Wikipedia talk:Verifiability/Archive 1

Miscellaneous
I see a number of possible applications of this rule, and I agree with some of them and not others.

Say if a statement is added about a fairly obscure topic. The statement sounds perfectly reasonable and there's no reason to doubt it. Should it be removed immediately? Or should it be left indefinitely?

Someone comes along and, without any good reason even after being asked for one, challenges the statement. They don't claim that the statement is dubious, just that it is unverifiable. The person removes it to the talk page pending proof. Should that be tolerated, or reverted?

If someone claims that the statement really is dubious, it should certainly be removed pending investigation. I've done that often. But I'm not at all happy with the other two. The problem is that it takes time to obtain evidence of a statement, especially if there is no freely available online source (as is the case with most human knowledge). Hence the immediate removal of material which has not been referenced would lead to a permanent reduction in equilibrium content. I'm convinced that the improvement in accuracy is not worth the reduction in information. There are many errors in print and online sources. You can't eliminate them merely by enforcing a policy like this one in its most extreme interpretation (i.e. including the first two scenarios).

-- Tim Starling 15:40, Aug 2, 2003 (UTC)


 * Yep, this all makes sense. We need to use common sense, avoid extremes (at either end), and expect contributors to act in good faith - saying why they think some statement is dubious, or why they think it's important to verify it. Martin 16:03, 2 Aug 2003 (UTC)

I have been thinking about this issue because attribution is indicated to be very important in our neutral viewpoint policy.

Legal papers often have a tremendous amount of footnotes (though many would cite the same source repeatedly) - 100+ notes are certainly common, a paper with 300+ notes can be found without a big hassle. I don't think many people would like the idea - even as readers. Even sources might be better used in moderation.

Accepted academic practice in many of the social science peer-reviewed journals are in general much more liberal, I feel. They accept claims like "Wikipedia is often considered the most successful wiki so far," "the problem of vandalism is less than novice would imagine," without citation.

Considering this,
 * perhaps the standard could vary depending on the topic.
 * one should better be knowledgeable on the subject if s/he removes some text b/c of lack of attribution. One shouldn't remove everything unattributed, but should apply some common sense and accepted standard of some applicable field (if there is one).

Tomos 16:36, 2 Aug 2003 (UTC)


 * Wikipedia is often considered the most successful wiki so far

The only information this sentence imparts is that the author has a pro-Wikipedia bias and is trying to cover it up with weasel words like "often considered". "most successful"? How does one objectively measure success? Aren't most websites considered successful in their own way?

Facts first. Look at the size of wikipedia. Look at the traffic. Look at numbers of edits. Quote figures, draw graphs. Then, if you still need to include mere opinions about successfulness, attribute them to known, and respected, advocates. For example, quote experts on wiki, like Sunir Shah or Ward Cunningham. Don't quote anonymous fans of Wikipedia, because nobody cares - if I want the rantings of the massings I'll go read usenet ;-) Martin 17:13, 2 Aug 2003 (UTC)

I'm very new to Wikipedia, so I may be speaking out of turn here, but I'm not sure the above discussion deals with one of the main verifiability issues to come out of the Daniel C. Boyer discussion.


 * Say if a statement is added about a fairly obscure topic. The statement sounds perfectly reasonable and there's no reason to doubt it. Should it be removed immediately? Or should it be left indefinitely?

I agree that the statement should remain until challenged by another user.

Upon a challenge, if the author cannot produce a source, then it should be deleted.

But what happens, if upon a challenge, the author produces a source that is itself dubious. The Boyer discussion references a weblog called Sigg3.net and a website called filmmakers.com. Do these sources have the credibility necessary to represent verifiable sources?

Two scenarios come to mind. One, where a "dubious" source contradicts a "reliable" source. Two, where a "dubious" source makes a statement, and there is no "reliable" source to corroborate or dispute the statement.

I realize that the terms "reliable" and "dubious" are undefined here and can themselves be the source of controversy, but for the moment, think of it as a spectrum of reliability (e.g. The Washington Post is generally a more reliable source than sigg3.net)

In the first scenario, I think we can agree that a statement from a comparatively anonymous website does not have the credibility of an established media site. If a statement supported by dubious web site was in direct contradiction to a statement from an established media site, I think most of us would agree that, on the basis of that evidence alone, the latter statement would have to be viewed as more verifiable, and therefore, more factual.

It gets more complicated as there are more sources supporting or contradicting a statement. If reliable sources can be found on both sides of a statement, than the article probably demands acknowledgment of both POVs. But only if both sources can be viewed as reliable.

In the second scenario, an esoteric statement or a purely invented statement may have no discussion in another form of media. For example, if I was to create a hoax website that reported that an ancient tribe called the Wikiorians had been found, a search of the available literature would find no contradiction of my statements, since, of course, I invented the tribe. If I then claim my website as a source and write a splendid article on the stunning cultural of the Wikiorians, I have made a farce of the encyclopedia, and according to the rules as I understand them, my entry would not be deleted, since no one can find a source contradicting my statements.

Since the Web is itself anarchic and the truthfulness of any web source cannot be ascertained, I think the "reliability" of sources should start to be considered in determining the verifiability of statements. Whatever metric is used to measure reliability, I think that a source that is not considered "reliable" should not alone be able to support a statement. In other words, upon a challenge from another user, if the only sources presented as evidence of the truth of a statement are a weblog, random website, or such other "dubious" source, the statement should be moved to the talk page until a more reliable source can be found. If the statement really is factual, there ought to be some form of corroboration available.

Just an idea. Hope I'm not regurgitating or contradicting guidelines presented some place else. Cheers, SpeakerFTD 02:08, 5 Aug 2003 (UTC)


 * I strongly disagree with Eloquence's actions at Daniel C. Boyer -- i.e. removing something on the basis that it is not verifiable, rather than removing it because it is thought to be incorrect. That was precisely the action that motivated my initial comment. I might write more on this later but I'm out of time. -- Tim Starling 03:01, Aug 5, 2003 (UTC)


 * If you want to list only facts, I'm not sure there is a difference between an incorrect and unverifiable statement. Even if there is no evidence to directly contradict a statement, in the absence of a single, dependable source available to others that can corroborate a statement, there is no way for the public to differentiate an invented statement from a defensible one.  You might be right that it is incorrect to immediately remove an unverifiable statement, but there has to be a mechanism to challenge a statement that comes from dubious sources.  SpeakerFTD 14:07, 5 Aug 2003 (UTC)

Mr. Starling, can you clarify your disagreement with Eloquence? Do you mean that the statement that he removed was literally impossible to verify from published sources, or do you mean that you think it could have been verified if only someone had tried harder?

If you mean the former, then of course Eloquence was right to remove the statement. A true but unverifiably true statement is indistinguishable from a false but unverifiably false one. Do we just include statements on the basis of how plausible they sound? I could think up any number of plausible-sounding statements that you would never be able to ascertain the truth value of. Do you think they should be included? If so, then we will end up with an awful lot of false statements.

If you mean the latter, well, that's a trickier question. How much effort should people have to go to to check things? If we say that people have to go to a lot of trouble to check obscure things before they remove them, then people will rarely bother, and so the false statements will stay. I think we need to allow people to remove content if they have made a reasonable attempt (the meaning of "reasonable" is debatable, of course) to check. But I think that we need more definite rules to avoid conflicts. Here is my proposal:

How does that sound? -- Oliver P. 22:46, 6 Aug 2003 (UTC)
 * 1) If you feel the urge to remove a statement from an article, first check the bottom of the article for references.
 * 2) If there are any, check the sources. If you can confirm the statement using them, leave it in; otherwise, continue to step 3.
 * 3) Use your common sense to work out what other resources would help, and check them. If you can confirm the statement using them, leave it in; otherwise, continue to step 4.
 * 4) Move or copy (?) the statement to the talk page, explaining that you have not been able to verify the statement, and stating what sources you have checked.
 * 5) (Optional?) Check the article history for who added the statement in the first place, and leave a note on their talk page telling them that their statement is disputed, and directing them to the appropriate talk page.
 * 6) Anyone may now feel free to try to verify the statement and produce a reference on the talk page.
 * 7) If at step 4 you only copied the statement, wait a week (or other random amount of time), and if no-one has found a reference in that time, remove it from the article altogether. (Don't worry, it'll still be on the talk page.)
 * 8) If someone does find a reference, the statement should be put back into the article, with the newly found reference. If no-one finds a reference, the statement can remain on the talk page indefinitely.


