Wikipedia talk:Citing sources/Archive 13

From Wikipedia, the free encyclopedia

Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Contents

Intermediate sources: State where you got it...

Is it not necessary also to read the entire book? Checking it, as cited on a web page, could potentially take it out of context (as much, at least, as the web page itself would). We should probably specify this, ne? elvenscout742 22:48, 24 August 2006 (UTC)

Let me see if I understand your concern. Let's say I find a useful quotation on a web site. If it's a reliable site, I could just write a citation along the lines of "Smith as quoted on Jones.com." But maybe I'm not sure if the quote is accurate, so I get Smith's book and read the page in question. Then I just quote Smith, and don't mention Jones.com. The problem is I only read one page of Smith, and the citation implies I read the whole book.
I think the most we shoud expect is the editor who put in the citation should have read the normal amount for the type of work being cited. If it is a novel, that would be the whole novel. For a dictionary, it would be the entry for one word. --Gerry Ashton 23:21, 24 August 2006 (UTC)
Well, I wasn't thinking of a quotation. I was thinking of something more like a statement that could easily be misinterpretted when taken out of context, and a user reading only what's referred to on a website would read into it the kind of (potentially unreliable, and uncited) things on the website when they perhaps weren't intended by the original author and woul not be visible to someone reading the book without having seen the website. Something like one of Joseph Campbell's books. elvenscout742 09:58, 26 August 2006 (UTC)
You can't use rules to force people to 'correctly' interpret something. Material may be taken out of context by editors whether or not they have read the whole book. What is important is that the source is given, so that readers can check it out for themselves. If you see something that you believe is misleading because it has been taken out of context, you are free to supply the context (all properly sourced). Just don't be surprised if your choice of what is the proper context is challenged, however. This is a wiki, after all.-- Donald Albury(Talk) 11:39, 26 August 2006 (UTC)

Problem

Can anyone fix the problem of the inline refs doubling up on Gastric-brooding Frog. Thanks -- Froggydarb croak 01:32, 29 August 2006 (UTC)

Is it fixed now? (If it is, see WP:FN for an explanation of what I did.) Gimmetrow 01:38, 29 August 2006 (UTC)
Thanks -- Froggydarb croak 02:31, 29 August 2006 (UTC)

Spacing in references

What is the consensus on writing footnotes like this:

"Lorem ipsum dolor sit amet," he said in a 2006 interview.<ref>
{{cite web
| last = Bloggs | first = Joe
| url = http://www.joebloggs.com
| title = Joe Bloggs' Blog
| date = 2006-01-01
| accessdate = 2006-08-28 }}
</ref>

I haven't seen this spacing very often, but I find it much easier to read and edit. Omphaloscope » talk 03:24, 29 August 2006 (UTC)

You can pretty much do it however you want based on whatever editors of the article in question can agree on. Although I wouldn't describe huge six-line jumps in the middle of paragraphs as particularly easy to read or edit. Christopher Parham (talk) 04:12, 29 August 2006 (UTC)

Request for comments regarding intermediate sources

See Wikipedia:Requests_for_comment/Policies regarding intermediate sources. Andries 17:04, 3 September 2006 (UTC)

"Should intermediate sources be sufficiently reliable (that is is there good reason to assume that the intermediate sources correctly quotes or have undistorted copies of reputable sources) or should intermediate sources be reliable in itself? SeeWikipedia_talk:Cite_sources/archive10#Intermediate_sources concerning wikipedia:Cite#Intermediate_sources:_State_where_you_got_it 17:01, 3 September 2006 (UTC)"
I don't understand where this discussion is supposed to take place. The RFC doesn't seem to indicate. Am I missing something? - Jmabel | Talk 06:03, 4 September 2006 (UTC)
HELLO! Will someone please explain where the actual discussion is taking place? - Jmabel | Talk 19:33, 4 September 2006 (UTC)
Is it the discussion at Wikipedia talk:Reliable sources#A different proposed addition? Hiding Talk 19:59, 4 September 2006 (UTC)
No, it took place here Wikipedia_talk:Cite_sources/archive10#Intermediate_sources Andries 18:48, 7 September 2006 (UTC)
See here [1] Andries 19:08, 7 September 2006 (UTC)

Andries is a former webmaster and current "Main Representative, Contact And Supervisor" for the largest site on the world-wide-web that opposes the Indian Guru, Sathya Sai Baba. Andries is trying to include links to his personal website in violation of an ArbCom ruling at: Wikipedia:Requests for arbitration/Sathya Sai Baba. He is attempting to cite a link on his personal website that was duplicated from a Yahoo Group and use that link as a reference for critical and selective (only gives a portion of the text and not the full text) information against Sathya Sai Baba. You can also view This Thread on FloNight's page. It doesn't matter what is agreed here. ArbCom specifically ruled that it is inappropriate for editors to insert links to websites (they are personally affiliated with) and that link to critical sites containing negative, personal stories and original research against Sathya Sai Baba. Perhaps the policy is different on other articles, but ArbCom made a specific ruling on the Sathya Sai Baba page and Andries is trying to circumvent it here. SSS108 talk-email 19:09, 7 September 2006 (UTC)

It is unrelated to the arbcom case. This was already a matter of dispute before the arbcom case openened as can be checked in archive 10. Andries 19:11, 7 September 2006 (UTC)

Good. I have you on your word saying that. The argument you are making here is what you have been attempting to argue on FloNight's page and on the Sathya Sai Baba talk page. Better safe than sorry. SSS108 talk-email 19:13, 7 September 2006 (UTC)

So is This a Hypertext Encyclopedia or What?

Of course, I'm kind of a latecomer to all of this citing stuff, but it seems to me that one of the worst misuses of the cite-needed templates (in this case, {{fact}}) is stuff like this (from Red Shift):

The first redshift survey was the CfA Redshift Survey, started in 1977 with the initial data collection completed in 1982.[1] More recently, the 2dF Galaxy Redshift Survey determined the large-scale structure of one section of the Universe, measuring z-values for over 220,000 galaxies; data collection was completed in 2002, and the final data set was released 30 June 2003.[citation needed](Note 1) (In addition to mapping large-scale patterns of galaxies, 2dF also established an upper limit on neutrino mass.) Another notable investigation, the Sloan Digital Sky Survey (SDSS), is ongoing as of 2005 and aims to obtain measurements on around 100 million objects.[citation needed](Note 2) SDSS has recorded redshifts for galaxies as high as 0.4, and has been involved in the detection of quasars beyond z = 6. The DEEP2 Redshift Survey uses the Keck telescopes with the new "DEIMOS" spectrograph; a follow-up to the pilot program DEEP1, DEEP2 is designed to measure faint galaxies with redshifts 0.7 and above, and it is therefore planned to provide a complement to SDSS and 2dF.[citation needed](Note 3)


Notes:

  1. The fact refers to the final dataset release of the "2dF Galaxy Redshift Survey", but there is a link to an article on that survey in the same sentence. (I'll concede in advance that the fact isn't referenced in the 2dF Galaxy Redshift Survey article, but I don't think that damages this argument—the solution seems to me to be to put the fact in that article and cite it there.)
  2. The fact refers to the measurement goal of the "Sloan Digital Sky Survey" but, yet again, there is a link to an article on that survey in the same sentence.
  3. The design goals of DEEP2 are marked cite-needed and, yet again, there's a link to an article on DEEP2 right there.

(This is just the latest place I've seen this type of thing—but I've seen this "cite-needed-near-obvious-link" a lot.)

It seems to me that if article MNOP contains 15 facts, all cited, and article MNOP is linked from 200 other articles, all of which are required to do individual cites for the same 15 facts, that this encyclopedia is going to be buried under millions (billions? trillions?) of duplicative cites that no reasonable planetary population can maintain and that will drive the average reader crazy.

I guess I'm wondering what's the point of having a hypertext encyclopedia if every article is required to stand alone? CoyneT talk 01:24, 7 September 2006 (UTC)

Articles get deleted, merged, split, moved (renamed). There is no guarantee that linked articles have appropriate sources, either. So it is our policy to require reliable sources for everything in Wikipedia, and to not regard other Wikipedia articles as reliable sources. -- Donald Albury 03:19, 7 September 2006 (UTC)
There is no guarantee that a specific citation in an article references a reliable source, that the source is available to the reader for verification, that the source actually states the fact in the article, that the context of the source matches the context of the article (water boils at 100 C, but not on Mars), that the fact is still considered true, that edits to the article have not changed its meaning to the point that the citation is no longer valid, etc. By festooning our articles with dozens of footnotes we are creating an aura of authenticity, not necessarily the reality. The reality requires a substantial investment in checking each citation and each subsequent edit. This is far more likely to happen in a primary article on a topic than when a fact central to that primary article is incorporate in another article with its own citation. We totally lack the tools and editing staff to maintain footnotes at anywhere near the rate they are being created. There are thousands of well-sourced articles that are quite stable, watched over by a cadre of editors and unlikely to be deleted, merged, split, or moved. We need to leverage that irreplaceable investment to keep citation maintenance manageable. If current policy precludes that, it is a recipe for disaster. Do the math. --agr 21:47, 7 September 2006 (UTC)
Dalbury - (I think) you misunderstood the question. It is not being asked if we should treat Wikipedia articles as reliable sources (of course we shouldn't), but where we should put the citations to reliable sources that we should have. I strongly not duplicating citations if at all possible; of course the date of the final dataset release of the "2dF Galaxy Redshift Survey" should be cited in the article on it, not in another article that mentions that. However, there is nothing wrong with adding a {{citation needed}} tag to the other article - what should be done is to make sure the fact is cited (in the correct article), that the correct article is linked to the one with the cite needed tag, and then to remove the tag. I don't think there's a problem, here, although it may be good to clarify this. JesseW, the juggling janitor 22:34, 7 September 2006 (UTC)
Citing sources should keep the reader in mind. As a reader I would expect to find a citation to the appropriate original source at the point where I first encounter the material. In the Redshift article there is a note with a proper citation of the source of the discussion of the CfA survey; there is no such note for the Sloan Digital Sky Survey or the DEEP2 Red Shift Survey.
I can see going without a note if the reference were mentioned in the immediately preceding paragraph and it were clear that this is the source for the later material. It doesn't seem fair to expect the reader to know that the reference is in a totally different article. Unlike JesseW, I see duplicating links to original sources as essential to a dynamic and changing Hypertext encyclopedia. Yes, it takes work, but failure to document sources in a derivative article leaves this reader with the strong suspicion that the editor has merely copied material from an existing Wikipedia article without reading the original sources. If an editor read the sources, she should have no problem citing them; if he didn't read them, he shouldn't be writing an article. --SteveMcCluskey 14:46, 8 September 2006 (UTC)
See m:Wikicite, intended to have one definition of a source which would be invoked within needed articles. (SEWilco 17:47, 8 September 2006 (UTC))
Interesting. I was thinking of something similar, with a different implementation:
When Wikipedia was born (or shortly after; I don't know), the only citation style available was for web pages, but it was simple; you just write the URL in brackets, and the links are automatically numbered in order:
  • URNs and URLs are both forms of URI, and can be used in similar ways.[http://www.w3.org/Addressing/]
We could go back to this reference style for all references, not just web sites, by using URNs, and get rid of this difficult-to-use, difficult-to-read Cite.php clutter. For instance, To cite a book, you should just be able to do the same thing with the book's URN:
  • USB cables contain four wires; VBUS, D+, D−, and ground.[urn:isbn:0-471-37048-7]
Then the software would determine the citation information, either from an internal database (I was thinking it would be part of wikisource, but they call it "wikicat") or an external database like isbndb.com, and generate the citation dynamically according to whatever style the user selected in their preferences. — Omegatron 18:21, 8 September 2006 (UTC)
URNs would be a good fix, when they become implemented. But the URN article doesn't refer to the specificity of a range of pages for citations.
One problem, among many, is that early books do not have ISBNs and I don't know how the URN program plans to handle that. I'm a bit worried that writing a URN (in its full specificity) may become extremely difficult. For an example, how would an editor easily determine the URN of Albert Einstein, "Zur Elektrodynamik bewegter Körper", Annalen der Physik, 17(1905):891-921. At the moment, the best may be the enemy of the good.
When one considers the 220 plus pages in the Chicago Manual of Style dealing with the problems of citations in different fields, implementing the URN program and developing software to convert it into human readable citations looks like something that is far from being implemented.
Finally, I don't see why the Cite.php format is described as cluttered. It only becomes cluttered if one tries to implement it in Citation Template form, which I agree is a mess. Writing citations as footnotes or Harvard in-line citations is really quite simple, once you get in the habit of doing it. --SteveMcCluskey 18:50, 8 September 2006 (UTC)
If Omegatron's suggestion could be implemented, it would deal with some references, but not all. Some books are too old to have ISBNs. Software does not have ISBNs. Magazines have ISSNs, but these refer to the periodical as a whole and do not identify a particular issue or article. None of these numbers provide page numbers that a particular fact or quote came from.
Also, the citation style should be chosen on a per-article basis, not a per-user basis. Still, I favor finding some approach that makes it easier to edit text. --Gerry Ashton 18:52, 8 September 2006 (UTC)

For an example, how would an editor easily determine the URN of

I was imagining a pop-up search utility. Type in one or more fields in the search box, like title or author, and it would present a list of works. Select the one you intended and it would either give you the URN to copy and paste or insert it for you.

Finally, I don't see why the Cite.php format is described as cluttered. It only becomes cluttered if one tries to implement it in Citation Template form, which I agree is a mess.

No, it's really quite <ref>"On the Electrodynamics of Moving Bodies". ( fourmilab.ch web site): [http://www.fourmilab.ch/etexts/einstein/specrel/www/ Translation from the German article]: "Zur Elektrodynamik bewegter Körper", ''Annalen der Physik''. '''17''':891-921. (June 30, 1905) </ref> cluttered and difficult to <ref>Lorentz, H. A. (1904) "Electromagnetic phenomena in a system moving with any velocity less than that of light", ''Proc. Acad. Science Amsterdam'', '''IV''', 669-78.</ref> read around, even without citation <ref> {{cite conference | first = BalaSundaraRaman | last = L | authorlink = | coauthors = Ishwar.S, Sanjeeth Kumar Ravindranath | date = 2003-08-22 | title = Context Free Grammar for Natural Language Constructs - An implementation for Venpa Class of Tamil Poetry | booktitle = Proceedings of Tamil Internet, Chennai, 2003 | editor = | others = | edition = | publisher = International Forum for Information Technology in Internet | location = | pages = 128-136 | url = http://citeseer.ist.psu.edu/balasundararaman03context.html | format = | accessdate = 2006-08-24 | doi = | id = }}</ref> templates. It's atrocious. The whole concept of a wiki is that it uses simple, easy-to-use markup, accessible to the most technically-challenged Internet users. If we're going to continue using something in the style of Cite.php, it really needs to be reconstructed with input from the people who are forced to use it every day. For one, the content of the references needs to be in the References section, where it displays, so you can edit them directly, instead of hunting through the entire article for the first occurrence of each. And the links in the midst of text need to be small and unobtrusive.[cite:Zur Elektrodynamik] (Or something like that.[cite:Ein05c]) I've proposed changes like this elsewhere already.

None of these numbers provide page numbers that a particular fact or quote came from.

For page numbers in a journal article, you use the SICI number in the URN instead of the ISSN, for page numbers within a book, you'd use the BICI instead of the ISBN. For instance, to cite Einstein, the code would be:
Maxwell's electrodynamics, when applied to moving bodies, leads to asymmetries 
which do not appear to be inherent in the 
phenomena.[urn:sici:0003-3804(19050630)17<891:ZEBK>2.0.TX;2-B]
As opposed to:
Maxwell's electrodynamics, when applied to moving bodies, leads to asymmetries 
which do not appear to be inherent in the phenomena.<ref>"On the Electrodynamics 
of Moving Bodies". (fourmilab.ch web site): 
[http://www.fourmilab.ch/etexts/einstein/specrel/www/ Translation from the German 
article]: "Zur Elektrodynamik bewegter Körper",  ''Annalen der Physik''. 
'''17''':891-921. (June 30, 1905) </ref>
The SICI is still kind of a long citation, I guess, but nowhere near as long as writing all the info out by hand, and I've seen much longer URLs as web citations. If wikicite plans to have a central database to edit the citation and insert it in article with a shorter code, I wonder what identifier they were planning to use.

One problem, among many, is that early books do not have ISBNs and I don't know how the URN program plans to handle that.

it would deal with some references, but not all. Some books are too old to have ISBNs. Software does not have ISBNs.

True. On the other hand, we need to think of WP:V, too. You can't cite an ancient book that no one can get a copy of. ISBN covers everything printed or reprinted since 1966, and SICI covers every journal article, so that's most of what anyone would be citing. But yeah, there would have to be some mechanism for the exceptions, like a custom namespace.[note:''Title of article'', Author, year]

Also, the citation style should be chosen on a per-article basis, not a per-user basis.

No, it should be uniform from article to article. Since some people prefer styles that they are used to, it should then be a user preference and be consistent for that user from article to article. — Omegatron 23:44, 8 September 2006 (UTC)
Omegatron and I certainly have a different concept of simplicity. Compare the clarity and simplicity of the standard footnote style:
<ref>Albert Einstein, "Zur Elektrodynamik bewegter Körper", ''Annalen der Physik'', 17(1905):891-921.</ref>
with the obscurity and complexity of:
[urn:sici:0003-3804(19050630)17<891:ZEBK>2.0.TX;2-B]
As an aside, I cannot and will not defend the citation template style, of which Omegatron gave a typical example. Citation templates clutter articles and make editing of content extremely difficult. One could even add that citation templates suffer from much the same problem of the URN proposal, forcing everything into the same Procrustean bed of standardized format. Wikipedia includes scientists, computer experts, classical scholars, and fans of popular music and films. Writers in these fields have different ways of citing things, and the format of each article will reflect those varieties. Whether an article cites references in Harvard style or footnotes, I see no reason to change it. --SteveMcCluskey 15:10, 9 September 2006 (UTC)
All reference styles which include the citation information inline with the text clutter articles and make editing of content extremely difficult. It's not citation templates; it's the entire paradigm of Cite.php.
You wouldn't be typing in the URN by hand; you'd be given it by an easy-to-use search tool. (And books would be shorter, anyway.[isbn:0-306-40615-2])
If wikicite is implemented, it would need some similar type of identifier to indicate that something is referenced. I wonder what they intend to use?
There is no reason to use either Harvard style or CMS footnotes, when we have infinite possibilities for citation formatting. Why force everything into a horrible little abbreviated list of references, completely distanced from the text that references it, with barely enough information to find a copy, when we have no space or formatting constraints? The print editions of these articles can certainly use such formats, but for web viewing we have much broader possibilities.
What we really want is what Kirill referred to as a "more sophisticated piece of metadata", so the software can format the references in whichever way is preferred by the user. The identifier will be unique to a specific work, allowing people to cite the same thing uniformly, without any missing information, throughout the encyclopedia without having to search for and properly format the complete reference information each time, allowing people to create (Mediawiki, Firefox, ...) extensions for enhanced functionality without going through and adding things to each reference individually. If you want a list of 30 books at the very end of an article in the order they were cited, with partial author information, you can have it.
I, on the other hand, will select a style that allows me to view the references next to the facts they are cited by, as I'm reading, to view extended information that wouldn't be included in standardized styles like Harvard refs, with relevant excerpts, links to copies of the book/article I can read online, links to related publications by the same author, a list of other Wikipedia articles that cite the same reference, etc.
I was trying to think of a short, well-defined, standardized way to identify references for this purpose and discovered the URN concept, which fits perfectly. If you have a better idea, I'd love to see it, but Cite.php and standardized, abbreviated, "we've always done it this way" references are awful. See also URNs, bibliographic citations in web authoring. — Omegatron 18:34, 9 September 2006 (UTC)

Wikipedia:Citing_sources#Page_numbers

What does everyone watching this page think about the removal of material that is sourced from a book, but the editor did not provide page numbers. To clarify the material was added from a book which is a reliable source first. Some six months later page numbers are being asked for with the idea that if they are not produced the material will be removed. Is the removal of this material acceptable, unacceptable, or borderline?--Birgitte§β ʈ Talk 19:39, 8 September 2006 (UTC) The material in question does fall under one of the situations decribed as Page numbers must be included in a citation that accompanies. . . emphasis mine.--Birgitte§β ʈ Talk 20:05, 8 September 2006 (UTC)

Hey, that's not what I said. The situation was different. The question is would a citation count for a specific fact if there were no page numbers but just a general reference to a book? The question of whether a book provided at the end of the article is sufficient for every fact in the article is quite a different debate. I'm sorry if you were confused by this. JASpencer 20:24, 8 September 2006 (UTC)
I may still be confused. I asked if you would be satisfied by adding a footnoted ciation and you replied "If it had a page number for the claim, yes" [2] So if we add a footnoted ciation to the fact from the book, you will not be satisfied but not so unsatisfied you would still remove it. Is this correct?--Birgitte§β ʈ Talk 20:33, 8 September 2006 (UTC)
I would wait a bit longer but I would argue that they don't meet WP:CITE and so aren't citations. As to removal, well it would go through the talk page as at present. The implication of your previous post was that the assertions were individually cited and that I was complaining about those fictional citations because of the lack of a page number. Your position is more friendly to citations than the others on the page as you do seem to accept that a book announced at the end of an article is sufficient for a controversial article. JASpencer 20:58, 8 September 2006 (UTC)
Well I started off asking the question generally, because am really more interested in general interpretation rather than what happens to one particular article. Then realized it would be important to know that whether someone had just added the the statements, or if they had been sitting there for some time first so I looked up how long ago the statements we were discussing were added and didn't realize how that changed the framework of my question. I apologize for misrepresenting you, it was inadvertant.
I don't think a book announced at the end of an article is sufficient, but I would be more hesitant to delete than you and probably wait much longer. I would rather see a book at the end of the article worked into a better style of citations. Honestly if I couldn't contact the original editor, I would order the book myself through the libray before removing another's work (unless it smelled funny). If you look at the history on that talk I was trying to better clarify my postion as you replied, so I reverted myself to maintain the integrity of your answer.
I really just want see where consensus is on the removal of assertions that are individually cited but missing page numbers. I know that is not the current situation we are involved in. But I believe most of the statements you list could be brought up to this level with little effort. I would rather draw the map on what everyone agrees on first before jumping in however. I am not at all interested if people think your actions leading up to now are acceptable or not. Which unfortunately led to me paying too little attention to how I represented them. I just want to figure how we move forward through the current situation, which the first step is determining how to deal with individually cited assertions which are missing page numbers.--Birgitte§β ʈ Talk 21:24, 8 September 2006 (UTC)

(Unindent) As I read the guideline, any specific quotation or paraphrase should have a page number, or an equivalent method of making it easy for a fact-checker to find the relevant passage in the source. Any editor could remove a quotation or paraphrase on that basis. I personally would weigh how harmful or controversial the paraphrase or quotation was before removing it. I would be inclined to remove it if the paraphrase or passage

  1. Appears to be false or misleading on an fairly important topic
  2. Disparages a living person
  3. Disparages a dead person who played an important role in a present-day religion, political party, etc.
  4. Replaces a passage in an article, if the previous version were better-sourced

--Gerry Ashton 21:46, 8 September 2006 (UTC)

I think it would be simpler to create a tag someone can put into the text requesting a page number (without removing the quotation) and state that if a page number is not provided within a reasonable amount of time (we can either debate that here, or leave it up to editors working on an article) the quote should be removed to the talk page until a proper citation is provided. Slrubenstein | Talk 21:51, 8 September 2006 (UTC)
I oppose tags for something like this - such a request should go on the talk page. It isn't worth defacing the article over missing page numbers when the source has already been given. If someone has trouble finding the material in the book they can usually use the book's index or in many cases search the book online via Amazon or Google Print. dryguy 23:04, 8 September 2006 (UTC)
Unnecessarily strict when the book is short or has a reference of subjects or key words. Andries 22:39, 8 September 2006 (UTC)
And extending the WP:BLP for dead people is I think wrong. Andries 22:40, 8 September 2006 (UTC)
I agree that a page number is not the only acceptable method for finding a quote or paraphrase, and a citation-checker should make a reasonable attempt to use whatever means are available for a particular work.
as for extending WP:BLP to dead people, I only said I would be inclined to delete passages that disparage certain dead people. For me, it would depend on the degree to which disparaging the dead person also disparages the organizations that the dead person was associated with. --Gerry Ashton 23:09, 8 September 2006 (UTC)
If an organizations is stupid enough to have its reputation hinge around the reputation of a deceased person person living or dead then it should suffer the consequences for that. Andries 23:33, 8 September 2006 (UTC)

{unindent} It looks the removal of sourced material missing page numbers is a borderline issue. One idea is that it depends on the source (i.e. if it is indexed or not) I think that should not matter. If someone has the book in hand to know if it is indexed, they can then provide the page numbers. Another view is the removal is based on the nature of the material. Actually I think everyone can agree that if assertion is contraversial or if someone actually disputes the accuracy then it should be removed to the talk page until someone can provide the page numbers. If we just talk about mundane material. Assertions that are not actually disputed. That are sourced from a reliable reasource but were added some time before pages numbers were asked for. I personally think such material should not be removed. It is sourced after all and I don't think the convience of fact-checkers should outweigh the benefits of actually having the material in the first place. I think there could be some kind of tag put on the reference itself where it will not clutter up the reading of the article. I just don't understand the reason we would want remove sourced information that is not disputed by anyone, just because we may not have the page numbers in the near future. --Birgitte§β ʈ Talk 01:08, 9 September 2006 (UTC)

WP:V says "Any unsourced material may be challenged and removed." Not "Any unsourced material that has not been judged by another editor to be mundand may be challenged and removed." Any unsourced material.
As far as whether the material is sourced or not this is proved by whether it is cited. These are not cited.
I see no problem with removing the material to the talk page until someone can cite it. To turn the convenience argument around, the convenience of individual editors should not bend the rules of WP:V.
JASpencer 10:25, 9 September 2006 (UTC)
I completely agree with the removal of unsourced material. However we are talking about sourced material. Material that was taken from a known reference and added to and article before that article had any citations using page numbers. WP:V does not say uncited may removed, just unsourced. Perhaps this is the real question "Can material be sourced if it is not cited?" I believe the answer is yes. Although the refusal to properly cite a source when adding new information should be viewed as unacceptable, we can not declare that all previously sourced material must now be considered "unsourced" because the article is being converted to citation-style.--Birgitte§β ʈ Talk 11:04, 9 September 2006 (UTC)
But where does it stop? You could enter some really dubious material and then say that it's from the book, or a book, "at the end". Far easier to cite the information. Otherwise it can go to the talk page. JASpencer 11:07, 9 September 2006 (UTC)
I agree that this may yield some problems in the case the book is lengthy an unindexed, but this is somewhat exceptional. Andries 11:10, 9 September 2006 (UTC)
You reply too fast. I was changing my mind slightly :P If someone has been using the "book at the end" approach convert it to citations. What is the problem with that? We have everything but the page numbers and the first person to fact check the article can add those. Where it stops is when the the arrticle is converted to ciatation-style. Where is stops is when the info is disputed or contraversial. You cannot add dubious information and say it is in the book at the end. No one is supporting that. --Birgitte§β ʈ Talk 11:16, 9 September 2006 (UTC)
A small problem with that is that Noel is not claiming that all the uncited facts come from the biography, but that some (most?) do. So if you or I entered them in we would not know that they were coming from where we were claiming that they were from.
However as you said before, if you have an indexed copy of the book - how hard can it be to put the citations in so that it meets WP:V?
JASpencer 12:53, 9 September 2006 (UTC)
I don't what you are getting at above. A review of Noel's edits are certainly in order and Noel is around to answer any questions. Maybe we will find problems, maybe we won't; it is hard to say without doing the review. I am not suggesting that we should mark all the uncited facts as being from that book without any further inquiries. I do not understand what you mean by However as you said before, if you have an indexed copy of the book - how hard can it be to put the citations in so that it meets WP:V? Obviously if someone has a book on hand get the page numbers, I am only questioning what to do when months have past since someone had a book open adding info to WP. I don't see what page numbers have to do with WP:V. What do you exactly mean when you say "meets WP:V]", because that policy is mainly talking using reliable sources and that the editor adding the info should give their reference, which has been done.--Birgitte§β ʈ Talk 21:58, 9 September 2006 (UTC)
I have started a section for such a review if anyone here is interested in following how this works out. Talk:Marcel Lefebvre#Noel's edits--Birgitte§β ʈ Talk 22:23, 9 September 2006 (UTC)

cite sources that represent a fair and balanced view

if it's ok with everyone, i would like to propose that we add a section entitled "cite sources that represent a fair and balanced view" The problem is that various topics have one major viewpoint and a minor viewpoint or multiple viewpoints. I've run into wikipedians who are citing a onesided viewpoint or citations from a minor viewpoint and trying to state that it is the absolute truth... hopefully this will prevent a one-sided view of the world...Kennethtennyson 17:24, 9 September 2006 (UTC)


  • The specific phrase "fair and balanced" is pretty loaded, and at this point has become identified with a news source many see as slanted. Without entering that debate, would it make sense to change the terminology to something else? "Objective," "even-handed," "neutral" might work? Jameson 23:44, 2 October 2006 (UTC)

Difference between Cite journal and Cite confrence

I have not been able to figure out the difference between Template:Cite journal and Template:Cite conference. I realize the second is meant specifically for the Procedings of confrences, but what, in terms of template arguments, or, output style, is the difference? I presume there is one, which is why I'm posting here. Thanks! (copied to Wikipedia talk:WikiProject Wikicite also) JesseW, the juggling janitor 05:41, 12 September 2006 (UTC)

GA: Looking at the documentation contained on the talk pages for cite journal and cite template, I found the differences listed below; I put the different attributes at the end of the table.

conference template journal template
first first
last last
authorlink authorlink
coauthors coauthors
date date
year year
month month
title title (REQUIRED)
pages pages
url url
format format
accessdate accessdate
doi doi
id id
booktitle —
editor —
others —
edition —
publisher —
location —
— journal
— volume
— issue

I don't recommend templates; I was just curious to see how citations between the two might be different. --Gerry Ashton 17:57, 12 September 2006 (UTC)

Thanks; but I'd put most of those differences down to a lack of development of cite confrence; it's still not clear to me what the difference in style might be. JesseW, the juggling janitor 23:39, 12 September 2006 (UTC)

Concordia Libraries link

I'm getting nothing if I click on the wikilink of this in Technical issues with footnotes section:

  1. Many of today's style guides require not using or recommend against using footnotes and reference endnotes to cite sources (Concordia Libraries)

Not sure how to fix it.qp10qp 01:30, 14 September 2006 (UTC)

This is an attempt at Harvard referencing, where the author's last name and the year of publication are put in the text in parenthesis, after a quotation or paraphrase. If you look at Template talk:ref_harvard you will see there is a template available to use in the text instead of simply typing the reference information as plain text. There is a matching template, ref_label, that is supposed to go in the references section. The project page has an entry for Concordia Libraries, but there no ref_label in the Concordia Libraries entry in the references section, so no connection is established. If everything had worked as intended, you would be able to click back and forth between the reference section and the text where the reference is used.
I have not used templates for Harvard references, and I don't know how well they work, or which method is best. I don't know if it is worth fixing or if it would be better to just make it a plain Harvard reference without a template. --Gerry Ashton 02:10, 14 September 2006 (UTC)
Oh, I see. (It didn't look like a Harvard reference; they usually have a date and page number, but maybe that's just for books, I don't know.) Without being blue, it will just look like something random in brackets, won't it?qp10qp 03:10, 14 September 2006 (UTC)

Reference neccesary for short articles without controversial information?

Would it be appropriate for a biographical article that contains nothing more than DOB, names of spouse/children, and education and key positions, with dates to not contain a single reference?

This information is obviously taken from somewhere, and the truthfulness of the facts could be easily verified by a simple web search. But the information is also not controversial. Does an article that contains just this basic information need a reference? Patiwat 18:37, 14 September 2006 (UTC)

Yes, all articles need references; this really isn't negotiable at this point. Kirill Lokshin 18:44, 14 September 2006 (UTC)
Sure does -- here's an example of a short article that most likely will not grow any larger and just has the basic info (and nothing controversial) but still has its sources -- it's an article I created yesterday: George Rogers Clark Floyd --plange 18:49, 14 September 2006 (UTC)

Where to place citations in a paragraph?

Where should the citations for the following paragraph be placed?

This sentence summarizes the paragraph. This fact came from source A. This fact came from source B. This fact came from Soure A.

Like this:

This sentence summarizes the paragraph.A B This fact came from source A. This fact came from source B. This fact came from Soure A.

or like this:

This sentence summarizes the paragraph. This fact came from source A.A This fact came from source B.B This fact came from Soure A.A

or like this:

This sentence summarizes the paragraph.This fact came from source A.A This fact came from source B.B This fact came from Soure A.

or like this:

This sentence summarizes the paragraph. This fact came from source A. This fact came from source B.B This fact came from Soure A.A

or like this:

This sentence summarizes the paragraph. This fact came from source A. This fact came from source B. This fact came from Soure A.A B

-- Patiwat 10:09, 16 September 2006 (UTC)

Depending on the length of the paragraph, I'd go with either the second option or the last one; as a general rule, a citation must come after the material it applies to, not before. Kirill Lokshin 15:22, 16 September 2006 (UTC)
My preference for the last option rises in direct proportion to the number of facts cited from each source. It gets ridiculous to have a footnote on every sentence in an article in cases where one or two footnotes per paragraph are sufficient. dryguy 15:34, 16 September 2006 (UTC)
One point to note would be that the last option works best if the footnotes are pure citations; if there are discursive remarks also present, combining them for an entire paragraph is likely to be confusing. Kirill Lokshin 15:40, 16 September 2006 (UTC)

Thanks a bunch for the suggestions, dryguy and Kirill Lokshin!

