fans of the topic themselves and will set a low bar. And what if reliable, non-self-published material is available? For example, Donkey Kong (video game) uses no self-published sources, no fan pages, no blogs, no Usenet (except for Killer List of Video Games, which is used as a primary source). But what if someone wants to go in and add some random fact they found on a fan page? With this new wording that would be acceptable. — BrianSmithson 09:09, 13 October 2006 (UTC)

What you are talking about occurs right now. And right now you may delete anything you believe to be false and believe is inadequately sourced. This proposal changes neither of those two things. What constitutes adequate sourcing is an ongoing debate whose details have no place in written policy. WAS 4.250 09:55, 13 October 2006 (UTC)
It occurs right now, but at least we have a tool against it. "You can't use that Geocities fanpage; it is a self-published, unreliable source." I fail to see the benefits of weakening Wikipedia's standards just because people are having a problem finding information on a minor character from a video game or whatnot. The proposal does change things; it says that pop culture topics are held to a lower standard than everything else. That's regrettable. — BrianSmithson 11:25, 13 October 2006 (UTC)
Frankly, I think this section is wrong, wrong, wrong; we will have a policy line that allows us to use blogs? What the... If no reliable source is available, don't add the material.
I realize that pop culture articles already do this, but that doesn't mean we should suddenly sanction it. Marskell 11:29, 13 October 2006 (UTC)
I agree; it is too difficult to determine which fan-sites are reliable and which are not. A clear line is better than a blurry one. Andrew Levine 15:50, 13 October 2006 (UTC)

No original research

I've created an NOR section because a few editors have expressed concern that the NOR aspect isn't prominent enough, and I restored the synthesis of published material example, because it's something that a lot of editors have difficulty with. SlimVirgin (talk) 12:13, 12 October 2006 (UTC)

Looks good. Are there oher areas of NOR that have recurring disputes about? ≈ jossi ≈ t@ 15:28, 12 October 2006 (UTC)
That's the big one. I can't think of anything else on the NOR page that isn't here in some form or another. Can you think of anything essential that's missing? SlimVirgin (talk) 15:34, 12 October 2006 (UTC)

Tinkering with problematic sources

I made three changes to this section. I broadened #1 to include all problematic sources, as all such sources are reliable primary sources about themselves. I broadened #4 very slightly to clarify something that was already implicit. Then I restricted #5 - the relevant areas of popular culture can get by on #1 and #4. Phil Sandifer 13:39, 12 October 2006 (UTC)

The changes look good, Phil. SlimVirgin (talk) 15:34, 12 October 2006 (UTC)

Added caveat about "witness account of a traffic accident"≈ jossi ≈ t@ 15:27, 12 October 2006 (UTC)

Reliable sources are not "people", but "published materials whose authors are generally regarded as trustworthy or authoritative regarding the subject at hand". An authoritative person whose opinion/research is not published in a reliable publication, is not a reliable source. Changed accordingly. ≈ jossi ≈ t@ 15:36, 12 October 2006 (UTC)

Jossi, I'd say that reliable sources can be people or the material they produce. The New York Times is a reliable source, but so is Seymour Hersh, whether published in the NYT or on his own blog; admittedly less so if on his blog, but still acceptable for our purposes. SlimVirgin (talk) 16:14, 12 October 2006 (UTC)
I think this is a semantic issue. Since Seymour Hersh can't live in the Wikipedia servers (as in Tron), he can't be a Wikipedia source, only his publications can be Wikipedia sources. --Gerry Ashton 18:10, 12 October 2006 (UTC)

I'm moving to the talk page the following exception to using self-published sources, because it's not clear what it's referring to, and it could create loopholes. Does anyone else have concerns about it? SlimVirgin (talk) 15:34, 12 October 2006 (UTC)

Self-published sources that have been reviewed and recommended by independent reliable sources

Other entities can sometimes fulfill the editorial function that self-published sources lack. When this is done, the primary objection to such sources evaporates. Be cautious, however, that the independent source is qualified to make the judgement about the source you are using, and note that it is best not to rely on the judgement of a single review.

Yes. Maybe the editor that added this can clarify this exception. It is not clear. ≈ jossi ≈ t@ 15:37, 12 October 2006 (UTC)
I wonder if this isn't trying to make a general case for what is a specific case in popular culture. This may be a good idea - in general, if there's a specific exception that needs to be made, there's probably a flaw somewhere in the general principle. See below for more details. Phil Sandifer 15:52, 12 October 2006 (UTC)

Are there areas other than popular culture that need self-published sources from people without traditional credentials? Technology issues, perhaps some hobbies? I don't really know here, but I can think of a bunch of things that might. I mean, I honestly have no idea if our articles on, say, quilting require reference to some widely respected quilting website. Or maybe the absolute best information on DVD players comes from some compendium of technical specs on them. I've no idea, but I'm skeptical that we could possibly come up with an exhaustive list, and even if we did, it would be likely to change under our feet. Should we perhaps carve the exception so that it can be adjusted on the fly when one of these comes up? Something like "Many topics, particularly those that do not recieve substantial coverage from academic and mainstream presses, have used the Internet to create excellent self-published secondary sources. These sources are rare - most topics will not have more than one or two - but they are generally reliable. Consult specific WikiProjects for guidance on using these." Phil Sandifer 15:52, 12 October 2006 (UTC)