 * You mean we don't do that now? :-) That is more or less the procedure I currently follow, so it sounds good to me; there is a remarkable amount of junk, especially in older articles, where people have apparently tried to work from memory, or from the junkiest web page they could find. Leaving a note on talk page should definitely be optional; if someone really cares enough to competently defend their statements, they'll be watching the article anyway. If we had a more organized system for references (at one point I was considering articles for all refs, but that's heavyweight), it would be easier to keep track of the articles unsupported by any reference material. Stan 23:21, 6 Aug 2003 (UTC)

I mean the latter. And I should have been more careful than to lay the blame solely on Eloquence -- a more thorough re-examination of the archives would have shown me that one of the statements I was particularly worried about came from Martin:


 * "The book received a favorable review in Spam magazine, a local zine in the writer's home town."


 * I removed this as unverifiable. Did anyone find a reference for it? Martin 15:38, 1 Aug 2003 (UTC)




 * If it's not verifiable, that's fine - it just can't go in the article. Martin 19:48, 1 Aug 2003 (UTC)



Hence the onus of proof is placed on the original author, not the editor. "Verifiability" is not defined as what can be possibly determined, it's defined as "what can be determined by searching Google". The idea that it received a favourable review in Spam magazine is plausible, it could be verified by tracking down the publisher and contacting them, so what's the problem? Should we remove everything in Wikipedia which can't be verified by x hours work, where x is whatever the Wikipedian in question feels like spending that day? Eloquence later laid down a challenge:


 * Even if particular facts can be verified now, it is unlikely that they will remain verifiable 10 years from now. So there is a substantial lack of verifiable information, but the article fails another test: Can it be reasonably expanded into something useful, or will it, after the removal of all unverifiable information, be an "unfixable" stub? It appears that the latter is true.


 * My offer to Daniel would be this: Put a copy of "Tailgating Spinster" online under a free license. If it's of any interest, and mirrored on a few sites, it will probably remain on the web, and we can at least give a summary of the book and its author on this page. Other than that, there's simply not enough useful data for a Wikipedia article. These simple rules -- verifiability and expandability -- are in my opinion enough to settle most questions like this.&mdash;Eloquence 00:10, Aug 2, 2003 (UTC)


 * I fail utterly to see why there is any responsibility for an author to digitise all his works, to satisfy the unstated rule, reflected in Wikipedia over and over, that if a source is not online, it is therefore valueless. Despite repeated challenges I've not received any response justifying this.  I don't understand the "why" of this challenge.  Why is the book itself not regarded as "data"?  Unanswered, unanswered, unanswered.  --Daniel C. Boyer 19:11, 1 October 2005 (UTC)

Is ensuring verifiability in 10 years time really the original author's responsibility? We editors, right now, have the opportunity to make these facts verifiable in 10 years time simply by documenting our sources. What could be better verification than a statement from a number of independent Wikipedians saying "I have seen the magazine, I can confirm that the article is correct" -- a statement which in 10 years time will no doubt be mirrored all over the world in time-stamped backups. If such a statement is not sufficient, should we also remove all accounts of historical books which are no longer extant? Unfortunately my knowledge of history isn't good enough to come up with really good examples... how about Diatessaron?

Now as for Oliver's 8 step process, it's all fine by me except for one minor detail, and that is the editor's motivation for removing the material. The test should be plausibility, not pseudo-verifiability. If Daniel had said "the book was critically acclaimed, receiving widespread praise from critics around the world", then we might say "yeah right", and remove it. But everyone knows how local media likes to talk up the acheivements of the locals. What reason do we have to doubt this statement?

Of course, the statement may well be removed on the basis of lack of importance. The entire section on The Octopus Frets in this revision, or indeed the entire article, could have been removed on that basis alone. Importance is subjective and I'm always willing to listen to community opinion on that issue.

-- Tim Starling 01:11, Aug 7, 2003 (UTC)


 * I think my take on the responsiblity of the community is a little bit different than Tim's. When he says...
 *  What reason do we have to doubt this statement? 
 * My natural reaction is to answer that, since this is an encyclopedia, and by my definition, it should only contain defensible statements (i.e. facts), the community has a responsiblity, or at a minimum, a right, to doubt every statement. Like a scientific hypothesis, a statement should be able to withstand a challenge from critics or it should not be included.  If a statement deserves inclusion in the document, the least the author can be expected to do is provide a source.  Oliver P.'s mechanism above seems like a reasonably sound approach.


 * The only question I have about that mechanism is whether or not it takes into consideration the quality of the source. Unless a source is independent of the author, I do not think it should qualify as reputable source.  If an author can create a website then claim that website as a source, the verifiability problem is not solved.  This issue has arisen in Talk:Collage - several sources are listed by Daniel C. Boyer, but all the sources seem to have a relationship with the author.  I would suggest that Oliver P.'s mechanism be modified to require an author, upon a challenge, to produce sources that are independent of the author.
 * SpeakerFTD 04:44, 7 Aug 2003 (UTC)


 * Daniel Boyer did provide a source, and the statement was defensible. The author of a statement is not always contactable. Should we remove a statement if we can't find the author, even if we could, in principle, verify the statement ourselves? -- Tim Starling 07:34, Aug 7, 2003 (UTC)


 * I don't think providing a source is enough. As Martin alludes to below, I think the author needs to provide a source that can, with a reasonable amount of effort, be checked by others.  As I understand it, "Spam Magazine" has a tiny, local circulation, meaning that it is almost impossible for any one else to  check the source and see whether a) an article actually exists, b) the article actually states what the author claims it states, and c) the article isn't self-authored or authored by someone with a personal interest in forwarding a POV.  Future readers who are unaware of a controversy surrounding a statement will likely take it at face value as a fact (that's what I do when I read an encyclopedia).  If the community cannot verify the source of a statement, then the possibility that the statement is a falsehood should preclude the statement from inclusion (the statement can continue to exist on the Talk page, of course). SpeakerFTD 05:50, 8 Aug 2003 (UTC)

Regarding my removal of the "Spam magazine" factoid. Daniel seemed to me to have a record of including information that is true, but incomplete and possibly misleading. For example, mentioning a film and neglecting to mention its three minute running time. Given that, I didn't want to go by his word alone - the review might be but ten words long, or it might have been a self-review, for example. I spent about half an hour or so following refs and trying to verify, without finding anything useful.

Daniel's response was ''What would you like me to do to verify this? (Spam magazine is not on the World Wide Web, the popularity of which its existence pretty much predated.)''. I interpreted this, perhaps wrongly, as an admission that he personally could not verify this information.


 * I have personally mailed Martin a photocopy of the article in question. Hopefully this will put this issue to rest.  --Daniel C. Boyer 14:26, 15 Aug 2003 (UTC)


 * This is difficult for me to interpret as other than a deliberate distortion of my meaning. If anyone will look at this for less than a second it clearly means anything but what Martin says.  What I meant was, would Martin like me to mail him a copy of the review, or do something else to prove to everyone that it exists?  --Daniel C. Boyer 18:56, 9 Aug 2003 (UTC)


 * If it was a distortion, it was not deliberate. Hanlon's razor applies...
 * Btw, thanks for the snail mail. :) Martin 09:41, 11 Aug 2003 (UTC)

If he couldn't, then realistically, I couldn't, so I left the information out of the article and on the Talk page. Probably not an ideal approach, mind you.

Oliver's eight step process makes a lot of sense. For large articles it can be difficult or near-impossible to find the original author of a specific sentence - and in any case they may have moved it from elsewhere - so leave that optional, I think. But certainly that's been my approach, in general - for example, removing some dubious quotes from Christianity and anti-Semitism. Unverified statements are also difficult to check for accuracy and neutrality, so these issues are linked, in my opinion. Martin 21:02, 7 Aug 2003 (UTC)

Oliver's procedure
I copied Oliver's procedure to the article. Needs tweaking, but it's a good start. I added checking the talk page, which seems obvious, but should be said.

Personally, my procedure is slightly different, I've realised:


 * 1) Check with sources in article, talk page, and other easy stuff.
 * 2) If can't verify easily, copy statement onto talk page - ask if anyone has an easy way of verifying this.
 * 3) If no response after a few days, make a further effort to verify - this might involve emailing someone, or heavy googling, etc.
 * 4) If still can't verify, move statement to talk page - say that I removed it as unverifiable.