One followup question. If work A is used as a reference for a large part of an article, and works B-F are used as references for some specific parts of an article, then where should I put the citation for work A? At the end of every substantial fact, at the end of every substantial article, at the end of just some substantial facts? Patiwat 17:52, 16 September 2006 (UTC)

In my opinion, it is best to reference specific facts, assertions, or whatever as you go along, rather than trying to make a single reference cover a large part of the article at once, because the reader tends to associate footnote marks with particular sentences or paragraphs in the text. So the sequence of citations might go AAAABAACAAAADEAFAAAA. I would reference A at the end of information culled from one page or sequence of pages in a book (12-16) rather than at the end of every sentence. The reader checking the source will grasp that intuitively.qp10qp 18:08, 16 September 2006 (UTC)
Can you cite any serious publication that makes a practice of repeatedly citing the same reference on several sentences in a row? I believe that is non-standard for a very good reason - it is distracting to the reader to have footnotes applied continuously throughout an article when one or two footnotes at the beginning of an article and on key points will do just fine. No serious publication that I am aware of puts a footnote on every sentence. A far more common practice is to group the relevant citations and place them on one or more key sentences in the introduction, something like this.1-6 dryguy 15:22, 17 September 2006 (UTC)
I just re-read what you wrote - obviously I missed where you said put the citations at the end of the paragraph. dryguy 15:32, 17 September 2006 (UTC)
Also, unfortunately, we have to be more citation happy than a publication who knows who they hired to write the article. Citations help editors combat misinformation being slipped into an article, as it should be easy to spot new info going in without a cite and require they provide one. --plange 16:15, 26 September 2006 (UTC)

Which template for streaming media

Hi, can someone tell me which cite template to use when linking to a streaming video/audio? -- Lost(talk) 17:12, 25 September 2006 (UTC)

When not to cite

IMHO mechanical application of WP:CITE and the desire for a cureall results in absurd situation, e.g. the recent idea that WP:GOOD articles do require by all means in-line cites. I very much hold that there a situarions where giving in-line cites is not only unnecessary, but misleading. Top example are overview articles about established area of science.

In special relativity, should we give a cite for:

two events happening in two different locations that occur simultaneously to one observer, may occur at different times to another observer

This can be found in literally every textbook on SR, possibly hundreds of books. If giving just one book reference, a contributor would not only promote a specific book without reason, but leave the possible impression, that only this book supports the sentence, other books possibly not. Also for this example the verifiability argument seems rather silly: If someone needs a reference to verify the correctness of this sentence, I would rather prefer someone other to check the correctness of the article.

Pjacobi 15:41, 26 September 2006 (UTC)

We could really use some feedback from seasoned editors here on the new changes to the GA criteria, which all seems to be pointing back to issues that have to do with WP:CITE..... --plange 14:05, 27 September 2006 (UTC)
WP:CITE is a guideline, the policy that you are trying to follow is verifiability. There are many different ways to cite information, from simply listing a general reference at the end of the article which supports a large part of the information, to actually providing footnotes to the reader. In situtations where an article contains many facts that are basic to the subject and broadly accepted, it might be worth listing a few general references at the end of the article, i.e. some widely used undergraduate texts, and not bothering with footnotes. When such material is challenged, you still ought to provide a specific citation for it, but that need not be done with a reader-visible footnote. You could use {{inote}} or an HTML comment, or simply note the source on the talk page. Christopher Parham (talk) 22:31, 27 September 2006 (UTC)

Date wikification in cites

I'm not sure exactly where this belongs, please point me elsewhere if this discussion either belongs or has taken place before. In many articles (such as Tupac_Shakur which is where i noticed it), the publication dates of referenced books etc are wiki links to the date in question. This is neither helpful or consistant as far as i can tell, but i cant see how to edit it, and it seems to be done in a lot of articles so maybe i shouldnt anyway, can anyone clarify this? Provider uk 18:28, 26 September 2006 (UTC)

If you wikilink a date in ISO format (YYYY-MM-DD), user preferences allow the date to be displayed 4 different formats, which is why wikilinking dates is considered useful. (See http://en.wikipedia.org/wiki/Special:Preferences#prefsection-4 .) Some templates wikilink certain date entries automatically. Armedblowfish (talk|mail|contribs) 21:26, 26 September 2006 (UTC)
the problem with wikilinking dates is that it colorizes and highlights them, thus misleading users into thinking the editors recommend clicking on a link. That tends to disfigure an article and mislead the users. Rjensen 00:05, 27 September 2006 (UTC)

Citing newspapers articles which are accessed indirectly

I got this in Factiva:

Starbucks Sued for $10 Million by Glen Cove, N.Y., Woman By Jamie Herzlich, Newsday, Melville, N.Y. Knight Ridder/Tribune Business News 717 mots 30 décembre 2003 Newsday (KRTBN) (N.Y.) Anglais Copyright (C) 2003 KRTBN Knight Ridder Tribune Business News

It looks like KRTBN aggregates the contents from many newspapers: so the date (December 30, 2003) is probably the date of the original article in Newsday, but as I accessed the Newsday article only through KRTBN, should I write, in the References sections, something like "through KRTBN"? Or only Newsday, December 30, 2003? Apokrif 16:33, 27 September 2006 (UTC)

When not to cite Part Deux

We're still debating over at GA our attempt to apply this guideline to our GA criteria. Can we get clarification from the seasoned editors here on whether the following is acceptable?

statements which are undisputed within an academic field need not be inline cited

as that's what we're being asked to make an exception for. We'd appreciate any feedback we can get on this heated debate, thanks! --plange 20:03, 27 September 2006 (UTC)

That's utterly ridiculous. Even if we ignore the necessity of citations for avoiding accusations of plagiarism, copious inline citation is still necessary so that readers/editors from outside the academic field in question can figure out where certain statements are being drawn from.
To give a somewhat more useful example than the silly "the sky is blue" stuff some editors there are throwing around, here is a statement that is entirely undisputed with a certain (but quite obscure) academic field:

Having sacked Brescia, Gaston de Foix returned to the Romagna and laid siege to Ravenna.

Even though the statement is undisputed, its correctness (and even what it refers to) is not at all obvious to someone not familiar with the topic in question; without a direct citation, it becomes extremely difficult for a reader to determine whether this is a valid statement, or merely something a random vandal inserted into the article. Kirill Lokshin 20:15, 27 September 2006 (UTC)
I happen to agree with you, but we're getting hammered over there on GA about this very thing. We're getting flooded by physicists from WP Physics who are upset with our new criteria to require inline citations and are saying we're going overboard on what we're asking. Now I'm at work and so sneaking in time (shh!) so haven't had a chance to look at special relativity to see if what is asked for is out of bounds or not, but I will as soon as I get off so I can weigh in. I do know the editors asking for cites are working in good faith to apply our new criteria though. --plange 20:30, 27 September 2006 (UTC)
That's more of a Catch-22 issue with the idea behind GA itself, I think. The only way they can produce consistent results using single-reviewer methods is to spell out all the criteria explicitly; but the need to do so requires that changes go through an absurdly bureaucratic adoption process (which are prone to the sort of try-to-vote-away-fundamental-policy behavior we're seeing here), rather than simply evolving as needed. (Compare this with FAC, where the explicit criteria only make some mention of "appropriate" inline citation, but where, in practice, an undercited article hasn't a snowball's chance of passing.) Kirill Lokshin 20:46, 27 September 2006 (UTC)
It is overboard. Show me any serious publication that applies inline citations to every fact. It isn't done because the harm to readability would far outweigh any minor benefit gained from adding citations for trivial, non-controversial statements. dryguy 20:38, 27 September 2006 (UTC)
Show me any serious publication that "anyone can edit," then! Wikipedia is something fundamentally new; trying to apply traditional standards of citation to it tends to be somewhat unproductive. Kirill Lokshin 20:46, 27 September 2006 (UTC)
In the same vein, applying standards of quality is essential. Quality encompasses readabiliy, verifiability, grammar, style, etc. Sacrificing all other aspects of quality in the name of verifiability isn't the right path. dryguy 20:56, 27 September 2006 (UTC)
The neat thing about cite.php references, of course, is that it would be quite easy to add a CSS feature to hide them entirely, for those readers who preferred not to see them; having more of them has, at worst, a minor and temporary effect on readability. Kirill Lokshin 21:02, 27 September 2006 (UTC)
Sound very interesting. Is someone working on an implementation? dryguy 21:04, 27 September 2006 (UTC)
The CSS code itself is quite trivial. If you add the following to your monobook.css (or other skin CSS of your choice), you'll hide the footnote numbers and the rendered footnotes (although there may still be an empty "Notes" section in the article):
.reference {
   display: none;
}

.references {
   display: none;
}
In terms of having this more functional, though (e.g. being able to click a button to show/hide the notes), I'm not sure; JavaScript is not really my strong point. Kirill Lokshin 21:35, 27 September 2006 (UTC)
Special relativity is indeed the prime example. It gives an overview over a stable field of science. Standard textbooks are listed in the reference section. Next any assertion in the article can be found in next to any textbook. And: Someone who hasn't read one of these textbooks or has had a course in SR, isn't good in deciding whether or not some insertion in the article is a valid statement. You know, (a) vandals can make up references which are hard to disprove and (b) there are any number of physics cranks in the wild who can be used as sources fro invalid statements about SR (not all of which can be easily detected by being just some AOL-homepage).
It's an unfounded illusion, that everybody can verify articles on every topic.
Pjacobi 20:39, 27 September 2006 (UTC)

Citing every sentence will make Wikipedia into something other than an encyclopedia

The argument that because Wikipedia is an encyclopedia that everyone should edit so every statement in Wikipedia should be cited is ludicrous. If that's really the case then we should be fact tagging every sentence in Wikipedia. I'm beginning to believe that there is a group of editors that actually believe this. Citing every sentence is not only amateurish, it is a completely different task to the one of writing an encyclopedia. It's a reference librarian's dream but it makes normal, content-focused editors sick to their stomach. Citing controversial, foundational, or novel pieces of information should be encouraged. Citing mudane and elementary facts should be discouraged. --ScienceApologist 21:30, 27 September 2006 (UTC)

I suggest we edit the page to include the statement that elementary facts should not be cited. There are reasons for this. Not the least of which is we should not cite a fact that has dozens, hundreds, or thousands of potential references because we would be biasing our referencing to particular sources. That's unacceptable, in my opinion. If we can find a thousand books that illustrate the pythagorean theorem, it would be inappropriate to choose one that does it to the exclusion of the others. --ScienceApologist 21:34, 27 September 2006 (UTC)

I'm with you on not citing every sentence, but are you seriously suggesting that a significant fraction of the reliable sources explaining the pythagorean theorem present significantly different views on the subject? dryguy 21:48, 27 September 2006 (UTC)
No, he's just saying that, if there are thousands of sources, why would you cite a particular one? -- SCZenz 22:00, 27 September 2006 (UTC)
If they are all the same, why does it matter? If you are really concerned, choose a notable one. In the present example, cite Pythagorus or Tufte or both, but there is no reason on Earth to cite every work that explains the pythagorean theorem. dryguy 22:06, 27 September 2006 (UTC)
But if I cite "Mathematics by Jane Smith, PhD" and not "Mathematical Exercises by Jane Doe, PhD" I'm tacitly endorsing Smith's book over Doe's book. It's basically an advertisement for Smith over Doe. Why should we be in the business of endorsing textbooks? --ScienceApologist 18:35, 28 September 2006 (UTC)
Name one serious publication that when citing a reference, cites every single reference in existence that applies to the topic. It isn't done because it isn't necessary. It would not add value. These are not advertisements or endorsements, they are references cited to back up a statement. The practice of citing references is used by literally thousands of reputable, reliable publications. Show me one that has stopped doing so over concerns about tacit endorsement. dryguy 18:47, 28 September 2006 (UTC)
Most serious publications when citing references only require them for specific ";;;citable" points not general knowledge facts. That's why we need to clarify when citations are not appropriate. For example, if I write a paper for publication in a physics journal and write "A Galilean transformation cannnot be used in situations where the relevant velocity is a significant fraction of the speed of light", no "serious publication" will demand a cite for this reference. However, there are editors here who would tag this statement with {{fact}} and demand a reference! It's these sorts of elementary, non-controversial, and "common knowledge" facts that we are talking about here, not the points that would normally get cited in "serious publications". What many editors don't seem to realize is that there are a huge number of statements that "serious publications" absolutely would not reference and would criticize authors for attempting to reference. There reasons for this are that it's amateurish and that a reference to a single source to the exclusion of other sources serves as an unnecessary endorsement of one particular treatment of a subject over another. When the statements of fact are about general, common knowledge it is considered improper to cite these statements! --ScienceApologist 20:29, 28 September 2006 (UTC)
This is true because of space limitations and printing costs for paper journals. Also, for journals that use inline citations, citing trivial statements hurts the typography unneccesarily. I'm with you on the need to limit inline citations of trivial, non-controversial statements. I'm especially with you regarding the nefarious {{fact}} tag. However, given that Wikipedia is not a paper journal, the concerns about space limitations don't really apply, and unless you can state some other specific objection, I see no harm in allowing people to include citations for anything, as long as they don't bomb the article with inline junk. Finally, I challenge you to show me the author guidelines for even one publication that discourages citation on the basis that it would be improper endorsement. No publication does that. dryguy 22:33, 28 September 2006 (UTC)
I cannot point to author guidelines that state anything as specific as Wikipedia regarding citations, so this challenge is a bit of a red herring as it is not very helpful for either of our opinions. If I'm mistaken in this, point me to author guidelines that are as detailed as this guideline page with respect to citations only. Anyway, the nature of the beast is that Wikipedia is much more explicit about its rationale than most serious publications' guidelines. Let me put it this way, I'm certain that if an author submitted a journal article for review with inline references to a single textbook for elementary facts, this would not be looked upon very favorably by reviewers for reasons close to the improper endorsement reasons I outlined. Even if the journal was an online journal with unlimited space. --ScienceApologist 04:34, 29 September 2006 (UTC)
I've never seen an article in a research journal that provided inline citations, or even general references, for facts that would be found in an undergraduate textbooks. But I can't imagine that the reason has anything to do with endorsing one source over another; many facts that are worth inline citations can be found in more than one source, but I've never heard of any rule that if you cite one source, you have to cite them all. Indeed, this is such an extrordinary claim that I think it is up to ScienceApologist to prove that such a policy exists, and it is fair for everyone else to presume that no such policy exists until proven otherwise. --Gerry Ashton 04:50, 29 September 2006 (UTC)
I agree with ScienceApologist. Citing every sentence in articles on stable scientific subjects is neither feasible nor, in my opinion, particularly desirable. -- SCZenz 21:57, 27 September 2006 (UTC)

As my reputation is ruined anyway, I can do away with politeness: The main problem is, that not everybody can contribute (to the content) of every article. Or has somebody here discovered a magic pill which will do away with the necessity formal (or even informal) education? In-line cites are not this pill. For quite a large number of articles, no amount of inline cites can make an article 100% verifiable for somebody without prior knowledge of the topic. For one, you cannot judge wether the topic is misrepresented by simply leaving important things out? Next obstacle will be the ability to judge on the reliability of the sources. And without some practical training, you may not even see the equivalence between two different formulation. --Pjacobi 22:10, 27 September 2006 (UTC)

No one is an expert in all areas, but fortunately, for most areas, there are experts contributing to Wikipedia. I don't see the need to make every fact in every article reviewable by non-experts (nor am I convinced that would be possible). The experts who write and maintain the articles are usually doing a good job already. If the physics community at Wikipedia is opposed to inline citations on every fact, perhaps it is for a very good reason - they already have a style that works for them. For non-experts that need to verify something, chances are that they are going to have some work ahead of them no matter what style of citation is used. Physics can be a hard subject, and sometimes there are no shortcuts to learn the difficult parts. dryguy 22:34, 27 September 2006 (UTC)
The same is true in mathematics. The content of most math articles is often covered by any of several textbooks on the topic, and mostly in the introductory chapters at that. Rarely a research paper or two might need citing. Verifying that a cited statement is justified by the source text often requires some training in the subject. Articles on elementary math, on the other hand, are generally common knowledge. Inline cites should only be used where they are helpful for the reader and that, in my opinion, demands judgement, not one-size-fits-all dicta. --agr 23:23, 27 September 2006 (UTC)

I increasingly wonder whether people who object to dense citations on the basis of readability are being frank. I agree that the present system <ref>, </ref>, <references/> system creates typographic ugliness and interferes with readability, but that's a technical and a UI design problem and it can be solved.

Yet people keep saying "let's have fewer citations" instead of "let's find a better way to present densely-cited material."

I can think of all sorts of workable approaches. Here's one that's used in books, but doesn't carry over well to Wikipedia. Laura Hillenbrand's Seabiscuit: An American Legend is very densely cited, and perfectly readable. It was a bestseller that was made into a movie. It uses a simple technique I've seen in other books. There are no visible footnotes in the text itself. Instead, at the back of the book, are notes, organized by chapter and page number, with each simply introduced by the phrase it is referencing. If you don't care about the references, you don't even need to know they exist. If you do, you can look them up virtually as quickly as if there were numeric superscripts.

I'd like to see a system in which reference markers in the markup identify the whole range, not just the end, of the passage being referenced. I'd like to see references displayed, not with superscripts, but with subtle color differences... which could be turned off by a user setting. If it's turned off, the text appears uniformly black. If its turned on, referenced text is black, unreferenced text is some darkish color that is distinguishable from the referenced text, but just barely. To go to the reference, right-click on the text.

Is that feasible? I don't know. But the existing reference system seemed like an impossible dream until it was implemented a year ago.

Let's be clear about this. Are the people objecting to references really objecting to it on the basis of readability, or is there a faction that actually likes the idea of an encyclopedia that rests on the personal authority of anonymous contributors? Dpbsmith (talk) 23:46, 27 September 2006 (UTC)

There is a component of irritation that you want us to spend five times as long finding spefic page numbers as it would take to simply improve a stub to a quality article, and that a result Wikipedia will have five times fewer quality articles in our area of expertise. We're not relying on our personal authority, and I'm insulted at your implications—we're simply saying that noncontroversial facts are adequately verified if they appear in general references at the bottom of the article. To claim non-experts should be able to verify every Wikipedia article, line by line, is rubbish; we could fool you if we wanted to, even if we did cite every souce. So yes, you have to trust, collectively, the people at Wikipedia:WikiProject Physics not to fuck you over: you're doing that whether you know it or not, and all that you're doing with the rigid inline citation requirements is making more work for everyone. -- SCZenz 00:00, 28 September 2006 (UTC)
Dpbsmith - believe it or not, my motivation is concern for the quality of Wikipedia. If you feel my motivations are otherwise, then state why you think so and give me a chance to address your accusations. I don't appreciate having you imply that my motivations are to hurt the verifiability of Wikipedia. I also believe that the majority of people voicing frustration with the overuse of inline citations are doing so in good faith.
Quality encompasses more than verifiability. I do not object to numerous or dense citations that are done intelligently. I object strenuously to dense inline citations that are applied to every sentence, or worse, multiple times per sentence. I don't object to people adding citations in support of trivial facts. I object strenuously to people who insist on adding an inline cite or a {{fact}} tag to every statement lacking an inline cite. Such behavior is absurd, and we don't need rules that actively encourage it. I strongly object to the addition of rules to require such a ridiculous practice. We would be better off with guidelines that gave people a clue about what quality actually means: verifiability, reliable sources, intelligent prose, sensible typesetting, etc. Wikipedia has been widely criticized for poor quality, and has responded well in the area of verifiability, but in my opinion, needs some attention to many of the other factors that determine the quality of an encyclopedia. dryguy 00:34, 28 September 2006 (UTC)
In defense of Agne who put those fact tags on the article in question, she was asked to on the talk page - she did not swoop in and fact-bomb the page as has been implied elsewhere. --plange 01:41, 28 September 2006 (UTC)
No, she didn't. First she informed people that the article would be "failed" and no longer "good" unless they provided inline citations, without giving any specific issues. This was even more problematic. -- SCZenz 01:44, 28 September 2006 (UTC)
True, though she was just a messenger and acting in good faith. Let's move past the method of how you were informed and focus on finding consensus. WP:V is a core principle and though you might be a good mentor for the page in question, and know full well that the whole article is backed up by the references at the bottom, what happens if you leave?? The article might start accreting pseudo-science (shudder) and by the time someone else comes along to monitor the page (who might not be a phyicist) they will have to look in each reference at the bottom to find which backs up each fact/opinion. Now, instead, if there were cites for these, the editor only has to check that one to see if it backs it up and that someone hadn't snuck one by us. The example that Kirill gives in the section above (as it relates to history) is a good one. Also, I'd like to test our review process so if you could submit the Special Relativity page for review so we can see if it does meet the new guidelines, that would be great. --plange 01:58, 28 September 2006 (UTC)
I don't think anyone intends to attack Ange. I agree wholeheartedly that we should move past that specific incident, but some hard-working editors did take (unintended, but very real nevertheless) offense at the generic skim-for-inline-citations-and-announce-immininent-delisting approach taken by the Good Articles WikiProject. We need a clearer approach to what GA asks for, in order to avoid such problems in the future. But yes, we should try for consensus; we're not there yet, though, and I don't think GA requirements should reflect things without consensus support. See also my comments here; they are heartfelt, including the apology, and I think working more calmly through the confusion and mutual frustration is very important at this point. -- SCZenz 02:07, 28 September 2006 (UTC)
Cool. So what do you think of Kirill's example above as an illustration of why cites are needed for things that are only common knowledge to experts in the field? --plange 02:13, 28 September 2006 (UTC)
I looked at the example, and the point is well taken; but I disagree on the conclusion of what the best way to move forward given Wikipedia's current state is. Given the relatively limited number of experts here, in my field at least, I think there isn't time to cite every fact that might seem to be of uncertain veracity. Nor does citation guarentee that a non-expert can/will verify a statement; who's to say a vandal won't cleverly fake a citation, perhaps by referring to a very difficult paper for an incorrect assertion. Believe me, even I can't read most string theory papers; thus I'd rather have some string theorists watching string theory than citations for every line. I think experts' time is more effectively spent improving poor content (of which there is much, in physics), than adding the citations at this time. That's a statement of wanting to use limited volunteer time effecitvely, rather than laziness; perhaps I'll reconsider my position some day if the facts on the ground change. -- SCZenz 02:23, 28 September 2006 (UTC)
Also note that often even properly-cited claims aren't verifiable to people without university-level computer access. For example, in my rather puny stub Crystal Ball (detector). (It isn't in-line cited, but it could be, since the ref section includes which references refer to which part, and it's only a few sentences.) I've included proper citations and web-links, but the papers can't be read without paying thousands of dollars to Elsevier or being at a university that does. -- SCZenz 02:33, 28 September 2006 (UTC)
Ease of verification isn't a requirement, however, just that it ultimately is. Yes, vandals will find ways to get around things, but it's a lot easier to spot when things have citations.... Also, can you guarantee that you will always have a string theorist in perpetuity watching your article? I understand your point on the other (re: your time). If you think your time is best spent improving content, that's great. There's no requirement that says you have to be "GA-stamped" in order to do so. What we're attempting to do at GA is bring our process more in line with what is required at FAC so that GA articles have a better chance there ultimately (after a good peer review, etc). If there are so few editors for your area, perhaps participation in the GA process shouldn't be a priority for you at this time. Once you get things out of stub status and feel caught up, then perhaps it won't feel so overwhelming to go the next step, etc. Thoughts? --plange 02:44, 28 September 2006 (UTC)
I have several thoughts. I'll bullet and sign them separately for easier discussion of them separately (if desired). -- SCZenz 02:59, 28 September 2006 (UTC)
  • I agree that verifiability is the ultimate issue; that's why I claim that "verifiable by any subject expert" is sufficient for well-settled facts in which the references are available but not in-line cited. It seems peculiar to me that you have no problem with "verifiable only by people with obscenely expensive website access" but object to my position. Can you explain? -- SCZenz 02:59, 28 September 2006 (UTC)
  • Obscene amounts of money being (somewhat) easier to acquire than an expertise in string theory, the former is more verifiable than the latter. (The ideal version, of course, is for everything to be verifiable by a decently educated layman through generally available sources; but this may not always be possible, obviously.) Kirill Lokshin 03:18, 28 September 2006 (UTC)
  • I think it should be verifiable by a non-expert. WP:V says that "any reader" should be able to verify. At least if it's cited, they know where to go, no matter if it is expensive, or only in D.C.'s National Archives. We're writing for non-experts, not experts. Experts on special relativity are not going to be looking up special relativity on WP. Writing for experts is the domain of peer-reviewed journals and the like. How do I know that such-n-such is an established fact in your field and not some pseudo-science slipped in if I'm reading what special relativity is about and what makes it so special? It might seem silly and embarassing to you to cite it because you live and breathe this stuff every day, but we don't and the average reader of your article won't either. --plange 03:32, 28 September 2006 (UTC)
    Any reader can read the textbook introductions to special relativity. It's hard work, but so is going to the National Archives. Looking at SR fact-by-fact without an understanding of the whole does not enable effective verification anyway, I don't think. -- SCZenz 03:39, 28 September 2006 (UTC)
  • If we lose our current expert volunteers (who include string theorists), the string theory article is in trouble regardless. One way we might lose expert volunteers is by telling them that articles they thought they've done a good job on and could leave for a while are now officially not "good" anymore. -- SCZenz 02:59, 28 September 2006 (UTC)
  • No one has delisted articles yet - we were only notifying of a potentiality not an eventuality. --plange 03:32, 28 September 2006 (UTC)
  • GA has changed, and I don't think it's for the better. It used to be a significant step below FAC, now a new generation has made it into a category of "ready for FAC" articles. I don't know that the recategorization is a good thing, especially since it's frustrating to article authors. It is also very bureaucratic, another source of frustration. A year ago we used Wikipedia:Peer review to prepare for a FAC. -- SCZenz 02:59, 28 September 2006 (UTC)
  • I get the sense that some people interpreted the negative reactions to the original GA idea as resulting from standards which were too weak, and have been trying to rewrite them into something unnecessarily complicated, when the real objection was that the standards—whatever they happen to be this week—are being inconsistently applied. Kirill Lokshin 03:18, 28 September 2006 (UTC)
  • I think bring it closer to FA is a good thing. Right now, GA is not held in very high esteem (or even good esteem) because of the lax rules. We're attempting to improve this so that the GA stamp will mean something (as well as help editors get closer to FA). --plange 03:32, 28 September 2006 (UTC)
    But GA doesn't mean anything if it's the same as FA, and it seems to be very close to that right now. -- SCZenz 03:39, 28 September 2006 (UTC)
  • It's not the same, it's just that we're trying to make it actually fit in actuality and practice, where it is placed in the 1.0 assessment scale. Above B and below A. Cites are basic things to require, but the nuances and fine-tuning of articles to get them to pass FA still makes FA more difficult to attain, as it should be. We're trying to say what's good, not what's fantastic. Fantastic covers what's good, but also includes nice tight prose, well-crafted and balanced leads, etc., etc. Peer reviews are so much easier to sink your teeth into and help perfect when the article is a good solid G or A-class article. --plange 04:30, 28 September 2006 (UTC)
  • Participation in GA isn't a priority for us. For my part, if people really want GA to be pre-FAC, and FAC is demanding in-line citations, then WikiProject Physics might do better to use its own internal rating system (which exists, and which the GA label can be taken out of easy). However, in my personal view if things go in that direction we should ask that physics articles not receive GA-related notices of labels of any kind; if the system can't be applied properly (and if we're not sure it's needed at all), then we shouldn't have to deal with generic, possibly ill-thought-out GA-related requests on talk pages. -- SCZenz 02:59, 28 September 2006 (UTC)
  • So use the "A-Class" rating in your in-project assessment scale as the mark to aim for and ignore the GA process entirely, then. (But keep in mind that if any physics articles are taken to FAC, a reasonable level of inline citation will almost certainly be asked for.) Kirill Lokshin 03:18, 28 September 2006 (UTC)
    Fine. FA's should be superb; I've made the effort to produce one, it was hard work, and I would expect nothing less in the future. -- SCZenz 03:39, 28 September 2006 (UTC)
  • One further point that I think people may have missed: inline citation does not mean, per se, that every sentence must be cited. It would be quite possible, for example, to have a single footnote for an entire paragraph or section (e.g. "The material in this section is summarized from Doe (1943), Smith (1972), and Jones (1996)."); while this would not be as rigorous a citation as may be desired (at least for some topics), it would at least give readers and other editors some indication of where to look if they want to track down the sources. Kirill Lokshin 03:25, 28 September 2006 (UTC)
  • This appears to be a misunderstanding on the part of the GA bureaucracy, one member of which claimed Mathematics may not be adequately cited. It has a cite in just about every single section, I believe. Also, the requests on special relativity were not handled in this way, and our reaction might have been far less frustrated if it had been. Perhaps GA criteria should be amended to make this clearer. -- SCZenz 03:39, 28 September 2006 (UTC)
  • Those were only added to the talk pages of pre-existing GA articles. GA editors are not asking that all articles go through us, etc. If you guys do decide to go this route (not participate) then if you want to avoid us reviewing your articles that already have a GA label on them, then you might want to remove them. Otherwise we're going to assume you do want to participate, etc. --plange 03:32, 28 September 2006 (UTC)
To add to Kirill Lokshin's points, I've been reading some of the arguing here and elsewhere, and I must say that this chimera of a reference for every line has been invented: there's no policy or good-article principle that requires it, there just isn't; it is a straw man (OK, a straw chimera). Referencing is an instinctive thing; you know when to reference and when not to. The secret lies not so much in the text of the article as in the text of the referenced book: you may put in one reference to cover one or more paragraphs; then as soon as someone checks the reference, they see straight away that the material intervening between two tags is actually covered. In other words, ten sentences might have gone by without another reference mark appearing in the article, but those ten sentences may all be covered in the referenced text. I see nothing wrong, by the way, in giving a range of page numbers (Smith, 2004: 67-76); this prevents the article from becoming clotted with tags and is similar to some entries in indexes where page ranges are given: you find the spot and read on.
I can't think of an article where, say, six successive sentences would require six tags to six different reference works, since the articles here usually represent blocks of knowledge that whole books or articles have already been written about (an exception, I admit, is contemporary POV-war political articles, perhaps, which are indeed often ruined by tag pile ups since no agreed version exists). No-one's asking for a reference a sentence, honest. Even where Agne, perhaps because she was asked to on the Talk page, put some fact tags on the special relativity article, a smart editor could cover several of those in a single inline reference if they were covered in a particular section of a referenced book.qp10qp 03:58, 28 September 2006 (UTC)
I say this in all honesty: the impression I have gotten from some GA editors is that citation of individual lines is precisely what would be required in some cases. If we can put a link to the same textbook at the end of every section in special relativity, with a different page number range each time, then that's very different than what I (and others) thought you were asking for. It's really very similar to what we were already doing; giving the source(s) that could be used for the article's basic facts at the end. -- SCZenz 04:12, 28 September 2006 (UTC)
(edit conflict)Another straw dog I've seen in some of the arguments (not on this page, but in other areas where this is being argued) is that we're requiring a certain magic number of cites per article (a certain density), or that since we're not, we can't even require them in the first place. One person even said that since we don't specify a number, then since special relativity had two inline cites, then it should pass since it has some inline cites. This had nothing to do with whether the article had references where it was needed. And if a reviewer goes overboard in their requests for cites, well, I'll keep saying this until I'm blue in the face I guess, but we do have a review process for cases where an article is failed and the nominator doesn't agree with the reasons. This would be precisely the time to protest whether the reviewer was being unreasonable and allow others to judge if this is the case. --plange 04:15, 28 September 2006 (UTC)

In response to Kirill saying "it would be quite possible, for example, to have a single footnote for an entire paragraph or section (e.g. "The material in this section is summarized from Doe (1943), Smith (1972), and Jones (1996)."); while this would not be as rigorous a citation as may be desired. . .", you could make that style rigorous enough by adding the page numbers to those three books. The Chicago Manual of Style has examples of how to do this at any level of precision and detail you wish in a single footnote.qp10qp 04:20, 28 September 2006 (UTC)

Actually, my point about the one-per-section citation not being as rigorous was more related to the fact that greatly spaced-out citation make it more difficult to detect later, unreferenced additions. Depending on the nature of the article, this may or may not be a significant problem. Kirill Lokshin 04:27, 28 September 2006 (UTC)

In reply to Plange's comment about density as a "straw dog" (after edit conflict) Well, the wording on WIAGA is entirely ambiguous in this regard, now isn't it? Some of the cite requests we were given looked totally random, and did not reflect the controversial or specific statements in the article, so density was the only thing I could think of that was being asked for. This is why we should be clarifying what requires a cite and what doesn't. -- SCZenz 04:23, 28 September 2006 (UTC)

In reply to SCZenz (not that comment, the one before: help, I can't keep up!), the point is that, yes, it is similar in principle to what you were already doing in that the article was already genuinely verifiable by the sources you named at the bottom of the page; but in practice, inline citing with page references marks an advance because anyone who wants to check specific parts of an article against specific parts of the reference text is enabled do so. It is however a nightmare trying to check a part of an article if you have to read through an entire list of reference works blindly trying to find which bit of which book references which bit of the article: believe me, I've actually done that on occasion. qp10qp 04:35, 28 September 2006 (UTC)

Above, Dpbsmith 23:46, 27 September 2006 (UTC) implies that, without numerous inline citations, the encyclopedia "rests on the personal authority of anonymous contributors". The basic point of many of us in the sciences is that encyclopedia will always rest on the authority of the experts that contribute to it, no matter how carefully the articles are cited. I disagree with Plange's statement that "Experts on special relativity are not going to be looking up special relativity on WP. " In practice, enough experts read through the articles as part of the editing process that errors are found and important citations added. It's not that random people off the street are writing articles on etale cohomology. Krill worries about false unreferenced additions being added; this is exactly what the history and watchlist systems prevent.