That is an excellent forumlation, Phil, and will avoid lost of problems with these type of articles, and reduce the need to dilute the policy to accomodate these. ≈ jossi ≈ t@ 16:17, 12 October 2006 (UTC)
Whatever sources experts use are reliable sources. WAS 4.250 16:58, 12 October 2006 (UTC)
Two points: I don't believe publishing houses have always played the role they do today. It isn't my field, but are there old self-published books that are respected and used by experts today?
Also, the self-publication definition now includes organizations. If the organization employs substantial editorial control, we can consider it a reliable publisher, and its publications would not be self-published even if no author is listed. But we may not have insight into the organization's editorial process, and may have to rely on external indicators of the reliability of the organization or the publication. --Gerry Ashton 18:21, 12 October 2006 (UTC)
Agreed for the non-academic subjects (crafts, hobbies, much of technology) as Phil Sandifer proposed. I also think that prior to some point in the 1800s or early 1900s almost all publication was effectively self-publication. I know that as of a generation before Charles Darwin wrote the normal arrangement in the UK was for an author and a printer to decide on a project, jointly solicit subscriptions, and then to print the book. The foundational works in geography were published in this manner. I'm not sure how to address this, but we need another commentary here, that mentions "be aware that in many fields these sources are reliable to discuss historical understanding of a subject but are no longer considered to accurate on that subject." GRBerry 22:09, 12 October 2006 (UTC)
"Many topics, particularly those that do not receive substantial coverage from academic and mainstream presses, have used the Internet to create excellent self-published secondary sources." Grammatically this is saying "topics…have used the Internet", which is poor wording. I think you intend to say "Subject-matter experts have used the Internet to create excellent self-published secondary sources for many topics, particularly those that do not receive substantial coverage from academic and mainstream presses."
However, we always have to keep the crackpot pseudoscientists in mind; their self-published ramblings are ignored by academia and the press more often than not for a good reason (e.g., their claims have been debunked). So I suggest slipping in some qualifiers to ensure that you don't accidentally open the door to sources for refuted claims.
Also something that should be addressed is the fact that Wikipedia itself, especially when it comes to pop culture topics, is increasingly becoming the most popular Internet-based self-publishing venue. Subject matter experts often don't want to bother self-publishing info on their own websites when they can instead publish it on Wikipedia (so long as they avoid original research), where it will have the benefit of (over time) extensive peer review and refinement by other experts in the field as well as random contributors who can fill in gaps. We need to make sure it's clear to the reader that the sources you're talking about don't include Wikipedia in large part (I assume) because WP articles don't currently indicate which statements have been authored and/or reviewed by experts. —mjb 02:07, 13 October 2006 (UTC)

Again, no, no, no. Who decides what's reliable? Clearly, the person I've been arguing with at Talk:Mami Wata thinks his or her organization is the source on all things related to this topic. But it's self-published, non-academic gobbledygook, and I will never stand for a change that could even potentially allow stuff like this to be used to write Wikipedia articles. This is a slippery slope. Self-published sources have been forbidden on Wikipedia for a good reason; let's not open the floodgates now. — BrianSmithson 09:16, 13 October 2006 (UTC)

The experts are the proper judge of what is and what is not a proper source for what. WAS 4.250 09:50, 13 October 2006 (UTC)
Who decides who the experts are? In the debate I linked to, I and others have argued against questionable sources on the basis of their being self-published. However, the person on the other side of the debate asserts that their sources are the real experts. When you get into lesser-documented areas like this, it's quite a gray area. With the "no self-published sources" proviso, you eliminate the guesswork. Here we're weakening that ability. — BrianSmithson 11:22, 13 October 2006 (UTC)
It is a gray area! It's a site of legitimate ambiguity and judgment calls. It can't be determined by a black and white procedure that does the work for you. That's why judgments need to be deferred down from the general policy page to pages more specifically tuned to different areas of the encyclopedia. We cannot hope to write a page that provides usable black and white guidance for every one of our million+ articles on a topic this rife with subtlety. Phil Sandifer 16:49, 13 October 2006 (UTC)

Problematic sources

Trödel, I'm not keen on the extra material in that section. [3] There's no need to explain what an unpublished source is, especially given that we don't use them; ditto anonymous and dubious sources. SlimVirgin (talk) 16:11, 12 October 2006 (UTC)

Sorry about reverting the changes you'd made to the first part, by the way; I didn't mean to do that. SlimVirgin (talk) 16:16, 12 October 2006 (UTC)

Is anyone else troubled with the statement that we allow cites from unpublished materials in archives? My concern is that materials in archives aren't readily available: usually someone who wants to consult items in an archive has to jump thru a lot of hoops to gain access to specific documents -- & the next person to want to see those documents may not be able to successfully repeat the performance.

My suggestion is that materials from archives are only permitted if they are also scanned & posted in a publically-accessible fashion, say at Wikisource. The same for rare books that can't be obtained thru normal means like Inter-Library Loan. -- llywrch 01:32, 13 October 2006 (UTC)

I dislike the idea of such a widespread ban of this material unless, although I also dislike relying on archives when more reliable sources are available. However there are topics which have simply not been the subject of more reliable sources. The suggestion of scanning and posting these materials is unlikely to be of great use. Although I think it would be great to get these things on Wikisource, I have prized but unfortunately extremely short list of libraries which will allow their archives and rare books to be scanned. Since we are mainly talking about material that is out of copyright in these cases, the libraries only means to control these materials is by controling access to them. Some libraries even make quite outrageous claims of copyright to what copies they will allow you to pay them to create. I do not understand what you mean by people needing to jump through hoops for access. Generally you must show up to the rare book room between 9-4 Monday - Friday, occassionally you must call two weeks ahead and make certain the head archivist will be in that day in order to be allowed acces to some particularly fragile documents. On the other hand permission to photogragh or scan these items is often explictly forbidden. I do not mean to imply there are no libraries that allow such things, but I wish it were more common. --Birgitte§β ʈ Talk 03:55, 13 October 2006 (UTC)
Because the word "archive" was not defined, I assumed it mean collections of unpublished materials kept by government or corporate bodies -- & sometimes collections in research libraries. Because access to these archives are restricted to only serious students -- e.g., academics & graduate students -- someone applying for access at the minimum needs a letter of introduction from a person that archive will accept as a reliable reference. And in some cases, access to certain papers is even more restricted, & may require permission from the person (or persons) who donated the material.
As for "rare books", would you believe a statement I made that is based on a work that no one else can access? If I quote directly from a book that exists in only two copies, both of which are under lock in key on another continent, wouldn't you suspect that I am fabricating my source? If you insist that you need to cite this source in this article, then provide us a copy of the source so we can verify it; otherwise, find a secondary source that repeats those facts or quotes the unpublished materials. I honestly can't think of any reason not to follow this rule; as it has been said time & again, we aren't engaging in original research here, so we don't need to permit access to these sources. Until someone can provide a real-life example for ignoring this rule in a Wikipedia article (so far we've been tlaking in hypothetical situations), allowing unpublished material from archives, or rare books only serves to help the original research none of us wants to slip into Wikipedia. -- llywrch 21:37, 13 October 2006 (UTC)

Unacceptable sources

I added back a few things on blatantly unacceptable sources. Some sources should be used carefully--but some sources, such as material which is unsourced, unpublished, or likely fraudulent--shouldn't be used at all.