Oliver's procedure has my stages 2 and 3 the other way round, but I think my way is often more efficient, especially if an article is being edited by lots of people who know more about the subject than me, or if the author is still around. Martin

In general, I find that I agree with most of what Eloquence and Martin have said. I like Oliver Pereira's checklist, though shortcuts may well be appropriate in some cases. And I can see Tim Starling's point in many areas, especially that material should not be removed simply because someone happens along and can't verify it in the first two minutes of checking. I also see his point about long-term verifiability. I have this to add:


 * Fact checking is tedious, time consuming, and not particularly rewarding. It is unfair to make later editors dig for sources, particularly when the initial content is questionable.  Those who write articles likely to be deemed in need of fact checking (because of obscurity, implausibility, novel nature, controversy, author conflict of interest, or author reputation) should expect to assist by providing references, ideally when the article is first written.


 * For an encyclopedia, sources should be unimpeachable. An encyclopedia is not primary source material. Its authors do not conduct interviews nor perform original research.  Hence, anything we include should have been covered in the records, reportage, research, or studies of others.  In many, if not most, cases there should be several corroborating sources available should someone wish to consult them.  Sources should be unimpeachable relative to the claims made; outlandish claims beg strong sources.


 * Similarly, sources should be reasonably readily available. They needn't be on-line, but a check at a suitable library (perhaps a specialist library for in-depth articles on certain scholarly topics) should turn up a suitable source. Interlibrary loan is widespread, and it should be possible to check anything of remotely encyclopedic interest that way.

Kat 23:13, 11 Aug 2003 (UTC)

A discussion of the issue of verfiability with regard to dates is at Wikipedia talk:Verifiability/Dates.

A general comment on verifiability -- we seem to focus too much on "google verifiability." I don't think you can seriously call a fact "unverifiable" unless you've visited a library, tried to look it up, and failed to find it. There's a lot of important stuff that doesn't come up on google. --Delirium 09:18, Aug 22, 2003 (UTC)

And I suppose we should note the converse of that. As well as there being true stuff that isn't on Google, there's stuff on Google that isn't true. Obvious, really, but it's easy to forget sometimes. There's a lot of information on random webpages that's actually rubbish. Some of it repeated over hundreds of pages! So how do we tell? I'm afraid to say that I've added stuff to articles before, having found some random webpage that claimed it was true, and then I've later found out that it wasn't. I suppose that (a) I should be more careful, and (b) I should attribute information from dodgy sources, so that people can tell it's dodgy... -- Oliver P. 09:38, 22 Aug 2003 (UTC)

Falsifiability?
Shouldn't the name of this policy be Falsifiability?

Since Karl Popper "Verifiability" is considered as a nasty way to come to a semblance of scientific allure, not to scientific NPOV.

Maybe it could be considered like this: "Verifiability" relates to "Falsifiability" like "Objectivity" to "NPOV". "Objectivity" is rightfully rejected in favour of "NPOV" in Wikipedia guidelines. I think the same should apply for "Verifiability" in favour of "Falsifiability".

This maybe doesn't even mean that the content of the "Verifiability policy" page has to be altered all that much: most of it will stand as "Falsifiability" guidelines as well. I suppose the difference is in the details, exactly those details that in my view are most effective in enhancing NPOV, like Falsifiability implying that something is "less" proven when there are no opposing views.

Naming this policy page "Falsifiability" should in my view also allow to give a low-treshold definition of this concept, which has been so important for science since early 20th century. The concept appeared to me quite abstract and unpractical the first time I heard it (in a philosophy course), but now I'm convinced its extremely practical (e.g. to avoid single-minded POV's), AND I'm convinced it can be explained in easily understandable terms. It would allow many people to learn something extremely practical, which they probably wouldn't learn from reading the Falsifiability article, which mainly explains it as a technical philosophical issue, in connection with "demarcation" (duh-mark-whatsion?), not the kind of thing people would find practical for their own lives.

--Francis Schonken 09:31, 29 Aug 2004 (UTC)
 * Absolutely. IMHO this policy tries to be "falsifiability", but gets confused. I will move the page in a day or so, if there are no objections. ··gracefool |&#9786; 01:24, 6 Sep 2004 (UTC)


 * I object. I also object to replacing policy tag with "proposed policy/thinktank". This has been accepted as a policy long ago. Andris 18:52, Sep 7, 2004 (UTC)


 * To explain my position, Wikipedia deals with different types of statements then Popper's philosophy. Popper deals with universally quantified statements, like scientific theories (or "All cats are black"). Such statements are indeed falsifiable (by exhibiting a non-black cat) but not verifiable (no matter how many cats you've seen, there can always be another cat that is not black). A typical Wikipedia statement is "Such-and-such person is born on January 12, 1951". That is a proposition, rather than universally quantified statement. Here verifiability applies (include the statement, if you have a reasonable evidence), rather than falsifiability (include the statement, unless you have an evidence he was born on a different date). Andris 19:33, Sep 7, 2004 (UTC)


 * (Returning here after a very long time):
 * What I understood in the mean while: it's not exclusively about scientific standards on Wikipedia, several other motifs are valid. The most important are probably "interesting" and "significant/important", as explained in the NPOV tutorial, "space and balance" section, for which "verifiability", on the whole, appears very practical as a concept.
 * I still think on the verifiability page some link with the "falsifiability" concept could be made, in the sense of: falsifiability can further enhance the "scientific value" of articles (about which there is much ado nowadays, e.g. in the form of questions like: "how reliable is wikipedia"?). In other words the difference between "proposition" vs. "universally quantified statement" as explained above by Andris appears, to me, not quite to the point. The question whether Antonio Vivaldi was yes or no a catholic priest, or whether "red priest" was just a nick-name for the guy, will be answered the more scientifically correct, the more falsification technique is used in the methods of trying to find that out. That he was called "red priest", is something very easily "verifiable". And maybe that's enough for wikipedia. Ascertain more than that, will ultimately always go back to scientifically researching contemporary documents or other references. And scientifically researching by definition comprises what Popper intended by falsification (or the research could not be called scientific at all). Of course it's not the "original" scientific research that is supposed to be done by wikipedians - "falsification" can be performed by wikipedians in choosing their sources: giving precedence to those sources that show to have applied the most broad research (in the Vivaldi example: taking an authoritative Vivaldi biography based on original sources above otherwise unverifiable CD leaflet notes).
 * I'll probably still think the falsifiability issue (and it's relation to verifiability) over a bit more, and when I feel inspired, initiate a paragraph about it on the project page.
 * --Francis Schonken 16:02, 25 Jan 2005 (UTC)
 * PS: in the mean while the Vivaldi article had been rewritten (the example above referred to the time when only the "red priest" nickname was mentioned) --Francis Schonken 15:25, 27 Jan 2005 (UTC)
 * PS2: see also Village_pump_%28miscellaneous%29 (topic started 25/01/2005) --Francis Schonken 15:25, 27 Jan 2005 (UTC)

I agree wholey with Andris, and disagree vehemently with Francis Schonken. The Popperian concept of falsifiability applies to scientific hypotheses and has nothing, nothing at all with what we are doing here. Slrubenstein 20:41, 27 Jan 2005 (UTC)


 * I agree; see my comment below. However, Schonken has posted at the village pump that he intends to introduce us to falsifiability-light, which I await with great pleasure. ;-) SlimVirgin 20:43, Jan 27, 2005 (UTC)

Copy from a debate about an obscure topic about first hand accounts of events
I copied this from the obscure topic talk:criticism of Prem Rawat because I thought it gives a good example of a debate of this difficult question. I hope that general guidelines can be expanded soon to cover first hand accounts. Andries 10:28, 12 Sep 2004 (UTC)


 * [..] Are there any standards for what is quotable or not quoatable? If tomorrow an exfollowers alleges this or that and writes in Usenet a "testimony", is that quotable? --Zappaz 15:40, 10 Sep 2004 (UTC)