In my field, mathematics, a perfectly cited article could easily contain many false statements and false interpretations. Thus we question why there is a push to go beyond the plain language of WP:V by requiring the basic facts to be cited inline rather than cited by a reference at the end of the article. The more recent threads above, about allowing one inline citation per paragraph, miss the point: the original argument was that no inline citation (with a page number) at all should be required for basic facts. References at the end of the article are, in the minds of many, in agreement with WP:V. Adding a sentence like Standard references are Smith [2000] and Jones [1985]. is easy enough, given that these book are at the end of the article, and I would view that as an acceptable compromise on inline citation. Looking up page numbers in the books takes precious time and only gives the illusion of more accuracy in the article. CMummert 10:44, 28 September 2006 (UTC)

I completely agree with the above statement. Providing an abundance of inline citations for what are generally accepted or universally acknowledged facts is amateurish, undergraduate, and unencyclopedic and it runs counter to the principle of implied authorty that accumulates to an article through the rigours of the editing process. Eusebeus 14:18, 28 September 2006 (UTC)
You guys are missing the point. There is no implied authority. In an encyclopedia where anyone can edit, we have no guarantee that you or anyone else is an expert, and we're therefore asking you to make appropriate citations. I happen to be a trained historian, but I modify my writing style here to conform to the nature of the beast-- no one here knows who I am or that I am trained. We're anonymous. We're unvetted. We must cite, therefore, because someone coming along after me has no idea who I am and where I might have gotten my information. Yes, experts may be reading your article as editors, but that is not whom you are writing for. In distilling some of your points about why it seems distasteful to you to make you cite things it sounds like you would be embarrassed that these are cited for other experts reading them would find it funny. That doesn't matter. That isn't who you're writing for. You are writing for us dumbos :-) I'll also repeat (I guess I'm still blue in the face), that if someone goes overboard and asks you to cite something as basic as the sun sets in the west, which I don't believe anyone has asked you to do, you do have the WP:GA/R process to go to to protest this. --plange 14:30, 28 September 2006 (UTC)
I disagree. I don't think we are missing the point at all. Any article that is the subject of active and involved editing (which is presumably a sine qua non of getting to GA status in the first place) will, as a matter of course, receive the attention of enough people knowledgeable in the field to vet the issues that you raise which effectively confers the kind of implied authority through aggregated editing and discussion that is achieved elsewhere through authorial standing. No one is suggesting that sources should not be cited as needed; but basic facts do not require inline citations. Because you are not necessarily aware of the basic facts in one or another field does not make them any less basic for that. Eusebeus 15:14, 28 September 2006 (UTC)

Suggestion for common knowledge exception

It was suggested to me that a proposal, expressing the point of view of editors of scientific articles that have been affected by recent changes to WP:WIAGA, should be put forward here. I cannot claim to speak for the community, but here is a suggestion. Please leave comments and assume good faith. I will not respond until tomorrow.

WP:V states that facts must be verifiable, not preverified. I suggest that a clarification be made in WP:CITE that facts that are obtainable in the generic undergraduate textbook on a particular topic need not be given a specific inline citation, unless one is explicity requested on the talk page. Moreover, editors acting in good faith should not request such citations unless the fact is believed to be disputed, not included in the generic textbook, or is a quote or attribution of opinion of a particular source or person. Thus lack of inline citations for uncontested facts is not a defect in the article.

CMummert 01:17, 29 September 2006 (UTC)

I don't think this is an unreasonable guideline. However I don't know what it would really solve, as it will just push the aside the issue until someone tries to take an article up to Featured Status. Often this happens as soon as GA is achieved. --Birgitte§β ʈ Talk 01:43, 29 September 2006 (UTC)

I am totally in favor of a common knowledge exception being made explicit in CITE guidelines. This would solve numerous headaches and allow us to get back to that sword of Damocles of writing the encyclopedia. --ScienceApologist 04:26, 29 September 2006 (UTC)

Proposal about common knowledge exception

Above, it was proposed that we make explicit that inline citations to elementary facts are not requirements. However, it is useful, because there are people who are under alternate impressions, to illustrate where science and math editors actively support inline references. After all, not everything is an elementary fact that doesn't need citation. In fact, there are elementary facts which actually may require citations! I'll try to illustrate these issues with common knowledge in this section.

The common knowledge facts that do not require a citation are "consensus" facts. There are other facts that require citation, namely foundational ideas, proprietary knowledge, or uniquely synthetic measurements/observations/derivations/proofs. That is to say there are a lot of things that may be "common knowledge" that should still be cited. I'll take examples from redshift since I've been working on that article for nearly a year now:

  1. Foundational ideas: The Doppler effect is named after Christian Andreas Doppler who proposed the effect in 1842.[2] this is common knowledge, but because there is a foundational statement attributal to Doppler himself, it deserves an inline citation to this work. What would be inappropriate would be to cite a textbook that said this fact. The primary source is of citable interest for this fact, but since there are hundreds if not thousands of textbooks which state this fact, citing one of them as a source for this quote is inappropriate.
  2. Proprietary knowledge: It is also the dominant cause of large angular-scale temperature fluctuations in the cosmic microwave background radiation (see Sachs-Wolfe effect)[3]. as a sort of extension to the foundational criterion for citation above, if a principle or concept is known as being proprietary to a particular group it should be cited as such to the research that established this or, barring that, a reference that gives a proprietary nod to the people considered. Again, general texts that describe this fact would not be appropriate to cite here.
  3. Uniquely synthetic measurements/observations/derivations/proofs: The largest observed redshift, corresponding to the greatest distance and furthest back in time, is that of the cosmic microwave background radiation; the numerical value of its redshift is about z = 1089 (z = 0 corresponds to present time), and it shows the state of the Universe about 13.7 billion years ago, and 379,000 years after the initial moments of the Big Bang[4]. this is common knowledge, but because it was a unique measurement accomplished by one collaboration (WMAP) an inline citation is appropriate. This would also apply to, for example, unique mathematical proofs attributable directly to a particular mathematician or group (e.g. from Poincaré conjecture: "In late 2002, Grigori Perelman of the Steklov Institute of Mathematics, Saint Petersburg was rumoured to have found a proof.[5]")

Of course, direct quotes and paraphrasing of ideas must be cited as per usual guidelines, but all other elementary facts and derivations that are considered common knowledge and are easily found or derived from relevant texts would be considered inappropriate to cite for a few reasons:

  1. Choosing a particular text is a means of endorsing that text to the exclusion of others.
  2. Elementary facts are applicable and sourced in a variety of contexts. Citing a basic noncontroversial elementary fact to a particular source will often have a context that is misleading (or worse controversial) since "trivial" facts are often assumed and not explicitly stated in more advanced and accurate works. For example, statements regarding the conservation of energy might be most easily cited only to very basic science texts where the surrounding text is outright inaccurate or misleading. Textbook errors and even intentionally introduced misconceptions are extremely common in basic science treatments. This might cause problems for people trying to research the ideas associated with the particular elementary fact. An extreme example of this might be citing a creationist who states that a fundamental principle of science is observability. The unwanted contextual baggage would be detrimental to an article written about subjects other than creationism.
  3. Sometimes elementary scientific or mathematical facts are not easily citable in the same way that opinions, historical facts, or political platforms. That's because the precision of mathematics and scientific theory is found in its formal presentation (the rules of mathematics and logic) rather than precise statements in a verbal language. So while finding a precise citation to the historical fact that the Battle of the Bulge started on December 16, 1944 is straightforward, finding a precise citation that "A reference frame is a point in space at rest, or in uniform motion, from which a position can be measured along 3 spatial axes." (from special relativity) or "\nabla_b \, {T_a}^b = \partial_b \, {T_a}^b + {\Gamma^b}_{cb} \, {T_a}^c - {\Gamma^c}_{ab} \, {T_c}^b = 0, where \nabla_b is a covariant derivative." (from general relativity) might not be tenable because the supporting references might offer trivial restatements of these facts that may actually be only indirect citations. If they are truly elementary facts, it is inappropriate to subject individual lines to citation lest it is assumed that the statement should be found directly in the text. This is beyond paraphrasing, this is a matter where simple mathematics or logical manipulation may be required of textbook statements in order to get to the text being cited. The solution to this quandary could be to clarify that citations in articles might not directly state the elementary fact they are supporting, but I think it is better to simply not cite the elementary fact if a standard textbook in the references would have the elementary fact implicit in its text (perhaps throughout the entire text!)

Notes

  1. ^ See the official CfA website for more details.
  2. ^ Doppler, Christian, "Beitrage zur fixsternenkunde" (1846), Prag, Druck von G. Haase sohne
  3. ^ Sachs, R. K.; Wolfe, A. M. (1967). "Perturbations of a cosmological model and angular variations of the cosmic microwave background". Astrophysical Journal 147 (73). 
  4. ^ An accurate measurement of the cosmic microwave background was achieved by the COBE experiment. The final published temperature of 2.73 K was reported in this paper: Fixsen, D. J.; Cheng, E. S.; Cottingham, D. A.; Eplee, R. E., Jr.; Isaacman, R. B.; Mather, J. C.; Meyer, S. S.; Noerdlinger, P. D.; Shafer, R. A.; Weiss, R.; Wright, E. L.; Bennett, C. L.; Boggess, N. W.; Kelsall, T.; Moseley, S. H.; Silverberg, R. F.; Smoot, G. F.; Wilkinson, D. T.. (1994). "Cosmic microwave background dipole spectrum measured by the COBE FIRAS instrument", Astrophysical Journal, 420, 445. The most accurate measurement as of 2006 was achieved by the WMAP experiment.
  5. ^ By April 2003 the press was reporting on these developments Mathematical Digest

There are numerous ways to deal with these issues, but they are real issues that are currently not discussed in this guideline. We need to either modify the statements to deal with elementary facts being not necessary for citation, or make some sort of detailed explanation as I have outlined above.

PS... if any other editors want to modify this essay, please do. It's meant to be didactic and illustrative of the problems science and math editors have been having with "CITE" issues, nothing more.

--ScienceApologist 05:22, 29 September 2006 (UTC)

Let me address the three numbered items above. For item foundational claims, I think that if the references include a paper by Doppler dated 1842, anyone interested in verifying the claim of the article would know to look at that paper; no inline cite is needed if the reference is included in the references section. For proprietary knowledge, I don't see why just a wikilink isn't enough. For experiments and proofs, I agree that specific results of experiments should have inline cites; these facts should be treated like direct quotes. And a claim that a particular person proved something should be cited inline, because it is an attribution, although I don't think that a link to a newspaper that also claims the person proved it is helpful for verifying that the person actually proved it. A link to the actual research paper, or to a peer reviewed paper or book, would be more trustworthy.
I don't think the advertising for one textbook argument about inline citation is very strong. I would just pick one or two sufficiently general texts (which probably makes them graduate level) and include them in the list of references at the bottom.
The common knowledge exception isn't just for science articles. For example, Henry I of England has no inline cite for the years of his reign, and the exception I am proposing would say that no such inline cite is needed so long as the years are not disputed. That article has references at the bottom which I am sure could be used to verify the years if somebody wanted to. CMummert 12:40, 29 September 2006 (UTC)
I'm glad you mentioned the Henry I example here, because it illustrates a point I've tried to get over in these discussions, one which if grasped would I think make this whole worry go away. If you referenced a historian for the reasons the White Ship sank, which are disputed, the book used would certainly count also as a reference for the dates of Henry I's reign. The same principle is true for a science article: without overtly referencing a consensus fact, you incidentally do so when appropriately referencing a related, consequent, or dependent fact.
I think ScienceApologist's suggestions are fine and would satisfy good-article requirements. There's nothing to stop science or maths editors from agreeing common-knowledge exceptions amongst themselves, but of course they mustn't expect that to be Wikipedia policy for all topics and they mustn't expect good-article status if they don't inline cite at all (a negative policy of the form "You do not have to..." is never going to make it to the list of requirements because it would create more problems than it solves). I'm sure, however, that articles following the suggestions of ScienceApologist above would be more than able to achieve good-article or featured-article status, because we all know that no-one would withdraw qualification if common-knowledge facts weren't cited inline so long as the appropriate ones were. Smart citing can cover even common-knowledge facts incidentally, since they would be assumed in the text referenced for the fact that was cited (this is a simple little point that I've tried to get over before, because it takes away this chimerical notion that consensus facts have to be specifically cited yet satisfies all verifiability principles on Wikipedia).
And I must go back to the basic point (which I suspect isn't fully addressed here) that this isn't about verifying scientific truths at all, but about referring to texts that say the same thing that we say in the article. The idea is to ring-fence Wikipedia from accusations that because it can be edited by anyone it is an unreliable source of information. qp10qp 13:14, 29 September 2006 (UTC)

I agree with qp10qp's comments above. I think it is possible to work within the existing GA and even FA system. I think most science and mathematics editors are basically on the same page as most of their editors – it's just that some recent events have had a bit of a chilling effect. Here is the rough standard I use for inline citations and which I think most Wikipedians would find reasonable:

  • At the start of the article, cite a prominent review or textbook or two from which many of the (relatively speaking) "elementary" facts in the article can be mentioned.
  • If speaking about a particular discovery or experiment, or is named after someone, cite the relevant paper, always. (This covers ScienceApologist's 1 or 3 above.)
  • If a wikilinked topic is mentioned in passing (as in ScienceApologist's 2 above, the Sachs-Wolfe effect) it is often best not to provide a reference if the Wikilinked article is appropriately referenced (i.e. the Sachs-Wolfe effect already provides a reference to the original paper). This eliminates redundancy and is in the spirit of not sending people out of Wikipedia for material that can be found inside.
  • Similarly, a section written in the summary style usually does not need footnotes if the daughter article has them.
  • Finally, articles that can be verified entirely using one or two books, papers or reviews do not need inline citations. (I know that these sorts of articles lack "balance", but on some subjects it is hardly a problem, and certainly better than no article at all.)

I think that one of the great things about Wikipedia is that anyone can figure out to edit it easily. Articles that are too densely cited are difficult for all editors to work with, and can look particularly daunting to new editors. –Joke 13:58, 29 September 2006 (UTC)

I find the guidelines just above to be perfectly reasonable. The difficulty is that these common sense guidelines are not part of any policy (official or unofficial) and may people on the GA discussion pages feel that, because of this, they are justified in asking for an inlince citation for things such as the reign of Henry I. When pressed, they say that WP:CITE supports them, when at best WP:CITE is unclear about the issue. Does nobody else feel that it might be worth adding some clarification? CMummert 14:34, 29 September 2006 (UTC)

New proposal

I suggest we add the following text to the citation article:

"Common knowledge facts do not need to be directly cited when they are verifiable by a large number of sources."

What do people think?

--ScienceApologist 21:08, 29 September 2006 (UTC)

--ScienceApologist 21:08, 29 September 2006 (UTC)

If an editor feels that some of the audience for an article would benefit by having a citation for a common knowlege fact, the editor should feel free to add the citation. You can't legislate readability. --Gerry Ashton 21:21, 29 September 2006 (UTC)
No, but you can have a comment in the guidelines that will help fend-off actions such as this [3]. --ScienceApologist 21:32, 29 September 2006 (UTC)
(edit conflict) I'm not quite sure what you wanted to propose, but the text as written is utterly ludicrous. We may, in some cases, not require citation; but under no circumstances should we ever actually forbid it. Kirill Lokshin 21:23, 29 September 2006 (UTC)
This statement has been modified to avoid the equivocation you caught. Thanks, Kirill. --ScienceApologist 21:32, 29 September 2006 (UTC)
(edit conflict) Shouldn't you discuss that on WT:VERIFY? Personally, I'm not going to go around removing things that look like unverified but probably/certainly true facts, but in the grand scheme of things (think entire world, including not only first-world countries, but also countries like Sudan, Iraq, Egypt, China, Brazil, etc.) I don't think there is much "common knowledge". If I caught an unverified but "common knowledge" (in my own culture) fact, it might stop me from supporting the article in an FAC, or lead me to add a {{fact}} tag (unless I just added a citation as Gerry Ashton suggests). Armedblowfish (talk|mail|contribs) 21:30, 29 September 2006 (UTC)
This proposal doesn't prevent you from doing anything you state. It's simply a statement to avoid problems like we were having over the issues at Wikipedia Talk: Good articles. --ScienceApologist 21:34, 29 September 2006 (UTC)
The statement is pretty vague in that it neither defines "common knowledge" nor says anything about what kind of sources are being discussed. In other words: "common knowledge" verifiable from 20 "Introduction to Science" books is quite different from "common knowledge" verifiable from twenty papers about particle accelerator design, even though both are "verifiable by a large number of sources." Kirill Lokshin 21:44, 29 September 2006 (UTC)
How about this:
"Inline citations are not required on non-controversial statements that are corroborated by more than three of the sources cited by the article."
dryguy 22:36, 29 September 2006 (UTC)
I don't think it's a simple matter of counting sources, I think it's a matter of how hard is it to look up. If an article references one or more well indexed textbooks, that suffices to look up many common knowledge facts. A reader who has the background to read the article in the first place can probably figure out which facts are common knowledge and look them up in the index of the textbook. Dryguy's version does not mention the fact being common knowledge. Suppose an article references 28 newspaper articles, and also mentions an obscure fact that is mentioned in 3 of the 28 newspaper articles. It would be a substantial burden on the reader to figure out which newspaper article the fact is drawn from, because there is no way to guess which, if any, of the newspaper articles is a comprehensive source of general knowledge about the topic.—Preceding unsigned comment added by Gerry Ashton (talk • contribs)

How about this? --plange 23:02, 29 September 2006 (UTC)

Inline citations for common knowledge items are not necessary. Common knowledge facts are those that appear in at least three general reference textbooks for the field, all of which are either listed in the references section or in a General Reading section of the article

I think three textbooks is overkill; one is enough. University professors who actually teach in a certain field may receive piles of new textbooks for review, but those of us who have to pay for them will probably only have one on each topic. Also, editors have better things to do than verify that everyday facts are mentioned in each of three textbooks. --Gerry Ashton 23:17, 29 September 2006 (UTC)
Okay, maybe all three don't have to be listed, but we need some threshold to define what is common knowledge. I know in graduate school I didn't have to cite something if it was present in at least 3 books, etc., so I think if we're going to go down this road, we need to have a definition of common knowledge that applies to all types of articles and disciplines and so saying that a person can consider something common knowledge if it is present in at least 3 general survey textbooks-- that way some crackpot can't complain that some wacko psuedo science is common knowledge since it's in their POV text book they somehow got printed. --plange 23:22, 29 September 2006 (UTC)
A fringe textbook can be brought into question as such, by showing it contradicts widely accepted textbooks, or that the publisher has a reputation for publishing junk. Without a doubt, someone will include information from fringe books, with or without inline citations. That does not mean the rest of us should have to reinforce our computer tables just so the table will hold 100 kg of books. --Gerry Ashton 23:33, 29 September 2006 (UTC)
Why not just "multiple" textbooks? After all it is conceivable that there are some points found in single textbooks, and if they are by my criteria above they are probably proprietary and should be cited anyway. But if the facts are in "multiple" textbooks, then it's pretty clear they are common knowledge. --ScienceApologist 23:39, 29 September 2006 (UTC)

Why is any action needed here? Isn't this well covered within the How and where to cite sources section?

  • When writing a new article or adding references to an existing article that has none, follow the established practice for the appropriate profession or discipline that the article is concerning (if available and unquestioned).
  • An article's previous content contributors usually know the established practice - if possible, follow their lead if the article already has references.

Unless i'm missing something WP:V does not require that all sources need to be provided as inline citations within the article. If an editor challenges a specific claim, and it would look ridiculous to add a citation to the article, then what is wrong with simply providing a source on the talk page? If an editor does not challenge any specific statements in the article but is just looking for a larger quantity of citations, then why not just mention the "How and where to cite sources" section and point out that the article citations are there to serve the reader. The ability to provide a source for every claim of the article is awaiting a technical solution.EricR 23:36, 29 September 2006 (UTC)

Oh that it were so simple. Except the main editor of the very page you cite as ridiculous has been fact bombing us and putting up citation notices due to this very issue. We want it spelled out crystal clear in the guidelines so this kind of problem is less likely to emerge in the future. --ScienceApologist 23:39, 29 September 2006 (UTC)
If editors are reading the policy to mean that every statement of the article needs an inline cite, or that some arbitrary number of citations are required for an article to be "good", then yes some clarification is needed. It's probably important to distinguish between WP:V and WP:CITE though. If i were to challenge some "common knowledge" statement, take the reference frame example above (don't the axes need to form a basis?), you might say i have no idea what i'm talking about (and be right), but in the end i should still get a cite if i really wanted one. That doesn't mean though that the statement needs a citation in the article.EricR 01:14, 30 September 2006 (UTC)
I agree with this sentiment completely and I think that the proposed wording doesn't contradict what you are saying. --ScienceApologist 15:39, 30 September 2006 (UTC)
1) Which is easier: to check that a fact exists in three cited textbooks and then do nothing... or to check that it exists in one and cite it?
2) How does any reader know which uncited facts have been found in three textbooks? An uncited fact found in three textbooks doesn't look any different from any uncited fact added yesterday by someone who didn't bother to check any sources.
3) How does a reader know which textbooks were the source, if the reader wants to check? What happens if the textbooks used as the source are subsequently removed from the list of references?
I've seen articles that specifically said, above a list, that no name should be added to the list unless the linked article confirmed that the person was a member of the list. On checking, about half of the linked articles said nothing whatsoever about the fact in question. Someone thought a name should go on the list, so they just added it, and they or someone else typed square brackets around it. And when it's done, it looked just the same as the linked articles that said the person was a member of the list... without citing a source... which, in turn, looked just the same as the linked articles that said the person was a member of the list, and cited a source.
And lists that do not have any citations have a strong likelihood of including at least a couple of dubious or prank entries. Dpbsmith (talk) 23:42, 29 September 2006 (UTC)

It should be noticed that "Common knowledge facts" is a culture-dependent notion. For instance, some mathematical elementary facts are asserted and proved in almost all school books in some country while not even mentionned in the school books of some other country. How should such facts be quoted ? pom 14:56, 30 September 2006 (UTC)

There is always a context to what is common knowledge and what is novel, but this is true whether you include an in-line cite to the fact or not. If you are arguing that the threshhold for what is and isn't a common knowledge fact is unclear, let me point out to you that no one is prohibitting people from citing "common knowledge" facts or discussing and coming to consensus what facts are common knowledge or not, we just are saying that citing common knowledge facts is not required. --ScienceApologist 15:36, 30 September 2006 (UTC)


I boldly added the common knowledge exception to the article. Pray to the gods of consensus. --ScienceApologist 15:42, 30 September 2006 (UTC)

While I am open to one general, catch all cite for common knowledge at the end of a section for the text book stuff, I am concern about the ambiguous nature of "Common Knowledge". There is a difference between the "Sun rises in the East" which chances are your average Joe would know and "Einstein developed general relativity to apply the principle generally, that is, to any frame, and that theory includes the effects of gravity. Special relativity doesn't account for gravity, but it can deal with accelerations." which any science major would undoubtly assume is common knowledge but would get blank stares if asked during a "JayWalking" segment. One can truly be considered "Common Knowledge" while the other is "Common Specialized Knowledge" which places a burden on the reader to either have a specialize background or to have any of the "3 or more textbooks" that this information is common in. Again, I have no problem with not having every single line cited but I do think we should give the reader something that is at least tied into the pertinent section of information. Agne 16:36, 30 September 2006 (UTC)

Also, there is a slight difference between "Common Knowledge" and "Folklore", which corresponds to well known facts that are either too obvious or that have been re-discovered so many times that no one knows where they come from. pom 22:56, 30 September 2006 (UTC)

User: Count Iblis recently made the following change (added words in bold):

Inline citations for common knowledge items are not necessary and should be avoided. This makes articles easier to read and does not hamper verifiability.

I have removed the new words because the new passage only considers the issues of verifiablity and readablity. It does not consider another merit of inline citations: telling the reader where to look for more complete material on the subject at hand. An editor might very well decide that letting the reader know precisely where to look for more information is more important than a slight decrease in readability caused by the presence of a note number. --Gerry Ashton 23:08, 30 September 2006 (UTC)

I replaced "all of which are either listed in the references section or in a General Reading section of the article"[emphasis added] with "all of which are listed in the references section of the article"[emphasis added] "Further reading" (or "general reading") is for material which has not been used as sources for the article (see here). I also specified "uncontroversial" - WP:BLP still applies, and inline citations are especially helpful for WP:NPOV issues.
I also added "If someone requests a citation, for example by adding a {{fact}} or {{verification needed}} tag to it, take this as a sign that the fact was not as "common knowledge" as you thought." No sense in arguing whether or not something is common knowledge when you can easily add a citation that is already listed in the references section.
Personally, I think this is an especially bad idea in cases where all of the supporting references are inconvenient to access (offline or requre a subscription). And if a source is used to write most of the article, it shouldn't be that hard to cite it once a paragraph (or once a section, if you want to stretch it), unless you are still new to Cite.php formatting. It may also be a bad idea if you have a lot of references for the reader to look through. In addition, I would really be more comfortable if we said to avoid doing this for facts that are only common knowledge within certain cultures or areas of expertise.
On the other hand, inline citations are not technically required by WP:VERIFY, but I do believe they become increasingly useful as the size of the article and number of references increases.
— Armedblowfish (talk|mail|contribs) 00:21, 1 October 2006 (UTC)
While inline citations are becoming more and more popular, they are more important to some research cultures than others. The problem is when the cultures that love large numbers of citations begin to demand a level of citation that is considered unreasonable by another research culture. Inline citations are great and used all the time in science and math explanations, but they aren't used in the same way as other communities. There are currently people working on writing guidelines for science article citations. --ScienceApologist 07:18, 1 October 2006 (UTC)
@Gerry: Your argument is: It does not consider another merit of inline citations: telling the reader where to look for more complete material on the subject at hand. But this is a total different issue than WP:V. For purposes of WP:V the in-line cite can be just one authorative, reliable source, saying exactly the same as the paragraph in the article. Regarding WP:V this is in fact the most favorite case. But regarding the curiousity of the reader, it doesn't help him in any way.
As an example, see this sentence which earned a {{fact}} in this controversial edit by Agne27:
Einstein's theory combines Galilean relativity with the postulate that all observers will always measure the speed of light to be the same no matter what their state of uniform linear motion is
Is the curiousity if the reader addressed if we give a reference to textbook which just states the same? In one sense yes, if the reader starts reading a good textbook on SR, many remaining questions should disappear. But this is not by virtue of on-line citing, but the point taken by WikiProject Physics anyway: There are good textbook given in at the bottom of the article, start reading these for a deeper understanding.
But there is indeed something missing in the sentence tagged by Agne27: Good hyperlinking.
Einstein's theory combines Galilean relativity with the postulate that all observers will always measure the speed of light to be the same no matter what their state of uniform linear motion is
Hoping that (or working on) the linked articles resp. sectiosn are well written, these hyperlinks will help to satisfy the reader's curiousity what's meant by this sentence.
Pjacobi 11:54, 1 October 2006 (UTC)

Strong oppose against "common knowledge"

I am not intending to delve into the discussion or to monitor it further, but I just wanted to express my very strong opposition to making explicit exceptions for "common knowledge". This is potentially very dangerous, as it leaves the door ajar for all kinds of abuse. Almost anything could be argued to be "common knowledge" (even considering some clarifications of the term that appeared of late), and it might lead to endless debates in talk pages without a proper solution.

I would leave the matter to the common.... sense of the editors - if somebody would require a citation for anything that is so "common knowledge" that there is no need to cite any source (e.g. that Celine Dion is a singer), then there are a myriad of more or less formal ways for the community of editors to sort that out, including also the possiblity of finding out that the fact is not so obvious and "common knowledge". Giving a formal argument for one side to permanently obstruct the debate is not a good idea. And providing an online citations when the community would decide it might be advisable won't hurt, especially if there are really so many sources that would confirm a given fact.

I am posting all that to make sure that there would be no claim of consensus regarding the "common knowledge" issue. I've heard quite many arguments for such "exception rules" before and they all boil down to the fact that some editors do not want to have to do this little bit of work more, and want "their" articles to be treated differently. Please consider this objection when making any decisions, and please sign below if you also object to such exceptions. Bravada, talk - 12:13, 1 October 2006 (UTC)

I actually don't understand your objection. We should indeed leave it to the editors of an article to decide by consensus if a certain fact is common knowledge or not. But we should also give guidlines about how to settle potential disputes. E.g. one can say that something is common knowledge if in textbooks written on the same level the facts are presented without citation.
Giving a citation for common knowledge facts can hurt. It can make a sentence look stupid, e.g. if you cite that "Celine Dion is a singer". That won't happen in case of an article about Celine Dion, but it could happen in case more technical articles, like special relativity as User:Agne27 demonstrated. Lay people won't see that it is stupid, but people who are more familiar with the subject would laugh if they see such a citation.
So, we are not motivated to get this sorted out because we are lazy, but rather because we don't want to be forced to make stupid looking edits to retain GA status. Also, this is not about not willing to make articles readable to lay people. Most articles on technical subjects are meant to be read by lay people also. They can educate themselves about the subject, but they may not be in a position to make judgements about whether or not to cite some fact. Count Iblis 12:57, 1 October 2006 (UTC)
I believe that leaving it to editors to determine what is common knowledge or not is exactly what may lead to endless disuptes in talk pages. To give you an example from another field, while I was creating/editing some articles on Chrysler Europe/Talbot vehicles of late, I had to mention the fact that Chrysler Europe was taken over by PSA and brands like Simca or Chrysler replaced by Talbot (etc etc), which is more or less common knowledge to people within this field of knowledge, and you can find this information in a myriad of sources. That said, I saw nothing wrong with referencing that, or for that matter, each and every fact in those articles, by means of inline citations.
Sometimes, almost the entire article would consist of information that is "common knowledge" and can be referenced to one source. I just put a cite to the source that contains it all at the end of every paragraph, which, I believe, is sufficient. Moreover, it allows future reviewers and editors check where exactly I got those info from. Bravada, talk - 13:55, 1 October 2006 (UTC)
I agree that this can be an acceptable solution. My personal writing style is to make an article as self contained as is reasonable possible. When User: Agne27 put in the fact tags in the relativity article, I was thinking that perhaps the article should explain these things a bit better. User: Agne27 explained on the talkpage that the reason why the tags were added was because the so-called facts were not clear. Adding refs would not have been useful, because all you'll find in textbooks is more of the same. Count Iblis 18:11, 1 October 2006 (UTC)
I think that the version of WP:CITE appearing right now would be acceptable for a majority of editors in the mathematics project. It does not prvent anyone from adding inline citations, does not prevent the use of the {fact} tag, and does not prevent editors from removing uncited information so long as they act in good faith. Thus the tools are still in place to deal with crank authors (the sciences have plenty of them) without unduly burdening hard-working editors with the task of looking up basic facts such as Celine Dion is a singer. The new policy clarifies what I believe has long been consensus within several scientific projects: an appropriate selection of reference texts, listed at the end of an article, is enough citation for articles that contain only basic, undisputed facts. CMummert 13:24, 1 October 2006 (UTC)
What would be the problem then with just referencing the paragraphs containing the basic theories and other "common knowledge" facts, such as physics or mathematics principles, to those reference texts? Bravada, talk - 13:48, 1 October 2006 (UTC)
Celine Dion is a singer is a bad example. Specifically, Celine Dion is a Canadian singer. So, how many people in Germany do you think know that she is a singer? how many people in Japan? in China? in Iraq? in Sudan? Celine Dion is a singer is a culturally-specific common knowledge fact - treating it as a common knowledge fact period is an example of systemic bias. Armedblowfish (talk|mail|contribs) 15:53, 1 October 2006 (UTC)
I dare say that the fact that Celine Dion is a (Canadian) singer is as common knowledge in developed economies as the theory of gravity, if not more, actually. I am not sure about China, Iraq or Sudan, of course, but then the scientific theories some people want to make exceptions for are far less common knowledge, only known to a small group of professional physicists or mathematicians, so the bias is even stronger. OK, the bottom line is - let's not make needless "exceptions" that would push us to discussing what is and what is not a delivery vehicle (see below). Bravada, talk - 17:12, 1 October 2006 (UTC)

Again we have ambiguity in what Common Knowledge is. I do not think the requirement of having college level text books equates to Common Knowledge for the masses. (You may be able to make an arguement for high school level). As the exception currently states, only those with access to said text books are privileged enough to have this particular "common knowledge" and not need the benefit of the in-line cites or a reference to a particular text book. I don't think that type of exclusion is what Wikipedia is about. Agne 13:25, 1 October 2006 (UTC)

Common knowledge applies to the editors not Wiki readers. If someone does not know what's in the textbooks on a topic that person is not likely to be a helpful editor. The textbook rule works in most fields but not current events. Common knowledge (for editors) for current events is what a well informed reader of the major press will know. Rjensen 13:33, 1 October 2006 (UTC)
The new policy does require references to the particular books:
Common knowledge facts are those that appear in multiple reference textbooks for the field, all of which are listed in the references section of the article.
I could never imagine using anything except a college level text, graduate level text, or research paper for a reference. High school textbooks never include the sort of material that I write about, and in the fields that they do cover high school books are often factually inaccurate and (in different ways in different countries) politically biased. CMummert 13:56, 1 October 2006 (UTC)

The notion of "common knowledge" seems unavoidable to me. The notion of what is "common" is usually obvious from the level of technicity of the entry. To make everything clear, one could add some commonly used sentence like "for basic facts about Y, the reader is refered to ...". This avoids to put inline citations for everything. Usually, the level of technicity of a part of an article is a good indication on what should be locally considered as "common knowledge". pom 14:00, 1 October 2006 (UTC)