On a related-though-separate note, perhaps a general "consensus" exception might be useful, along the lines of:

Exception: This policy may be overridden with respect to a given usage of a source, if a consensus of editors on the article's talk page believe it appropriate to do so. A separate consensus is required for each article that the source is used on.

The rationale behind this is obvious. Rather than trying to enumerate case-by-case all the exceptions here (or on a subpage), or having a policy which is a straitjacket, or appealing to blank checks like WP:IAR--I think it reasonable that a consensus of editors on a topic should be able to override the rules here. When a consesus cannot be developed, this policy applies.

Thoughts? --EngineerScotty 16:07, 12 October 2006 (UTC)

ES, I'd oppose that. We've had people arrive at V and NOR trying to introduce that clause; one of the LaRouche editors tried to get it inserted at one point, as I recall. Policy can't be overridden by the consensus of any group of editors on a single page. SlimVirgin (talk) 16:19, 12 October 2006 (UTC)
Consensus of editors is a can of worms. You may have 3 editors atending to an article, and their "consensus" is not a good basis for adding a source that would be unacceptable to a different group of editors. 16:20, 12 October 2006 (UTC) ≈ jossi ≈ t@ 16:27, 12 October 2006 (UTC)
To address both complaints; I would further suggest that such discussions be announced to the wider Wikipedia community (i.e. an WP:RFC must occur on the issue), and that consensus may be withdrawn at any time if additional editors arrive to dispute it; such decisions are not a fait accompli when done. Getting such an exception should be very difficult.
OTOH, if nobody ever challenges a given questionable source to begin with...
--EngineerScotty
There'd be no point in having policies if a consensus of editors on any given page could decide to override it. We'd end up with neo-Nazis overriding it to write in glowing terms about themselves, and so on. SlimVirgin (talk) 16:32, 12 October 2006 (UTC)
Which is why a strong consensus, and notification, should be required. Unless the entire userbase of Wikipedia were neo-Nazis, I doubt a faction of neo-Nazis could achieve the consensus. As it stands, articles on controversial subjects are watched like hawks by users supporting the various sides of the controversy. If a cabal of neo-Nazis currently WP:OWNs a page due to nobody else paying attention; then I suspect that it says just what they want it to say, policy be damned. --EngineerScotty 16:36, 12 October 2006 (UTC)
I strongly disagree with consensus. While it is true that there may be a "good consensus" on some pages, I doubt there are watchdogs on all 1.5 million pages. An encyclopedia is based on good fact, and should be from reliable sources, not a consensus. (You wouldn't build a house from a consensus of random persons off the street. You would build a house based on the design of an engineer/architect.) This is what is always dangerous about Wikipedia to me; there must be published, good sources to support fact. Olin 20:21, 12 October 2006 (UTC)

I agree that this is not a good proposal. However, some way should be provided to break out of the one-size-fits-all problem with the policy in the case of subject areas that don't fit the policy well. I propose that instead of a concensus of editors, what is needed is a concensus among the area specialists. And that consensus would need to be established by a "reliable source". We ought to be able to use whatever sources the experts in the field regard as the best sources for that field. This opinion of the experts would itself be stated in the article with an appropriate source. 222.229.239.224 13:28, 13 October 2006 (UTC)

Self-published sources

Jossi, regarding this edit: "Same applies to websites of organizations, political parties and enterprises": the websites of organizations or political parties aren't self-published sources, unless they're tiny. The NYT website or Republican Party websites aren't self-published, because they have editorial oversight, fact checkers, legal scrutiny, lots of people involved. A self-published source is one man and his dog, or thereabouts, where no one, or hardly anyone, stands between the writer and the act of publication. SlimVirgin (talk) 16:30, 12 October 2006 (UTC)

We should make the distinction clearer between "man+dog" operations and entities like Encyclopaedia Britannica; both of which self-publish, but only one of which employs rigorous editorial process. However, it would be mistaken to give organizational self-publishers a free pass; many of them are quite good at producing impressive-sounding bullshit.  :) OTOH, most newspapers frequently self-publish, as the same guy who signs the editors' paychecks also signs those of the writers.
The key issue, really, is the level of editorial restraint which the publisher requires of the author. For certain entities (large publishing houses, many academic publishers, major news outfits, Britannica), the level is known to be acceptable. For others (most blogs), it is known to be deficient. The self-published/not-self-published dichotomy is really a proxy for this issue, as we often cannot tell just what scrutiny a given work was placed under, but we can tell if the author and the publisher are different entities.
--EngineerScotty 16:42, 12 October 2006 (UTC)
Yes, we need to make a distinction and avoid giving the imprimatur of reliability to self-serving material BS published by organizations, political parties, etc. Any ideas on how to express that? ≈ jossi ≈ t@ 16:48, 12 October 2006 (UTC)
It goes to conflict of interest. I wouldn't quote Britannica's opinion on Britannica. WAS 4.250 17:03, 12 October 2006 (UTC)

My opinion, for what it's worth: If ALL you find is a self-published source, then use it and let the reader decide if it is reliable. I refer to the article on Inglewood, California, wherein most of the history of that little town was written by one local historian and published with no supervision or editorial oversight by the local historic society. There can be plenty of other examples in which self-published books are good sources for an encyclopedia. Sincerely, GeorgeLouis 11:37, 13 October 2006 (UTC)

Primary/Secondary Mess

The logic on primary and secondary sources is still a mess. First of all, it's still suggesting that sources are ontologically primary or secondary, instead of contextually. A New York Times account of the President's speech is a secondary source for the President's speech. It's a primary source for the New York Times's coverage of the Presidency. This is a vital issue, especially as it relates to questionable sources.

Second of all, the bias against primary sources is far too strong. We ought not be favoring a NYT account of the President's speech over a transcript for reporting the content of the speech. It's just silly. And in terms of literary criticism, I have no idea how to even make the distinction work. Is Paul De Man's famous reading of Yeats a secondary source on Yeats, or a primary source for the section on literary analysis of Yeats?