 * Zappaz, I think the guidelines are out there in the collective consensus more or less amorphously, even if not explicitly stated. From my background I tend to approach things from the standpoint of legal evidence rules, and whether or not this material if it were offered as evidence could be shown to a jury. Hence I refer to those "indicia of reliability." We can see the difference between the Dettmers allegations, which I support for inclusion, versus the notion of Prem Rawat knowing of Jagdeo's activities, which I don't. With Dettmers we have someone whose organizational title gives plausibility to his having had personal access and knowledge, and who tells a fairly detailed story of specific events, for example a drunk and abusive Prem Rawat offending a particular woman on a particular occasion. I didn't include that here because it's not particularly informative, but it means Dettmers is daring to put out specific items that he knows are subject to being specifically tested and disputed, by counter-testimony, corroborative documentation, discrepancies in the details he gives (similar to an alibi), whatever. He says, "I was there and this is what I saw." Now, the guy may be lying through his teeth, but we can't say he's assembling conclusions from other people's info without a basis in his own personal experience. With the Jagdeo thing, by contrast, there just is no one and nothing that pins down Prem Rawat to knowing about it. That's the kind of inference we keep out. Even in Dettmers' stuff, I left out his characterizations&mdash;Prem Rawat is a coward, Prem Rawat is abusive, there is a climate of fear around him&mdash;because those are of no help to us; instead, we can read the specific events Dettmer claims to have witnessed and decide for ourselves. I did put in one characterization, about Prem Rawat being not cynical but instead believing himself beyond restrictions, because that characterization goes against the more obvious one of hypocrisy being drawn by other ex-premies, and so the very fact of it coming from our witness is somewhat informative. So, to (finally!) answer your question, if an ex-premie gives a testimony on usenet tomorrow, before including it we would ask, Is it specific? Is it testable? Does it come from his personal observations, and not someone else's hearsay? Do the testimony's or the person's situation make it plausible that he might be telling the truth? Is it corroborated? [..] --Gary D 23:13, Sep 10, 2004

I agree that this is a very useful comment. But -- helpful as it is, it is of limited use. Different sciences for example have different standards than a law court (indeed, in the US criminal and civil courts have different standards). We can draw on the above for ideas, but we can't apply it strictly to Wikipedia, which is not a court. Applying these evidentiary standards might be appropriate in some cases, but if we try to use this for articles in social science or the humanities we will only stir up unnecessary conflict. Slrubenstein 17:35, 27 Jan 2005 (UTC)

Falsifiability and verifiability
I have a concern about the issue of falsifiability. The way the term's being used above is not the way I believe it's normally used. A proposition is falsifiable if it is not necessarily true i.e. if it is contingent. All necessarily true propositions (i.e. tautologies) are regarded as meaningless. The importance in science of the principle of falsifiability is therefore connected to the importance of ensuring that scientific propositions (if they want to be called scientific) are meaningful, and are not just closed systems. On that basis, for example, Freud's theories could not be regarded as scientific, or in any sense meaningful, because they are not falsifiable and amount to a closed system. I'm wondering what any of this has to do with verifiability in Wikipedia? SlimVirgin 20:24, Jan 27, 2005 (UTC)


 * First I'd like to emphasise I want to go easy on this one, I mean: trying not to rush, neither trying to get something done for which there's clearly no consensus at the moment. So be assured, nothing's going to happen if people think as they do now. Just a few thoughts:
 * I don't think Popper said anything near to "non-falsifiable" theories being "void of meaning". Yes, Popper didn't like Freud's theories. The only point he attacked them on was that these theories were not scientific, because that was what his falsifiability was all about: demarcation between the scientific and the non-scientific (not between the meaningful and the meaningless). Yes I read in one of the wikipedia articles this error that Popper would have said that Freud's theories were meaningless - he didn't: he only said they were non-scientific, WHILE STRESSING that non-scientific is not the same as meaningless (amongst the meaningful but not scientific he considered e.g. art, music, religion, ..., psychoanalysis, ...). Note from history: from the second half of the 20th century on, many of Freud's theories have been subject to scrutiny along scientific methods (i.e. involving the falsifiability technique): some of Freuds theories have proven false (or: at least too limited); other theories have been scientifically confirmed, while probably most of Freud's theories are still out of reach of scientific scrutiny. I suppose the situation could be summarised, in the early 21st century: Freud set the world a-thinking on the topic of psychology, though he was probably wrong in many issues (e.g. when he got involved in anthropology).
 * My view on wikipedia: scientific where possible/meaningful - in all other cases being "interesting/informative" and "significant/important" (in other words: "non-scientific meaningfulness") suffises more than well.
 * But note that when "reliability" of wikipedia is an issue now; when there's talk about quality standards, etc... then the question of how wikipedia sees its scientific level can not be avoided (and I don't think "falsifiability" has been surpassed as a "level of scientific credibility" criterion).
 * About my allegedly "unorthodox" interpretation of falsfiability: I don't think it's all that unorthodox; equalling "non-falsifiable" with "meaningless" is much more unorthodox w.r.t. Popper's theories.
 * --Francis Schonken 09:53, 28 Jan 2005 (UTC)


 * I do not think it has anything to do with Wikipedia. The crucial word in your commenht is "proposition."  Wikipedia, not being a place for original research, is also not a place to propound propositions.  I think you are reacting to an error made by Frances Schonken.  He saw the word "verifiability" and, assuming that this word can have only one meaning and use, understood it in terms of Popperian philosophy.  Since Popper is proposing to substitute "falsifiability" with "verifiability", Schonken thought that we should do the same.  Alas, in English as in most languages words have several meanings, and "verifiability" is a perfect example.  I think anyone reading the policy should be able to tell that it is not being used in relationship to empirical propositions or to laboratory experiments. Slrubenstein 20:54, 27 Jan 2005 (UTC)


 * "proposition": of course in the "wikipedia:" namespace (where we are now) policies and guidelines are proposed (and e.g. not "imposed"): see How to create policy - this has no relation with reservations against "original research".
 * No, I don't assume "verifiability" has only one meaning & use. Yes, verifiability got an additional connotation as something Popper rebutted in the science/no science debate. Popper said "verifiability" is not enough if one wants scientific credibility.
 * Popper proposed to substitute "verifiability" with "falsifiability" in the science/no science debate (not the other way around!).
 * Popper wanted to disenchant the idea that a scientific approach is guaranteed by "empirical propositions or laboratory experiments", the empiricists (which Popper was fighting) had been hiding too long behind this kind of external appearances as a proof for scientific level. So, the fact that verifiability doesn't indulge in "empirical propositions or laboratory experiments" has nothing to do with whether or not wikipedia aspires to enhance its scientifical level.
 * --Francis Schonken 09:53, 28 Jan 2005 (UTC)

Slrubenstein is right, Francis. The theory of falsifiability has nothing to do with checking sources for an encyclopedia entry. But I have to ask you about a couple of things you wrote: (1) Which of Freud's theories has been "scientifically confirmed"? (2) Can you quote me a section from Popper that shows he believed psychoanalysis was in any sense meaningful? The whole argument against closed systems is based precisely on their lack of meaning; in that, by explaining everything, they explain nothing. Read Gellner (who was a Popperian) for more on psychoanalysis, closed systems, and fraudulent thinking. SlimVirgin 11:03, Jan 28, 2005 (UTC)


 * I was just extending what I tried to say about "propositions", but there was an edit conflict. Here's the text I wanted to add above (and which somehow is an answer to the rightful remark of SlimVirgin):
 * But probably Slrubenstein meant "propositions" in the main namespace (the "encyclopedia") - there also nothing but "propositions": that's how a wiki-system works: somebody "proposes" a text for a new article, and unless the page gets blocked (which is and should be a rare exception), others extend the proposition, until nobody changes it any more (which should mean every wikipedian agrees with the proposition). The "no original research" principle has nothing to do with yes or no working with "propositions", only that such propositions are not weighed against original data, but against primary and secondary sources, as explained in no original research.
 * Sorry for having been too quick
 * --Francis Schonken 11:32, 28 Jan 2005 (UTC)


 * Still about: Can you quote me a section from Popper that shows he believed psychoanalysis was in any sense meaningful?.
 * My answer:
 * No, I can't, at least not presently, because I have none of Popper's books at hand. I only have a secondary source which is a course on science philosophy. The course is by Herman Roelants, and is in Dutch.
 * You see how quick you go in "falsifiability" mode yourself? I make an assertion about what Popper intended, and (while you doubt that assertion is true), you propose a "falsification scheme": go through all of Popper's books and bring forward the paragraph(s) where he confirms psychoanalysis to be meaningful. "Verifiability" would be limited to quote someone who had it printed that Popper thinks psychoanalysis meaningless. So, I think that, in fact, you got already part of my point.
 * --Francis Schonken 11:47, 28 Jan 2005 (UTC)