In my opinion, any attempts to make common-knowledge exceptions a part of this policy are doomed, because rules become a nit-picker's nightmare unless they apply universally. For example, it's no good saying that delivery vehicles can ignore no-entry signs, because then you'd have arguments about what a delivery vehicle is. If science editors or maths editors want to create exceptions amongst themselves, they are free to do so; but they shouldn't try to make Wikipedia policy follow suit, and they shouldn't expect the good-article judges to change their criteria on that account.
Simple, clean policies are the best, I think. Anyone who has spent time referencing articles knows that when to cite is instinctive, not done according to policies. You don't cite for the fact that Celine Dion is a singer but you do for the fact she had to stop singing in 1996 for throat problems, which incidentally confirms that she's a singer (and which incidentally I made up; did you spot it?). You don't make a fence entirely out of fenceposts.
On the question of what books to use as references when there are plenty to choose from, it seems to me obvious that you reference the best books available; it is usually fairly clear which those are. Where there are several of equal authority, I'd reference those in print and readily available at a reasonable price, if possible. A quick glance at Amazon indicates which these are. qp10qp 16:30, 1 October 2006 (UTC)
I oppose too, for the same reasons as Qp10qp - Also, I don't find the argument that it makes the article look silly to experts a valid one-- we shouldn't be concerned with making our editors look silly to experts; the article is for lay people. This isn't the New England Journal of Medicine. --plange 17:06, 1 October 2006 (UTC)
I think I agree with Bravada. While I am sympathetic to the idea that GAs shouldn't have to be the perfect article, a good guideline or policy should describe what the perfect article is in a way that acknowledges that many articles will and do fall far short of that standard. Not making it clear that inline citations for everything is preferred, whether at the end of the sentence or the end of the paragraph, fails to describe the perfect article. Besides, the current wording fails to make it absolutely clear that if anyone does ask for a citation, you should just add one rather than arguing over whether it is "common knowledge". Armedblowfish (talk|mail|contribs) 17:51, 1 October 2006 (UTC)
I don't think it is reasonable to demand that "if anyone does ask for a citation, you should just add one rather than arguing over whether it is "common knowledge"". Besides the objections that have already been raised here, you would also make make life much more difficult for the people here who are trying to make articles on pseudoscientific theories, such as e.g. Heim theory more NPOV. There is an AFD going on right now for this article because some editors think that it is just impossible to maintain the article in a NPOV state. The supporters of Heim theory would be able to challenge any simple refutation of some statement by demanding a ref. even if such refs don't exist because it is so trivial.
I think that it is a bad idea to give certain editors a tool to force an article to make certain changes without having to discuss it. Count Iblis 18:40, 1 October 2006 (UTC)
No WP policy or guideline states that inline citations for everything are preferred. The entire point of this discussion is that many feel that they are not preferred. WP:PERFECT says that facts must be cited, but not that they must be cited inline. CMummert 20:44, 1 October 2006 (UTC)
By "not preferred" do you mean "not worth the time" or "actually worse"? So if someone did take the time to cite a nice, happy, reliable source inline for a "common knowledge" fact, do you think it should be removed? :/ ... :( Armedblowfish (talk|mail|contribs) 21:41, 1 October 2006 (UTC)
By "not preferred" I mean not preferred universally over other sorts of citations such as listing references at the end of the article. For some types of information, such as disputed facts or quotations, inline citations ware certainly preferred. For other sorts of information, such as Hydrogen is an element, I would say that inline citations are in not preferable to listing references at the end of the article. The difficulty at WP:GA is that some editors feel they are preferred, while many scientific editors disagree. CMummert 22:42, 1 October 2006 (UTC)
It depends on whether that citation improves the quality of the article. In some cases it was already clear from the context that a fact or argument is from a theory cited at the beginning of the article. E.g. in case of the article on special relativity, it is clear that everything that follows is based on that theory. You then don't need to cite the same textbook over and over again. If you do include such an "extra " citation, then that suggests that that particular fact is something special. E.g. it could be that an elegant proof of a certain fact that you are presenting in the article was firs given in the Ref. that you are citing.
I think one should simply follow the conventions of writers of similar texts as the article. Count Iblis 22:02, 1 October 2006 (UTC)
Are you talking about the information that the sentence or paragraph actually contains, or the information that that information is built on? A citation for the former is really enough. Or perhaps I misundertand you? Do you mean the same text book is used to provide a large amount of different information in the article? Why not simply repeat the reference once-a-paragraph using <ref name="NAME" />? Also, I think that Wikipedia, not being peer-reviewed, should use a higher standard of referencing than other encylopedias, which are peer-reviewed, since good citations allow readers to peer-review articles for themselves. Armedblowfish (talk|mail|contribs) 22:13, 1 October 2006 (UTC)
The idea that inline citations allow readers to peer review for themselves is commonly presented but I don't believe it is accurate. Many of the references I use (in mathematics) are out of print, only available at university libraries, and beyond the level of the intended reader to comprehend; the article itself is written at a lower level that is (I hope) understandable to the reader. Thus a reader who is unfamiliar with the material in the article will probably not be able to understand what the references are saying in the first place, while a reader who is familiar will not need to look up common facts. Thus I believe that listing books from which the reader can educate himself or herself is the most appropriate way to ensure verifiability of common facts. It would help me understand the opposite position if somebody could point out where my reasoning goes astray. CMummert 22:42, 1 October 2006 (UTC)
If the reader is unfamiliar (and so in need of cites), it helps them out tremendously to know where to start for that bit/section they don't understand than making them go to the bottom list of general references (that have no page numbers) and wade through these to understand that bit/section. --plange 22:47, 1 October 2006 (UTC)
I am adamantly in favor of careful inline citation of facts for which this care is necessary, but I don't think that you addressed my point. Verifying references in scientific and mathematics articles, except for isolated numerical facts such as the mass of the sun and direct quotes, is not a matter of looking up the fact on a particular page of a book. It requires reading some initial segment of the book, understanding the material, and then deciding whether the statement in the article is accurate. A page reference will not help a novice perform this process, and the sort of fact that is under discussion is the sort of fact for which a person familiar with the subject needs no reference. For example, the article Machine that always halts is carefully cited and completely verifiable (it has other problems which are not relevant to this discussion). Let's make this discussion less hypothetical and discuss only things in that article where a detailed page-level citation would really benefit a reader who wants to learn more about the subject. Please don't just say "page numbers always help," because that is the very point I disagree with. CMummert 01:05, 2 October 2006 (UTC)
You could also write that you are going to follow a certain ref. in a particular section. It should be clear where the information comes from. But this is not going to lead to problems. I.m.o., we need to focus on addressing the real problems that threaten the quality of articles instead of on hypothetical problems that simply don't occur. Articles that have made it to GA status are edited by many authors and are on the watch lists of many experts. The idea that some crank editor would be able to add nonsense to such articles that won't be reverted and that therefore you need a citation policy to put lay people in a position to guard these articles is just nonsense (and it wouldn't work anyway as CMummert pointed out above).
What are the real problems we are dealing with? Take a look at the AFD discussion about the article Heim theory. Crank editors are writing articles on pseudoscientific in a very POV way. A citation policy that demands that even well known facts should be cited will hinder us, not help us.
E.g., a crank editor makes a POV claim and adds a citation to a crank journal, say Physics Essays. I want to write that the claim contradicts well established laws of physics. But, I can't give a citation to an article that directly addresses that point, simply because very few scientists devote their time on debunking crank theories. All I can do is give a citation to a book about a basic physics topic which is relevant to that edit. According to the present citation policy, my edit could be "legitimately" reverted as the citation I provide does not prove my point directly.
Another problem is that we could be forced to include inapropriate refs that will make articles look ugly. This is how the discusion here started, but this may not be a very serious problem. Count Iblis 01:41, 2 October 2006 (UTC)

What about geography?

What about geographical articles? Someone has added {{sources}} to the stub on Stanbridge, Bedfordshire, which contains statements like "Stanbridge lies 3 miles east of Leighton Buzzard". I cannot see what is gained by adding a source for this type of statement, as it can be verified by consulting any suitable large-scale map. Giving a reference to one particular map would clutter the article, and might imply that that source is thought to be more reliable than others. In fact, if a statement like this survives for a year on Wikipedia without being challenged, it is probably a better sign that it is true than if it appears in just one external source. To create the reference requires a (small) amount of work, and it would discourage me (a bit) from filling out Wikipedia with stub articles. I am quite happy to add references for doubted or hard-to-find facts, and have done so when editing other articles. (Note: Since {{sources}} was added to Stanbridge, Bedfordshire, I have added a reference for the population estimate. This was because the 2001 census figure, which I assumed was a 'well-known' fact, was replaced by a more up-to-date estimate, which could differ depending on who makes the estimate.) JonH 13:59, 2 October 2006 (UTC)

I tend to agree with Jon here (as long as we understand that "a statement like this" really means "a statement like this"; in an article like Übermensch, who knows what crap might survive for a year): it is often a pain to find sources for obvious, trivial statements like this, and it is hard to see the benefit. - Jmabel | Talk 06:23, 17 October 2006 (UTC)

What about medicine?

This discussion has become too particular to two fields only -- math and physics -- while completely overlooking, thus far, some other scientific areas like medicine, in which multiple citations are routinely given for almost every line in journal articles. If I could claim common knowledge on any medical fact you can find three sources for, I can write some fantastic quackery articles. Have a look at attention-deficit hyperactivity disorder, which recently passed GA under the allegedly stricter requirements. It has a number of statements someone apparently thought were common knowledge, since the article is largely uncited, and a good deal of the under-cited article does not reflect medical consensus, and there are not legitimate, peer-reviewed medical cites for a good deal of the content there. Someone on GA thought it was well cited.

  • "Something that can be found in every textbook?" Not all fields are as static as some of the math/physics people here would have us believe. Facts need to be cited so we can verify the veracity of the source. In the articles I edit, I know the leading researchers, the medical consensus, and the well-controlled, replicated research from the garbage.
  • "Statements which are undisputed within an academic field need not be inline cited." Without citations, how is the reader to know what is or isn't indisputed?
  • "Show me any serious publication that applies inline citations to every fact." Read some medical journal articles.
  • "I suggest we edit the page to include the statement that elementary facts should not be cited." Less than 15% of people with Tourette syndrome have coprolalia. That's everyday, common knowledge to me - a very basic, elementary fact. Good, I won't cite it :-)
  • "But if I cite "Mathematics by Jane Smith, PhD" and not "Mathematical Exercises by Jane Doe, PhD" I'm tacitly endorsing Smith's book over Doe's book. It's basically an advertisement for Smith over Doe. Why should we be in the business of endorsing textbooks? " This is a silly argument. You cite any source you know to be the highest-quality source. In every academic field, there are recognized top-notch experts. Citations aren't recommendations to purchase something: they are acknowledgements of top-notch sources.
  • "No one has delisted articles yet." A physician or neurologist needs to look at attention-deficit hyperactivity disorder.
  • "A year ago we used Wikipedia:Peer review to prepare for a FAC." This is the most sensible thing I read in this entire discussion. We keep getting *really* REALLY bad GAs at FAC, and lately we've been recommending them to peer review. That's not very helpful, since no one is helping out at peer review. I reviewed 9 articles at PR this week; even found some bad copyright stuff there. Why don't we all start helping out more at PR?
  • "But GA doesn't mean anything if it's the same as FA, and it seems to be very close to that right now. " I pulled random sentences out of the extensive discussions above, so I don't know who said this, but I can't believe anyone who really believes this has spent a lot of time at FAC. The GAs coming through there are often a wreck.
  • "One further point that I think people may have missed: inline citation does not mean, per se, that every sentence must be cited. It would be quite possible, for example, to have a single footnote for an entire paragraph or section." Mountain --> molehill.
  • "To add to Kirill Lokshin's points, I've been reading some of the arguing here and elsewhere, and I must say that this chimera of a reference for every line has been invented: there's no policy or good-article principle that requires it, there just isn't; it is a straw man (OK, a straw chimera)." Reason at last.
  • " In practice, enough experts read through the articles as part of the editing process that errors are found and important citations added." Not at GA, where the ADHD article got through with inaccurate, uncited information.
  • "I suggest that a clarification be made in WP:CITE that facts that are obtainable in the generic undergraduate textbook on a particular topic need not be given a specific inline citation, " Almost ALL undergraduate and many medical textbooks on Tourette syndrome contain outdated, inaccurate information. If I see someone sourcing a TS statement to Comings, I know it's not well researched. In every field, certain journals are more respected than others. Facts in an undergrad textbook mean nothing. Comings is still one of the most-cited TS researchers, and most of what he believed has been disputed or refuted by subsequent research. But I can certainly find 3 textbooks which cite him.
  • "Common knowledge facts do not need to be directly cited when they are verifiable by a large number of sources." Diet is a cause and treatment of ADHD. You can find that in a phenomenal number of sources. But you won't find it in controlled, replicated, peer-reviewed journal studies from the top-notch journals and the top-notch researchers. Nope - I want to see who you're citing. The ADHD article cites Biederman and his group (hands down the top ADHD researchers) once, and it cites a lot of pop authors.
  • "Inline citations are not required on non-controversial statements that are corroborated by more than three of the sources cited by the article." See above, and who defines non-controversial? Folks of the POV that ADHD is caused/cured by diet don't consider that to be a controversial statement. The way we keep POV and OR out of Wiki is by requiring highest quality sources on technical articles.

Wikipedia is not only math and physics: a whole lot of this argument is just too focused on a few areas, and overlooking the need for citations in other areas. Not only that, citations do more than verify facts: they give Wiki readers a way to dig further into what they read here, helps us battle POV and OR, and builds the reputation and quality of Wikipedia. Math and physics are no exceptions: anyone read Atomic line filter lately? Sandy 01:38, 3 October 2006 (UTC)

Summary style and CITE contradiction about linked subarticles

I would like the policy to indicate some flexibility. I think, for example, the dictum against "citing other Wikipedia articles" is too harsh. I'm looking at a few summary style articles right now where there are demands for sources that it woudl be tedious to compile from the subarticles, many of which merely list sources in end-notes or external links. If an inline citation exists, it's easy to move, but in summaries or lists they can create "citation clutter". (Hey, if people can debate category clutter, even for useful categories saying there are just too many, why can't I debate having too many citations?) To my mind the referenced article is a subarticle, which handles the topic as a primary topic, and thus the place for citations about it. To be pertinent to the question at hand here, maybe we need to slip the words "editorial consensus" in there someplace. If editorial consensus is that there are enough citations ... after all, that is practically how it works, becaouse obviously one citation per word is too many, and I bet only a handful of articles even achieve one citation per sentence. --Dhartung | Talk 23:18, 29 September 2006 (UTC)

I think this is a separate issue, but one that needs to be discussed. Therefore I'm breaking it out as a new section. --ScienceApologist 23:35, 29 September 2006 (UTC)

I completely agree that adding citations when using summary style can create undesirable clutter. My only worry is that often the article that is being "summarised" is, in fact, in much worse shape (regarding content and referencing) than the parent article. I just added a sentence to Wikipedia:Reliable sources indicating that there is, in fact, no conflict. If nobody deletes it, then I guess there is no conflict. –Joke 00:55, 30 September 2006 (UTC)

Doesn't work in practice: articles change over time, and move away from the daughter articles they relied on. Former FA Hugo Chávez has become a POV wreck, the daughter articles have fallen apart, and the main article was never cited because supposedly the references were in the daughter articles. Maybe they were once, but they aren't now. Wikipedia is not a static document: because of the dynamic nature, all articles should be cited. Sandy 01:41, 3 October 2006 (UTC)

I think it is a stretch to say it has been shown that it "doesn't work in practice." It merely makes the job of maintaining a particular article harder, because you have to ensure daughter articles are kept up to snuff: but plenty of choices we make – such as the entire WP:V policy! – make the job of maintaining the encyclopedia harder. As the encyclopedia improves, and more and more articles are actively being monitored, I think it is a perfectly reasonable position and consistent with the philosophy that we shouldn't direct people out of Wikipedia for information that can be found within. Ideally, an article and the daughters move in lockstep, and if an article has well-referenced, quality subarticles I don't see the point of repeating the references in the main article, unless they are clearly necessary for some other reason.

Clearly, Hugo Chávez fell apart and was rightly demoted (however, it wasn't demoted merely because the daughter articles had problems, as you yourself commented on the FARC the page itself also had problems). That indicates to me that the process worked! –Joke 03:00, 17 October 2006 (UTC)

Video games

I've been talking to the mortal kombat project about their use of sources, the conversation can be seen here. At the current moment, there are multiple pages without a single source - however, they and I cannot work out how to source material when dealing with video games. Should the guidance be here somewhere? somewhere else? --Charlesknight 16:26, 30 September 2006 (UTC)

{{cite video game}} is the cite to use for in-game text; the recent Final Fantasy FAs tend to also quote the text in the citation (see Final Fantasy X for example). The video game project has more information and/or participants that can answer questions. Nifboy 18:28, 30 September 2006 (UTC)
Templates are only one way to do citations, there are several other options. Although some editors don't like templates, they do work for the situations that the template-writers have thought of. But if it is necessary to add some unusual information to the citation, templates don't work very well (for example, how to get a video game to display the names of the programmers on the screen).
Don't forget to sign comments. It's possible to add random text to the end of a reference thus:
<ref>{{cite mumbo-jumbo}} A stick, a stone, It's the end of the road.</ref>
That makes it a combination of a reference and a footnote. The text can be wikified, as well, which isn't possible inside the template itself. --Dhartung | Talk 10:19, 2 October 2006 (UTC)

Desperate help needed at the Black people article! Please get involved!!

This article is an absolute mess. It provides no coherent well sourced definition of a Black person and just rambles on and on about various people who were labled Black in different times, places, and languages, and tries to merge them all together as a coherent ethnic group. It would be like trying to merge Native Americans and people from India into a coherent article called Indian people. It makes no sense. We had requested mediation and the mediator said we should use the census as our source. Here's what the U.S. census says:

A Black is “ a person having origins in any of the Black racial groups of Africa. It includes people who indicate their race as "Black, African Am., or Negro,"or provide written entries such as African American, Afro American, Kenyan, Nigerian, or Haitian.”

Black Africa is a synonym of sub-Saharan Africa and all of the non-African groups mentioned (i.e. African-Americans, Haitains) are descendents of the recent African diasporas. And yet we still have editors insisting that South Asians be given equal weight in the article and be considered Black. These people provide no cited definitions or census classifications to defend their assertions, instead they cherry pick from different sources in different countries for examples of South Asians being labeled Black, often in different languages. But by the same logic, I could argue that the Black Irish are Black. The point is the people editing that article need to be forced to adheare to a coherent sourced authoritative definition of a Black person, or the entire article should just be deleted as POV and unencyclopedic.

Dictionary.com[[4]], the free dictionary online[[5]]., the U.S. census[[6]], and the British census[[7]] all emphasize the idea that Blacks are of African origin-in fact it is against the law for a dark-skinned person of South Asian or Australian origin to claim to be black in the census. An article by the BBC makes a clear distinction between Blacks and the dark skinned people of South Asian ancestry[[8]]. This article about race in biomedicines says “The entities we call ‘racial groups’ essentially represent individuals united by a common descent — a huge extended family, as evolutionary biologists like to say. Blacks, for example, are a racial group defined by their possessing some degree of recent African ancestry (recent because, after all, everyone of us is out of Africa, the origin of Homo sapiens)."[[9]]. I really need help getting the editors of that article to stick to a coherent definition, instead of just pushing their own POV. Editingoprah 06:10, 1 October 2006 (UTC)

How does the reader KNOW which facts have been verified to be "common knowledge?"

Even if I grant most of the arguments of proponents of a "common knowledge" exception, I still feel there's are two serious problems that are trivial to solve with inline citations, and almost impossible to solve with any other existing mechanism.

  • Let's suppose that a certain section of an article is introductory in nature.
  • Let's suppose that there is a team of editors who mutually acknowledge each other as knowledgeable about the issue.
  • Let's suppose that these editors review the section, and on October 1st, 2006 agree that everything in it can be found in ordinary textbooks.

1) How on earth does a reader know, just by looking at the text, that such a review has been performed? Are they supposed to comb through the talk page looking for a consensus discussion? How can they tell the difference between this text, which is reliable, and an uncited and identical-in-appearance section of another article, that was written off the top of someone's head and hasn't received such a review?

2) Once the review has been accomplished, what's to stop another editor from coming through and inserting uncited non-common-knowledge material? Should the section be frozen once consensus is achieved, and from that point on should further editing only be allowed after someone makes the case for a change? Is every reader supposed to go into the history and perform a diff between the current version and the last reviewed version?

The only obvious solution I see using existing technical mechanisms would be to add a single footnote at the end of a "common knowledge" paragraph, with text such as "It has been determined by Wikipedia editors that as of October 1st, 2006 the contents of this paragraph consist entirely of common knowledge," with a link to the discussion at which the determination was made, with an HTML comment saying "This paragraph should not be edited from the consensus version without first discussing and gaining consensus in Talk." That seems crazy to me, but it seems to me the logical consequence if one assumes that readers should be able to judge the reliability of the source of information in Wikipedia, but that common-knowledge items need not be sourced.

But, really, once someone has found that a piece of information is contained in a standard textbook, I'm completely baffled as to what, exactly, is gained by not including an inline citation to the textbook. Dpbsmith (talk) 21:40, 1 October 2006 (UTC)

It's only make full scale sense, after the software feature of tagging peer reviewed versions is switched on.
Unfortunately, in-line citing doesn't sove the problem trivially. Gaming the system of verifiability by in-line cites has entered the repertoire of POV-warriors and sneaky vandals.
Pjacobi 21:34, 1 October 2006 (UTC)
OK, not trivial. But what happens when you click on the citation? Do you get a phony reference? It seems to me that even in the case of a print source the deception would be found fairly quickly and would be just cause for a ban of the person who did it.... Dpbsmith (talk) 21:40, 1 October 2006 (UTC)
Also, these "POV-wariors" could place "citation needed tags" on refutations of their POV edits. Often these refutations are not easy to find in the literature because scientists don't usually publish articles debunking pseudoscience. You have to use the results of well established theories to refute, say, Heim theory, but you won't be able to find a single article with the Title "A refutation of Heim theory". Count Iblis 21:50, 1 October 2006 (UTC)
How to game verifiability by in-line cite? Many possibilities. Not all references are online (for comparison: de: requires all featured article references to be print, for longetivity and better reliability). So giving an obscure but not immediately falsifiable print reference is an option. Or use references which aren't really reliable, but a layman wouldn't immediately see the difference. Articles in Galilean Electrodynamics, Hadronic Journal, Physica Scripta, Physics Essays, or Foundations of Physics. Referencing result which have subsequently be disproved. Etc. --Pjacobi 21:52, 1 October 2006 (UTC)
Dpbsmith asks why an inline citation can't be included for information that was looked up. The issue is that the sort of information we are talking about doesn't need to be explicitly looked up by the experts who are writing these articles, because it is so well known. Take, for example, the definition of a group (mathematics); this is common knowledge and I would not look it up if I were writing the article from scratch. I can tell you, from the top of my head, several books where I know this definition can be verified. This definition is not a direct quote of any one person and is not disputed. Thus I would list several good references at the bottom of the article but not include an inline citation for the definition. The opinion of many editors of scientic articles is that this is an adequate way to cite these standard definitions. Also see my post of 22:42, 1 October 2006 (UTC) above explaining why, in some cases more complex than this, inline citations will not actually aid verifiability. CMummert 22:50, 1 October 2006 (UTC)
Wikipedia articles are not, however, scholarly journal articles and the same practices are not appropriate. Jjournal articles are identifiable as the work of a specific real-world person, with verifiable credentials (or by a team of collaborators, with the senior author presumably having responsibility for the entire article). The identity of the contributors in turn is guaranteed by the editor of the journal.
Wikipedia articles are nothing like that.
First, how am I supposed to know whether or not the author of any particular edit is an expert? If it's the 1911 Encyclopedia Britannica it's easy. In the article on "Radioactivity," I go to the end and see the initials "(E. Ru)." I know that the entire article was written by E. Ru, whoever he is, and that he's responsible for all of its content. I then flip to the front of volume 22, and on p. vii I see that E. Ru. refers to one Ernest Rutherford, F. R. S., D. Sc. LL. D., Ph. D., Langworthy Professor of Physics, University of Manchester, Nobel Prize for Chemistry, 1908, Author of Radio-Activity; Radio-Activity Transformations; etc. Fine. Now, this particular encyclopedia article contains no less than fifty-four inline references, but I'm reasonably happy to assume that the stuff in between is either from one of the five "general treatises" or from Rutherford's head. But the point is, it's Rutherford's head and I know who Rutherford is. There's no possibility that someone else could have injected plausible-sounding dodgy material into the middle of it.
Now, I see that User:CMummert describes himself as "a professional mathematician, with a PhD in mathematics." In all likelihood, this is true. CMummert sounds like the abbreviation of a real name, possibly that of Carl Mummert, described by Stephen G. Simpson in a 2006 paper as "my recent Ph.D. student at the Pennsylvania State University." I have no way of knowing this for sure--Jimbo doesn't ask for ID cards at the door--but it's likely. I'd guess there is a 99.44% chance that User:CMummert is the real-world Carl Mummert, has a Ph. D. in mathematics, and makes extremely trustworthy edits on mathematical topics.
As a reader, I'm still faced with the problem. Even if C. Mummert were certified by Jimbo Wales as an expert in mathematics, much as Rutherford is certified by the editors of the Encyclopedia Britannica, when I read an article on mathematics, how am I supposed to know which uncited material is the work of C. Mummert or of equivalently competent contributors?
I'd add that in the areas in which I consider myself to have personal knowledge, I also have some books within arm's reach. If I'm called on to cite a "common knowledge" item, such as Jack London was alcoholic, I'll blink with surprise a couple of times—doesn't everybody know that?—but it really doesn't take me long to reach down any of half a dozen books that say so, and add a citation. It's just not that big a deal. Especially if you do it at the same time as you're making the edit. Dpbsmith (talk) 00:38, 2 October 2006 (UTC)
P. S. When I do write down material from my head, and then check it subsequently, it's really amazing how often I need to make a small change... or a not-so-small change. I'm sure that's just me, of course... Dpbsmith (talk) 00:48, 2 October 2006 (UTC)
It is not that editors in the sciences wish to avoid citations, or hide their sources. We want our articles to be clear, correct, and verifiable. Any article on a significant topic will be vetted by experts in the project and unlikely to contain facts that need citation but do not have it. The mere presence of citations does not actually ensure that the facts are correct, for various reasons, and furthermore the references might be so specialized that they are unreadable to a hypothetical untrained reader who wishes to look up one fact without understanding the basic ideas behind the article. Thus, although it cannot be guaranteed that everything is vetted by experts, it is also not guaranteed that pedantic citations lead to correctness and verifiability. A sense of balance is required.
My purpose in coming here was to seek some clarification in WP:CITE that might lead to more humility on the part of untrained reviewers, so that they seek consensus from the editors of an article about whether it is adequately cited rather than unilaterally declaring that it is not. CMummert 01:30, 2 October 2006 (UTC)
In my opinion, this all seems to be somewhat of a red herring. The presence of citations are not there to ensure "the facts are correct", but for verifiability. Then there is the assumption that the references are so specialized that they are "unreadable" to an "untrained reader". But why does one assume that? The reader (or reviewer) may be able to understand the reference. Thirdly, it seems that there is an assumption that reviewers are more than likely will be "untrained" and will ask for "pedantic" citations without being reasonable or civil. This has to be shown for the case of GA and FA reviewers. It seems like an lot of effort is being put into defining a vague standard for "common knowledge" to solve a nonexistent problem. RelHistBuff 08:05, 2 October 2006 (UTC)
I'm all with you, I hope. If we can assume that a review may need some domain knowledge -- which I until recently thought to be consensus -- most of these troubles don't exist. But backing their position on the "any reader" clause in WP:V, some editors recently attacked several attacks in the mathematics and sciences for be being unverifiable for laypersons. Their belief, that verifiabiliyt by layperson is always achievable and desireable, started all this. --Pjacobi 08:11, 2 October 2006 (UTC)
I am all in favor of verifiability, but I don't believe that verifiability can only be achieved by page-by-page references for basic facts. Verification in science and math articles is usually not a matter of checking direct quotes, but a matter of understanding the definitions and seeing whether the claims made in the article actually follow from them. By giving some good general references, an article can give the reader a place to start learning more about the subject so that the reader will be able to verify the content. I assume that many of my references in my field are not accessible to a layperson because I have read these references and know that they require a strong background in mathematics to read, while anyone with a background in the subject area is not going to need page-by-page citations for the basic facts that are the subject of discussion here. This is less of an issue in basic articles such as Pi and more of a matter in articles such as cobordism. Remember that this page, WP:CITE applies to all articles, not just GA and FA candidates. CMummert 10:39, 2 October 2006 (UTC)
I too think the lay-person thing is a red herring. Certainly, it is useful for readers to be able to check sources, but we know that most of them won't or can't (the latter not only because of the concepts but because they don't have the books in the house). The greater consideration is that of Wikepedia's standing; so long as journalists and critics can pick holes in our articles, Wikipedia is vulnerable to the accusation that it is an unreliable source of information. Inline citing is therefore a form of transparency, of auditable accounting. qp10qp 15:10, 2 October 2006 (UTC)
No one is going to shoot holes in articles because of missing refs. They'll only do that if there are actual inaccuracies in articles. Articles that are edited by scientists here are unlikely to contain serious inaccuracies. There are a number of bad articles on pseudoscientific topics, such as Heim theory and Cold fusion that can be improved more easily if the citation requirement is relaxed a bit. The problem here is that we scientists should be able to write that a certain pseudocientific claim is ruled out or inconsistent with some well established theory. However, we can't give a direct citation, because almost no scientists devote their time on debunking crank theories. In science, the burden of proof lies on te person making extraordinary claims. If you level the playing field then that's going to lead to more POV pseudoscientific aticles. Count Iblis 15:38, 2 October 2006 (UTC)
The lay-person issue wasn't my invention; see the post of Pjacobi 08:11, 2 October 2006 (UTC) above. If you aren't trying to make the articles verifiable by a lay-person, you should agree that detailed citations are only needed for facts that deserve special attention. CMummert 16:16, 2 October 2006 (UTC) (this message was accidentally deleted by another editor and has been restored)

"No one is going to shoot holes in articles because of missing refs." On the contrary, if they are like me, that's exactly what they'll do; they will say, "why should I believe this?"

Not being a scientist, when I read Bill Bryson's "A Short History of Nearly Everything", I was dubious, as I knew he was a mere journalist. However, the fact he peppered the book with notes and references, despite a long bibliography, gave me more confidence in what he was saying. To three pages on special relativity, for example, he bothered to append seven endnotes and a footnote. In the text, for example, he says, "On the Electrodynamics of Moving Bodies is one of the most extraordinary papers ever published". That could be classed as common knowledge, to say the least; however, Bryson's note says " '...is one of the most extraordinary scientific papers ever published': Ferris, Coming of Age in the Milky Way, p193." Other notes to basic content include: " 'Even a uranium bomb...releases less than 1 per cent of the energy it could release': Bodanis, E=mc2, p 77", and " 'it is undoubtedly the highest intellectual achievement of humanity': Bourse et al., The Atomic Scientists, p 53." In short, Bryson doesn't say anything referenceable without referencing it. To a scientist, that book may be laughable; but it is aimed at a general reader, and so, at least in their introductory and concluding sections, are articles on Wikipedia. Just as people might say, "Who is Bryson to tell us about science?", so might they say, "Who is Wikipedia to tell us about science?" Notes induce more confidence in the reader.qp10qp 16:59, 2 October 2006 (UTC)

No one here is saying that we should never use inline citations for more commonly accepted facts. I'm in favor of giving such citations were appropriate. However, if you say that you must always give such citations, then that will aid the crack editors here who will demand citations to refutations of their crack theories that simply do not exist in the literature. And that hurts the reputation of wikipedia much more. Count Iblis 17:15, 2 October 2006 (UTC)
When I've wanted to include material balancing fringe theories, I haven't had any particular difficulty finding sources to cite; e.g. "Skeptic Martin Gardner characterizes Bates's book in Fads and Fallacies in the Name of Science, as 'a fantastic compendium of wildly exaggerated case records, unwarranted inferences and anatomical ignorance.'"
I don't know how to deal with bad behavior or POV-pushing, but ditching the verifiability policy is not the way to do it. Dpbsmith (talk) 18:00, 2 October 2006 (UTC)
Seems to me you could use WP:RS to combat the crank stuff as I'm sure it's easy to prove that the publication or book as a whole is not reputable. --plange 18:55, 2 October 2006 (UTC)
O that it were so easy. When cranks claim that the scientific establishment is biased against them and that we therefore should ignore all that we learn in that context and instead listen to them, it is somewhat difficult to use WP:RS to force them to agree. They'll just point to WP:NPOV and demand that we discuss their argument which may be notable enough to warrant a Wikipedia page but incorrect enough to be easily refuted by common knowledge. For an extreme example, see Time cube. --ScienceApologist 21:41, 2 October 2006 (UTC)
Same/similar can be seen at Asperger syndrome, autism, and attention-deficit hyperactivity disorder. And precisely what keeps these articles somewhat sane is requiring that every edit conform to this. Wiki is not only math/physics. Sandy 01:45, 3 October 2006 (UTC)

Examples of "Common Knowledge" that is not so common at all

To use the "3 text books" example, I have 3 books that are considered text books for some stage of completion for Master of Wine certification. Karen MacNeil's Wine Bible, Oz Clarke's Encyclopedia of Grapes, and Jancis Robinson's Wine Course. In there you can find such common knowledge tidbits as....