Ideally, I think we should do away with the primary/secondary distinction entirely, as it leads to many more exceptions than clarifications. Phil Sandifer 16:31, 12 October 2006 (UTC)

We do say that primary sources are fine for descriptive claims, so it'd be okay to use the transcript of the speech over the NYT's account of it, so long as we're describing and not analysing.
I agree that the description we give of the primary/secondary distinction is simplistic. The problem is that it's a complicated distinction; whether something is primary/secondary depends on context and on where you stand in relation to the source. It's too complex to explain properly for this policy. I've also thought about leaving it out entirely, but it can sometimes be useful. What do you say otherwise to an editor who wants to base an edit on the New Testament? It's useful to be able to ask for a secondary source for whatever the claim is. SlimVirgin (talk) 16:40, 12 October 2006 (UTC)
Well, obviously, it depends on the edit. If it's "Jesus is the Son of God (citation to the New Testemant)," I point out that there is substantial dispute on that point, and it has to be cited more precisely as "The New Testament claims..." If it's a matter of, say, evolution and someone is trying to introduce the Book of Genesis, I point out that it's not a reliable source - primary or secondary - in the field of science, which is the field in which we're dealing. Either way, the issue doesn't seem to me to be one of primary/secondary but of reasonable/unreasonable. Phil Sandifer 17:04, 12 October 2006 (UTC)
Another example is that I recently had someone on Animal rights who wanted to use the Bhagavad Gita to show that Sikhs had always supported animal rights. It's an appropriate source to use for descriptive claims, but not interpretive ones, because it's a primary source. I don't know how I'd have gotten that across without reference to the primary/secondary distinction. SlimVirgin (talk) 17:31, 12 October 2006 (UTC)
In general, just because a given scriptural work X advances theological claim Y; it is incorrect to use X to support the claim that X-ists believe Y. Many things in the Bible are not observed in practice by many Christian denominations; or by many Christians as individuals. The Bible would be an inappropriate source for the claims that "Christians are against adultery"; many Christians in fact enjoy it very much.  :) --EngineerScotty 17:37, 12 October 2006 (UTC)

I think a key distinction is that secondary sources, in whatever context, are usually analytical. At a minimum, they present and vouch for the authenticity and relevance of a primary source. More useful secondary sources provide more in-depth analysis, often using multiple primaries and drawing more complicated inferences then would be possible from examining primary sources in isolation. In many ways, WP:NOR (and the passages here that seek to replace NOR) address the issue; only information/claims that are manifest in a given source may be included in Wikipedia. For primary sources, in many cases, that ain't much. (And many primary sources are not published or even publishable; which makes them inappropriate references to begin with). --EngineerScotty 16:48, 12 October 2006 (UTC)

But, especially in fields like philosophy and literature, those secondary sources are also primary sources, because analysis of the analysis is vital to the process. And novel synthesis of secondary sources is just as bad as novel synthesis of primary ones. The issue is the novel synthesis - no building something new and putting it on Wikipedia. Phil Sandifer 17:04, 12 October 2006 (UTC)
As suggested above; I'm for getting rid of the distinction between the two, or at least trimming it consistently. Prohibitations on original research and use of unpublished material address the bulk of the issues people have with "primary sources". --EngineerScotty 17:09, 12 October 2006 (UTC)

I think we should rely on insisting on sources that directly support what someone wishes to add to wikipedia rather than indirectly through analysis. A source that does that analysis can itself be used as a source (and it is a primary source for that analysis). I think the use of a no conflict of interest for controversial claims criteria combined with insistance on direct support will allow getting rid of all talk of primary and secondary sources. WAS 4.250 17:15, 12 October 2006 (UTC)

I think I could go for that. As Phil said above, it's the novel synthesis that's the problem, regardless of whether it's of primary or secondary sources. I wrote something like this not long ago on NOR or V, which I'll try to find. SlimVirgin (talk) 17:36, 12 October 2006 (UTC)
  • What conflicts of interest are you referring to? Cases where it's obvious that a given researcher has an axe to grind--i.e. the Tobacco Institute? Any researcher will have an investment in his/her own theories and writings; this can sometimes lead to pathological science and error, but it would be absurd to suggest that researchers should be disqualified from their areas of expertise because the effect of publishing is to make them bias. I'm certain this is not what you are suggesting--explain a bit more. --EngineerScotty 17:28, 12 October 2006 (UTC)

Here's what I wrote about this on NOR talk last month:

Would it be possible to avoid discussion of the primary/secondary distinction in the policy? We must use material from reliable published sources. The important questions therefore are: (1) is the source published by a third-party? and (2) is it a reliable source that's appropriate for the subject matter?
The issue of whether it's primary, secondary, or tertiary doesn't really come into it. We do want to discourage editors from using primary sources, but only because, in so doing, they are more likely to misinterpret how the source should be used, and what its limits are. But that can happen with a secondary source too. It's not the type of source that's the issue, but the misuse.
The sentence that says articles relying on primary sources should make no interpretive claims is problematic, because if the primary source makes interpretive claims, and if the source is reliable and appropriate, then the article can make them too, citing the source. For example, if the day after 9/11 the New York Times published stories by a bunch of journalists who were present when the planes hit the WTC — i.e. eyewitness accounts, primary-source material — there would be no reason we couldn't use those articles to repeat interpretive claims that the journalists made, so long as we felt they were appropriate sources.
I may be missing some other point about making the primary/secondary distinction, and if I am, I apologize. SlimVirgin (talk) 16:59, 15 September 2006 (UTC)
I think what I wrote about interpretive claims was off the mark, though. SlimVirgin (talk) 17:40, 12 October 2006 (UTC)

I've added some language WRT this discussion to the policy; please review. If we like the language, I'd love to shoot the paragraphs on primary/secondary sources in the "key principles" section in the head. (I tried to kill it once already, but it came back from the dead; I'll ask here before killing it again.) --EngineerScotty 17:50, 12 October 2006 (UTC)