 * Third part of the answer: the quote appeared easier to find than I thought, by this marvel called internet (googled on Popper + Freud, found this reading published in Conjectures and refutations, 1963): http://www.stephenjaygould.org/ctrl/popper_falsification.html - I quote from section 2 - 4th and 5th paragraph:
 * The two psycho-analytic theories [...] were simply non-testable, irrefutable. [...] This does not mean that Freud and Adler were not seeing certain things correctly; I personally do not doubt that much of what they say is of considerable importance, and may well play its part one day in a psychological science which is testable. [...] These theories describe some facts, but in the manner of myths. They contain most interesting psychological suggestions, but not in a testable form.
 * At the same time I realized that such myths may be developed, and become testable; that historically speaking all - or very nearly all - scientific theories originate from myths, and that a myth may contain important anticipations of scientific theories. Examples are Empedocles' theory of evolution by trial and error, or Parmenides' myth of the unchanging block universe in which nothing ever happens and which, if we add another dimension, becomes Einstein's block universe [...]. I thus felt that if a theory is found to be non-scientific, or "metaphysical" (as we might say), it is not thereby found to be unimportant, or insignificant, or "meaningless," or "nonsensical." But it cannot claim to be backed by empirical evidence in the scientific sense - although it may easily be, in some genetic sense, the "result of observation."
 * I never read Gellner, but it might be I falsified some of his assertions.
 * --Francis Schonken 22:52, 28 Jan 2005 (UTC)

And which of Freud's theories has been "proven false" or "too limited"? SlimVirgin 11:07, Jan 28, 2005 (UTC)
 * I refer to (amongst others):
 * "sociale en culturele antropologie" by Eugeen Roosens, ISBN 9033410974 (in Dutch alas), explaining that few of Freud's anthropological assumptions survived field testing, where field testing is some kind of a scientific method (of course, in field testing also non-scientific parameters like the "anthrolopogist's experience" enter the game).
 * Daniel Goleman described in his book Vital lies, simple truths. The psychology of self-deception an "extended model" of the human mind, where he uses Freud's model as a starting point, and extends it with what has come up since in scientific research (e.g. measuring electric brain pulses when under stress). This is part 2 of the 2001 edition of that book (I only have the ISBN of the Dutch translation: ISBN 902549708x)
 * there are other examples.
 * --Francis Schonken 11:32, 28 Jan 2005 (UTC)

Uh, I wouldn't give much importance to a refutation of Freud's "anthropology." If the book in question is referring to Totem and Taboo, Freud wrote it rrelatively late in his life and admitted it was mostly speculative (the vice of old scholars); moreover, most of the ethnographic claims were based on what anthropologists knew at the time. I think it is more accurate to say that those anthropologists were refuted (or corrected) by later studies. In any event, it seems we all agree that falisification has nothing to do with the topic of this page. Slrubenstein 16:58, 28 Jan 2005 (UTC)


 * Whatever "old" or "speculative" or "copycatting" Freud was, most of "Totem and Taboo" was not scientific at the time he wrote it (while, simply, not falsifiable - I don't need a report of Freud's mental status or of the lacking scientific stature of his contemporaries for that), and most of it was also refuted afterwards (except, probably, for some odd diehard Freudians). But indeed, this was not the topic, merely an example. I suppose I spent time enough elaborating these examples, to the point that they should be clear by now.
 * The examples were only intended to show:
 * Basicly, falsifiability is an acceptable criterion for "scientific value", and falsifiability is in some context usable in wikipedia.
 * Meaningfulness (which, also expressed as "importance" plus "informativeness", is the basis for inclusion in wikipedia) is not tampered with, if applying falsifiability to some articles in addition to verifiabilty.
 * Falsifiability can enhance scientific value of wikipedia articles, which should have a positive effect on wikipedia's over-all reliability.
 * Falsifiability, if applied in wikipedia, can only be applied for falsification against existing sources and not for falsification against raw or original data, see no original research. Raw or original data have to be fiably interpreted and/or described outside wikipedia, before they could have any effect on a wikipedia article.
 * Not discussed above, but an important side-remark: For wikipedia most of the extensions later added to Popper's original falsifiability principle have few importance (e.g. "paradigms"), while these extensions generally only have importance for original research (which is not the case for wikipedia).
 * Anyway, no support for these ideas presently, so I think I'll do as the previous time: do more thinking (e.g. on how to make this practical), and return here after some while, I'd be glad to read more comments.
 * --Francis Schonken 22:52, 28 Jan 2005 (UTC)

Francis, I didn't realize when I first saw your posts that you are Dutch. I think the problem may be that we are using words differently. I hope what follows isn't too long, but I see this issue has cropped up before on this page, so perhaps it would help if I explained how I'm using the various terms.

A proposition in the sense I used it has nothing to do with proposing something ("proposing" as opposed to "imposing", as you wrote). A proposition is simply a statement about the world that has a truth value. An example is: "The cat is on the mat." Assuming for the sake of argument that there is only one cat and only one mat, and that the cat is indeed on the mat, the truth value of this proposition is that is contingently true. We know it is contingently true because we know it is falsifiable i.e. it could be false. If we were to see the mat without the cat on it, or the cat without the mat underneath it, the proposition would be contingently false. Another way of putting this is: "There is a possible world in which the cat is not on the mat."

"The cat is on the mat" is therefore regarded as a meaningful proposition in the Popperian sense: it tells us something about the world. Contrast that with: "Two plus two equals four." This proposition is necessarily true. There is no possible world in which it is not true. That is, it is not falsifiable. That means it is not meaningful. It covers every possibility and therefore has no explanatory power.

Popper suggested that the important question to ask of any theory that purports to be scientific (and, as I understand it, meaningful) is to ask "Could it be wrong?" and only if the answer is "yes" can the theory be regarded as conveying a meaning. That is what the principle of falsifiability is based on. An example of why Freud's psychoanalytic ideas might be regarded as meaningless is as follows: A man who denies he is homosexual may be regarded as "in denial," and the more he denies it (by marrying a woman, denouncing homosexuality, attacking gays with clubs), the deeper in denial he is deemed to be, and the more he is "projecting". How could this man ever show that he is not gay if, the more frantically he seeks to deny it, the more he appears to confirm it? These ideas about denial and projection (and repression and all the rest) represent a closed system, and are therefore unscientific and devoid of meaning (both, not just the former).

The principle of falsifiability has no application when it comes to writing an encylopedia article. If I want to make this edit: "Wittgenstein wrote the Tractatus," I can verify this by checking with the publishers, going to amazon.com, asking his former students. But what purpose is served by attempting to falsify it? (And this I could never succeed at doing because he DID write the Tractatus.) The important question is simply: Is it logically falsifiable? - in other words, is there a possible world in which Wittgenstein did not write the Tractatus? - and of course the answer is yes, and so as a proposition about the world, "W wrote the T" is fine; it passes the "it's meaningful" test. That's all falsifiability is. It has nothing to do with verifiability, or whether something is actually true or false.

Your idea of music being "meaningful" is to use the word in a different way. Music may indeed convey something to you, and you may call that "meaning," but it is a private experience, and is not related to the word "meaning" that I used above, which refers to the denotation and connotation of words and the truth value of propositions. Similarly, people may find that Freud's work conveys something to them (personally, I love psychoanalysis), but it can't be said to be meaningful in the sense of the term used above. Let me know whether this horribly long post has helped. ;-) SlimVirgin 23:14, Jan 28, 2005 (UTC)


 * Just to add this: I see you found a section from Popper saying psychoanalysis isn't necessarily meangingless (though I'd like to know what was said before that sentence). He's simply advising against throwing the baby out with the bathwater, which I agree with, and probably doesn't define how he's using "meaningful." Read Gellner for a good attack on psychoanalysis as a closed system: The Freudian Movement: The Cunning of Unreason. SlimVirgin 00:14, Jan 29, 2005 (UTC)