  • Müller-Thurgau is not a cross of the Riesling grape and Silvaner, despite the popular myth. (Clarke pg 142), (Robinson pg 106), (MacNeil 516)
  • Despite being classified as a White grape, Gewürztraminer is actually a pink grape. (Clarke pg 103), (Robinson pg 109), (MacNeil 284)
  • Zinfandel can trace its heritage to the Italian Primitivo grape. (Clarke pg 289), (Robinson pg 148), (MacNeil pg 404)

And I could go on. While most of the articles wiki-linked above (with the exception of Riesling) are not adequately referenced at all, the above 3 items are clearly common knowledge to a good wine shop owner, Sommelier and certainly a MW wine expert. They are not, however, common knowledge to the masses--the wine lovers and curious readers alike. It would not be fair to expect the reader to have the same "common knowledge" as someone who specializes in a particular field. It is not fair to put the burden on the reader to attain this "common knowledge" with purchasing 3 of these "hypothetical college level text books" in order to truly access and verify an article-ANY article be it science or literature or wine. I do ardently believe we can come to a compromise to minimize the number of "Common knowledge" in-line cites but due to the very nature of this being specialized common knowledge, we can not leave the reader with absolutely nothing. Hence the reason why I oppose the current common knowledge exemption. Agne 00:46, 2 October 2006 (UTC)

I don't understand your point. The three facts you list seem unlikely to be disputed, and if I saw an article on wine grapes and wanted to learn more or was confused I would just pick up one of those books and read it (the "popular myth" part of the first fact deserves attention, but not just the fact that the grape is not a cross). Please don't point out that without the citation somebody could claim Zinfandel is a descendant of a (nonexistent) Positivo grape, because anyone could add a fake citation that says it is. The only way such a fraud would be detected is if an expert read the page. I don't suggest leaving the reader with nothing; I suggest leaving the reader with a good set of references that can be used to learn more, unless the editors working on the page believe a more detailed citation is necessary. CMummert 01:45, 2 October 2006 (UTC)
The point is, these items are not really "Common Knowledge" despite being in 3 textbooks of a specialize field. They are only common to individuals within those specialize field and such do a diservice to the vast majority of readership to an article. As Qp10qp alludes to below, there is more reasons for an inline cite then consideration on whether it it is "defraudable". As WP:CITE notes, cites serve the benefits of showing it's not original research or plagrarism, allows easy verifiability by any reader or editor, steers readers towards other worthwhile resources on the topic, and most importantly "To improve the overall credibility and authoritative character of Wikipedia.". Agne 21:41, 2 October 2006 (UTC)
I was about to respond to this last night, but instead it made me open a bottle of wine, and so I never got round to it. I think I was going to say that all stipulations are defraudable but that that's no reason not to have them at all. qp10qp 15:29, 2 October 2006 (UTC)
I agree it's no reason not to have them at all; I think it is a reason that page-by-page citations are only necessary when facts are particularly likely to be challenged or otherwise need to be verified. Right now, several people are making an argument to the effect that Spain is a country in Europe needs an inline citation, with page reference, because someone who has never heard of Europe might think they are being fooled by the sentence. My goal is to get clarification of exactly which sentences do not need a page-numbered cite, if you agree that Spain is a country in Europe is such a sentence. CMummert 16:23, 2 October 2006 (UTC)
Speaking about Spain is a country in Europe needs a citation, I remember this true story:

One day, a Texan asked me a question when I lived in U.S.. The question was "How many hours does it take to go to Japan by car?". He didn't know where Japan is, and even before that, he didn't know that Japan is an island. And then, I thought. "What kind of world map is pictured in his mind?"

– from Fool's World Map, [10]

Back to the real question. Do we need an inline citation if "Spain is a country in Europe"? Maybe yes, if there's somebody ask about it. Let's be constructive. Cheers. — Indon (reply) — 18:32, 2 October 2006 (UTC)
Do we need a citation if we claim that "Wikipedia is an encyclopedia"? pom 22:13, 2 October 2006 (UTC)
Probably not to that exact statement but if you have a paragraph about Wikipedia in general with other claims being made about the subject, I'm pretty darn sure you can find a cite that will back up those claims in the paragraph as well as happen to mention that Wikipedia is an encyclopedia for the skeptical. (Which is the case in the current Wikipedia article) An important facet of using in-line cites is the incorporation of common sense and making good choices. The best cites come from references that back up several of the article's claims and can be placed in the most opportune spot for verification-where you get the most "bang for your buck". As others have said, it is a bit of an instinct but really it comes down to the skill of the editor. A good editor can have a well referenced articles with applicable in-line cites for WP:V benefit with potentially as little as one cite per paragragh. It's all in finding the right references to cite. Agne 22:48, 2 October 2006 (UTC)
What's so difficult to find these citation?
  1. Wikipedia is an encyclopedia.[citation needed]] → Jimmy Wales, "Wikipedia is an encyclopedia", March 8, 2005, <wikipedia-l@wikimedia.org>
  2. Spain is a country in Europe.[citation needed]] → CIA - World Factbook -- Spain. CIA.
Come on guys, I'm a scientist also. When my manuscript is being peer-reviewed to a journal, sometimes reviewers ask more citations in a statement that I think everybody who reads my article already know about it. The point is one common knowledge for somebody maybe not for others; and I believe a reviewer asks for that is not going to make the article cluttered with citation junks, but to clarify and to make better understand to whoever read that article. When I'm reading a WP article, I tend to skip citation numbers where I do understand the statement, but I click that citation number if I raise my eyebrow whether this statement is true or not. — Indon (reply) — 00:09, 3 October 2006 (UTC)

Do not change the guideline before you reach consensus!

Dear All,

The current discussion shows clearly that there is no consensus at the moment regarding the "common knowledge exception". A change to a guideline requires a consensus among WP users. Therefore, please do not make changes to the guideline before one is reached regarding them. Bravada, talk - 09:11, 2 October 2006 (UTC)

Reference to common knowlege has been removed. What is your objection to what remains? By the way, please see this thread at WP:V talk: Wikipedia_talk:Verifiability#Lack_of_Consensus_as_Grounds_for_Revert. Also see Help:Minor edit - you shouldn't mark policy changes as minor. dryguy 12:17, 2 October 2006 (UTC)

I agree that there should be solid consenus for any major change in policy. (like the "Common Knowledge" exception). I, personally, did not have a problem with Dryguy's addition. I don't think anyone was arguing for every statement needing an in-line cite, especially when an entire paragraph can be covered with a single cite. It seemed to be a cosemetic clarification. Agne 21:44, 2 October 2006 (UTC)

User:Indon claims above that Spain is in Europe should get an inline citation if it is requested, which seems to me to imply that pretty much everything needs one to match that reading of WP:CITE. At least one person has claimed that one cite is not enough for an entire paragraph, because page numbers are required for every fact. But User:Agne27 disagrees. I find it very difficult to find out exactly what the consensus is on this issue, because nobody in favor of more citations has put forward an opinion about exactly what needs to be cited.
Here is a new proposal for a straw poll. I think it is a basic point that we will have to reach consensus on before we can move farther along.
Proposal. Every fact in every article should have its own inline citation with a specific page number where the fact can be found (online citations must have a paragraph number instead).
Please register your opinion below. CMummert 22:51, 2 October 2006 (UTC)
Support
  1. Of course I do not actually support this obviously nonsensical proposal, but this poll has just been used to justify including redundant statements to the guideline. Therefore, I vote support, knowing that this proposition will fall through, just to make the claims that this poll unanimous false. Bravada, talk - 01:41, 3 October 2006 (UTC)
Since this is a straw poll aimed at measuring consensus, not a vote, saying you don't support while listing yourself as a supporter doesn't give weight to the idea that the consensus favors the proposal. CMummert 01:47, 3 October 2006 (UTC)
Oppose
  1. CMummert 22:51, 2 October 2006 (UTC)
  2. Count Iblis 23:12, 2 October 2006 (UTC)
  3. Oppose. I am also opposed even if the page and paragraph number requirements are removed. No serious publication puts an inline citation on every statement because it would seriously harm readability. Any benefit in the form of instant gratification for those attempting to verify a statement is far outweighed by the harm to appearance. Most statements without inline citations can be quickly Googled, so the increase in convenience would only be marginal anyway. dryguy 23:17, 2 October 2006 (UTC)
  4. Vsmith 00:27, 3 October 2006 (UTC)
  5. ... but if the proposal is "Inline citation is required for every fact that is being asked its verifiability", then I'll strongly support. You don't get it, do you? — Indon (reply) — 00:54, 3 October 2006 (UTC) Note: I pulled my vote here, as this proposal is intended for playing a foolish twisting-the-word game (see thread below).
Neutral
  1. Honestly, this "every line" part is a red herring. No one has argued for that. I think Indon's comment only related to if someone specifically asks for a cite for the "Spain is in Europe" comment. Similar to my reply about Wikipedia, I will point out that most likely one (if not all) of the other references used to support the article (like Spain) will also be able to verify the "Spain is in Europe" part. If someone ask for a cite to that, at most you would need to modify an existing cite (maybe change from page number to chapter number, etc) and you've simply and easily fulfilled that request without adding another cite. Agne 23:22, 2 October 2006 (UTC)
It might be that the consensus on "Which facts need citation?" is "every fact does, no matter how trivial." If it is, this poll will tell so. If the consensus goes the other way, we can then discuss which lines do not absolutely require their own citations with page numbers. I hope that this poll will be a more productive way to build consensus than the discussions above, which show that each side has strong opinions but didn't seem to be progressing. I think that the statement in this poll is the first stepping stone forward. CMummert 23:33, 2 October 2006 (UTC)
Actually, the "every fact" thing came up recently on WP:V over Marcel Lefebvre. Last time I checked, editors at that talk page were claiming that WP:V demands a cite for every fact. I think clarification of WP:CITE on this topic will really help in these cases. dryguy 00:28, 3 October 2006 (UTC)
  1. Neutral / in favor
I am neutral on this straw vote on a proposal that was set up as a straw man to be shot down. For example, how many web pages have numbered paragraphs, as is assumed by this proposal.
I am in favor of the idea that every fact that is not common knowledge should require citation of specific sources that can be readily located by the reader. It's standard practice in the discipline of History. --SteveMcCluskey 00:32, 4 October 2006 (UTC)
Wikipedia:Discuss, don't vote
  1. Dpbsmith (talk) 23:55, 2 October 2006 (UTC)
See WP:DDV#Straw polls. The goal here is to judge consensus on a basic issue, not to vote on policy. CMummert 00:04, 3 October 2006 (UTC)
CMummert is correct. A straw poll is appropriate after extended discussion results in little or no progess, as is the case here. See Wikipedia:Consensus#Consensus_vs._supermajority. dryguy 00:13, 3 October 2006 (UTC)
It's impossible to make a formula for what needs citing, because one article is different from another. I've said twice already that this is an instinctive thing, and I've tried to give examples and methods of how it can be done, why it is a good thing, and why you don't need a reference per line. Every fact or assertion without a note, including that Celine Dion is a singer or that Spain is in Europe, will be verifiable from the citations to the more specialised facts because those not cited are assumed in the ones that are. If you can verify three paragraphs of an article from two pages of a particular book, for example, the reader who checks a particular reference note will see all the related or dependent information without having to be signposted by a note to every sentence.
OK, if I have to spell this out, the following is my opinion about what exactly has to be cited: You should normally note quotes, historical facts, vivid or surprising facts, disputed or controversial facts, highly particular or precise facts, hyperbolic description or claims ("he was the greatest scientist of the twentieth century") etc; other material doesn't need to be referenced but should be covered or understood in the parts of the books referenced. qp10qp 00:40, 3 October 2006 (UTC)

I neither support or oppose what is written above: I support doing exactly what we've been doing, with no changes to the requirements. Anything that someone requests a cite for needs to be cited: what is common knowledge to one is not common knowledge to another, and citations help avoid POV and OR. If the problems are over at GA, they should be solved at GA. If they don't want articles cited, they should handle that over there. I will oppose any weakening of current standards for inline citations, but it is equally absurd to say every line has to be cited. Sandy 01:49, 3 October 2006 (UTC)

nobody in favor of more citations has put forward an opinion about exactly what needs to be cited

This above statement from CMummert is probably the most relevant issue to tackle (certainly more then the red herring "Every line" part). There is no magic number of cites and the only requirements for what needs to be cited are those laid out in WP:V and WP:CITE. From WP:V

1. Articles should contain only material that has been published by reputable sources.
2. Editors adding new material to an article should cite a reputable source, or it may be challenged or removed by any editor.
3. The obligation to provide a reputable source lies with the editors wishing to include the material, not on those seeking to remove it.


Yes the basic assumption is that everything should be cited and verifiable (to prove that is not OR and has been published by a reputable source) but that does not mean that every line needs to have a cite. As I stated above, there is a degree of common sense and basic skills of a good editor that play a part. In your goal of writing a well reference article you look for the best references that can verify as much of the article as it can. It is quite possible to have an entire paragraph exceptionally well referenced with just a single cite at the end pointing to a Chapter or a range of pages within a source. If it's an exceptional claim then you would probably want an individual cite tied to an individual page/paragraph. But for the bulk of an article's info, generally a chapter listing or a page range will suffice. Agne 23:45, 2 October 2006 (UTC)

In my proposal above, I say every fact not every line; you have misquoted the proposal twice now. I think everyone here knows and agrees with WP:V but we interpret the citation requirements differently. Since a lot of conversation above made no progress in reaching a consensus, I hope the poll above will move us forward by telling everyone whether the most strict reading is the consensus one. If it is not, you are free to propose a change to WP:CITE that makes it clear which facts should have inline, page-numbered cites and which facts are OK without them. CMummert 23:58, 2 October 2006 (UTC)
Don't you have better things to do? Now, while every fact should be cited, this can easily be done by putting a citation at the end of the paragraph in case most or all of the facts in it can be sourced from, well, one source. In case there are specific facts in the paragraph that don't come from that source, a ref to the relevant source should be put by the very statement sourced from "elsewhere". I have found this more or less the going custom and I guess this sorts out 99% of cases I can think of. Bravada, talk - 00:08, 3 October 2006 (UTC)
As a first approximation, things should be cited when someone asks for a citation, by inserting a "citation needed" tag (or by some other means). The appropriate response to such a request is simply to supply the citation. Inline, replacing the tag.
There's nothing much in the way of rule-writing that can be done to counter bad faith or bad behavior or gaming the system, but in the normal run of things if someone tags something it's because they think it needs a citation. Once tagged, there's no particular hurry about doing anything, but as has been pointed out, if something is really common knowledge it is usually not very hard to find a place where someone has said so.
If someone one slaps "citation needed" on "the sky is blue," then the appropriate response is to spend ten minutes and add
"the blue sky is so commonplace that it is taken for granted" Schaefer, Vincent J.; John A. Day (1998). A Field Guide to the Atmosphere. Houghton Mifflin Field Guides. ISBN 0395976316. 
(I use that example because that's the book that happened to be closest to my hand. You may have Minnaert's "Light and Colour in the Open Air," or a book of physical optics, or you may have easy access to a university library with a book by Lord Rayleigh. Or a book of poetry. Or Irving Berlin's lyrics to Blue Skies. It doesn't matter. Cite it and move on.) Dpbsmith (talk) 00:11, 3 October 2006 (UTC)
Wow, look what's online these days Dpbsmith (talk) 00:18, 3 October 2006 (UTC) "It is now well established that the luminosity and blue colour on very clear days and at considerable altitudes above the sea-level can almost be accounted for by the scattering of light by the molecules of air, without postulating suspended particles of foreign matter." R. J. Strutt (Lord Rayleigh), Proceedings of the Royal Society of London. Series A, 94(662), June 01, 1918, pp. 453 - 459.

Okay going into the interpretation of WP:V and WP:CITE, can you demonstrate or explain how an article with no in-line cites (or entire sections completely absent of cites) can pass statement #1 of WP:V listed above? How do I know that info listed in the section Paradoxes relating to false assumptions of the Physical paradox article have been published in a reliable source? How do I know that the references at the bottom of the page have anything to do with that section? Or which ones? Or even where in the reference to find them, if the information is there? I think of in-line citations of the bridge connected point A to point B. You don't need 15 bridges connecting the same point A to the same point B but there needs to be at least one. Without the bridge, you just have one big canyon of space between the information and the source of the information. Agne 00:15, 3 October 2006 (UTC)

No individual is likely to have the expertise necessary to verify the citations in every article. Fortunately, Wikipedia has experts that are contributing in almost every field. In general, they have already established what style of citation works well in their field. We should respect that and not impose inline citations where they are not needed. dryguy 00:22, 3 October 2006 (UTC)
But we should provide them whenever they are requested. (Barring bad faith or WP:POINT, of course). Dpbsmith (talk) 00:23, 3 October 2006 (UTC)
We should provide references when requested. Wether they are provided inline or not should be determined by what would best serve the quality of the article. If an article already has inline cites for all key points and controversial points, and the fact in question is already supported by the existing references, then extra inline cites are not warranted. dryguy 00:33, 3 October 2006 (UTC)
We can never guarantee that we will always have experts actively watching every article. Every editors lasting legacy is their edits (and their cites), not their "omnipresence". A good in-line citations tied into a solid reliable source will stand the test of time longer then any "expert presence" and is more accessible for the average editor. We certainly need experts and we need them to leave the lasting legacy of well referenced and cited articles. We can not bank on them being here forever to verify things for us. Agne 00:29, 3 October 2006 (UTC)


Continuing the conversation that was going on before is likely to result in the same lack of consensus. If there are editors who believe that every fact needs a citation, as some comments seem to imply, I encourage them to register an opinion in the poll. So far, not a single editor has registered the opinion that every fact needs a citation! CMummert 00:36, 3 October 2006 (UTC)

I think it is because the more pressing issue is this "interpretation" of WP:V. The policy is clear is that everything needs to be verifiable and such attached to a published reliable source. WP:CITE explains how to connect point A information to point B source of information. That is all a cite is. So is the strawpoll asking you to decide between whether or not we should have info in the article not attached to a published reliable source? Hence my neutral stance because the strawpoll, as worded goes in opposition of WP:V and that wouldn't get any progress either. Agne 00:44, 3 October 2006 (UTC)


It is very simple, really. If an editor challenges a fact that is included in an article and that fact is not supported by a reliable source, then you need to provide that reference, otherwise the material may be removed for violation of WP:V. So, the issue is not about "every line" or "every paragraph" needing a reference, only those that are challenged by an editor wanting to make the article verifiable. There are tens of thousands of articles without any references and citations, and that is OK. But if an editor comes across such article, he/she is in his/her right to request sources to substantiate claims or support facts, and if these are not forthcoming, that editor will be on his/her right to remove the material. ≈ jossi ≈ t • @ 00:49, 3 October 2006 (UTC)
I agree that is the basic and most simplest aspect of WP:CITE. This in-line cite brew-ha actually started from a change to the Good article criteria where articles need to be well referenced including in-line cites. Yes, you can wait till an editor "challenges" an item but is an article with entire sections of info without in-line cites really "well referenced"? Agne 00:54, 3 October 2006 (UTC)
I think there has to be an exception for commonly accepted facts here. For most articles no one would challenge a commonly accepted fact. In case of scientific articles you can imagine a lay person challenging a fact. Giving an inline citation can make the article look stupid to insiders. But this isn't a very serious problem, I don't see dispute about such a things arising very frequently. There are other ways of making clear that the fact is indeed a commonly accepted fact and that citation is not warranted. So, where is the problem?
The real problems occur in case of articles where different editors are not cooporating and fighting revert wars. If you have a policy where someone has the right to force changes in the article by inserting a "citation needed tag" then that will be misused by people who can't get consensus on the talk page.
A good example of such articles are the various articles on pseudoscientific topics. Many of these are written in a POV way that gives undue weight to the position of the proponents of these theories that are rejected by mainstream science. What should happen in these articles is an explanation why these theories are flawed (and not just a one sentence statement that they are rejected by mainstream scientists). But such explanations are not allowed under the current rules, because there are no relevant references to the literature (most scientists don't write articles debunking crank theories, precisely because the issue is already settled).
Of course, you could give a citation to the accepted theory on which the debunking is based, but you need to do a little work to derive the arguments fot the debunking from the accepted theory. So, although it is a commonly accepted fact that these arguments are valid, they are nowhere to be found literally in textbooks or peer reviewed articles.
One has to take into consideration here that if a fact sounds rather trivial to scientists then that will bar you from getting it published in a peer reviewed journal. In particular, I would be wasting my time trying to get an article published in Phys. Rev. Lett. explaining why Heim theory is nonsense. Given that this is the case, the lack of such articles should not be used to challenge the validity of the aguments debunking Heim theory. Count Iblis 01:37, 3 October 2006 (UTC)

One is never enough

One citation per article claim is just not good enough. There should ultimately be at least two, more depending on the statement. Just one may work for direct quotes, then again how often are people misquoted? Anyway, more is better. But why must they all apear inline within the article itself? Much of the above discussion appears to assume that since WP:V requires that article claims must have already be published in a reliable source and those claims may be removed if they cannot be cited, then each claim must have a footnote. But is that really the way to provide the best article for the reader?

If an editor places a {{fact}} in an article, must it be replaced with an inline citation? What's wrong with adding a cite to the talk page and removing the tag if that's what is best for the article. What about reference sub-pages, m:Wikicite, or detailed bibliographic essays? wouldn't these meet the requirements of WP:V without having footnotes in the article? Shouldn't the policy be to have references onhand for everything, but only add citations to the article where it make sense? Whether or not a particular claim of the article needs a footnote, etc. would just be an editorial decision like any other.EricR 01:04, 3 October 2006 (UTC)

One source can be enough. An example would be where a reliable source is not protected by copyright and is imported into Wikipedia verbatim. --Gerry Ashton 01:11, 3 October 2006 (UTC)
Would it be then an article in wikisource? — Indon (reply) — 01:18, 3 October 2006 (UTC)

Are you honestly saying that we need multiple citations for (for example) the spelling of every name in an article? for the date of a highly public event? for the publication date and ISBN of a book??? (the book apparently not being its own adequate source for that.) How many overt citations do you feel we would need each time we mention that George W. Bush was president of the United States in 2003, or that Tony Blair was UK prime minster? This is absurd.

Or, in another matter, try finding multiple independent citations on most facts about early Yiddish theatre. There are topics where, if you want every fact doubly cited, it is almost unimaginable that articles will ever get written.

While we sit here quibbling about this sort of thing, people are out there writing steaming piles of crap, going on at enormous length about things they know little about and not providing any citations worth a damn. Fish where the fish are, folks. Instead of demanding things that are unlikely ever to happen (and probably shouldn't) I'd really like to see more people out there actually trying to get some decent level of citation into articles like liberalism or Israeli-Palestinian conflict. - Jmabel | Talk 06:37, 17 October 2006 (UTC)

You may have mistaken the point i was trying to make. As an example: we can find published sources which state that an apple fell from a tree and thus Isaac Newton discovered universal gravitation (or something along those lines.) Stopping there with one such source would leave the reader with a low quality article. What's wrong with using multiple sources for article claims? Authors make mistakes, fail to tell the whole story, or provide differing interpretations. Looking to multiple sources to back up the article text is a good thing isn't it?
I wasn't trying to require anything of anybody, but simply trying to point out that all the sources used to develop an article shouldn't require an inline citation. As far as your article suggestions go, no thanks, go ahead and fish next to the sewer pipe if you like, but those aren't the only fish to be had.EricR 07:20, 17 October 2006 (UTC)

Going around in circles

My impression from reading the mass of discussion above is that everyone is violently agreeing with each other and happily cooperating in tearing down the strawman du jour. As far as I can tell, the essential point—which I don't think anyone is disputing—is just WP:V:

The obligation to provide a reputable source lies with the editors wishing to include the material, not on those seeking to remove it.

What does this mean, in practice?

  1. Is an inline citation required for every X (fill in: fact/line/sentence/paragraph/section/article)? No, not necessarily; the statements must be sourced, but there is no requirement that any particular technical method be used to accomplish this.
  2. Must a source for any statement be provided on request? Yes, it must, even if the statement is "the sky is blue"; that's the whole point of WP:V.
  3. But must that source then be cited? No, not necessarily; it is, technically, sufficient to provide it. While a reader-facing citation is one possibility, talk-page discussion, HTML comments, or other methods are also possible; remember that WP:V is a policy for editors, not readers.
  4. Can that source be cited? Sure, subject to the normal consensus of the article editors. Having more citations is rarely a bad idea; at the very least, it will prevent the question from coming up again and again and again.
  5. Is a lack of citations indicative of poor sourcing? Not necessarily, but, statistically speaking, the vast majority of articles with no citations are not in that state because expert editors have examined each claim and do not believe it requires citation. At this point in Wikipedia's development, it is quite reasonable to assume that an article with few or no citations should have more; it is incumbent upon the article's editors to demonstrate that it does not (per points 2 & 3).

Is there anything above that people disagree with? Or anything else that needs to be debated? :-) Kirill Lokshin 01:55, 3 October 2006 (UTC)

Looks good to me. Sandy 01:58, 3 October 2006 (UTC)
Looks good to me too. Well summarized, Kirill. ≈ jossi ≈ t • @ 02:08, 3 October 2006 (UTC)
See above ;) Bravada, talk - 02:16, 3 October 2006 (UTC)
If you feel these points represent consensus, please propose (or make) changes WP:CITE to make them clear to all. Points such as 1,2,3 have been vigorously debated above and at WP:GA. I have no objection to the ideas in 1 to 4. As for 5, I think that the majority of significant scientific articles have indeed been examined by experts, although you will find many editors above who dispute this. CMummert 02:10, 3 October 2006 (UTC)
Hmm, maybe just have a FAQ section on the page, then? ;-)
(More seriously, I'm not convinced that anyone has actually argued against these points so much as that the convoluted discussions have caused some statements to be misinterpreted at times.) Kirill Lokshin 02:14, 3 October 2006 (UTC)
(after edit conflict with Kirill) I believe that if these points represent consensus, it means that there is no need to make any changes to either WP:CITE or WIAGA. This is just what the current (i.e. from before the rampant changes made of late) guideline says. The recent change to the guideline seems to indicate that some users long for changes for changes' sake. No need to make any. Bravada, talk - 02:16, 3 October 2006 (UTC)
I can't speak to WIAGA (uncited articles are still getting through), but I concur there is no need to make any changes to WP:CITE (as it stood before this discussion). Sandy 02:20, 3 October 2006 (UTC)
(butting in) - Sandy, if you see an article with GA status and with insufficient or nonexistent references, simply delist it, citing ;) lack of those, or, if you do not want to make such decision yourself, submit them for GA/R. Thanks! Bravada, talk - 02:34, 3 October 2006 (UTC)
I have (one), and I will (another). Eventually. Sandy 02:35, 3 October 2006 (UTC)
Then how do you explain the results of the poll above? dryguy 02:22, 3 October 2006 (UTC)
Well, they seem to match neatly with my point #1, above; is there something else you have in mind? Kirill Lokshin 02:37, 3 October 2006 (UTC)
Then why was this line removed from WP:CITE? It is the same as your point #1 and the poll results.
Inline citations for every statement are not required, provided the statements are corroborated by the references cited in the article.
dryguy 02:45, 3 October 2006 (UTC)
Because as a blanket statement, it's not always correct, it will surely be misapplied, and it violates WP:V. In some articles, inline citations may be required for every line. Sandy 02:49, 3 October 2006 (UTC)
More circles. You agreed with Kirill above who says the exact same thing in point #1. How do you see this as different? dryguy 02:55, 3 October 2006 (UTC)
It doesn't say the same thing at all. There's that little word "necessarily", and the context, which is that a cite must be provided, per WP:V, anytime another editor requests it, and the Kirill's wording deals with how the cite should be provided. A blanket statement that every line doesn't need to be cited will be quoted as such and abused of. What was removed from the article was contrary to policy WP:V. What Kirill wrote is not. Sandy 03:00, 3 October 2006 (UTC)
(Cascading edit conflicts...) People who deal with more contentious topics are, presumably, wary of adding any text that could provide more loopholes for some of the not-so-nice editors to exploit. Perhaps a better (safer) way of expressing (what I think is) your point would be:

Editors must provide a source, which should be listed in the "References" section, for every statement in an article; while this does not necessarily need to be tied to the statement in question through an inline citation, it is often appropriate to do so.

How about that? Kirill Lokshin 02:57, 3 October 2006 (UTC)
It is good. I hope someone adds it or the equivalent, beacause I'm done for tonight. Bye! dryguy 03:01, 3 October 2006 (UTC)
If an editor has the source, why not add it inline? Wiki is dynamic: talk pages go into archives, subsequent editors inadvertently (or not) delete general references, articles endure. Inline cites attach the reference to the statement, where it won't get lost in the shuffle. Sandy 03:08, 3 October 2006 (UTC)
To the contrary of Bravada's claim, the motivation of these discussions is that several people from WP:GA claim that inline citations are required, which disagrees with items 1 and 3 above. They moved the discussion here claiming the consensus at WP:CITE supportes their changes to WP:GA. So either items 1 and 3 are consensus, or the GA people are correct, but not both. That is why an edit to WP:CITE is needed, so that everyone can tell which it is. CMummert 02:38, 3 October 2006 (UTC)
Actually, I would suggest that what started this whole GA mess was nothing more than point 5. The GA people saw a bunch of articles with a relatively small number of citations—at least when compared to GAs in other subject areas—and started leaving messages regarding that point. There's no reason to blame them, I think; as I said, it is the responsibility of the article editors to show—by whatever means they choose (although actually providing some more citations is a pretty obvious possibility)—that the articles are adequately sourced. Kirill Lokshin 02:43, 3 October 2006 (UTC)
Thank you Kirill. This is the point I tried to make over at the GA debate. That if the article editors thought an article was well-sourced, and a GA reviewer disagreed, then they can bring it to WP:GA/R so that it can be hashed out and established that it was indeed well-sourced (or, that it wasn't after all). --plange 02:49, 3 October 2006 (UTC)
So you are saying (point 3) that if, in response to the GA query, the editors just listed citations on the talk page, the GA people would accept this as adequate sourcing? This seems unlikely based on discussion above. What if they continue to protest that they want inline citations? And, as a factual matter, the articles in question have thorough references at the end of the article to cover the claims made (point 3) but not inline references. I expect that, without a clarification, reviewers at WP:GA/R will continue to press for inline cites, even with point 3 above. CMummert 02:56, 3 October 2006 (UTC)
That's something to discuss with GA, really. In theory, they could ask that all articles have a picture of a pink teddy bear; in reality, they're likely to be quite reasonable people who can discuss the best level of citation with you as needed. As far as inline citations go, it's probably best to determine an acceptable level on a per-article basis; but once you actually have all the citations available, it becomes trivial to add them in (either with real citations or simply with HTML comments) as the editors feel is appropriate. Kirill Lokshin 03:01, 3 October 2006 (UTC)
Why would someone list citations on the talk page, and not just add them to the article? If you've got the source, add it inline (see above). Mountain --> molehill again. Sandy 03:08, 3 October 2006 (UTC)
It might be possible to make a cogent argument for having (at least part of) the sourcing done through HTML comments, {{inote}}s, or something else not visible to the reader. Admittedly it's more of a technical consideration of what is and isn't "inline citation", but I don't think it's necessary to beat this point to death too much. Kirill Lokshin 03:14, 3 October 2006 (UTC)
(edit conflict) ah, I see. On the other hand, if an editor has the source, it would seem easier to just add it with the most commonly-used ref mechanism, to avoid problems years later. Sandy 03:24, 3 October 2006 (UTC)
That would be my personal recommendation, yes; but I don't think it's a significant enough issue to be worth arguing over, at this point. Kirill Lokshin 03:30, 3 October 2006 (UTC)
When I read But must that source then be cited? No, not necessarily; it is, technically, sufficient to provide it. I think this means a reference at the end of the article is sufficient, but the GA reviewers do not. It is not sufficient to discuss it with them; they refuse to accept the fact that they are interpreting WP:CITE themselves, and insist that they are merely following what WP:CITE says. Thus the editors of an article may feel that citations are available and included in the references section at the end of the article, while GA reviewers say these are not adequate citations. I think I may not be succeeding in explaining the POV that the science editors have about the role of WP:V. CMummert 03:20, 3 October 2006 (UTC)
They aren't interpreting it themselves: they are applying WP:V. Sandy 03:24, 3 October 2006 (UTC)
(edit conflict) Meh, read point 5 again, please! :-) The point is that the article editors (you) must demonstrate—explicitly!—that every statement in the article actually comes from the listed sources; this involves a bit more than just saying, "Oh, sure, everything here is from those three books; run along now." Most of the time, it's expected that the editors have—or can produce on demand—specific sourcing for every single statement. Whether you choose to then indicate that through inline citation is open to discussion; but if you already have the relevant statement-source linkage available, it seems trivial, to me, to actually add some (possibly quite broad) indication of its presence into the article. Kirill Lokshin 03:29, 3 October 2006 (UTC)
But read point 3 again, please! :-) This says that source need not be explicitly cited.
The interpretation of WP:V favored by several GA editors is at odds with the interpretation common among science editors. We feel that a list of references at the end of an article is sufficient for verifiability; since WP:V only says sources must be cited, it is WP:CITE that is at question. CMummert 03:32, 3 October 2006 (UTC)
It doesn't matter what GA says, or what you say is common among science editors (which I say is in opposition to medical articles, which I believe are science). Joe Bloe can come along, independently of "science editors" or GA, and ask for a cite for any statement, per WP:V, and that's the bottom line. Just because GA is letting 'em through, doesn't mean other editors will, because WP:V is policy. Sandy 03:35, 3 October 2006 (UTC)
OK, I'm really done for the night after this :). WP:V most definitely does not require inline citations, as you keep saying. dryguy 03:43, 3 October 2006 (UTC)
(edit conflict) That's not a question of what it means to cite, but a question of what it means to "provide a source". I would argue that saying something like "It's in Newton's Principia somewhere" is not specific enough to qualify (or, in a more limited sense, such a poor style of referencing that the article could not be considered "good" anyways). At the very least, providing some indication of provenance within a long work seems appropriate, regardless of whether any footnotes are involved. The idea of adding footnotes is merely to take care of the "provide a source" bit in advance for particular statements.
(Beyond that, the specific requirements GA wishes to impose need not match actual policy. I'll point out that there's no policy that says articles have to be "well-written" either, but that's a requirement too.) Kirill Lokshin 03:41, 3 October 2006 (UTC)
To Sandy: For many science (by which I mean physical sciences and mathematics) articles, if someone asked for a cite, they would be given a specific reference from the references section and told to read through it. Point 3 above says that this is enough; there is no need to edit the article because some random person asks for a citation for a well known fact. All that your questions indicate to me is that the 5 points above do not actually represent consensus. CMummert 03:45, 3 October 2006 (UTC)