I don't really regard those 9/11 witnesses as primary sources for Wikipedia; they're primary sources for the newspaper. I think it's muddling to include that example alongside the White House summary-of-a-speech example in the text, because the first type of source is perfectly usable and the second not (until it too appears in a newspaper). I've never agreed with policy bracketing all types of primary sources together; nothing brings a history article to life better than quotations from primary sources (quoted or collected in a published book, of course); if interpretation is required, you just report the interpretation of a historian.
Phil is right that secondary sources can be considered primary sources for their own point of view, but we should, in my opinion, merely recommend the normal scholarly caution on that one rather than addressing the complexity of it in the policy text.
I think we should depend mainly on the "published" criterion and, as SV suggests in her second paragraph above, not bother too much about definitions of primary and secondary sources which may only serve to confuse. qp10qp 18:38, 12 October 2006 (UTC)
I agree about the 9/11 witnesss thing. I got that wrong. SlimVirgin (talk) 18:45, 12 October 2006 (UTC)
Qp, you wrote that if interpretation is required, we just report the interpretation of a historian. But that's the same as saying interpretive claims need secondary sources, which is the bit we're talking about removing. I do wish we could get leave the distinction off the page, but I wonder whether it's more useful than it is confusing. SlimVirgin (talk) 18:48, 12 October 2006 (UTC)
It's simply saying that interpretive claims need sources (reliable ones, that is), the word secondary in the above is completely unnecessary. It doesn't matter whether this historian is categorized as a primary, secondary, or zillionary source :); just that he is a reliable source for the interpretation offered. (As opposed to, say, a random drunk in a bar; whose interpretations we can certainly do without). --EngineerScotty 18:53, 12 October 2006 (UTC)
But the historian is only a reliable sources for interpretive claims if he's a secondary source on the topic, ES. If his work was a primary source on the topic, we couldn't use him for interpretive claims. That's an example of why the distinction is useful. SlimVirgin (talk) 19:05, 12 October 2006 (UTC)
I think the key here is that primary sources generally need to be interpreted to be properly understood, and when people make use of primary sources in articles, they inevitably end up interpreting them in some way, whether explicitly, or via re-wording/translation, or by placing them in a specific context. For that reason I feel the distinction is useful. Jayjg (talk) 18:56, 12 October 2006 (UTC)
[edcon]I'd be in favor of language stating such, i.e. "many primary sources need to be interpreted to be properly used" somewhere in the OR section (or adding it to the "problematic sources"); and leaving it at that. The long dissertation on primary vs. secondary, which we all seem to agree is flawed, doesn't improve the article. --EngineerScotty 19:01, 12 October 2006 (UTC)
The issue with the 9/11 accounts seems to me to be that our article is about a geopolitical event, not about the personal experiences of people. This is a subtle point, but it is the one that I think matters. It's not that personal accounts of 9/11 are unreliable - it's that they're irrelevant. Phil Sandifer 19:00, 12 October 2006 (UTC)
Not if these journalists reported important observations.qp10qp 19:06, 12 October 2006 (UTC)
No original thought: it's as simple as that. Nothing wrong with using a primary source so long as you take it from a secondary publication. And since we don't use original thought, it follows that interpretation of a primary source must come from a secondary source.
By the way, this bit in the text
An eyewitness account of a traffic accident published in a newspaper is a primary source.
amounts to the same as the 9/11 example, doesn't it? For our purposes, it's not a primary source, only for the newspaper's. The White House one is the nasty one: that is an excellent example of a primary source unrefracted through a secondary medium.qp10qp 19:12, 12 October 2006 (UTC)
And given that the newspaper is a RS, the quote (assuming it is on topic) is acceptable to cite. As opposed to a Wikipedian interviewing a 9/11 witness and using that as a source. One is OR; one is simply depending on the OR performed by someone else (a newspaper). No need to worry about primary/secondary. The distinction is useful in analyzing the policy; far less so in expressing it. --EngineerScotty 19:17, 12 October 2006 (UTC)
To explore the issue a bit more. If we cite the newspaper interviewing a witness; our source--the newspaper--is a secondary one. We're not using the primary source, any more so that we are using secondaryprimary sources if we cite a historian's book. The historian's book is a secondary source; the fact that the historian doubtless drew on numerous primary sources is irrelevant to us. The secondary source--in either case--is vouching for the authenticity and relevance of the primary source, even if no further analysis or synthesis is being performed. --EngineerScotty 19:20, 12 October 2006 (UTC)

(unindent)The problem with "many primary sources need to be interpreted to be properly used" is that "primary source" is not really a description of a specific type of source, but rather a description of the use of a source. Many things can be used either as primary or secondary sources. I understand that the intention is to describe things which cannot be used as secondary sources, and are therefore limited to being used as primary sources. However there has to be a better way to describe this. The only thing I can think of is "documentary source", but is there any established terminology for what we are describing?--Birgitte§β ʈ Talk 19:23, 12 October 2006 (UTC)

Good distinction, Birgitte. SlimVirgin (talk) 20:09, 12 October 2006 (UTC)
The closest thing we have is "unpublished source". I think that suffices for most cases. People quibble about blogs, popular culture etc. but I think the "published" criterion will cover most eventualities. However, as Phil has pointed out, it seems there are types of subject matter and article where "unpublished" sources should not be ruled out. qp10qp 19:32, 12 October 2006 (UTC)
"Unpublished source" is really a different concept from "primary source". There are many reliable published primary sources, for example, you can buy microfilm of raw census data from the U.S. National Archives. --Gerry Ashton 19:48, 12 October 2006 (UTC)
I know. Birgitte was looking for a term to describe things which cannot be used as secondary sources and are therefore limited to being used as primary sources. I don't think there is one term to describe that.qp10qp 20:01, 12 October 2006 (UTC)
Birgitte hit the nail on the head in pointing out that "primariness" or "secondariness" is not an inherent property a that some given source just possesses. Rather, it is a quality which describes that source's relationship to the reason for which it is used. That's hared to make clear in a few words, and a main reason why including that terminology in whatever policy or guideline we have is usually more confusing than helpful. Gene Nygaard 21:28, 12 October 2006 (UTC)
And, what that really means is this: it is totally improper to say, for example "primary sources (e.g. posts on bulletin boards, blogs, and Usenet)" because you can never know whether any one of those is a "primary source" or a "secondary source" unless you also specify what it is being used to show. "Primary" in not some inherent property of any document or recording or whatever. Gene Nygaard 21:37, 12 October 2006 (UTC)

I've never really been comfortable with the primary/secondary/tertiary analysis. We have to judge reliability of all sources. In some fields, primary sources are nearly always preferred over secondary sources or tertiary sources. Genealogy is an obvious example - primary sources are almost always preferred. (We aren't a genealogy site, so this specific example is of limited relevance.) I think the primary/secondary/tertiary source discussion would be an excellent one to move out to an essay (or guideline). The policy should emphasize reliability, the essay(s) should explain how primary/secondary/tertiary sources are more or less reliable in a given field. Given the different usage of the term in science and history, we may well need one essay for science, a second for historical topics, et. cetera. GRBerry 22:17, 12 October 2006 (UTC)