 * Let's agree this is a very interesting discussion. So: no, I don't think your contribution too long. I might be introducing some subtitles at some point, just for readability, not because I think this talk should be cut up in any way.
 * I'm a Dutch-speaking Belgian (so I'm not Dutch if referring to nationality).
 * I'd like to know what was said before that sentence: the hyperlink to the full text (which doesn't take more than 10 minutes to read) was given above: http://www.stephenjaygould.org/ctrl/popper_falsification.html
 * Really, I'd like to stick to Popper's terminology when talking about falsification (otherwise, I think, we'll continue to use terms in a different meaning, more than when I would start to write in Dutch). So I give another quote of that same text hyperlinked two lines above. This is the last paragraph of that text by Popper:
 * Thus the problem which I tried to solve by proposing the criterion of falsifiability was neither a problem of meaningfulness or significance, nor a problem of truth or acceptability. It was the problem of drawing a line (as well as this can be done) between the statements, or systems of statements, of the empirical sciences, and all other statements -- whether they are of a religious or of a metaphysical character, or simply pseudo-scientific. Years later -- it must have been in 1928 or 1929 -- I called this first problem of mine the "problem of demarcation." The criterion of falsifiability is a solution to this problem of demarcation, for it says that statements or systems of statements, in order to be ranked as scientific, must be capable of conflicting with possible, or conceivable, observations.
 * So, I want to draw your attention to:
 * Falsification (in Popper's original vision) is not about truth vs. non-truth/lie; it is only about science vs. non-science (where "scientific" is in no way a synonym to "meaningful").
 * Above, I linked this discussion with Village_pump_%28miscellaneous%29, where wikipedia community appeared to be adamant that also Wikipedia is not about truth vs. non-truth/lie. In this respect I consider "Popper's original falsifiability concept" fully compatible with wikipedia.
 * Wikipedia is completely about meaningfullness, as far as I can see in the same sense Popper used the word. From which I conclude Falsifiability can not be used as an inclusion criterion for Wikipedia (while meaningfullness is the accepted inclusion criterion).
 * How much falsification technique is used in wikipedia (on top of "verification") has, in my view & IMHO, a direct relation to how high scientific level wikipedia aspires to acquire. That is the real topic I wanted to come down to: how do wikipedians see their asprirations (or not) to reaching (any) scientific level? That's what I'm interested in (of course, for myself, I think you might understand I think many articles might benefit from being more scientific in the Popperian sense, not in the Vienna Circle scientific method sense; although probably for most of the wikipedia articles there is no need to go any further than verifiabilism, I suppose - but maybe I already said too much, I'd really like to hear what wikipedia community thinks on this issue).
 * Anyway, for myself I use falsification technique more than often, and I'd like to give an example (which is of course falsification light, while not speaking of the big eternal laws that govern the world):
 * Example: When writing about Erik Satie I discovered he was not the first to use "funny" titles for short piano pieces. About half a century before him, also in Paris, Rossini had done likewise. So I developed the idea ("theory") that Satie might have been influenced by Rossini in this respect (there were some striking similarities). Apart from lists of the respective compositions, and descriptions of them, each in their own right, I was not able to find any author making any link between Rossini's and Satie's compositions (but then, maybe I had only 0.1 % of all literature available regarding Satie). Still I wanted to falsify my little theory, if possible. So I contacted the Archives Erik Satie in Paris, whose director, Ornella Volta, is the authority on Satie. In a "verifiability" logic I could have contacted any of an abundance of "Rossini clubs" flourishing around the world. I feared they might've been too glad to confirm Rossini as a precursor to Satie. Ornella Volta, on the other hand, gave me what she knew: that Satie had stopped to use funny titles for short pieces around the time his friends of the Ballets Russes had re-discovered Rossini's "péchés de vieillesse" in Italy (where they had moved after Rossini's death, without being published before that date). The information I had been looking for appeared not to have been published in any Satie or Rossini biography but in books on the Ballets Russes and Respighi's Boutique Fantasque (which was the Ballets Russes production derived from the Rossini "funny" pieces).
 * So for me falsification light is (amongst other things) about going to look for information where it is most likely an idea is contradicted. That's the kind of "reliability enhancing" approach I'm looking for.
 * You're right in Popper not using the examples of "art" or "music" for non-scientific meaningfull stuff, I suppose because he was only speaking about "theories". I could've used "myths" as an example instead of art and music. Sorry, when using those examples I was just writing from memory, before reopening my old course and discovering the Popper text on internet. Apart from those two words I think I was fairly correct in my recollection of the issue.
 * --Francis Schonken 11:11, 29 Jan 2005 (UTC)

A new example
A few days ago someone added the following to Gymnopédies in the opening paragraph of the Erik Satie article:
 * The term is borrowed form the ancient Spartan ceremonial dances by naked boys, and is a not-so-subtle allusion to his own homosexuality.

Well, sure, I knew we would have to tacle the topic of Satie's sexuality one day or another. I didn't get involved in this subject here in wikipedia before today. I mention it here, before getting involved (which I'll do in a minute or two), because *maybe* (I'm not sure yet) this might clarify some more about "verification" compared to "falsification" approach: Now, that's my suppositions. Let's see whether it works (I mean, I'm going to try the second approach).
 * Applying mere "verifiability" would probably lead to amply listing references "pro" and "contra" Satie's homosexuality, with a struggle over who has the more "valid" references. Would take a long stretch, I suppose, before anything near to a NPOV treatment of the topic of Satie's sexuality would end up in the article.
 * I suppose a "falsifiability" approach would go quicker to reach consensus over a NPOV statement on this tacky issue, in the sense of a short list of the most valid arguments pro and contra, with references for both type of arguments, avoiding to state any of the arguments in an "unfalsifiable" format. ("...not-so-subtle allusion to his own homosexuality" is of course a statement in a kind of non-falsifiable format).

--Francis Schonken 13:34, 29 Jan 2005 (UTC)

I see no problem with the phrasing. The phrase "is a not-so-subtle allusion to his own homosexuality" either has a source that can be cited, or it doesn't. The Verification policy is a way of compelling whoever wrote this to provide a source.

The effect that we want is for people to cite sources. The verifiability policy is one way of ensuring this effect. I do not see how a "falsifiability" criteria is in any way helpful. Not only is it unnecessary, it also has the potential to confuse too many people, since you are using the term in such an ideosynratic way. Slrubenstein 18:16, 29 Jan 2005 (UTC)


 * I just realized that I gave an example above of a contradiction and called it a necessary truth. I've corrected it. I hope you'll both be gentlemen and won't go and look. ;-( SlimVirgin 18:55, Feb 1, 2005 (UTC)


 * I'm fully supportive of cite your sources, always was, sorry if I didn't stress that enough. Recently I've added more source references to Erik Satie, Gymnopaedia and Gymnopédie, up to the point that for the Gymnopaedia article a comment was given that it was out of balance while giving more source references than other text (see Featured article candidates/Gymnopaedia/archive1 - comment of 15:04, 4 Feb 2005). So I suppose wikipedians in general experience an "upper limit" to citing sources too, which might be cast in a recommendation like: the sources list should not become longer than the body of the article. For me, I don't want to be limited by such recommendation. As far as I'm concerned the sources list can be trice the length of the article, as long as the references are relevant and interesting.
 * But my point remains: verifiability is probably more than enough in standard wikipedia practice (and yes, it would be great if more wikipedians saw the value of citing sources), nonetheless verifiability is not enough when (and if!) aiming at a higher scientific standard. The if I mentioned between brackets in the previous sentence still has to be determined, and as said, that was my foremost reason for bringing up the falsifiability issue.
 * Returning to the Satie example:
 * Satie was gay
 * is a non-falsifiable statement, while with the current state of publications on Satie, Satie's homosexuality can only be demonstrated when based on a Freudian type of analysis, not on known and published facts about Satie.
 * Making the statement:
 * Verifiably, Satie was gay
 * only further obfuscates the topic, from a scientific/historical point of view, while indeed it can be verified in published books that Satie was gay. Here is the reference:
 * Larivière, Michel, Homosexuels et bisexuels célèbres: le dictionnaire, preface by Pierre Bergé, drawings by Jean Cocteau - Paris, Delétraz, 1997. - 393 p., ill., 22 cm. - ISBN 2-911110-19-6
 * Note that the "Verifiably, Satie was gay" statement has *exactly the same* falsifiability flaw as "Two plus two equals four". The proposition is necessarily true, while it is alwas possible to find Freudian confirmation/verification for such statement.
 * So for treating the (full) non-falsifiable statement introduced by someone in the Erik Satie article:
 * The term [gymnopédie] is borrowed form the ancient Spartan ceremonial dances by naked boys, and is a not-so-subtle allusion to [Satie's] own homosexuality
 * I approached this, in the gymnopédie article, with an attempt at falsification that Satie knew about the "naked boys". Of course the falsification attempt was extensively referenced in wikipedia style (published texts by Lewis, Smith, Fertiault and Höjer for that matter), I never said falsification would even be conceivable without citing sources. So, that in the end I could rephrase the original non-falsifiable statement in a falsifiable format, as I put it in the top section of the gymnopédie article:
 * As the term gymnopaedia is borrowed from [...] ancient Spartan ceremonial dances by naked boys, this is seen by some as a not-so-subtle allusion to Satie's supposed homosexuality. The present article also goes in to the question whether that is necessarily so.
 * This rephrasing, and the rest of the article, boils down to following more general falsifiable statement:
 * Up to February 2005 it can not be proven Satie ever was gay
 * (which, maybe disappointingly, reduces the whole thing to much ado about nothing - the only reason for keeping this in wikipedia is IMHO that it gives an interesting answer to a question that keeps popping up, where the "19th century perception of antiquity" is a relevant contribution)
 * I don't know whether this additional example, that I further elaborated here, was able to clarify more comprehensively what I intended by falsification (in a Popperian sense!)
 * --Francis Schonken 11:55, 6 Feb 2005 (UTC)

This discussion is going nowhere. First of all, what you claim is a "verifiable" sentence completely misses the point and violates our verifiabilty policy.
 * Verifiably, Satie was gay

You think "verifiability" is a feature of a proposition. You are coming at it from the perspective of logic. This is inappropriate as our policy is not addressing logical propositions. The above statement is not in accordance with our verifiability policy. If I were editing this article, I would delete it -- but not because if is not falsifiable, I would delete it because it violates our verifiability policy.