(unindent) Since I'm one of the GA people that supported the new requirement for inline citations, I'll weigh in. As Kirill states in point 3, if the source for some statements are questioned by a reviewer or reader, you need to have some way of providing them per WP:V. It had never occurred to me before, but his point that they could be made in HTML comments would be a good example of satisfying that it is, in fact, sourced, and that would be inline. This would allay your fears (outlined in some arguments above) of looking silly to your peers who might be reading your articles and still satisfy GA reviewers perhaps (I'm just speaking for myself obviously). Keep in mind, this doesn't let you off the hook if Joe Blow comes along and wants a cite (since they wouldn't see the HTML comment). Since I wrote the above, more comments have come in, so I'll add in response to "if someone asked for a cite, they would be given a specific reference from the references section and told to read through it" -- can you not then add it as an HTML comment so it won't keep coming up again and again, wasting editors' time? Asking for inline citations is our way (at GA) of making sure it is verified and, again, I'll repeat, that if you run into a situtation where a GA reviewer was being unreasonable in what they're asking, please bring it up at WP:GA/R so we can discuss specifics instead of generalities. Knowing when and how to cite is an art form, and should be handled on a case by case basis. The way WP:CITE currently stands seems valid to me on how to implement the WP:V policy. --plange 03:59, 3 October 2006 (UTC)

Having read the many kilo(mega?)bytes of discussion above, I'm very glad to see Kirill bring this towards a consensus. IMHO User:qp10qp also made some excellent points. I would summarise my views as possible:
  • Until we have a complete system for article validation (i.e. fact-checking of every single fact against an authoritative source) in place, perhaps along the lines of Wikipedia:Pushing to validation and m:Wikicite we have to work with a commonsense (dangerous word!) compromise. That means we need to be restrained in our {{fact}} tags, but if a key statement seems dubious or unsupported in existing refs/footnotes we should request a source.
  • At the same time we need to make it clear (for both a non-expert reader and an editor) where they can verify information on a topic. I could (perhaps) write a lot of articles off the top of my head, but a citation clarifies the fact that I didn't.
  • Regarding common knowledge — of course commonsense (that word again) dictates that I'm not citing lots of things "everyone" knows — but if that "common" knowledge lies at the heart of the topic then it's certainly worth a reference or two (not 57 refs!). Even when I worked on the definition of chemical substance (which as a chemist I should know by now) I cited sources to make sure I got the nuances right. This turned out to be critical as I had in fact argued that the previous definition (uncited) used was flawed; my earlier uncited edit was reverted, but my new cited one was better and it was even supported by the "reverter."
  • Citing any source has a bias towards a particular publisher; but Wikipedia policy requires us to verify facts so we can't avoid that. I've never had anyone remove a published reference I've given on the grounds that it showed publisher bias.
  • I think inline references are the most specific (=>powerful) way to cite sources, but general refs have their place too.
  • I do not think that having an occasional superscript spoils an article's appearance; from recent FAs I think most agree with that.
  • I think what the reviewers at GA require is up to them; if the consensus at GA among reviewers is to require inline cites, that's GA policy. Reviewers aren't perfect, and they will make mistakes. We need to assume good faith, remember those reviewers devote many hours of their time reviewing your articles. If you don't like the policy they use, that's fine too, just nominate it to your WikiProject for A-Class status instead (that classification is deliberately more flexible and is ideally used with peer review).
  • Maybe I'm an optimist, but that's why I'm a Wikipedian. I think Wikipedia is based on the premise that things will evolve towards a commonsense (again!) consensus that works well. I think the Good Article project is still evolving, and it will reach a stable state which (almost) everyone is happy with.
  • Outside of GA, I see no reason to change the wording of existing general policy for citing sources. This is the same wording that has allowed people to interpret whatever is appropriate for their article on physics, military history, pop singers.
In summary, we can agree to allow a flexible interpretation of general policy (in some cases agreeing to disagree on interpretation), and meanwhile the GA team can separately agree on their interpretation of citation policy. Let's rally round Kirill's commonsense summary above, and draw this (important) discussion here to a close. Walkerma 05:43, 3 October 2006 (UTC)
Rallying around the flag! Thank you Sandy, Kirill and Walkerma for helping to clarify the guidelines! --plange 05:52, 3 October 2006 (UTC)
To respond to Walkerma: my goal here is to find a way to get the science articles to work with GA, not find a way to ignore GA and replace it with a per-project assessment. The point of my poll from yesterday is to verify what I always thought was consensus: it is not necessary to have an inline citation for every fact in an article. It seems that the GA editors are content with a vague policy that doesn't define which sorts of facts need inline cites and which do not. I had hoped that some clarification would be forthcoming. CMummert 10:27, 3 October 2006 (UTC)
While only having scanned the above, I'll just jump in with one point that I find myself repeating a lot lately: GA is not a policy. It does not create policy and it is superceded by the policies that exist. It is a process that many have attached themselves that is basically in limbo. It can generate content improvement, but as often as not it generates hundreds of k of wasted talk time. This effect, judging from above, seems to be spilling over. Marskell 11:20, 3 October 2006 (UTC)
I just now tried to follow the discussion. I confess I don't know exactly how this discussion started, but Sandy's opinion ("If an editor has the source, why not add it inline? Wiki is dynamic: talk pages go into archives, subsequent editors inadvertently (or not) delete general references, articles endure. Inline cites attach the reference to the statement, where it won't get lost in the shuffle.") fully expresses me. Don't be afraid of inline citations! Provide them even to confirm that "the sun is shining". Inline citations along with good references make the article more credible.
For me it is also important the level of research of an article. Especially those going for FA and GA status. I believe and I have said that the level of research in such articles is not always as good as it should be. If you are interested in this topic check Wikipedia talk:Featured article candidates#Are the FAs that pass really FAs?.--Yannismarou 15:22, 3 October 2006 (UTC)

When should citations not be included?

Some people here have taken the position that one should always provide for references if requested. The failure to do so could lead to the removal of the uncited text. An obvious flaw in this rule is the case of simple facts that can't be found in the literature precisely because of its trivial nature.

I can live with a rule that says that references should always be provided upon request. For common knowledge facts only if such references exist. It has to be established by consensus that the reason why the refs don't exist has to with the trivial nature of the fact and not because it is not an established fact. Count Iblis 12:50, 3 October 2006 (UTC)

Again, that's not a question of citations, but of general sourcing; but your point is quite unacceptable, I think. The whole intent behind WP:V is that everything in Wikipedia—no matter how commonsense or trivial—needs to have come from an external source; we should never be the first party to publish some fact. (But really: do you have any example of something that has never been published before, but that we would still want to include?) Kirill Lokshin 13:08, 3 October 2006 (UTC)
IMO things that qualify as "obvious" to the point of lacking a citation are general points that will be cited elsewhere in the article on a more specific level. You don't need to cite "the Sun is very hot", if a paragraph later you provide (and cite) the actual temperature. If the obvious point is a singular mention, however, it should be sourced. Marskell 13:55, 3 October 2006 (UTC)


I agree that in most cases this doesn't happen and you can find citations to very trivial things. I gave the example of some pseudoscientific topics in a reply to someone else above where this does happen. You want to be able to give the reasons why some pseudoscientific claim is wrong and of course you must then cite the sources on which that is based. However, in most cases you won't find the actual arguments presented in the literature, because (debunking) pseudoscience is not an active topic of research.
Some time ago User: Agne27 included a citation wanted tag in the special relativity article for the claim that simultaneous events in one frame is not, in general, simultaneous in another frame, even though it follows straightforward from the theory. In that case the tag was "innocent", such references do exist, but including an inline citation would make the article look silly. But if we can be challenged about such straightforward consequences of established theories, then that would bar us from presenting the scientific POV on pseudoscience here. Wikipedia's reputation already suffers enormously because we can't write such articles in an acceptable way, e.g., according to wikipedia, cold fusion is a viable option, Heim theory is a promising candidate for the theory of everything just like string theory, etc. Count Iblis 14:08, 3 October 2006 (UTC)
To provide the right context, she placed the tags in the article because she was requested to do so to help identify where there may be some potential problems. It was not a challenge! Any differences of opinion could have been settled by quietly talking on the Special Relativity talk page rather than raising it on CITE. This is going round in circles again. Kirill Lokshin and Walkerma has summed it up nicely. Let's get back to working on articles. RelHistBuff 15:37, 3 October 2006 (UTC)
"...such references do exist, but including an inline citation would make the article look silly", says Count Iblis (great name, by the way). But why would a little blue number make an article look silly? It might look superfluous, if that's what you mean, to a science expert. In my field, history, if a see a footnote mark on something I know inside out already, I don't see it as silly—I just read past it, accepting it might be helpful to someone else. If people want to argue that inline cites are unnecessary, that is a viable point of view, though I don't agree with it; when people argue that inline cites are actually damaging in some way, I find that, to use Count Iblis's word, silly. qp10qp 15:48, 3 October 2006 (UTC)
It's a matter of taste whether or not it looks silly. That's why this particular case isn't a very serious problem. But, this maters becomes serious if for some reason no referenses exist, as I pointed out above. Count Iblis 17:41, 3 October 2006 (UTC)
Ditto. Claims that inline cites in an encyclopedia that anyone can edit are silly ... are silly. Sandy 15:51, 3 October 2006 (UTC)
Suppose that the consensus of editors of the special relativity page is that the article is adequately sourced after a few of the fact tags are replaced with citations, but not all the fact tags are replaced by citations; would a GA reviewer be justified in still claiming that the article is not adequately sourced? Both sides would say that WP:CITE is on their side. CMummert 17:25, 3 October 2006 (UTC)
I feel citations are needed within articles for many reasons. It;
  • Shows Wikipedia can be a reliable source, and checks the facts.
  • People can look at the inline citation sources fully and read them in context.
  • Shows that Wikipedia is an encyclopaedia, and not a drop off point for others to vent their ideas. Original research material should be reserved the appropriate publication, not here. LuciferMorgan 18:31, 3 October 2006 (UTC)

This is my last comment on this issue for the time being, because I think everything's been said. I believe it's time for CMummert, Count Iblis, and others to take something away from the discussion, even if they will never be convinced of the value of inline cites:

  1. They don't have to do inline cites if they don't want to.
  2. They can still seek validation for their articles elsewhere than GA.
  3. No-one doubts their good work or good faith or that of their colleagues.

(Wipes brow and leaves room) qp10qp 19:12, 3 October 2006 (UTC)

Point 2 above may point to the right solution to this conundrum. Maybe we need a separate endorsement process for technical articles, say "Good Technical Article" (GTA), where the needed reviews would be carried out by editors sensitive to what is appropriate in these articles. This would include appropriate citations along with, perhaps, other criteria that the GA process might miss, like well formatted equations or appropriate level of detail in sketched out proofs. I think it would be very interesting to hear what people in other technical disciplines say about a given technical article. As with everything else on Wikipedia, we would rely on editors self-selecting for this role. Yes, Wikipedia is written for the general reader, but that doesn't mean people with no background in writing about technical subjects should dictate how those articles should look. --agr 19:44, 3 October 2006 (UTC)
They could utilize the A-class designation and set it up as a Peer review vote, as WP:MILHIST does for theirs. --plange 20:47, 3 October 2006 (UTC)

As a more of less objective third party voice (since they are not GA reviewers or members of the Physics/math projects), I think Kiril, Sandy, Qp10qp, and Walkerama have made many (similar) summaries of consensus that I have no problem acknowledging. It's been stated elsewhere but there is no obligation to for an article to carry a GA tag. There are many other peer review/assessment systems that can be adopted (or news ones like GTA created). At GA we strive to identify articles that are well referenced meaning, as WP:V notes that any editor can verify the information and all information is clearly drawn from a published reliable source and there is no OR. GA caters more to the average reader then to the specialist and if a GA reviewer can not verify the information, then we can not, in good conscious pass the article for GA. (As plange has noted many, many times WP:GA/R can always be used if a reviewer was potentially unfair) If an article doesn't have a GA tag, it doesn't mean it's a bad article, it just means that it is beyond the scope of GA consideration and criteria at this time. Agne 21:58, 3 October 2006 (UTC)

Inline citations for every sentence are not required

What is controversial about this? --ScienceApologist 21:49, 3 October 2006 (UTC)

As I believe Sandy noted, there maybe an article that would actually require an in-line cite for every line. I can't readily think of an example off the top of my head but I could visualize a 5 line stub with some extraordinary claims that would feasibly need 5 in-line cites, one for each line. It's certainly an article by article basis and I think the controversy is that line would put itself in potential opposition to WP:V. Agne 22:01, 3 October 2006 (UTC)
I believe the onus is on you, then, to find such an article if you are claiming one exists. --ScienceApologist 22:10, 3 October 2006 (UTC)
Almost one per every line, and many lines require more than one. (Except the lead, which is a summary). Tourette syndrome, completely typical of journal-published medical articles, which often include up to five inline citations for every single sentence. Sandy 22:56, 3 October 2006 (UTC)
There is no onus to find anything really. Even if such an article doesn't exist now, it doesn't mean that it will never exist or more importantly that it could not exist. The fact that your insertion could directly oppose WP:V for an article that would need and in-line cite for every line is what the controversy is. Agne 22:14, 3 October 2006 (UTC)
Hypothetically, then, there might be an article that never requires any inline citations. Does that mean we should delete this guideline? Absolutely not. Trying to eke out extreme hypothetical abstractions doesn't support your case very well. In any case, I have modified the wording so there is no need to be interested at the article level. We can talk about the entire project. --ScienceApologist 22:17, 3 October 2006 (UTC)
No, that hypothetically can never exist because WP:V clearly says that everything must be verified. You could have a sloppy stub with 3 lines and then a source listed at the bottom instead of an in-line cite. That doesn't mean that an in-line cite is not needed, it's just not being utilized. Agne 22:30, 3 October 2006 (UTC)
Now, please, Agne27, don't go attacking this stob as "sloppy". Just because they hypothetically snubbed your inline citation pet doesn't mean that this is an inappropriate or unverified work and certainly it is inappropriate to claim that an in-line cite is needed just because that's what you think. In my hypothetical article there is no in-line cite needed because the sources listed are global in nature and the article is written solely on the basis that there is some reason it is notable to know that the sources themselves exist. Therefore, there is no in-line citation required. If you can have a stub that requires a citation on every line, I can have a stub that doesn't. --ScienceApologist 22:38, 3 October 2006 (UTC)
(edit conflict) Blanket statements tend to be. ;-) There are certainly cases where inline citations are required for every (or nearly every) sentence.
In addition, the addition omits the critical point that the statements must not only be corroborated by the references, but must be explicitly linked to specific references on request, even if such links are not ordinarily present in the article (or visible to readers). Kirill Lokshin 22:02, 3 October 2006 (UTC)
Forget the (or nearly every) because that's not what the sentence says. The addition doesn't need to include the "criticial point" because the critical point is already included in the guideline. Since the points do not contradict each other this objection is hollow and irrelevant, based out of hypothetical fears and a citation-chauvanism rather than a realization that there is more than one way to write a well-referenced article. Inline citations are not required for every sentence in Wikipedia. This is a statement NO ONE has yet contradicted. --ScienceApologist 22:10, 3 October 2006 (UTC)
And not one person, including several contributors here, was willing to sign a poll in support of that statement. I see no consensus that cites are required for every fact, so the sentence There is not consensus that every fact must have an inline citation would be a fine addition to WP:CITE. The real question, though, which I emphasize nobody seems willing to answer, is exactly which facts do not require inline citations. It is this issue that WP:CITE needs to clarify. CMummert 22:17, 3 October 2006 (UTC)
The issue has always been WP:V (It's clear that everything needs to be attached to a reliable source that can be verified, regardless if it's done on a line by line basis or with an entire paragraph covered by one cite). I don't think the editors here can make that sentiment more clear. Crafting a strawman poll that could potentially put itself in opposition of WP:V doesn't solve or clarify anything. If you have info in an article, it needs to be verified. It's simple. Agne 22:27, 3 October 2006 (UTC)
WP:V does not require inline citations. It does not specify any citation style. That is the job of WP:CITE. dryguy 22:31, 3 October 2006 (UTC)
(edit conflict in parallel with dryguy) WP:V does not say any inline citations are required; it says facts must be verifiable and need some sort of citations, then points to WP:CITE as a guideline for how to cite things. Also, WP:CITE only indirectly says inline references may be required (and this is only because the three citation methods it lists are inline styles). Changing how WP:CITE sugggests a fact should be cited does not changed whether that fact is verifable and thus does not affect WP:V. CMummert 22:34, 3 October 2006 (UTC)
No, the issue has NEVER been WP:V, and we have said this time and time again. The issue is whether WP:V REQUIRES in-line citations. It is clear that it doesn't. In-line citations are helpful and useful for complying with WP:V but they are not necessary. No one is arguing that we should "unverified information". We are arguing that not all verified information lends itself to in-line citation. --ScienceApologist 22:38, 3 October 2006 (UTC)
As WP:V says, any editor can remove information that they can not verify. You can say an item is verifable all day long but an in-line cite is more empirical proof of that then just your word. Agne 22:43, 3 October 2006 (UTC)
That was the point of my post of 17:25, 3 October 2006 (UTC) which nobody replied to. CMummert 22:48, 3 October 2006 (UTC)
As WP:V does not say, any editor cannot remove information simply because it doesn't have an in-line citation. Claiming an in-line citation is "empirical proof" is a slap in the face to people who deal with real empirical proof. An in-line citation is just someone who knows how to use a citation system and it is still based on the word of the person who placed it at that point that the cite is accurate and verifiably supports the statement it claims to. Just because someone used a citation doesn't mean that the statement is verified. Hell, I could finish all my in-line citations requested for Redshift right now by putting in a reference to an obscure textbook no one could access because it is out-of-print or some such nonsense.
Basically what you are arguing for is lazy editting. As long as someone [1]puts [2]a [3] reference tag in the article, you think WP:V is satisfied. It's not. That's why WP:V doesn't say anything about not removing statements that are cited because cited statements can be just as unverified as non-cited statements. Simply having a cite does not make the statement any more or any less likely to be verified. --ScienceApologist 22:50, 3 October 2006 (UTC)


There is also a common sense application that is being overlooked. Just because there isn't a footnote tag at the end of each and every sentence doesn't mean that every sentence in a paragraph isn't supported by an in-line cite at the end of the paragraph. You could have a 10 line paragraph with a cite at the end that supports each and every one of those lines. Every line is cited (and thus passes WP:V) but there is only 1 footnote instead of 10. Agne 22:20, 3 October 2006 (UTC)

Even more reason to include our sentence in the guideline! --ScienceApologist 22:22, 3 October 2006 (UTC)
Except for the articles where it would violate WP:V not to have an in-line cite for each line.  :) Agne 22:31, 3 October 2006 (UTC)
Once more, in unison: WP:V does not require inline citations. dryguy 22:41, 3 October 2006 (UTC)
But it doesn't say that there are statements in every article. You are adopting an informal linguistic fallacy in trying to claim that it does. Just because not all P are Q does not mean that Q doesn't exist. --ScienceApologist 22:40, 3 October 2006 (UTC)
No, as someone who has done stub sorting, it just accepting the nature of Wikipedia. Anyone can create an article with a much or as little information as they can leaving you with plenty of P's and Q's as well as a R's, B's and Z's for that matter. :) Agne 22:47, 3 October 2006 (UTC)
I'm not getting through to you. Your previous statement was an invalid argument. In no way was it ever said that "every article has statements that don't require in-line citations". Which is what you are claiming we said (All P have Q). Instead what we said was (Exist P have Q). See the difference? --ScienceApologist 22:54, 3 October 2006 (UTC)
I tried re-reading the entire Citing sources style guide to see whether this addition about inline citations was in the right place. My overall impression is the "style guide" looks not so much like a "style guide" as a former battlefield littered with debris. The whole guide needs to be done over from scratch with a view to being helpful to the beginner. --Gerry Ashton 22:31, 3 October 2006 (UTC)
Unfortunately, Gerry might be very right here! Bravada, talk - 23:09, 3 October 2006 (UTC)
I can see the merit in that but I think it would be prudent to wait till this discussion dies down and some cool, objective heads that have not been involved in this more recently debate could take the time to re-work it. Agne 23:14, 3 October 2006 (UTC)

An example

Let's look at this example. Where do I put the requested references? I explained on the talk page of this article that at most I could give a ref to some textbook at the start of this section. But since such books require more background knowledge to understand than what I wrote, you won't be able to verify the correctness of the derivation using the book. If you can understand the book, then you wouldn't need to verify the derivation I gave in the first place. Count Iblis 22:23, 3 October 2006 (UTC)

I would assume that there are textbooks that show this derivation, no? You could just put a single footnote at the top of the section (after the first sentence, perhaps) saying something like "The derivation here roughly follows that given in Doe (1986), Jones (1992), and other standard textbooks." Kirill Lokshin 22:32, 3 October 2006 (UTC)
This is, then, citation for citation sake. It doesn't help in verifiability at all. It's just mindless clutter. --ScienceApologist 22:42, 3 October 2006 (UTC)
Not really. It doesn't necessarily help to actually verify the statements; but its presence does serve as an indirect indication that someone has examined them. In other words: if a statement appears in an article with no citation, someone wandering by has no idea whether that statement (a) has been examined and found not to need one or (b) has never actually been checked. If there is a citation, then there is, at the least, some assurance that another editor has examined the statement and believes it to be sourced. Kirill Lokshin 22:47, 3 October 2006 (UTC)
Baloney[4]. Are you telling me that simply by including the citation I have indicated I actually checked/verified my "baloney" statement? Anybody can insert a citation. That's not an indication that any examination has occured. It is very dangerous to make this kind of argument by association. Just because there are many articles that are well-verified and have a substantial amount of inline references does not mean that all articles with a substantial amount of inline references are well verified. Likewise there exist articles with few inline-references that are well-verified and some that are not. This is basically confusing concepts and indicators, and it should not be done. The inline reference is a concept, a fairly meta-concept that isn't present on poor articles because poor articles often don't have very many meta-concepts in general. Just because it is a meta-concept does not mean that it is not an indicator of a verified-article.
To illustrate this another way, consider the fact that most articles with images tend to be well-verified. That doesn't mean that in order to meet verifiability policy we should include images in articles. This is an argument by association and it is clearly not a valid one.
--ScienceApologist 23:11, 3 October 2006 (UTC)
Are you suggesting that you have a habit of inserting nonsensical citations? ;-)
I would think that, as a general rule, people inserting citations have actually checked the source; while it's possible that some (bad-faith) editors just make them up, they are likely to be responsible for only a statistically insignificant portion of citations. Kirill Lokshin 23:19, 3 October 2006 (UTC)
Have any data to back this up? You are confusing correlation with causation right now. The general ruel that most people inserting citations actually checked the source only works because most people inserting citations are doing so because that's what they like to do. As soon as you start telling people that the only way to make sure an article is verifiable is by including citations, you'll get plenty of bad citations. After all, we should be in the habit of checking up on people's citations anyway. Just because a citation exists doesn't mean the statement is verified. --ScienceApologist 23:25, 3 October 2006 (UTC)
The reason why people wouldn't make up citations wholesale is pretty much the same reason that most editors don't vandalize, try to introduce bias into articles, or make personal attacks; most editors are here to improve the encyclopedia, not to play silly buggers with it.
(You're correct, actually, in that we should check up on people's citations. It is somewhat difficult to do so when there aren't any, however.) Kirill Lokshin 23:31, 3 October 2006 (UTC)
So the big issue is that someone saying that WP:CITE is the key to WP:V is really not correct. We must verify citations. Simply having citations doesn't imply anything about the verifiability of an article, despite what some editors here have intimated. It might make an article more easily verifiable/criticizable, but it might also be a way to lie or stretch the truth in order to claim verifiability when a point actually isn't verifiable. Or, in a more honest case, including a citation to a derivation may be a matter of picking-and-choosing that is best left up to a discerning reader rather than an editor closing their eyes and pointing towards a random source. --ScienceApologist 23:38, 3 October 2006 (UTC)
(edit conflict) Maybe you're right, but you need to give the reader the chance to verify and understand it. Without a cite, they have no idea where to start. There are smart people out there who might not be physicists and so wouldn't be immediately familiar with that section, but might be smart enough to read the relevent pages you reference and deduce the same thing. This argument that we wouldn't understand so there's no need to cite just isn't valid. You need to assume we can. --plange 22:43, 3 October 2006 (UTC)
Are we to assume that the reader doesn't know to start with the reference section of the article? If they are that dumb, how are they going to know what to do with a citation in any case? --ScienceApologist 22:44, 3 October 2006 (UTC)
And so with a list of 10 books of varying degrees of scope and breadth, where would I find this particular nugget of information that I need to verify? At the very least a cite at the end of the paragraph directing me to a particular book and particular chapter (maybe even page), gives me a bridge to where I can find the information. Agne 22:49, 3 October 2006 (UTC)
If you need to verify this "nugget", you're going to have to read the entire book. No shortcuts in physics, I'm afraid. --ScienceApologist 22:51, 3 October 2006 (UTC)
Sorry, for butting in a comment here, but timezones make it impossible to respond. Physics is NO EXCEPTION. Yes, the field is difficult, but it is understandable, maybe not to all, but at least to many. I do believe, having gone through from undergrad texts to doctorate, that verification can be done by checking a textbook by someone with a medium level of intelligence. RelHistBuff 06:57, 4 October 2006 (UTC)
Physics is a different subject than others, far more mathematical than anything except, well, math. Is it so impossible to believe that the readability of the texts is therefore different as well? Your accusations of elitism are not going to move the discussion forward. -- SCZenz 08:12, 4 October 2006 (UTC)
I agree it is more mathematical, but I repeat, verifiability of physics material is possible, again maybe not all persons can do it, but at least many can. Therefore physics should not be treated differently in Wikipedia. On the point of elitism, I must say that the claim that physics is different and should be treated differently from other fields, well, I won't speak for anyone else, but that claim certainly appears elitist to me. When one make claims like that, it's no wonder that most people think that we don't live in the real world. I say "we" because I am a former physicist. RelHistBuff 09:11, 4 October 2006 (UTC)
This is basically an argument against nobody. We aren't asking that physics be treated "differently" as a means of making a hierarchy of norms, we're asking that the context of articles be considered and the fact that articles are written differently be taken into consideration when evaluating them. Wikipedia, inasmuch as it is NPOV, does not skew favorably toward one subject or another, but it doesn't treat all subjects the same. An article on Days of Our Lives is not, nor should not, have the same format, style, tenor, and organization as an article on the categorical imperative or complex analysis. That's how it should be, different subjects demand different treatments. Otherwise, we could just write a script and have artificial intelligence protocols write the articles for us. What we are saying is that citation, like many other aspects of articles, is appropriately going to vary from one kind of article to another. Trying to impose a monolithic standard is not beneficial to the wide range of subjects Wikipedia has articles about. --ScienceApologist 13:42, 6 October 2006 (UTC)
"What we are saying is that citation, like many other aspects of articles, is appropriately going to vary from one kind of article to another. Trying to impose a monolithic standard is not beneficial to the wide range of subjects Wikipedia has articles about" Wow, do we actually agree on something? That's exactly why we are resistent to add more language to the guideline (making it more monolothic), because doing so will make articles conform more to some magic formula of # of cites, etc., when right now, its current flexibility allows intelligent and common sense folks to determine for that article if it is well referenced. --plange 20:08, 6 October 2006 (UTC)
You fail to understand, though, that the kind of phrase the editors you are fighting against want included is wording to the effect that there is no monolithic standard so that we can avoid the imposition of well-meaning but misguided editors who think that there is such a thing. If CITE were to explicitly state this then we wouldn't have any issue. --ScienceApologist 13:52, 8 October 2006 (UTC)
Well this Wikipedia, not a physics class. You are placing undue burden on the reader and hampering verifiability. Agne 22:55, 3 October 2006 (UTC)
That's the nature of the beast. Some topics will be so obscure as to require a heavy amount of work in order to verify them. Other topics are more easily verified. But there is no statement in all of Wikipedia's guidelines or policies that says the process of verification has to be easy, it just has to be doable. I'm sorry that physics tends to be difficult to reference without a very large context, but there you go. --ScienceApologist 23:03, 3 October 2006 (UTC)
Then what's the big deal with adding a cite per section that references these? And then you could put in an HTML comment at the start of the section that as of such and such date this section can be verified by the books referenced at the end of the section? --plange 23:09, 3 October 2006 (UTC)
What in the hell for? Why should we include an inline citation that references every source in the reference section? That's an absurd request! --ScienceApologist 23:14, 3 October 2006 (UTC)
Sorry, left right after posting to eat dinner (!) and just got back to all this. I'm glad you asked this as maybe this highlights where we differ in our views and this will give me a chance to explain my statement. If you and I were working for a reputable scientific journal I would completely, 100% agree with you and would laugh at such a notion. I mean, come on, the journal is from x (insert reputable journal name here), it's written by y (whose name is attached to it) and they have sources at the bottom, all of which means the reader can be confident of several things: 1) It's coming from x publisher that has a reputation for its rigorous peer review process and wouldn't let questionable suppositions through 2) the author is Y, and I know who Y is, they know their stuff and 3) it has a list of references listed at the bottom that I can trust supports their writing. Agreed? Now, what we have here is a completely different animal. For one thing, the article you write, that is based on all the references at the bottom, is not a stable, inviolate thing that is printed. Anyone can edit this article. Your name isn't attached. We have no guarantee that you or any of the people currently writing/editing at WP:PHYSICS will still be around, and you can't guarantee that either. It's not possible to guarantee that, and if you assert it is, you're being disingenuous. Because of this, WP has to do some things considered unconventional so that we can strive to present factual articles. So, to get to why what I said isn't crazy: Someone might come along later (when none of you are watching this article) and insert another paragraph in that section which is not based on the refs at the bottom but on something completely different. We make them cite it, but now we have a section that can start falling apart at the seams, because other statements can start accreting, and then we have no idea where the original paragraphs you wrote for that section came from because it's now part of a larger section (and there's probably more references at the bottom we'd have to sort through to figure out where it might have come from). However, if you had gone ahead and taken the few minutes to add a reference at the end of that section to say "As of this date, 3 Oct 2006, this section is derived from the following sources: (Brown, 1998) (Smith, 2002) then you save future editors a huge headache and save the integrity of your article. WP:FAR right now is dealing with this issue BIG TIME, because a lot of old FAs were promoted in the days when references at the bottom were fine. The problem is now, that the editors who wrote those are no longer here and we can't find people who are caretaking them or are SMEs and so we have to demote them and hope a) an SME will come along to reference it b) some nice editor is willing to crawl through all those references and figure out which statements in the article are supported and which crept in after the editor left and are not supported. Sandy could probably speak more to the FAR problem, but recently all current FAs were added to a list and the same notification process that GA did is happening for FAs. --plange 03:41, 4 October 2006 (UTC)
"we have no idea where the original paragraphs you wrote for that section came from because it's now part of a larger section" --> Um, sorry, no. That's a poor rationale. The nature of the Wiki is you can go back and look at what previous versions of the article were. If an editor has come in and is adding bad content, it's very easy to go into the article history and find out who added what content when. What you seem to be talking about here is the ease of reviewers, not the ease of readers. If a reader comes in and uses the references at the bottom, they will find that the section of the article is incorrect and proceed accordingly. If not, the reader isn't helped whether there is an in-line citation or there isn't. Including an in-line reference to a fact that uses all the references does no service to the discerning reader and the undiscerning reader is undiscerning: it doesn't matter whether we cite or not because they aren't going to check up. The ONLY rationale then for including this ridiculous in-line citation is so potential reviewers have "insurance" that the points are valid without doing the dirty work of verifying the text with the references themselves. What you are saying is, in effect, that, no, we don't need to check the references: we can see whether an article is verifiable just by counting references. Poor form. --ScienceApologist 21:00, 4 October 2006 (UTC)
At FAR, lack of inline citations is the hardest thing for an article to recover from, because you have to almost start over if the original authors go missing. Poor prose, bad article organization, a lead that needs to be rewritten, other factors that lead to article deterioration over time are easier to fix if we can get to the sources. You make a good point, which I thought was obvious to everyone, but maybe it's not to those who aren't dealing with articles that have two years' worth of deterioration, and not one editor who was involved in the original writing still on Wiki. Wiki is dyanmic, editors who got an article to FA status six months ago may be gone today. Inline citations don't guarantee an accurate article, but they make it easier for others to find/fix the problems down the road. Besides that, general references listed at the bottom of the article may get deleted by a subsequent editor: you have to have nerve to delete an inline cite attached to a specific sentence. Sandy 04:01, 4 October 2006 (UTC)
If at some point in the future we no longer have a cadre of editors knowledgeable in physics, then it will not be possible to maintain the bulk of Wikipedia's physics articles as freely editable. It will be necessary to freeze the articles at some stable version created when there were knowledgeable people around. The same holds true for math, and several other branches of science and technology. No amount of inline citations can prevent that. People lacking a certain minimum of training will not, in general, be able to verify the more technical articles even if pointed to a page number for each sentence. Conversely, someone with that level of training can verify most articles from any of a number of textbooks that cover the subject in question, just by using the index or table of contents. Fortunately there is every reason to believe that we will have such people editing at Wikipedia for the indefinite future, unless we make the process so burdensome by adding unreasonable requirements that we drive them off. That is what's at stake here.--agr 09:07, 4 October 2006 (UTC)
So there is, presumably, a single book that could be consulted (out of the, say, half-dozen listed references) for the given statement? Why not just cite that, then? ;-) Kirill Lokshin 22:53, 3 October 2006 (UTC)
Nope, EVERY book could be consulted. --ScienceApologist 22:55, 3 October 2006 (UTC)
Could, or needs to be? In other words, is there some minimal subset of the listed references consulting which would be sufficient to verify the statement? Kirill Lokshin 22:58, 3 October 2006 (UTC)
The minimal subset of references is one. You can choose any reference you like. They will all lead you in to the same end. --ScienceApologist 23:01, 3 October 2006 (UTC)