Your mention of genealogy suggests an obvious example, to help clarify that primariness is not an inherent property of a document: Is a death certificate a primary source or is it a secondary source? Gene Nygaard 05:55, 13 October 2006 (UTC)
The importance of the primary/secondary distinction (beyond the special case of the sciences mentioned above) is that the primary source documents the facts, & the secondary source documents the resulting discussion. An example from history would be to say Bede and the Anglo-Saxon Chronicle (ASC) are primary sources, which offer facts (e.g. the ASC states that in 751 something happened), while the writings of historians provide verification of the discussion over what the facts mean (e.g., according Sir Frank Stenton, the ASC is wrong, & that event actually happened in 749); as Jayjg noted above, we desperately need to observe this distinction when covering exegesis. However, if we observe this distinction it will cause havoc in some areas, such as our articles on literature: for example, The Cantos, which is a FA, includes a great amount of unattributed interpretation -- yet most of it can be verified by simply reading the work. -- llywrch 23:14, 12 October 2006 (UTC)
You are incorrect, Bede and the Anglo-Saxon Chronicle (ASC) are sources that can be either used as a primary or secondary source. The Diary of Anne Frank is an example that is even more likely to be used a primary source as it is an eyewitness account. However Bede is clearly setting out to write a history. Why do you consider the authors of the ASC any less of historians than Sir Frank Stenton? It is true we should avoid using the ASC as secondary sources, because it has proven to be inaccurate. However the ASC was certainly written to be used as a secondary source at the time. Whereas Anne Frank was intentionally writing to document her own experience rather tell the history of Netherlands during WWII. These terms are not as clear-cut as you believe them to be. --Birgitte§β ʈ Talk 02:20, 13 October 2006 (UTC)
While you are right that a document can fall into both categories, you seem to be missing my point -- which does not depend on a document belonging only to one or the other of these categories. -- llywrch 21:54, 13 October 2006 (UTC)

The most definite conclusion that can be drawn from this discussion, and long similar discussions before, is that the primary-secondary distinction is much more trouble that it is worth. For one thing, nobody can agree on what it means. Getting rid of it would be a definite improvement. 222.229.239.224 13:39, 13 October 2006 (UTC)

Clean that mess up on WP:OR and WP:RS, the sources of the mess here, and there may well be less of a perception of a need for change of the sort proposed in this discussion. Gene Nygaard 15:51, 13 October 2006 (UTC)
If by the "change of the sort proposed in this discussion", you mean avoiding the terms Primary and secondary source. I must disagree. The problems are inherent in how the terms are defined and used in the world at large not some special Wikipedia definition from the way they are used at WP:OR and [[WP:RS].--Birgitte§β ʈ Talk 16:09, 13 October 2006 (UTC)
I guess I wasn't very clear. I meant that fixing the problem in WP:OR and WP:RS may well obviate most of the perceived need for combining them and WP:V into a WP:Attribution in the first place. After all, one of major reasons cited for the WP:Attribution discussion was "WP:RS is a mess" or something along those lines. The primary/secondary stuff is a big reason for that being seen as a mess. Gene Nygaard 18:01, 13 October 2006 (UTC)

Self-published sources in articles about "themselves"

A couple of issues with this section.

First up, I've seen people get confused in the past about what exactly "themselves" means in this case. It should be a reference to the author of the source, but it's sometimes interpreted as a reference to the source itself. I'm not sure how best to clarify this, though.

Secondly, I'm of the opinion that the exception should be widened. Current practice seems to be that self-published sources are frequently used in articles on subjects which the authors are deeply involved with, but not directly about the authors themselves. It's a frequently cited example, but its still true: most of the sources on Scientology are self-published, either by L. Ron Hubbard or by the Church of Scientology, neither of which "are" Scientology, as can be witnessed by the fact that we have separate articles on each. By a strict reading of this rule, these sources are not permitted (although you could argue that Hubbard was an expert on the subject, I don't see a good argument that can be made for the CoS), but there is little doubt that they make the article better. Again, I'm not sure how best to achieve this. The suggestion I put in the text was reverted quickly.

Any comments? JulesH 19:42, 12 October 2006 (UTC)

Scientology is an interesting case--many of the commentators on it are either affiliated with the CoS, or sued by it :); few if any of the published sources on the controversies surrounding Scientology are very good. Scientology is an important enough topic that we can't ignore it; but it is terribly difficult to cover in a reliable, NPOV manner. Other than my throwaway comment here, it's one of several Wikipedia topics that I avoid like the plague.  :)--EngineerScotty 20:04, 12 October 2006 (UTC)
Actually, they don't litigate at the drop of a hat like they used to; nowadays they're more likely to "fair game" critics -- speaking from personal experience. But you're smart to avoid the topic. ;-) -- llywrch 23:18, 12 October 2006 (UTC)
It is both. Not only is someone's blog a reasonable primary source about the author of the blog, the blog is also a reasonable primary source for articles on that blog. This is, again, an essential point that cannot be lost. Phil Sandifer 16:50, 13 October 2006 (UTC)

remove of OR example

I removed an included example of original research (commented it out, actually), as it was three paragraphs long. This sort of thing might be better on a FAQ or example pages. If someone really wants to include it, feel free to revert. --EngineerScotty 20:00, 12 October 2006 (UTC)

The foundation concept of attribution

The foundation concept of attribution is that because "anybody can edit" we MUST, in every article, for every questionable claim, answer the question "Who says so?" in a way that makes sense to the reader. This has nothing to do with primary or secondary sources. This has to do with claims made by a source that we directly quote or use without alteration or analysis. If we quote an analysis it is a primary source for that analysis and it doesn't matter because what matteres is can we trust that author (source) for that analysis which is not a question a lot of talk about primary versus secondary adequately handles. This has to do with why that source should be believed. If there is a conflict of interest of the type that makes their claim suspect then that must be made clear ("Scientologists say ... about scientology." or "Cold fusion researcher says ... about his own work".) If they lack expertise then that should be made clear. If other experts disagree then that should be made clear. We are not going to clue in all our wikipedia editors into the all the details of proper sourcing no matter how many pages we write. We only need to tell our editers to provide for the reader the answer to "Who says so and why should I believe them?" and here are some wikipedia articles that do that and here are some external links into the art of proper sourcing for further study if you wish. WAS 4.250 20:21, 12 October 2006 (UTC)