Similarly, "falsifiability" in the Popperian sense applies to experimental hypotheses, not encyclopedia articles -- period. Even your best suggestion does not call for anything at all like "falsifiability," for two reasons: first, insofar as there is merit in your suggestion, it is already covered by our Neutral point of view. Second, your proposal as a whole violates our No original research policy. To be specific,
 * As the term gymnopaedia'' is borrowed from [...] ancient Spartan ceremonial dances by naked boys, this is seen by some as a not-so-subtle allusion to Satie's supposed homosexuality.

Okay, the use of the passive voice is atrocious, but the basic point "some people see this ..." is exactly what is called for by our NPOV policy. All you need to do now is say which people, and provide a citation. ( Verifiability and Cite sources policy -- note, you should see now why it is that you do not even understand our "verifiability" policy, because even here you have violated it. Verifiability is not the opposite of falisifcation, it mean that you have to have a cirable source when you make a claim like "Some people say...")
 * The present article also goes in to the question whether that is necessarily so.''

This violates our policy against original research. It is not appropriate for you to write an article "going into the question of whether this is necessarily so." The only appropriate follolw-up to your first sentence is "Others, however, say ..." again specifying which others, and providing citations -- thus complying with our verifiability policy. Do you get it, finally? Now do you understand what verification means at Wikipedia? As long as you try to twist it into a Popperian framework, you will misconstrue our policy and probably end up violating it.

I am not going to address this issue again. I know that you are acing in good faith and trying to contribute, but as long as you keep trying to redefine our words in a Popperian framework, you are simply wasting our time. Read our policies and figure out what they actually do mean, not what you want them to mean. Slrubenstein 15:01, 6 Feb 2005 (UTC)


 * Easy, easy, I start to understand some things (sorry if appearing slow-minded). Yes, "verifiability" is not the same in a Popperian framework and Wikipedia framework, you pointing at logic for the Popperian meaning finally lit my candle. I defend that words used in the "wikipedia:" namespace have as much as possible the same meaning as outside that namespace. Verifiability forks when starting to use Popperian terminology. I understand "cite your sources". I would prefer a more suitable word than "verifiability" to express that understanding, but there appears to be none.

I am glad! Slrubenstein


 * Sorry for using the atrocious passive voice.

It is a common problem -- but very important to avoid Slrubenstein 17:44, 7 Feb 2005 (UTC)


 * Regarding the sources for:
 * As the term gymnopaedia is borrowed from ancient Spartan ceremonial dances by naked boys, some see this as a not-so-subtle allusion to Satie's supposed homosexuality.
 * What I understood from User:Haiduc (see talk:Erik Satie) is that the first part of this sentence were his own musings, on the homosexuality claim provided by Larivière.

If this is indeed the case, there is only one thing to do: delete the sentence. Wikipedia is simply not the place for any contributor/editor to present his or her own "musings." Period. To be fair to Haiduc, you should first ask him/her to provide a citation for an appropriate source (this request is exactly what we mean by "verification"), and remind him/her about the "no original research" rule. If s/he does not respond or resists, delete it. Slrubenstein 17:44, 7 Feb 2005 (UTC)


 * Yes, this is about NPOV. And I use some sort of falsifiability (there too I'm at a loss to find a better word) because it helps me to create better NPOV.
 * Yes, Haiduc failed to provide a proper reference for his insert (which was originally in the Erik Satie article). Throwing that insert out because of that and without any negotiation is/was an option. Yes, I thought conflict avoidance (which is wikipedia policy too!) a better approach.

One cannot and should not avoid conflict. What one must avoid is unnecessary incivility (i.e. we have to resist the urge to call someone a moron sometimes), and we must be willing to listen to someone else's reasons, and invite them to improve their work to comply with our policies. Some conflicts -- really, involved disagreements between users -- actually play out in civil tones and results in better mutual understanding and an improvement to the article. This is the goal. But we must always strive to follow our policies. Slrubenstein 17:44, 7 Feb 2005 (UTC)
 * No, the section about "19th century perceptions of antiquity", is not in conflict with "no original research".
 * Yes,
 * The present article also goes in to the question whether that is necessarily so.
 * is in accordance with no original research too, while it is addressing the controversy without taking sides.
 * --Francis Schonken 19:09, 6 Feb 2005 (UTC)

I am not sure I understand you. My problem was not with the actual phrasing of this clause, but rather what I thought (maybe mistakenly) it suggests about the rest of the article. But what I meant was, this would violate our NPOV policiy if the "questioning" were on the part of the editor of the article. I infered this from your Popperian stance, but if I misunderstood you I apologize. It is not the job of us editors to question any of the sources used in an article (aside from asking if they are reputable sources being used appropriately -- i.e. we question the origin and use of the source, not the claims the source is making). What we can and should do is look for other sources that represent other views. So if you meant by "goes on to question" is that you have added to the article alternate views from other sources that you can cite, well -- that's okay!

I am sorry I was so blunt before, I am glad we are reaching some mutual understanding, Slrubenstein 17:44, 7 Feb 2005 (UTC)


 * Before starting a next section (you'll see it'll get even more interesting), I just still wanted to give the reference regarding the "conflict avoidance" I mentioned above: this goes back to conflict resolution, which states effectively, it is better to "avoid" a conflict before it even started, and use talk pages as a first resort to avoid a conflict getting out of hand: regarding the "example" of this section, this was extensively done on talk:Erik Satie and some other talk pages. --Francis Schonken 11:55, 8 Feb 2005 (UTC)

confirmability
Regarding the ambiguity perceived around terms like verifiability and falsifiability, as described above, I'd like to point out Jimbo used the term confirmability on a page linked from No_original_research, in a mailing list topic on Crackpot articles:
 * [...] *individualized* crackpot ideas, [...] should [...] be deleted, not for being false, but for failing the test of confirmability.

(see link above for full quote, hope I didn't abbreviate it too much)

Well, I like this one. IMHO, in wikipedia context, confirmability sounds better than either verifiability or falsifiability, and doesn't need much explanation, and probably links better to the cite your sources idea (my, whatsoever subjective, feeling is that verification is closer to original research than confirmation, but then, I'm no native English speaker).

As for a "definition" of confirmability, however, I was thus far only successful in finding one in a "scientific" context:
 * http://www.gifted.uconn.edu/siegle/research/Qualitative/trust.htm#confirmability

Notwithstanding that "flaw" I propose it should be acceptable to use this term in wikipedia context (on talk pages etc...), and that maybe a page like confirmability or test of confirmability could be initiated.