I was trying to follow but I lost the plot halfway through. So what's the problem with citing a source which contains the derivation in this section? Bravada, talk - 23:08, 3 October 2006 (UTC)

Every source contains the derivation, but not in one page or even in one section. Trying to make a citation to this is an exercise in absurdity. --ScienceApologist 23:15, 3 October 2006 (UTC)
Are you trying to say that this WP article is the one and only source in the world that includes this derivation in such a compact form? Secondly, why can't you just reference the original paper - I guess it would contain the derivation, or wouldn't it? Bravada, talk - 23:19, 3 October 2006 (UTC)
See Count Iblis's response below. --ScienceApologist 23:22, 3 October 2006 (UTC)

Kirill Lokshin, plange, the derivation I gave is different form what you'll find in textbooks, precisely because I wanted to give a derivation that can be understood by people lacking the necessary background. I could have copied and pasted a derivation from an online textbook and then even people who don't want to understand the derivation could verify it. I'll include a reference to a textbook, but I'll also say that this derivation is different. Count Iblis 23:18, 3 October 2006 (UTC)

If you did it yourself, and it's not in any textbook, it's original research. Sandy 23:45, 3 October 2006 (UTC)
Wrong. In science, one can illustrate arguments and build from first principles as required and it is totally unoriginal in terms of research. It is original arrangement, not original synthesis. --ScienceApologist 18:21, 4 October 2006 (UTC)
This is the problem. Although the derivation is verifiable by the references given, it isn't necessarily citable because the references didn't make this particular derivation. Nevertheless, if one bothers to actually read the references, the derivation is a trivial result of the references. This kind of articulation is not well-understood outside of math and science circles, but it happens in math and science writing all the time. --ScienceApologist 23:21, 3 October 2006 (UTC)
That's not a problem, actually. If you can point to a source that contains a proof that your derivation is correct, even if this source contains a more elaborate derivation you have shortened for readbility, it's just what you should cite. An example from another field - suppose we have an article on somebody's historic proposal of a, say, tax reform. And suppose some other prominent person wrote a lenghty treatise in which he or she would argue why the original proposal is bad and they disagree. I believe it is OK just to write in the article that "this and this person disagreed with the proposal" and cite the work of the abovementioned critic as a reference, without necessairly repeating it in full. It is the information, or facts, that actually matter, not the way they are presented. Summarizing is common practice in encyclopedias. Bravada, talk - 23:47, 3 October 2006 (UTC)
Every source contains a proof that the derivation is correct. So should we source every reference in an inline cite? --ScienceApologist 18:21, 4 October 2006 (UTC)
User:Count Iblis said: "the derivation I gave is different form what you'll find in textbooks, precisely because I wanted to give a derivation that can be understood by people lacking the necessary background." Are you saying that it was derived by yourself? How can a reader be sure that it is correct? Because you're a scientist? This is a kind of origininal research, isn't it? — Indon (reply) — 00:20, 4 October 2006 (UTC)
The reader just has to read through the derivation. I didn't give the standard derivation, because that wouldn't serve any purpose. What I did can be "distilled" from textbooks too, but not from the chapter on Planck's Law. If you go directly to that chapter then you'll miss all the prerequisites that has been explained in the previous chapters.
Wikipedia is actually an extremely suitable medium to give self contained derivations because, unlike a book which has a one dimensional structure, wikipedia with all its internal links is multi-dimensional. So, you can give internal links to some tools you need for the derivation. It is thus convenient to adapt your derivation by taking into account the existing wiki articles on related topics. This will inevitably result in a derivation which is unlike anything you'll find in books.
To verify that the derivation is correct, you'll have to read the textbook on the subject and then read the wiki article. You can't verify it by not understanding the derivation. Is this a bad thing? No, because what does verifiability mean? If you can't read English then there is little point for you to verify the accuracy of wiki articles, although you could theoretically do it by comparing the characters from a cited fact of the wiki article with those of the source, provided you can find it verbatim in the source.
Suppose you can read English, but you can't understand any maths, not even simple formula. Then you may have difficulties verifying the correctness of a single mathamatical formula in an article, e.g. Planck's formula for blackbody radiation. You could again compare the characters in the formula with those in a textbook but that won't work if, e.g., the variables in the formulae are rearranged relative to each other. But if you can't read formula's then why would you want to verify that it is correct?
Finally suppose that you can understand maths, read formula's but you don't have the patience to go through all the derivations in the article. Or perhaps you lack some simple mathematical knowledge to fully understand the derivation, analogous to someone who can read English, but at a very low level causing him/her to not properly understand some wiki articles. Then you may not be able to verify that the derivation is correct. But again, there would be little point or importance for you to verify it, if you can't understand it.
So, it is really a matter of being able to understand the relevant fact. In this case the fact is the derivation itself. Facts must be verifiable for those who can understand the stated fact. Count Iblis 01:37, 4 October 2006 (UTC)
You said, "To verify that the derivation is correct, you'll have to read the textbook on the subject and then read the wiki article. You can't verify it by not understanding the derivation", but which one? which page? You didn't point it in the corresponding section.
It's funny that you correlated the ability of reading math equation and English. It's irrelevant. This is English wikipedia that assume every readers understand English. — Indon (reply) — 01:59, 4 October 2006 (UTC)
Are you saying we should start a math Wikipedia and leave all the illiterates behind? --ScienceApologist 18:21, 4 October 2006 (UTC)
I'll give the ref. tomorrow, probably the book Fundamentals of Statistical and Thermal Physics byF. Reif. I'm of the opinion that facts must be presented in as simple form as possible, so that as many people as possible can understand it. But then to say that just because some people don't understand it, we must therefore not incude that fact in wikipedia, is wrong. Count Iblis 02:23, 4 October 2006 (UTC)
Good, but don't give it here. Put there in the article, as the reader who asked it is not going to come here. Now, isn't that easy to just put an inline citation? ;-) — Indon (reply) — 02:32, 4 October 2006 (UTC)
That's not easy at all. The citation is unhelpful and doesn't directly work for the derivation presented. It's putting a bandaid on a sore throat. --ScienceApologist 18:21, 4 October 2006 (UTC)

A reader must be able to find a published source readily

I changed

Inline citations are not required for every statement in Wikipedia, provided the statements are corroborated by the references cited in the article.

to

Inline citations are not required for each statement in Wikipedia, as long as the article provides some means by which a reader can readily find a published source corroborating that statement.

The point here is that inline citations are not sacred, and inline citations for every statement are not necessarily required, but the spirit of verifiability must be met.

My goal was to find language that blunts the attack of "{{fact}} nazis" who seek to make bad-faith abuse of policy in order to gain tactical advantage in a POV-war by tying the opposition up in busy-work—which seems to be one of the concerns here—but does not compromise WP:V.

At one extreme, sure, if it is patently obvious to the average reader that a statement like

[Beethoven] is widely regarded as one of history's greatest composers, and was the predominant figure in the transitional period between the Classical and Romantic eras in Western classical music

must be "common knowledge" in the music field, and if the article cites some general references (which it does), and it's obvious which references would be expected to confirm the statement ("Davies, Peter. The Character of a Genius: Beethoven in Perspective?"), there's no haste in providing an inline citation.

At the other extreme, if I'm faced with a statement that is not obvious, and there are a dozen general references cited at the end of the article, and it's not obvious which one of them would corroborate the statement, and the references are not online links and several of them are book that would not be expected to be in a medium-sized public library... well, it's not fair to expect me to obtain a dozen books and comb though each of them. What if I can only put my hands on ten of them? How do I know whether the corroboration is in the other two?

Personally, I'm not totally convinced that there's any good way to meet WP:V with present apparatus other than inline citations, but I'm willing to settle for stating the goal to be achieved rather than dictating a specific means for achieving it.

The reader has to be able to find a published source fairly easily. Dpbsmith (talk) 22:56, 3 October 2006 (UTC)

I fully support Dpbsmith's changes. dryguy 22:59, 3 October 2006 (UTC)
Meh, I suppose that will have to do; we're likely to sit here reverting for a long time, otherwise. Kirill Lokshin 23:01, 3 October 2006 (UTC)
Nope, it won't. I wholeheartedly disagree with the policy of steamrollering changes without consensus, just by implying the threat of another lenghty edit war or dispute. There is currently no consensus concerning any changes to the guideline, which means it should not be changed in any way. I would like to ask all parties involved to behave in a mature way. Bravada, talk - 23:04, 3 October 2006 (UTC)
Mmm, do you have any problems with the actual text that was added, or just procedural concerns? Kirill Lokshin 23:06, 3 October 2006 (UTC)
(reply to Kirill) Yes I do - it is absolutely redundant and litters the guideline to make somebody feel better. See also Gerry Ashton's comment far above. Bravada, talk - 23:13, 3 October 2006 (UTC)
As others have pointed out, not one person on the page has said inline citations are required on every statement. Until someone says so, we have consensus that they are not required. dryguy 23:07, 3 October 2006 (UTC)
(edit conflict) This is a false argument. The question in the poll was "Every fact in every article should have its own inline citation with a specific page number where the fact can be found" which no one really supports, but you can't reverse that and say ergo "Inline citations are not required for every statement in Wikipedia" because sometimes, in some cases, they will need to be. That's the difference.--plange 23:14, 3 October 2006 (UTC)
Yep, that's why I didn't respond to the poll, but I opposed the way it was written. Sandy 23:15, 3 October 2006 (UTC)
The most recent change to WP:CITE does not mention page numbers. dryguy 23:18, 3 October 2006 (UTC)
The two sentences are opposites of each other. Either every fact need an inline citation, or not every fact needs an inline citation. Could you explain your opinion that is is possible to simultaneously disagree with both? CMummert 23:22, 3 October 2006 (UTC)
One can, of course, also hold the view that this guideline should make no statement regarding whether or not an inline citation is required for every fact. Kirill Lokshin 23:24, 3 October 2006 (UTC)
Yes, but that is exactly the point that is at issue in these discussions. CMummert 23:26, 3 October 2006 (UTC)
The problem is that when we aren't explicit, we get some editors who think that CITE supports them in their declaration that particular articles are unverified simply on the basis that they don't have inline references. --ScienceApologist 23:29, 3 October 2006 (UTC)
(reply to Kirill) Egg, zactly. We get to use our brains, instead :-) Sandy 23:26, 3 October 2006 (UTC)
A very convenient way to disagree with anything while not having to provide a coherent argument as to why. dryguy 23:29, 3 October 2006 (UTC)
We do not have consensus; we have gone in circles and are back to the same point we were 24 hours ago. I have a problem with the current vagueness. " Inline citations are not required for each statement in Wikipedia, as long as the article provides some means by which a reader can readily find a published source corroborating that statement." So, these editors who don't want to cite can say, go read this 400-page textbook. We need exact page numbers (or ranges of page numbers). And once you've given page numbers, there is no reason just to not give the darn cite already. And, I have a problem with any wording that refers to "each statement", as it ignores the fact that sometimes entire paragraphs can be cited from one source, etc. Sandy 23:11, 3 October 2006 (UTC)
No one has contradicted the statement that not every sentence needs an in-line reference. That's pretty much a consensus that this statement is correct as far as I can tell. --ScienceApologist 23:18, 3 October 2006 (UTC)
Faulty logic, as explained above, and the sentence inserted says more than that. Sandy 23:21, 3 October 2006 (UTC)
It would be easier to focus on the semantics to everyone's satisfaction if you would just tell us your position. It is like you are constantly morphing your opinion. Do you think inline citations should be required on every statement or not? There is no logical flaw in ScienceApologists point. You either support this very specific requirement, or you don't. dryguy 23:25, 3 October 2006 (UTC)
My position hasn't changed, and is that the current wording is/was fine, no changes are needed, we shouldn't be making specific statements at all, and that no one here has yet proposed any new wording worthy of inclusion on WP:CITE that won't ultimately be a detriment to the project, allowing for speculation, ongoing edit wars over whether to cite, and continuing problems at GA (noting of course, that you won't have this problem at FAC - uncited articles won't pass there.) I supported an earlier summary by Kirill, which basically said the same thing the current policies say, hence no need for change. If you can construct new wording that leaves no wiggle room, I'll endorse. Sandy 23:30, 3 October 2006 (UTC)
You still haven't answered the question as to whether or not every sentence needs an in-line reference. Please take us seriously. --ScienceApologist 23:32, 3 October 2006 (UTC)
Perhaps you haven't kept up with the entire page: I and others have answered it several times. There may be an article which requires an inline site for every sentence, it is common for medical articles to have multiple cites per sentence, and there is no need to place any explicit statements about numbers needed in our guidelines, which are working just fine for most of Wikipedia. Sandy 23:48, 3 October 2006 (UTC)
I haven't seen a version that says "no article may include inline cites on every statement." I have seen plenty articles with multiple, redundant cites on every sentence that serve no useful purpose. I can only guess they were added based on a misguided interpretation of WP:V and WP:CITE. The clarification will help avoid such missteps in the future, thus improving overall quality at Wikipedia. dryguy 00:02, 4 October 2006 (UTC)
Wow, a lot of words but still no answer to the basic question. --ScienceApologist 23:57, 3 October 2006 (UTC)
  • Wikipedia:No original research - "Wikipedia is not the place for original research. Citing sources and avoiding original research are inextricably linked: the only way to demonstrate that you are not doing original research is to cite reliable sources which provide information that is directly related to the topic of the article and to adhere to what those sources say". This is official policy, and can't be amended by so-called consensus (which hasn't been met. LuciferMorgan 23:36, 3 October 2006 (UTC)
Well, this is certainly a non-sequitor. There is a difference between original research and original phrasing. Original phrasing in science and mathematics is fundamentaly equivalent statements that require contextual manipulation of verifiable references. Ironically, it was in order to combat pseudophysics that original research was developed in the first place, so it doesn't make much sense to attack editors who are reasonably dealing with how to indicate sources in science and math articles. --ScienceApologist 23:45, 3 October 2006 (UTC)
That is WP:OR. This is WP:CITE, which says how sources should be cited to comply with WP:V and WP:OR. If consensus is reached that inline citation is not necessary to quality as what [[WP:V] and WP:OR call citation then WP:CITE should be edited to reflect that consensus. CMummert 23:40, 3 October 2006 (UTC)

It seems to me we are back at the same straw poll I suggested yesterday. No editor has explicitly said that every fact needs an inline citation. Is there any editor who advocates that position? If nobody does, then it is reasonable to think the consensus favors the opposite opinion, that not every fact requires an inline citation. In that case, it would be reasonable to edit WP:CITE to say, simply, Not every fact requires inline citation. This doesn't mean no facts require inline citation, only that some do not. CMummert 23:40, 3 October 2006 (UTC)

How about articulating the fact that there's a range of opinion?

My personal opinion, which I do not feel was adequately captured in the poll or in CMummert's which-side-are-you-on-boys challenges, is that

  • Inline citations are always desirable, even for "the obvious," and are not that hard to supply.
  • Inline citations are highly desirable for "the obvious" when the obvious involves historical or political overtones. Or, generally, when "the obvious" is a belief that's so widely held that it is widely presented as a standard narrative, but where there is a substantial body of dissenting or unorthodox opinion (even nutcase or crackpot opinion).
  • Inline citations should be supplied whenever a good-faith request is made for them. Note that a "good-faith request" doesn't mean "I think this is wrong," it means "I'd like to see a citation for this."
  • Inline citations for "the obvious" are not needed until someone makes a specific request that a specific item be cited.

I do not like the idea of summarizing my opinion in the words "Not every fact requires inline citation." I really want to see some modifying language. I would be angry if someone removed good-faith "citation needed" tags with an edit comment like "Not required, see WP:CITE". I would be very angry if someone removed an actual inline citation, referencing this policy.

Perhaps the language should capture the fact that there's a range of opinion.

Not every fact requires inline citation. There is no bright-line rule for when facts do require inline citation. The spirit of verifiability must be met.

Dpbsmith (talk) 12:48, 4 October 2006 (UTC)

Modified Strawpoll

Two parts since this seems to the two separate circles we are going around in. And yes, voting is evil. :p Agne 23:43, 3 October 2006 (UTC)

This poll is quite odd. The second two parts of (2) mean the same thing - that not every fact needs a citation. And it is not coherent to say that although a fact is represented by consensus, that fact shouldn't be added to the policy. CMummert 23:45, 3 October 2006 (UTC)

Honestly, this is more about bringing the discussion to a conclusion so that we can all go back to working on the encyclopedia. These are the two points that keep being brought up "Does it need it change?" and "Well what about cites for every line?". The problem with your poll was that it combined the two assumption together. Agne 23:48, 3 October 2006 (UTC)

I think the argument that thepolicy doesn't represent consensus but doesn't need changed is quite weak. So if consensus says not every article needs a cite, it should go into the article. Arguing to keep the current article against consensus is not exactly good faith. But if the article does agree with the consensus, then clarifying this can't make the article worse. CMummert 23:59, 3 October 2006 (UTC)

Shall current guidelines be changed?

1.) Does the current guideline need a change?

  • Yes-complete reworking per Gerry Ashton's suggestion
  1. ScienceApologist 23:48, 3 October 2006 (UTC)
  2. Bravada, talk - 23:51, 3 October 2006 (UTC) - my heart is here, but I would wait with it for a few months until emotions cool off.
  • Yes-we need to clarify what to cite and what not to cite
  1. ScienceApologist 23:48, 3 October 2006 (UTC)
  2. --agr 03:08, 4 October 2006 (UTC) We need to think through what is best for our readers and editors, recognizing that the right answer may vary by discipline.
  • No-keep the status quo
  1. Bravada, talk - 23:51, 3 October 2006 (UTC) - for the time being
  2. SteveMcCluskey 00:46, 4 October 2006 (UTC) - It's basically good and there's too much danger of throwing out the baby with the bathwater.
  3. Sandy 23:54, 3 October 2006 (UTC)
  • No-I'm open to change but not with what has been be suggested so far


"every sentence needs an in-line reference"

2.) Does it?

  • Yes
  • Maybe. For some articles, every line may need to be cited. Sandy 23:59, 3 October 2006 (UTC)
    The question isn't asking about specific articles, it is asking if every sentence in Wikipedia needs an in-line reference. Stop equivocating and join the discussion. --ScienceApologist 00:01, 4 October 2006 (UTC)


  • No
  1. ScienceApologist 23:48, 3 October 2006 (UTC)
  2. Bravada, talk - 23:51, 3 October 2006 (UTC) - that's obvious, but this straw poll might lead to another changes like the ones we've seen, so I believe it would be better to close it ASAP. - sorry, wasn't thinking much. Bravada, talk - 00:05, 4 October 2006 (UTC)
Considering the small sampling here (compared to the whole project) an item would need near unanimous support to be even close to considered "consensus". It wouldn't be wise for anyone to change the guideline based on this. This is mostly for our benefit of discussion. Agne 23:55, 3 October 2006 (UTC)
That was not the consensus opion on WP:GA when the guidelines were changed there. Notice has been placed on many project message pages about this discussion, so there has been adequate chance for editors to participate. And WP:CITE must be on a lot of watchlists. CMummert 00:07, 4 October 2006 (UTC)
  1. CMummert 00:23, 4 October 2006 (UTC)
  2. dryguy 00:48, 4 October 2006 (UTC)
  3. Vsmith 00:53, 4 October 2006 (UTC)
  4. --agr 03:08, 4 October 2006 (UTC)
  5. No. No encyclopedia I have ever seen does this. Readers can trust the editors, and if they challenge one point then the reader can add a reference. Rjensen 03:14, 4 October 2006 (UTC)


  • Not needed, but encouraged
  1. A citation per sentence is not required in the sense that an article can meet WP:V without it; but editors should not be discouraged from thorough and exhaustive citation, and such citation can be required by optional non-policy processes. Kirill Lokshin 00:57, 4 October 2006 (UTC)
  • Should be decided on an article by article basis
  1. Agne 23:49, 3 October 2006 (UTC) It's all about WP:V. If an article needs a cite on every line in order to be verifable then so be it. Agne 00:00, 4 October 2006 (UTC)
  2. Maybe Depends on the article: it could. Sandy 23:53, 3 October 2006 (UTC)
  3. Bravada, talk - 00:05, 4 October 2006 (UTC) - obviously!
  • No one ever said it should apply to every article
  1. ScienceApologist 23:57, 3 October 2006 (UTC) It's all about people thinking it's all about WP:V when it's really about WP:CITE.
As I pointed out above, this position implies that not every fact needs a citation, if there is room for interpretation on a per-article basis. Thus this position supports the consensus that not every sentence needs a citation. CMummert 23:55, 3 October 2006 (UTC)
As I pointed out above, no it doesn't. Sandy 23:58, 3 October 2006 (UTC)
CMummert is right. If the choice is between every sentence requiring a citation and not, you are saying that every sentence does not require a citation. --ScienceApologist 00:00, 4 October 2006 (UTC)
WP:DONOT (shit, that's not an article?). Do not engage in this strawpoll. This is a game with a calculated intent. Marskell 00:02, 4 October 2006 (UTC)
WP:AGF. We are trying to assess consensus. This poll was placed by Agne, who favors inline citations, and a similar poll was placed by me, who doesn't. SO apparently both sides are interested to know. CMummert 00:05, 4 October 2006 (UTC)
Strawpolls should be an outcome of consensus (i.e., "we're here, what next"), not a means to find consensus. And, they should not be content specific unless you're polling specific content. I've done market research--you don't have clauses like "per Gerry Ashton's suggestion" unless you're polling re Gerry Ashton. I'd suggest: take it easy. Marskell 00:12, 4 October 2006 (UTC)
My "calculated intent" is that reasonable people will see that they can reasonably disagree on matters like this. As I mentioned to Bravada, it certainly wouldn't be wise to guage such a small sampling as consensus in either direction. But the group of 6-9 editors who have been discussing this on and off have pretty much been going in circles. If we can get a "break" in the circle, maybe we can bring this matter to a conclusion. (after edit conflict) LOL, I've done market research too and no this is not a market research poll. The "per Gerry" was simply because his comment best summed up that sentiment. Agne 00:15, 4 October 2006 (UTC)
Hm, if only in a cursory way, if you've done market research (i.e., designed opinion surveys) you should be able to design a question without a nod towards one option. In any case, this is a "we're tired--let's poll!" poll. The only consensus you'll be verifying (on one of the most important issues facing Wiki) is that between ten editors between Sept 20-something and early Oct 2006. This may need a haul into userspace and a poke at people ("What do you think?") before using this talk as some definitive measure. Marskell 00:27, 4 October 2006 (UTC)
I see your point. As I said before, I don't think the result of this poll should be used to the change the guideline at all. But you are quite right that an invitation to get more people involved is more ideal. The views of 10 editors should never decide anything as major as this. Agne 00:58, 4 October 2006 (UTC)
These polls seem to agree with WP:DDV#Straw polls. There is a real question of whether Not every fact needs a citation is consensus. I didn't write this poll, and I pointed out at the top of it that the answer choices overlap each other. (I doubt the first question has any weight, since it's hard to argue that policy should not reflect consensus.) CMummert 00:22, 4 October 2006 (UTC)
I only drop into this discussion on rare occasions, and I find this straw poll to be once again, posing a straw man. No one says every sentence should have a citation; but a no vore will lend support to those who oppose the idea that every fact should have a proper citation. I don't plan to dignify it by a reply. SteveMcCluskey 00:55, 4 October 2006 (UTC)
Nobody here has said that articles do not need adequate citations or that WP:V should be changed. The question is the role of inline citation. If you believe that every fact needs an inline citation, you should vote yes above. So far, not one person (including you) has made that claim. At best, some people refuse to make a decision one way or another, which is hardly a strong argument. So the consensus seems to be leaning towards Not every fact needs an inline citation, the literal opposite of Every fact needs an inline citation. CMummert 01:04, 4 October 2006 (UTC)

Question the usefulness of this straw poll

With all due respect to those who constructed this, I am removing myself from this poll, which has now become confusing and poorly-constructed, noting that my position remains simple: WP:V is policy; WP:CITE does not need to change, and we don't change polices and guidelines by straw polls involving an extreme minority who happen to be following a page at any given time. I reiterate my position: the policy/guideline is working fine for most of Wikipedia, a minority is opposed because of an issue that started and should remain at GA, and sources, per policy, are required throughout Wikipedia, independently of GA, for any statement another editor requests, whether that means every line or not. No change to any wording on WP:CITE is needed or justified. Those who don't like GA's tighter requirements should opt out, knowing they won't pass FA either. Sandy 00:29, 4 October 2006 (UTC)

I agree 100% Bravada, talk - 00:31, 4 October 2006 (UTC)
The claim has been made several times that articles without inline cites will not pass FA review. But the featured article earlier this week, Detroit is clearly lacking inline citations for many facts (in the intro paragraphs and lower down). I would wager that if I survey the featured articles, few of them have an inline citation for every fact. CMummert 00:33, 4 October 2006 (UTC)
As I stated above, this issue also arose with Marcel Lefebvre at WP:V when those editors brought it there. This is not isolated to the GA dispute. And seeing as not one person has yet stated that every statement should be inline cited, it is definitely not an "extreme minority"; quite the reverse. dryguy 00:40, 4 October 2006 (UTC)

Different standards for different disciplines?

I thought I'd point this discussion in a different direction. The discussion is caught between a one size fits all standard and an idea that every article sets its own standards. It's not individual articles, but particular disciplines that set their standards.

I notice that much of the active discussion here is by a mathematician User:CMummert. I think part of the problem arises from the different expectations in different disciplines. In history, which is an evidence-based discipline, we're used to citing a source for every fact, although we don't provide citations for inferences from those facts.

Mathematics, on the other hand, involves reasoning from generally accepted axioms. There's very little evidence that needs to be cited.

Most other academic disciplines fall somewhere in between on this spectrum.

In current events / politics/ biography there is much the same kind of need for citations that we find in history.

In the realm of pop culture, there are all sorts of factoids which are floating around -- but I don't pay enough attention to music, video games, etc., to know which of them are common knowledge. I don't know if there are enough points there which are likely to be disputed. Perhaps someone who edits there could provide some insights. --SteveMcCluskey 01:16, 4 October 2006 (UTC)

Your post is very reasonable, and the idea that different standards should apply to different disciplines is quite reasonable, but all efforts to reflect this in WP:CITE seem unlikely to succeed, since it is also reasonable for wikipedia to hold everyone to the same standard. CMummert 01:21, 4 October 2006 (UTC)
This is very unreasonable, actually, as it fails to acknowledge that we are editing an encyclopedia here, and not a collection of scientific papers from various disciplines. The rules of writing an encyclopedic article on a topic from maths are, and should be, the same as for an article from any other field of knowledge. Information science, which encompasses encyclopedias, is actually a branch of its own, and we are all practicing it working on articles here.
Where I'm getting at is that we cannot expect articles from different disciplines to be discipline-specific in their treatment of encyclopedic rules etc. For example, in political science, and also to some extent in history, adopting a POV is common practice, and writing flowery essays is probably at least as popular as dry "encyclopedic" papers. That said, making an exception would seem quite obviously improper to, I hope, most people here. I believe we should make no exceptions for no field of knowledge or branch of science.
On the other hand, I still believe that common sense is the rule number one we should apply here (by which I mean all the time we edit WP) and expecting a section of an article describing a mathematical derivation to be rife with citations is not quite reasonable. I trust the community overall to have enough common sense to see that, and if some user would still be of a different view, one can still ask the broader community for their opinion. In most cases, editors posing extreme or unreasonable demands are convinced to drop them facing community opposition.
Finally, while writing a paper in mathematics, physics or perhaps also some other branches of science does not require using inline citations, or references at all for some parts, this is still an encyclopedia, and we can reasonably expect all articles to hold to good referencing standards. To clarify what I mean, I believe an article explaining e.g. Planck's law can for the most part be referenced to Planck's original paper, and one citation per paragraph or even per section would suffice if the paper actually contains the knowledge or information given in the article in any form. Another clarification - if a mathematical derivation or another set of formulas is given in a different form in a source than it would be reasonable to present in an article, I see no problem with referencing to the source and using laws of mathematics to convert the formulas (I apologize for my nonexistent maths vocabulary in English).
The bottom bottom line - no need to make exceptions for any field, any article can be well-referenced under the current guideline, and actually WIAGA too. Bravada, talk - 02:21, 4 October 2006 (UTC)

Referencing Planck's original paper? Are you serious? What people who don't study math and science don't realize is that while in name science follows the great man theory of history, in practice, science is always up for modification and proceeds on the shoulders of giants. Planck's initial paper on the subject was incomplete, had misconceptions and errors that were subsequently adjusted, and is not by any estimation the most straightforward introductory scientific derivation, nor is it the one that belongs in a summary-style encyclopedia. Trying to get science to conform to the unjustified and ignorant prejudices of those who don't study science is exactly the problem with the current monolithic WP:CITE guideline. It lets good faith editors make very bad suggestions based out of ignorance and then, when they are rebuffed, it encourages self-same editors to continue to be tendentious in their clamboring for citations that are not reasonable. --ScienceApologist 18:32, 4 October 2006 (UTC)

I have been trying not to get involved in this discussion, since I long ago concluded that it is going in circles (it seems like some people agree) and that, if you're careful, it is possible to just ignore this debate and write well-referenced articles that make everybody happy. I have written a set of guidelines which I think are applicable to math and physics articles and conform to the spirit and letter of WP:CITE and WP:V. I don't think these articles need to be treated seperately – after all, we're not writing academic papers – but I think a set of guidelines emphasizing the similarities and differences between writing for Wikipedia and writing for a journal can be helpful. They are at Wikipedia:WikiProject_Physics/Citation_guidelines_proposal. Feel free to comment on the talk page. Unless an imbroglio erupts, in the next few days I'll ask members of the two WikiProjects (Physics and Mathematics) if the consensus is that they support the guidelines. –Joke 19:20, 4 October 2006 (UTC)
I just want to say I think those guidelines are pretty good. Anville 20:10, 4 October 2006 (UTC)
I think Joke has a good idea. His tactical idea of bypassing this non-productive discussion is briliant and his specific proposal looks OK too -- but then the last time I studied physics was when I was an undergrad in the early 1960s. Maybe I should start writing a citation guide for History. --SteveMcCluskey 21:00, 4 October 2006 (UTC)
It's a star among all those circles. Thank you for helping to create well-referenced articles that are truly a part of Wikipedia. I hope your proposal goes through. RelHistBuff 08:14, 5 October 2006 (UTC)

Back to basics

The real issue here is not special treatment of one field versus another. This conversation started because of disagreement over when inline citation is preferable over other forms of citation. Because WP:CITE is the guideline on how to cite facts, this is an issue for WP:CITE. Neither WP:V nor WP:OR says how facts need to be cited; the only say that citations are necessary and point to WP:CITE. Thus this is the correct place to discuss the issue.

Since the overall issue is when inline citations are required, a basic question is whether Every fact needs an inline citation. is consensus or not (note the word fact not sentence). If that bolded statement is consensus, then that would settle this discussion. This is not a question of whether a reader can ask for a citation, which the reader can certainly do, in which case a citation must be provided. It is a question of whether that citation must be given in inline form.

Here are three options:

  1. Consensus favors Every fact needs an inline citation.
  2. Consensus favors the literal opposite, Not every fact needs an inline citation. This means that there are some facts that do not need inline citation; it does not mean that no fact at all needs inline citation.
  3. There is not consensus on this issue.

In order for (1) or (3) to hold, there would have to be advocates for (1). Thus if nobody advocates (1) then (2) is, by default, the consensus view. It isn't possible to directly advocate (3); that option would hold only if there were some people advocating (1) and some people advocating (2).