Examples: An article on some obscure ancient village that concludes with "This article comes from the 1911 Britannica" or an article on some obscure fundamental physics particle that concludes with "This article's contents are are found in every physics textbook on the subject." adequately answer who says so and why should I believe them. WAS 4.250 20:30, 12 October 2006 (UTC)

Good points. SlimVirgin (talk) 20:33, 12 October 2006 (UTC)

Extremists

I was recommended to repost this here - i posted it on the talk for WP:RS. The current ability to discount a group for being "extremist" is a huge argumentum ad hominem. He points out that it's more specifically a reference to groups that are known to be dishonest - which is a better way of phrasing it. A group simply being considered "Extremist" doesn't mean they're wrong. 50 years ago groups agitating against Female Genital Mutilation and for women's rights would have been considered extremist. (Heck look up The Rape of Innocence - americans used to practice FGM!). In one of the subjects I edit on groups opposed to the practice get labeled "Extremist" left and right as an attempt to discredit them baselessly, when these groups are basing their arguments entirely off medical citation and human rights! Lordkazan 20:24, 12 October 2006 (UTC)

Until such groups have more widespread acceptance, however, they do not belong in Wikipedia. Wikipedia is not a platform for activism. Phil Sandifer 20:30, 12 October 2006 (UTC)
That's absolutely asinine! That's 100% argumentum ad hominem! you'd get laughed out of a university for that kind of a statement! If the groups are citing medical facts, etc then that's even MORE assinine! Mainstream vs Extremist is 100% argumentum ad hominem! (Godwin!) In Nazi germany a support of religious freedom would have been considered an extremist, does that make them wrong? NO! Give me a good reason wwhy this isn't an argumentum ad hominem! Lordkazan 21:04, 12 October 2006 (UTC)
There are lots of reasons to disbelieve claims. If the source is promoting something whether its pepsi or peace or women's rights; then their claims must be evaluated accordingly. WAS 4.250 20:35, 12 October 2006 (UTC)
Yes, but this isn't a valid reason to disbelieve claims - the "mainstream" considering a group "extremist" is merely argumentum ad hominem and completely arbitrary as I'm trying to point out - it's not a VALID objection. We're talking about reliability here and we're allowing logical fallacies to shoot down sources? that's simply unencyclopaedic! Lordkazan 21:04, 12 October 2006 (UTC)
Some sources are known to have lied--consistently--in the past; it is not unreasonable to exclude them. The reason we don't trust Stormfront is not because their ideology is offensive, it's because they have exhibited no regard for the truth whatsoever. Their material a long litany of "down with blacks, down with Jews, down with everybody but us".
We don't automatically exclude sources with an axe to grind--and groups biased to "mainstream" viewpoints are still biased, BTW. However, we ought to acknowledge those biases so we can properly consider the sources.
In the case of FGM--if zillions of doctors are on record deploring the practice, then don't quote anti-FGM advocacy groups--quote the doctors directly. Laundering neutral claims through political organizations is almost always a mistake. --EngineerScotty 21:11, 12 October 2006 (UTC)
It's not ad hominem, because advocacy groups are, by dint of being POV-pushing, suspect sources. And perhaps that's a better word than "extremist". Mangoe 21:13, 12 October 2006 (UTC)
I'm not specifically speaking of advocacy groups, i'm taking about any group that isn't considered "mainstream" - doctors who release studies that condmen circumcision for example are not considered "mainstream" - but their studies are a lot more solid research than the crap we hear trumpted in the news like Auvert et all (the study about HIV they raved about at world aids conference 2006) Lordkazan 21:15, 12 October 2006 (UTC)
Considered mainstream/not mainstream by who, exactly? Given that many people oppose circumcision/FGM, and many more don't give a crap either way--it seems out of place, in this instance, to label the anti-cutting side as "non-mainstream".
At any rate, I'm all in favor of clarifying the word "extreme". Simply labelling someone an extremeist is not a free pass; absent additional information, advocacy groups are advocacy groups. OTOH, if one can demonstrate that a particular organization is fond of telling whoppers, or simply publishes political screeds with no research backing it up, I'm in favor of limiting the scope of those sources use on Wikipedia. --EngineerScotty 21:30, 12 October 2006 (UTC)
The united states - where just bringing up objections to male circumcision often gets one ridiculed as "whiner", "cry-baby", "antisemite", and a million other things, but supporting female circumcision gets you called a mutilator and a million other things (rightly so) - quite the double standard.
This however is entirely irrelevant, i'm objecting to the "Extremists" thing because of the abusability of it and the fact that it's argumentum ad hominem Lordkazan 21:34, 12 October 2006 (UTC)
You seem to misunderstand, a bit, the meaning of the logical fallacy of argumentum ad hominem. Argumentum ad hominem is essentially "A says B; A is a moron/liar/whatever; therefore ¬B". What we here are saying is essentially, "A says B; A is a moron/liar/whatever; therefore disregard A". This policy says nothing about the ideas promoted by sources deemed unreliable; that an unreliable source promotes B should not, in any way, be construed as evidence against B. Do you see the difference? And should we explicitly state as such in the policy proposal? Or is that going to far? --EngineerScotty 21:39, 12 October 2006 (UTC)
the Term "extremist" is too easily applied - and it's completely subjective is the problem. Lordkazan 21:41, 12 October 2006 (UTC)

The "synthesis" section

I take issue with the sentence: "An exception is if the derivation of C is trivial, obvious, uncontroversial, verifiable by others, and requires no additional assumption or interpetation." "Trivial, obvious, uncontroversial" are inherently interpretable subjects. I know of the uranium trioxide edit war, what is "trivial" and "obvious" to one editor was (and still is) not "trivial" and "obvious" to another. "Verifiable to others" implies a source, which the policy already says there is not; that clearly needs clarification. "Requires no additional assumption or interpretation" is redundant, because only "A" and "B" are presented in the deductive argument. Olin 20:33, 12 October 2006 (UTC)