In a long discussion above, no editor has explicitly advocated (1). In two straw polls, no editor spoke in favor of alternatives similar to (1). Some editors refuse to take a firm stand, but I think a maybe response must mean that not every fact needs an inline citation. I would like to settle the issue of which of (1), (2), or (3) hold. This is a black-and-white question of fact, and we should be able to decide which of the three options is true. CMummert 12:30, 4 October 2006 (UTC)

This conversation started because of disagreement over when inline citation is preferable over other forms of citation. My recollection is that it started because some math/physics editors were upset when some articles were routinely notified that GA requirements on citations would be tightening up. Sandy 14:04, 4 October 2006 (UTC)
I don't think it's a black-and-white question of fact. You're logic-chopping. My opinion (as stated above, sorry for the repetition) is:
  • Inline citations are always desirable, even for "the obvious," and are not that hard to supply.
  • Inline citations are highly desirable for "the obvious" when the obvious involves historical or political overtones. Or, generally, when "the obvious" is a belief that's so widely held that it is widely presented as a standard narrative, but where there is a substantial body of dissenting or unorthodox opinion (even nutcase or crackpot opinion).
  • Inline citations should be supplied whenever a good-faith request is made for them. Note that a "good-faith request" doesn't mean "I think this is wrong," it means "I'd like to see a citation for this."
  • Inline citations for "the obvious" are not needed until someone makes a specific request that a specific item be cited.
Logically, this sorts into alternative #1, but to put it there is IMHO a misrepresentation of my opinion. I feel that you are trying to force me to agree to language which technically captures my opinion, but, if placed into the guideline as given would be interpreted by most readers to mean something far from my opinion.
Of the options you present above, the one which most fairly captures the current situation is 3, there is no consensus. Dpbsmith (talk) 12:55, 4 October 2006 (UTC)
I pretty much agree with everything Dpbsmith wrote here. The issue is not so much that "Not every fact needs an inline citation" isn't correct; it is, technically. The problem is that a number of people (myself included) feel that adding such wording to the guideline would be counterproductive, at best, and downright harmful, at worst, because it would be interpreted as justifying a refusal to provide inline citations. Kirill Lokshin 13:02, 4 October 2006 (UTC)
After all the circles, I have to agree with Dpbsmith's words. It's not a black and white issue and option #3 best sums that up. Agne 13:33, 4 October 2006 (UTC)
Agree with Dpbsmith, Kirill, Agne. Sandy 14:04, 4 October 2006 (UTC)
Perhaps you could be clearer about the specific reasons why you feel it is important
  • to include the words "Not every fact needs an inline citation" in the guideline
  • not to include any further explanation or qualification. Dpbsmith (talk) 12:57, 4 October 2006 (UTC)
I would be glad to see further clarification in WP:CITE beyond a single added sentence.
You say you are in position 1 but your fourth bullet puts you in position 2. This is the problem that the editors on my side of the discussion keep running into - we feel that the inline citation proponents won't make any precise statements that can be discussed. I get the impression that the opinion may be Inline citations are required for every fact, but I don't want to admit that. So I am asking you to make a precise statement: do you think Every fact needs an inline citation or not? Everybody agrees that the inline cites are desirable in certain circumstances.
About your third bullet: are you saying that it would be good faith for me to go through a dozen featured articles and put a fact tag beside every sentence? The article Detroit, on the front page earlier this week, has lots of unsourced facts. I would say that a good faith request for a citation includes the immediate plan to pursue the citation to actually verify the fact. I am curious to hear opinions about this issue, because it relates to the fact-bombing of articles such as special relativity.
CMummert 13:33, 4 October 2006 (UTC)
1) It would be good faith for you to go through a dozen featured articles and put a fact tag beside every sentence if you, CMummert, truly felt that for each of those sentences you had a specific reason why you, CMummert, wanted to see a citation for that particular sentence. That's pretty hypothetical, though. In real life if anyone actually did that I would suspect them of WP:POINT. Given your stated views, if you, CMummert were to do so, I'd strongly suspect WP:POINT.
I am aware of WP:POINT, which is why I haven't done it. I do plan to participate in the GA review process to enforce the new standards of citation that, apparently, are required for good articles these days. That is entirely in the spirit of GA review, using the community consensus on what needs to be cited. CMummert 16:02, 4 October 2006 (UTC)
2) "I would say that a good faith request for a citation includes the immediate plan to pursue the citation to actually verify the fact." No, I disagree, if you mean that the person making the request has an obligation to try to fulfill the request his- or herself. This is very much like the "burden of proof" statement in the WP:V. If you mean that when the citation is provided, the person requesting the citation has an obligation to verify the citation, I don't agree with that, either, and I don't even see your point. Unless you mean that courtesy and collegiality means that people shouldn't make work for others if they're not willing to do work themselves. But I don't think that can or should be legislated into a citation guideline.
3) I can't speak to "fact-bombing." I don't think a citation guideline can deal with dishonest behavior, which is what I assume you're referring to. But you think it would be helpful to have something in the guideline to specifically address "fact-bombing," why not put a section... near the end... that defines what you mean by "fact-bombing" and says that the guideline does not condone it? Dpbsmith (talk) 15:35, 4 October 2006 (UTC)
I can speak to the "fact-bombing" that he mentions and it's not a fair characterization as I've pointed out before. Agne was requested by one of that article's editors to do it so that they could see what she thought needed citing. That is not fact-bombing. --plange 17:54, 4 October 2006 (UTC)
You are correct that a request was made to point out what Agne thought needed citation. By fact-bombing, I mean adding cite tags for facts that you are not actually interested in seeing the sources for. As I said earlier, "I would say that a good faith request for a citation includes the immediate plan to pursue the citation to actually verify the fact." This common sense guideline would ensure that things that actually need citation get them as soon as someone asks,while avoiding situations like the one that developed at special relativity. Added as a clarification: In the context of FA and GA review, this would mean that a reviewer has to take responsibility to get their hands dirty, understand the material, look at references, etc., rather than just scanning over the references and declaring them suitable based on their outward appearance. CMummert 18:09, 4 October 2006 (UTC)
Here is an alternative set of options, which puts citation expectations in a continuum rather than as an either/or dichotomy, and seems to this editor to offer a more fruitful model to arrive at consensus.
  1. Consensus favors Every fact needs an inline citation.
  2. Consensus favors Most facts need an inline citation.
  3. Consensus favors Only a few (exceptionally controversial?) facts need an inline citation.
  4. Consensus favors Avoid inline citations.
  5. There is not consensus on this issue.
FWIW, I favor item #2. --SteveMcCluskey 13:29, 4 October 2006 (UTC)
6. Consensus favors Key points and controversial statements need inline citations. Repeated inline citation of previously cited sources is not warranted for subsequent trivial, uncontroversial, uncontested statements that come from the same source. dryguy 14:05, 4 October 2006 (UTC)
Or consensus favours common sense and we don't need a list. This is not "a black-and-white question of fact". Common sense: all quotes must have a cite; all negative criticism on a BLP must have a cite; all statistics must have a cite; extraordinary claims require more than one cite. Common sense: a paragraph with three or four facts can take one cite covering all of them (if none is extraordinary); a general level statement does not require a cite if a more specific example has one nearby in the article; paragraph topic sentences generally do not require cites if cites are to follow. And so forth. A blanket "every," "most," "few," "none" is unneeded and will only invite more wiki-lawyering. Marskell 14:37, 4 October 2006 (UTC)
Bravo! Amen, brother! What he said! Hear, hear! etc. Dpbsmith (talk) 15:38, 4 October 2006 (UTC)
I, surprisingly, agree with Marskell here - common sense is needed here much more than any of those definitive statements with "every" or "none" and a myriad exceptions. What Marskell said is all true, I would also like to add that actually all facts in an article can be cited without it being infested with blue superscript digits, by using citations for entire paragraphs. An example I can think of is an FA I worked on, the Talbot Tagora - AFAIK, it is 100% "inline cited" and somehow readable IMHO. Bravada, talk - 15:17, 4 October 2006 (UTC)
Re "surprisingly"—have we disagreed so often? ;) Marskell 15:25, 4 October 2006 (UTC)
This has been suggested before, and some editors responded that this was not suficient, that every fact must have a specific page-numbered inline cite, etc. So it seems to me that this is another example of talking in generalities so that no commitment to actual principles needs to be given. I agree that excessive legality is not needed, but some specific guidance needs to be. Do other editors agree with Marskell and Bravada that in many cases it is acceptable to have facts without a specifically page numbered citation? CMummert 15:28, 4 October 2006 (UTC)
I agree with everything Marskell and Bravada said, and nowhere do I see that either of them said anything about "in many cases it is acceptable to have facts without a specifically page numbered citation". Perhaps I missed a post. Sandy 15:46, 4 October 2006 (UTC)
I agree with Dpbsmith, Kirill, Agne, Marskell, Bravada, and Sandy. I made it to US time for the conversation for once. But I have to go now :( Hopefully by tomorrow this will all be settled. I'm not betting on it though. ;) RelHistBuff 15:56, 4 October 2006 (UTC)
When I read "Common sense: a paragraph with three or four facts can take one cite covering all of them (if none is extraordinary)" it makes me think that a footnote saying something like (see Soare [1987] chapter 3) would be enough to cite the whole paragraph. Would that be enough for you, or are page references requried for each fact in the paragraph? CMummert 15:53, 4 October 2006 (UTC)
As Marskell and Bravada have both stated (and I agree), the problem is with this attempt at making "every", "none", "most", "few", "always", etctera statements in a guideline. If Chapter 3 is 100 pages, no it wouldn't likely be enough. Marskell's and Bravada's point, and the guideline, is entirely clear to me: the reader has to be able to easily verify the information. Sandy 15:59, 4 October 2006 (UTC)
We seem to disagree about the meaning of Marskell's sentence "Common sense: a paragraph with three or four facts can take one cite covering all of them (if none is extraordinary)". I am looking for a general standard that an author can use to add citations while writing so that the final article satisfies WP:CITE. I think that there is informal consensus about what that standard is, but nobody yet has found a way to express it. The standard will probably not be black and white, but it needs to be less vulnerable to interpretation than Marskell's proposal above. CMummert 17:28, 4 October 2006 (UTC)
All this would just amount to having this same 300 kilobyte discussion at the talk page of any article where this issue comes up. There is consensus that tagging every statement is not a requirement and often not a good idea. The purpose of WP:CITE is to record the consensus to avoid future pointless discussions of this nature. Yet there is this objection to putting the consensus on WP:CITE for fear it will be misused. WP:CITE and WP:V are already misused and misinterpreted. We have the opportunity now to clarify WP:CITE for the issue at hand. The other issues mentioned above haven't been discussed for 300 kilobytes because they aren't currently in contention. Let's please address the one narrowly defined issue at hand and come up with a wording that will satisfy people's concerns about misuse. dryguy 16:37, 4 October 2006 (UTC)
Anything challenged or likely to be challenged needs an inline citation, as do quotes. SlimVirgin (talk) 07:56, 5 October 2006 (UTC)
Currently there is no such requirement for inline citations in any Wikipedia policy. dryguy 12:08, 5 October 2006 (UTC)
To CM: I have a feeling that anything said to you is going to be disagreed with. My last post wasn't a proposal for wording but an appeal to use common sense and an attempt to show that wiki-lawyering always invites caveats to the Nth degree. There is no wording change that is suddenly going to encompass every potentiality. But, as near as I can tell, there isn't actually consensus that a problem exists with the wording as it stands. You just want more rules.
Re the "three or four facts, one cite", consider: "OJ was arrested,[1] tried,[1] and found not guilty."[1] This is over-cited. Again, perfectly commonsensical. We don't need a sub-sub-point somewhere here to point this out to editors. Marskell 08:59, 5 October 2006 (UTC)
Of course no wording change will encompass every eventuality, but the article review process is legalistic by its very nature. You might pause to consider the reason that more specific guidelines were requested in the first place.
Unfortunately, I think the most likely outcome of this discussion is that a more specific set of citation guidelines will be developed for physics and mathematics (and maybe other projects will join in); see here, and these will become a set of "best practices" that we will make our best articles follow.
It would be a shame if the GA project fades into irrelevance in the sciences, which is a second possible outcome of the lack of consensus on the appropriate level of citation in articles. My goal in coming here was to prevent this, but I feel that there is little willingness to openly discuss the matter here. Some editors seem to not read posts carefully, to ignore direct questions that are posed to them, and to make little effort at compromise. CMummert 12:20, 5 October 2006 (UTC)
To reiterate again, in reference to your statement "You might pause to consider the reason that more specific guidelines were requested in the first place" is that if someone (a GA reviewer) went overboard and requested a cite for something like fish live in water (I'm purposely using this example, BTW, since it's one people have accused us citaters as wanting, taking things, as usual, to the extreme), and that it was failed because of that, you have WP:GA/R to go to to make your case where I have full confidence commonsense would prevail. I also doubt anyone of us would ask for a cite for "fish live in the water, that Helium is a gas, or that Carbon has atomic mass 12". It sounds like you guys are wanting to have a specific formula in hand so that your articles will be GA assessment proof before hand. Also, keep in mind that if you don't participate in GA, that's your own choice obviously, but it would be a shame for us, and also for you as I'd like to think we'd help get your article a step closer to FA. --plange 15:47, 5 October 2006 (UTC)
To reiterate again, again, the GA issue is not the only reason for this discussion. See for example Marcel Lefebvre, where some editors are insisting on inline cites for basically every trivial detail. Not to mention the practice of {{fact}} bombing that was raised in previous weeks, and the 20-30 example articles that were presented at that time. dryguy 16:29, 5 October 2006 (UTC)
(Moved up a few levels). Plange said "It sounds like you guys are wanting to have a specific formula in hand so that your articles will be GA assessment proof before hand." Yes, that is more or less what I want, as an author: a more or less specific standard of what I should aim for. Of course it can't be black and white, but some general guidance beyond "inline cites are required" would be extremely helpful. This is especially important because there is essentially no way to deny a cite request once one is made, even if it is made by a reviewer who has no desire to verify the information. Thus there needs to be some policy to encourage reviewers to request cites only for facts where cites are particularly important. CMummert 17:24, 5 October 2006 (UTC)
Actually, you cannot deny a good faith request for a cite. And the onus is not on the requestor to hunt up that cite, but on the editor that wants to keep it in there. --plange 19:51, 5 October 2006 (UTC)
That's why it's especially important... CMummert 21:46, 5 October 2006 (UTC)

More examples

"To reiterate again, again, the GA issue is not the only reason for this discussion. dryguy"

No, it's not a problem unique to GA: here are some older FAs which need a review with an eye towards thoroughness of inline citations. Perhaps the none to few provided are adequate, perhaps not, but they might be reviewed before they come before FAR. In particular, given that some of these articles have a list of References that is completely out of control, I'm particularly curious about how a reader is to find a citation for any particular fact among dozens to scores of references and external links:

  1. ATLAS experiment
  2. Caesar cipher
  3. Carl Friedrich Gauss
  4. Galileo Galilei (an *absurd* number of external links and references, but no inlines
  5. Game theory (references out of control, four inline cites)
  6. Infinite monkey theorem
  7. John Dee
  8. Margin of error
  9. Monty Hall problem
  10. Prisoner's dilemma
  11. Quantum computer (further reading out of control, but two inline cites)
  12. Regular polytope
    Just noticed that regular polytope does have some cites that could be converted, but unfortunately has also been overrun with external jumps. Sandy 21:50, 5 October 2006 (UTC)
  13. Robert Oppenheimer
  14. Speed of light

Sandy 16:50, 5 October 2006 (UTC)

What about Detroit? I have raised this example several times, with no response. The article clearly does not have inline citations for every fact. Does it also need to be put up for FA review? CMummert 17:24, 5 October 2006 (UTC)
Well, there are many older FAs that aren't adequately cited, and even some newer FAs that squeak through because of less than thorough review. I personally wouldn't nominate an article for review without checking with the original author to see if s/he will cite the article: I have no reason to believe the author of Detroit is resisting efforts to upgrade citation requirements on Wiki. Further, the policy on FAR is to not nominate articles that recently appeared on the main page, as often articles improve as a result of appearing on the main page. Sandy 17:28, 5 October 2006 (UTC)
So would you say that the particular article Detroit is adequately cited and sourced, or that it isn't? I have no idea when it went to FA, but I can look at the article and assess the level of citation. I didn't pick this article through some devious method; it was on the front page and so it was an easy non-technical FA to find. CMummert 17:35, 5 October 2006 (UTC)
Well, look at the talk page. It went to FA last spring. I just looked at it, and it's not nearly as bad as you're making it out to be. I added four cite tags, from a very cursory look starting at the bottom of the article, but I also found that it's more a matter of some of the cite tags aren't well placed. Some of the citations are given *before* the text they reference. Talk is cheap: I suggest you spend some time in there, examine the sources, add cite tags you feel it needs, and then see if the author adds them. S/he is still active. Doing that kind of work might contribute to some editors here being less critical of Agne. Sandy 17:52, 5 October 2006 (UTC)
But, the question you overloooked is just how we are supposed to verify facts on some of those math/physics articles that have dozens to scores of references provided. Are all of those really the most top-notch references, or have the lists just become dumping grounds? We've been told several times, just check the References, but these reference lists have apparently turned into hardprint versions of external link farms; another reason that specific inline cites are preferred. Sandy 17:32, 5 October 2006 (UTC)

The article on the ATLAS experiment only needs to cite the first ref. from the reference list in the text (perhaps just once or twice). All the other facts are referenced via internal wikilinks. Count Iblis 21:10, 5 October 2006 (UTC)

Wikilinks don't count as citations because Wikipedia is not a reliable source. --Gerry Ashton 21:26, 5 October 2006 (UTC)
True, but if a wikilink links to an article with many citations, it's like having however many citations are in the linked article attached to the sentence. --ScienceApologist
I don't believe the references found on wikilinks are considered references for the linking article, except summary type articles which have links to the main article for each heading. One reason for this is you can't put in a footnote that refers to a source in a linked article. --Gerry Ashton 19:53, 6 October 2006 (UTC)
Footones are not the requirement. Verifiability is the requirement. dryguy 20:02, 6 October 2006 (UTC)

My preference for citation

I just posted this on IRC, and I thought it'd be good to get on-wiki somewhere, so, here it is. I'm not making it a proposal or anything, just noting it down in case anyone else is interested.

The {{cite X}} templates will be very nice to have when we have some scripts to generate interesting reports from them(like, what are all the citations of this book in the 'pedia, what books by this author are cited, what is the avarage year of the books cited in this category, etc.); IMO, cite tags should go in a section called ==Cited works== (or ==References== if you like), while <ref> tags create footnotes, and should contain short (shorter than a full cite template) mentions of specific pages or subtle points in the text. They should show up in a section titled ==Notes== (or maybe ==References== if you like). That's my prefered way to do it... JesseW, the juggling janitor 06:56, 5 October 2006 (UTC)

I don't use the cite templates because they return sloppy results: I type out the info myself in the correct format between the ref tags. Sandy 07:34, 5 October 2006 (UTC)
Me too.
Juggling Janitor, I've come to the same conclusion. I think "Notes and references" would be a good heading for the footnoty ones and "Sources" for the second one. But sometimes a note has to combine a footnote with the cite. And I think the first footnote to a book should contain the full book reference, even if that is repeated again in the reference section. This can create oddities, though, in a small article, where you could have the footnote and full reference to a book, an abbreviated reference or two to the same book, and the same book reference in the reference section all in the space of a few inches. In such cases I think having two types of cite ref is excessive, so some combined system should be used. As soon as we can hide things, this will all seem less of a problem, and we may have five-hundred-yard footnotes, if we like. It will be like the red button on Sky: you don't have to inspect the hairs on Wayne Rooney's kneecaps if you don't want to. qp10qp 14:21, 5 October 2006 (UTC)

Cite templates are generally a useless imposition on editors; in my opinion they should be avoided like the plague. Jayjg (talk) 17:08, 5 October 2006 (UTC)

Could you and Sandy expand on this a bit? Are they an imposition on editors once added (taking up too much space in the edit window) or just an imposition on the editor who has to fill them out? Many footnotes are of the form: <ref>http://example.org/some_online_journal.html</ref>, using the cite templates seems the best way to provide a consistent look for the notes—fill in missing volume, issue, publisher, DOI, ISBN, etc. Also, as pointed out above the templates would make some bot work easier. What are the drawbacks to making use of the templates if there are editors that don't mind filling in the fields?EricR 19:24, 5 October 2006 (UTC)
I have several problems: they seem to really chunk up the KB on the article size (making it hard to edit when I'm traveling and on a slow laptop dialup connection), and they return messy punctuation, with ugly gaps. Also, lots of people use the wrong templete, citing a news report as, for example, a website. They don't do an adequate job at all on returning medical journal cites. See my footnotes at Tourette syndrome: I do them by hand by just edit copy, edit paste from the PubMed data. Sandy 19:30, 5 October 2006 (UTC)
In addition to Sandy's points, they're a specialized language that one must learn to provide citations; it's just as easy to learn how to cite properly. Also, they're inflexible, and make lengthy citations (e.g. from multiple sources, or containing quotations) impossible. Finally, they require extra typing which can be easily misspelled and which does not benefit the editor in any way, merely provides some possible potential future benefit to some bot. As for "<ref>http://example.org/some_online_journal.html</ref>" type references, they should be reverted and turned into inline links to the actual articles, since they're also useless. I do this myself when I see it. If one isn't going to provide some benefit with the "ref" citations (e.g. at least the name of the article, author, source), then one has no business converting a perfectly useful inline link into a "ref" citation. Jayjg (talk) 23:10, 5 October 2006 (UTC)
(oops :-) Sandy 23:15, 5 October 2006 (UTC)
I agree w/ the point's y'all have raised except for Jayjg's concerning inline links; if the article text does not actually mention the site in question then the link should go in a footnote (especially if the material has been published or at least an author's name can be found.) There are, though, articles out there that need some reference cleanup (those of us who can't write need something to do.) What should be done w/ the typical article which has a mix of cite templates, Harvard refs, manual cites, naked URL's, etc. I'm willing to use either manual or template citations, but really aren't the majority of editors currently using the templates? I've never seen any complaints due to putting a citation within one of the templates, but could well imagine some opposition to doing the opposite.EricR 01:26, 6 October 2006 (UTC)
I must heavily disagree with Jayjg on inline links here. While the usefulness of formal citation templates is an open issue, I think it's clear that citations need to be more than "here, here's a URL" because what happens when that URL breaks? Check out Anthony Pellicano for a quick example and look how many of the references are URLs to stories from the Los Angeles Times. Now many of those URLs don't work anymore. That would be no problem, because someone who wanted to verify that the story in question supports what it's supposed to could simply look it up through the Times archive using the title of the story -- that is, if anyone had ever added the title of the story to the reference, which they didn't. Turning an inline URL into a reference makes it a lot easier for any editor to see that it needs that attention. -- Antaeus Feldspar 16:19, 17 October 2006 (UTC)

On a somewhat related note, where is the preferred order of sections (Notes, References, See also, External links) listed, and should there be a link in Citing sources? --Gerry Ashton 17:41, 5 October 2006 (UTC)

Per WP:LAYOUT, See also, Notes, References, Further reading, External links. The order of preference moves from Wiki content to off-Wiki content. Sandy 17:49, 5 October 2006 (UTC)
I have to add a disagreement here because, despite initially being against them, I've come round to the cite templates. Consider: I diligently type them out manually based on PubMed; I get hit by a bus; the next user decides on their own lazy manual entry; a third user comes along and, not knowing which to pick, does it themselves a third way; a mess is created. If a cite format were used from the beginning in this hypothetical case, a grain of consistency would help avoid the mess. Yes, people can use the wrong template and they are clunky, but are we expecting people will nicely follow abstracts if no format to do so is provided? In the absence of templates, casual editors are more likely to do whatever they please. Marskell 23:24, 5 October 2006 (UTC)
What are the odds that your hypothetical "casual editor" is actually going to use a citation template, and use it properly? I'd say just about nil. Jayjg (talk) 23:30, 5 October 2006 (UTC)
That's just another way of saying that we should accept that they will always be done wrong. Perhaps they will be, but I suggest a kind of cite categorical imperative: always use a citation format that you think could be readily used everywhere and would always appear more-or-less uniform. Clunky as they are, the cite templates meet that. Marskell 23:35, 5 October 2006 (UTC)
No, what I'm actually saying is that "casual editors" won't use them, and they are nothing but a nuisance for experienced editors, so they're of no benefit to any editors. Jayjg (talk) 00:58, 6 October 2006 (UTC)
Well, this is a bit circular, but: the casual editor will do what they please most of the time, and definitely do what they please (e.g., simply add a link) unless some template on the page suggests doing otherwise. They definitely won't use them if they're not there, but having them present at least makes some uniformity possible. And they aren't that hard for the experienced editor. You know your T's are crossed when you use the same format on a page. Marskell 23:34, 6 October 2006 (UTC)
I just wanted to say I am using the citation templates too and find them pretty useful, they also remind me to include all info I should (ISBNs, access dates, authors etc.) and allow me not to worry about the layout. Perhaps the layout they generate looks funny sometimes, but I owuld guess somebody who has done that was familiar with the standards. I often find the official citation standards weird anyway, but consistency should rule. I also don't find the citations any more "cluttering" than simple refs, they also do not eat up THAT much kilos, unless somebody places a ref at the end of every sentence (which, I believe, we agreed is not best practice unless necessary).
But why am I replying is first and foremost that I infer from this discussion that we are over the "let's change WP:CITE" thing now, and I want to say I am really happy about it. Have a great Friday and Weekend everybody! Bravada, talk - 23:56, 5 October 2006 (UTC)
Well, we're momentarily over it. I don't know if the majority was ever really on it. Someone will inevitably show up to gobble more talk energy, if it seems kosher for now.
More importantly, have we just agreed twice Bravada? If I could give up my complete indifference to GA, we might make a wiki-duo. Marskell 00:02, 6 October 2006 (UTC)
I don't see how any editor can enjoy using the wretched cite templates; I mean, how hard is it to remember the sequence of information required? On the other hand, I do see the point about consistency, and I think the infernal things will one day be used extensively to standardise Wikipedia. However, there are people who enjoy going round standardising things, so let them take care of that in due course, if they like. Editors have got better things to do for the time being, though when you care about an article, you tend to gently standardise the references anyway (without cite templates, in my case), the way you straighten up the books on your bookshelf. But the goat owned in common dies of starvation. qp10qp 00:05, 6 October 2006 (UTC)
Like Bravada, mainly posting b/c it's nice to be talking about something different here, but I like the cite templates too :-) I put regular short notes in my ref tags (ease of citing, plus better readability) which appear in my Notes area, and then do the cite templates in my references section only. --plange 00:08, 6 October 2006 (UTC)
The templates are a horrible nuisance. They make the page hard to edit; they force editors to use one particular style; some of them aren't formatted properly. How hard can it be to remember Surname, first name. Name of book. Publisher, year of publication? SlimVirgin (talk) 01:03, 6 October 2006 (UTC)
I use "cite book," but I'm constantly being puzzled by questions like: what to enter for "first" and "last" for a book by H. G. Wells, or George Gordon, Lord Byron... how to handle works that have editors rather than authors... etc. Dpbsmith (talk) 22:10, 6 October 2006 (UTC)

(Unindent) One objection I have to cite templates is that they do not follow any particular style guide, as far as I can tell. If I come across a situation not covered by any template, I can't look in a style guide to see how to format the item manually. --Gerry Ashton 01:52, 6 October 2006 (UTC)

Why not? CMS covers every permutation known to man, as far as I can see. Or you could just fit in with the style used in that article.qp10qp 02:43, 6 October 2006 (UTC)
My point is that if an article uses templates, but I have a source that does not fit within the limitations of the templates, then I have to format the citation manually. However, since the citation templates implement some style, but there is no documentation of what that style is, I would just have to make it up as I go along. --Gerry Ashton 02:51, 6 October 2006 (UTC)

(Unindent) I just want to add something, maybe it's too late. I like citation template, because of its consistent style. If it gives an ugly looking now, then perhaps later somebody else will modify the template and improve its look and it will automatically change all articles that use the template. Isn't it nice? — Indon (reply) — 09:17, 6 October 2006 (UTC)

I fully agree. Templates is a way to have a common style for citations. However, there are some lack in the fields (for instance, you cannot properly describe a paper published in a series). However, the citations templates should conform to some citation standard for the publication types and main fields names (for instance BibTeX). pom 09:47, 6 October 2006 (UTC)
Agree too. Perhaps someone knows about template "programming" and modify the citation template to conform with BibTeX? — Indon (reply) — 11:44, 6 October 2006 (UTC)

Yet another example

I've been improving the article mirror matter since yesterday. A few more citation need to be included for some facts. However, if I were to follow the guidlines expressed by some here, then this article would need 100 more citations.  :) Count Iblis 13:59, 7 October 2006 (UTC)

If you really find that you need to add four and five cites to one sentence, then I suggest you review the recent and exemplary FA Daniel Boone to see how that can be done effectively. I didn't read the article because, on first glance, it looks more like a WP:POINT. Sandy 15:34, 7 October 2006 (UTC)
I don't understand why the article violates WP:POINT. Combining multiple refs to one will be done later because the article will be expanded and some refs will have to be referred to separately. Count Iblis 15:42, 7 October 2006 (UTC)
I didn't say it violated WP:POINT. Sandy 15:44, 7 October 2006 (UTC)

Citing Television (episodic) and film

Episodes of a given TV show should have a citation form too... shouldn't they? MLA gives a way of doing it so can't wikipedia also do this. While it might seem inherent to put Episode X, there are also anime in Japan that redo series, have multiple series for the same show, and even the US does remakes or have films with the same titles. Having a long line to clarify who did what adds clutter. It would be nice to have a template to add to the pages for TV and movies, conforming the MLA style that's not dependant on specific date seen (not like video, more like for TV episodes, films, etc). So if one film references another film that's not up, that the person can find out the date it was made and the director as well as the title, instead of having to expound it within the article. I looked around, but I couldn't find anything that had that (which was not dependant on "air" date, but more on "publishing" date) So one for film and one for TV (series, not journals) or combine the two... might come in handy later since some of the wikipedia stuff is self-referencing, but it's nice to have it kicking around. Less clutter would be nice....--Hitsuji Kinno 04:12, 8 October 2006 (UTC)

Where to put template for unreferenced or similar tag?

This section fails to give guidance on where to put unreferenced template tag. WP:CITE#Tagging_unsourced_material

So for the Unreferenced, Not verified, or Primarysources templates--where to put them? Top? Bottom? Do people care if the notice is at top or bottom of the article page? Yellowdesk 14:26, 9 October 2006 (UTC)

On the {{unreferenced}} page it says:
There is currently no consensus about where to place this template; most suggest either the bottom of the article page (in an empty 'References' section), or on the article's talk page.
My preference is to put it on the talk page. dryguy 21:50, 9 October 2006 (UTC)
I'd keep it atop the main article, as it warns the users that the article can contain OR and whatnot, and it really urges the editors to do something about it. Bravada, talk - 02:17, 10 October 2006 (UTC)

Help! Articles on obscure/poorly-documented subjects or esoteric practices

I've run into a problem and would appreciate some guidance. Fully aware of WP's guidelines for citations, I have noticed however that WP is rapidly turning into a treasure-chest of obscure and esoteric knowledge! In fact its become, to people like me, a primary research tool for exploring, say, the esoteric martial arts and spiritual practices. It has deepened my knowledge vastly and therefore been an immense boon to me and others. Without WP, I simply would not be able to research things in the way that I currently do.

The thing is, these things were never intended to be documented. By definition, esoteric, occult and Hermetic practices have been kept quiet through most of history, but this period has now ended. Up until now, the primary source that people have for them is experience or observation, not documentation and thus citation is in many cases impossible. For example the article on Kuntao has no citations and will probably remain so, becuase there is virtually nothing written about it.

If there is virtually nothing written about it, then are these? Dpbsmith (talk) 16:01, 11 October 2006 (UTC)

But it is still a known and practiced art.

So, what to do? If we delete such articles, then we diminish the knowledge of mankind,

No, we do not. If there's no source citation, and if the editor's real identity and credentials are not certified, then nobody knows whether it is knowledge or not. Also, Google searches more than Wikipedia, and if people have things they want to make available on the Internet that do not meet Wikipedia's verifiability criteria, there is no reason why they shouldn't be posted somewhere else. Wikipedia is not the "publisher of last resort." On the contrary, Wikipedia is only the integrator of information that has already been published. Dpbsmith (talk) 15:58, 11 October 2006 (UTC)

but if we leave such articles, they do not meet WP citation criteria. And if we cover them with "citation needed" tags, then we are stating the obvious but to no immediate benefit, if ever.

It confers several benefits. It alerts readers. It postpones deletion, perhaps indefinitely. As long as readers are alerted, there's no reason for haste in deleting material. It gives knowledgeable editors the opportunity to provide sources. In my own experience, such sources are often forthcoming. Dpbsmith (talk) 15:58, 11 October 2006 (UTC)

My own preference would be to have a "softer" citation criteria for things that are obscure, in the hope that over time people will emerge and add citations where possible. Meaning, even uncited articles have value and that value is worth preserving, however partial it may be.

For example, Windows Vista is a well-known, live and current topic about which a great deal of information is easily available. We should therefore expect this article to adhere strictly to the citation criteria. But the Kuntao article should, in my opinion, be given more lenient treatment.

Any suggestions or alternative views??? Is there a debate elsewhere on WP that I should know about??

Thanks Punanimal 11 Oct 2006, 13:09 (UTC)

Up til now? What do people have now that is not experience or observation? You suggest that there is something else. This is nonsensical - am I missing something? KillerChihuahua?!? 13:33, 11 October 2006 (UTC)
Yes I see your point, but I mean more simply that personal experience is not citable, but a book or article is. Punanimal 11 Oct 2006, 13:46 (UTC)
I think this is an issue for verifiability. I'm not sure what the standards for inclusion in Wikipedia regarding occult, heremetic, or spiritual practices should be as they are not normally included in academic or encyclopedic research. --ScienceApologist 14:08, 11 October 2006 (UTC)

Knowledge demands not only the inclusion of information, but discrimination between information and misinformation. A lot of monkeys typing for a lot of years will eventually include in their output every fact in the Britannica, but what they have produced is not a body of knowledge because it will also include corrupt, misleading, and contrary versions of every fact in the Britannica, with no way to tell which are the correct versions.

Which of these statements embodies more "knowledge?" Which is more useful to a reader?

Presidents of the United States who are still alive include: Gerald Ford, Jimmy Carter, Mickey Mouse, George H. W. Bush, Bill Clinton, and George W. Bush.
Presidents of the United States who are still alive include: Gerald Ford, Jimmy Carter, Mickey Mouse[citation needed], George H. W. Bush, Bill Clinton, and George W. Bush.

Dpbsmith (talk) 16:35, 11 October 2006 (UTC)

I agree. If there is nothing published about the subject, then it's probably either non-notable or simply untrue. Please search again, more thoroughly, to make sure whether there really is no single source to cite (chances are there is, but it might be quite obscure) - writing "from your own experience" is OR, and it's strictly forbidden in an encyclopedia. Bravada, talk - 22:15, 11 October 2006 (UTC)

Hmmm, what if some editor is in a trance and in contact with some ghost? Maybe what he is writing is "common knowledge" in the "ghost world" :-) Count Iblis 23:29, 11 October 2006 (UTC)

  • Not sure what your point is, but if a published book asserts that something is common knowledge in the ghost world, that book can be used a source for that statement in an article about belief in ghosts, and if that book includes, say, automatic writing author by a ghost, the article could say "in book XYZ medium QRS reported that a statement by ghost JKL, obtained by automatic writing, stated thus-and-such." I am sure we have already articles that reference published content said to have been obtained by supernatural means from deities, etc. Dpbsmith (talk) 10:11, 12 October 2006 (UTC)