The point is to allow trivial things like units conversions, "if X lives in New York, X must live in the US", and other really simple stuff. Feel free to wordsmith it better. If someone can reasonably dispute that it is simple and obvious, then it isn't trivial and obvious--it's OR and doesn't belong. --EngineerScotty 21:01, 12 October 2006 (UTC)
I'm uncertain of the phraseology; I can imagine concepts of "syllogisms involving verifiable fact" to cover some cases. It's less clear how to cover math issues. Interpolations (and even some extrapolations) of numerical data, for example, may seem "valid" to some, but in my view at least, should be excluded, but I could imagine other mathematics other than unit conversions that could be included (if the speed of a car is 14 mph and the car has been travelling 30 min, it has done 7 miles is not a syllogism, but would probably not be considered OR unless it advances a point). The advancing a point part, actually, does seem important here. thoughts ?Olin 21:43, 12 October 2006 (UTC)
JMabel imported text that I originally wrote. The cases I have in mind are exemplified by the 140-odd articles I added about the woredas of Ethiopia: in each I offer calculations for density & a percentage with the number of urban inhabitants. I feel these help the reader understand the data a little better; they helped me. And my practice follows the precedent of the Ram-Bot articles, based on data from the US census returns.
In other words, I assert that we do not slavishly copy out our sources; we are allowed to actually think about the material & make choices about how to best explain it. But we are also required to report our sources accurately & honestly. -- llywrch 23:30, 12 October 2006 (UTC)

I don't know of a better way of saying it than controversial vs. noncontroversial. WAS 4.250 00:37, 13 October 2006 (UTC)

I changed some wording. Infer is dangerous, and I still think the caveat about advancing a point has to be in there. Olin 15:24, 13 October 2006 (UTC)

While one can argue that certain topics on Popular culture are not well covered outside of Internet sources, the change from that to "Respected self-publications", along with the removal of caveats, just opens the floodgates to purely subjective and self-serving analysis about what is "respected". Jayjg (talk) 20:48, 12 October 2006 (UTC)

Yes. It does. Unfortunately, that's a fact of life - it's folly to think that we can actually come up with a prescriptive policy on this. Defer it to the WikiProjects. Otherwise, we get put into a situation of utter madness. The caveats might well need to be reinserted, but I'm very, very wary of deciding that popular culture is the only such area. We need to be careful, in making these guidelines, not to step on the needs of areas of the project that we don't interact with. Phil Sandifer 20:59, 12 October 2006 (UTC)
The experts are the proper judge of what is and what is not a proper source for what. If someone is judged an expert by anyone worth paying attention to then he should be quoted in wikipedia giving his opinion and the basis of his opinion. "Saint Paul said ...." WAS 4.250 00:43, 13 October 2006 (UTC)

Strange Removal...

Does this edit summary look strange to anyone else? It seems like a part of the page prohibiting something was removed with the edit summary "this should be prohibited"? Am I missing something? ---J.S (t|c) 20:53, 12 October 2006 (UTC)

I have to say, it looks like Phil was confused here, and the passage did not look confusing to me. I'd be in favor of restoring; if anyone can see what was unclear, then reword accordingly. - Jmabel | Talk 21:40, 12 October 2006 (UTC)
Restored. Maybe Phil could clarify the reasons, if the deletion wasn't a mistake. ≈ jossi ≈ t@ 22:05, 12 October 2006 (UTC)
Because the section on questionable sources has exceptions, this doesn't belong there. Unpublished sources are unusable, period, in all cases. They belong elsewhere, or just not explicitly mentioned, since they necessarily fail verifiability. Phil Sandifer 16:40, 13 October 2006 (UTC)

Added section on conflicting sources

though I may regret it.  :)

The essential point is that if reliable sources conflict, NPOV is how we resolve it. Not by playing the "my sources are better than yours game", which is fondly played in several edit wars I can think of.

The main exception is if the balance of sources is tilted in one direction or another; the existence of the Flat Earth Society and their publications aren't sufficient to include "according to some, the earth is flat" in Earth and other articles.

--EngineerScotty 20:58, 12 October 2006 (UTC)

Concur. This may deserve a guideline outside of the policy. Here is a useful example, drawn from our Merchant of Venice article:

Caveats may sometimes be in order, as when a noted actor says that "the tradition of playing Shylock sympathetically began with Edmund Kean in 1847" and we know from reliable sources that Kean died in 1833. Wikipedians are not mere copyists, bound to repeat simple statements without any scholarly judgment. In cases like this, cite the authority accurately and honestly, but mention (and cite for) the conflicting fact or facts in a way to suggest that the authority may be wrong; point to the problem, but do not attempt to solve it with your own arguments.

- Jmabel | Talk 21:39, 12 October 2006 (UTC)

Publication and archaic works

I'm repeating a comment made above, concerning how archaic works square with the publication requirement (many works from prior centuries were not "published" in the manner we speak of today).

I would think that with any archaic work (which I'll somewhat arbitrarily define as anything written before 1900--feel free to substitute a more appropriate guideline), such works fall into one of several classes:

  • Those which have been analyzed and commented on over the years by more recent scholars. These I have no issue with using (subject to the opinions of the scholars of their worth)--the lens of history can serve as a replacement for publication.
  • Those which have not. In some cases, an archaic work may have been lying dormant for years in an archive, recently discovered; in other cases, it wasn't considered worth commentary. In either case, the lack of historical perspective on these should disallow their use on Wikipedia; at least until modern scholars outside Wikipedia comment on and analyze them.

--EngineerScotty 22:19, 12 October 2006 (UTC)

'Archaic works' worth using are almost always available to us in modern annotated editions. We can then apply the reliability tests to the modern editions. And, as you say, if they haven't been published in an annotated edition or otherwise made avaialble with commentary and/or analysis from a reliable source, we shouldn't use them. -- Donald Albury 22:52, 12 October 2006 (UTC)
Hmm. I'm not at all sure I agree with that claim. For example, no one generally reprints critical editions of 19th-century newspapers, but 1860s New York Times issues would be sources for the events of their period of roughly the same quality as the New York Times today, which we don't hesitate to cite. And when something is being cited as a primary source—for example, as a citation for a quotation—there is no problem at all with the source being "archaic". Certainly one wouldn't normally quote a 19th-century science book for any reason other than to show the state of a science at the time, but (again for example) if you are writing about 19th century actors, it is likely that for any but the most famous, 19th century periodicals are the best sources there are. - Jmabel | Talk 04:19, 13 October 2006 (UTC)