--Francis Schonken 11:55, 8 Feb 2005 (UTC)

I don't think it is going to happen. The meaning of a word is determined by the speech community. I think most if not all of the members of this speech community (those who have made the most edits) all agree on what the word "verifiability" means. Why write a new policy when that policy is going to be identical to the one we have? I think your efforts are better spent going over the content of the Verifiability page to make sure that it clearly explains what we mean by verifiability. By the way, on avoiding conflict, I thought you were using "conflict" in the most general sense. If you were refering to our policy which addresses edit conflicts, then of course, the arguments should be made on the talk page, and we shoud try to avoid edit conflicts. Slrubenstein 17:28, 8 Feb 2005 (UTC)


 * Dear Slrubenstein,
 * I had a big laugh when reading ...most if not all of the members of this speech community [...] all agree on what the word "verifiability" means.
 * A bigger laugh when reading ...(those who have made the most edits)... (I'm one of "those", and I happen to think that "those" are often the most stubborn).
 * The naivity that speaks from your sentences is both charming & disarming, but that doesn't make them more to the point.
 * Reiterating an earlier point: thus far you (& SlimVirgin) failed to give a reference that "falsification" in the definition Popper gave of it has anything to do with the discipline of logic:
 * Logic
 * is about true/not true, an by extension about meaningful/not meaningful;
 * while falsification
 * is about scientific/not scientific, and, according to the reference by Popper I gave above, expressly NOT about true/not true, NEITHER about meaningful/not meaningful.
 * I don't know about (and don't feel responsible for) the confused mind that invented "falsification" had anything specific to do in the field of logic, but the least that can be said is that that would probably be a mind of lesser stature than Popper himself.
 * Nonetheless, if you're convinced Logic and falsification are related in the way you say, please come up with your references, of course otherwise I'll follow your good advice and remove the unreferenced stuff on this page to some archive or so (well, your actual advise was to erase it).
 * --Francis Schonken 07:48, 17 Feb 2005 (UTC)


 * Dear Francis Schonken -- I am glad I could give you a laugh. I may be misreading you, but I wonder if you felt I was patronizing you.  I wasn't.  I really did not think you were among the top editors -- I checked the top 1000 contributors and couldn't find your name on that list.  Actually, it seemed that compared to the top 1000 editors you have made relatively few edits.
 * Appears the statistics are unreliable, and not updated for over a month (2 months?): the last three months of 2004 I took a large excursion to Dutch Wikipedia, which topped me out of top 1000 (for English Wikipedia - the statistics don't add amounts of edits accross languages). The statistics don't show I'm fully back to English wikipedia for over a month now, still my over-all ranking according to that list is N° 1116 (which is still "top 3%", I believe, and competing with robots). See Full data in the csv for reference.
 * Well, anyway, it is a principle of mine to look down on people that bring the level of a wikipedia discussion down to counting numbers of edits. Generally this means someone trying to avoid discussion about the content of the issue at hand, and trying to "win" a discussion ("winning" is of no importance to me) by an argument of power. It also shows someone like that has some kind of disrespect for what is written on "wikipedia:"-namespace pages, for List of Wikipedians by number of edits is very clear that "number of edits" statistics can't be used in that way, while too approximative.
 * Anyway, the main reason I left Dutch wikipedia again (for the time being) is that that spirit of non-arguments was ruling too heavily there. To re-iterate something you said earlier: in an edit-war some Dutch wikipedian (in fact a "top contributor" of Dutch wikipedia) said "conflict resolution" guidelines did not apply, because it was an edit-war and not a conflict... Well all of that is stupidity in my eyes: a conflict is a conflict, often a conflict manifests itself in an edit-war, conflict avoidance is what you do before something becomes an edit-war (or any other form of "conflict"). "Conflict avoidance" is just good wikipedia practice, with or without the context of edit-wars, and it is in that vein the "conflict resolution" guideline is in all probability meant. So I think both you and the Dutch wikipedian I mentioned were completely missing the point.
 * So that makes me laugh at the unfalsifiable supposition that the main body of wikipedia contributors "all agree" about what the central policies and guidelines of wikipedia mean. That supposition is mythology in the full sense Popper meant it. That's why fine-tuning wikipedia guidelines (including the terminology used in these guidelines) is a difficult task that nonetheless should continue.
 * --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * But whether I missed your name or not, I was just trying to be helpful. You seem to be pushing an interpretation of the word verifiability that no one else to my knowledge shares.
 * I don't want to "push" anything. Popper's understanding of "verifiability" is something that exists, whether it is in your knowledge or not. And maybe that understanding can be useful for wikipedia. Or maybe not. Still trying to find out about that. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * My advice -- meant only to be constructive -- is to stop arguing about whether this is the right word, to focus on the content of the policy.
 * The terminology used (e.g. to avoid ambiguity) is extremely important w.r.t. the content. Generally, the better the terminology, the better the content. Counting user contributions, on the other hand, is barely of any use to improve guideline content. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * I still think this is constructive advice. I don't see what you mean by "naive"
 * Well, I think I explained that above. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * -- you yourself suggest that many editors here are stubborn, so you should have been the first to realize that your attempts to redefine the word would get no-where.
 * Ah, yes, but the "stubbornness" remark was directed at myself too. Thought you would've guessed that pun. Redefining words is however none of my business. But some of wikipedia guideline's troubles are undeniably that every now and then a wikipedia guideline redefines a term to a different meaning than the meaning(s) of that word in the non-virtual world. And as said above, sometimes that can not be avoided, but my stance is that is nonetheless better to avoid that wherever possible. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * Also, I never claimed falisifcation has to do with logic
 * Well, I quote what you wrote above:
 * You think "verifiability" is a feature of a proposition. You are coming at it from the perspective of logic. This is inappropriate as our policy is not addressing logical propositions.
 * Well I was never coming at it from the perspective of logic, I came at it from the perspective of Popperian falsification (and yes, Popper thought of "verifiability" vs. "falsifiability" as features of a proposition - but still in Popper's view nothing to do with the features "logic" defines about propositions), so the problem was in your mind indeed: you came down to saying that if features like "verifiability" or "falsifiability" are defined w.r.t a proposition, that one is in the domain of logic. Well that's a prejudice that seemed to be hampering the understanding of what Popper meant with falsification. That's a prejudice I wanted to do away with.
 * --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * (I did say that it has to do with experimental hypotheses, which is true.
 * The analogy I wanted to make is that when someone publishes something in a wiki system that can be edited by anyone else, this works in the same way as formulating a hypothesis. Popper did not exclusively write about "experimental" hypotheses, he wrote about hypotheses of any kind. Yes, he said something in the vein of: if you write a hypothesis down in a form that no "experiment" is thinkable that conceivably could have a different outcome than what the hypothesis predicts, than it is not a hypothesis (but mythology and the like). Now that is the way how we wikipedians work on a day by day basis. example: I think Vivaldi was a Roman Catholic priest, because I read it somewhere in a CD booklet (which is not such a reliable source always). I put that hypothesis in wikipedia, which is all that is needed to trigger the experiment - because there is maybe some clergyman in the Vatican that is a devout wikipedian, who goes to look it up in some dusty archive, and turns up with some evidence pro or contra. Or some other wikipedian has a biography about Vivaldi, assesses the capability of that biography to be a good reference for the wikipedia article, and adapts the article accordingly. Well, all of that is more scientific than looking at Vivaldi's picture and saying "he doesn't look like a priest, so he surely wasn't one" (which is a non-falsifiable approach). But even publishing that "unfalsifiable" hypothesis in Wikipedia, would be submitting it to experiment, i.e. the possible editing of other wikipedians: so wikipedia turns out to be a great tool to convert unfalsifiable hypotheses in falsifiable ones. Popper would be in seventh heaven. But the point I wanted to make is that there is nonetheless a good - probably even necessary - wikipedia guideline thinkable that discourages people to write "random" hypotheses in wikipedia, which somehow strain the system in a manner that is avoidable. "No original research", "Cite sources", and many more wikipedia guidelines work towards this goal in my view. But a good criterion to distinguish "quality" or "constructive" hypotheses from "random" hypotheses (and all degrees between those two extremes), is in my view the falsification criterion, as Popper expressed it (none of my invention): the more a hypothesis is formulated in a falsifiable format, the less other wikipedians would need to put their time in erasing such hypotheses as newbie vandalism, or crackpot theory, or put it through a VfD procedure, etc... --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * If you didn't know this, you better go back and study the scientific method more closely).
 * In what aspects Popper rewrote the scientific method about a century ago is maybe not yet so clear from the wikipedia article about the scientific method. Anyway, that's also one of my plans, to work on that article. This discussion here also contributes to that plan. It gives me much feeling about what are the sensitive points when writing about scientific method. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * You suggested using the word "confirmability" instead of "verifiability" and all I did was explain to you why that is not going to work.
 * Yes, I accept. Even if "confirmability" does not come from me, but from, lets say, an experienced wikipedian. The "why" however, I don't agree upon. Yes, I think too that the falsifiability discussion should split to another page. I'll do so whenever I've got some time for that. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * You keep changing the subject.
 * No, no, you do, from falsifiability to logic, and back; to counting contributions; etc... --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * And you know what? This is the talk page for the verifiability policy page.  It is not a talk page for a philosophy, popper, or scientific method page.
 * Yes, appears we have a bunch of spin-offs by now. I'll try to make work of these on the appropriate spots, whenever I can. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * I ask you sincerely: why do you keep trying to talk about philosophy? I ask in all seriousness, because I want to take you in good faith -- but by refusing to discuss the actual article, it seems that you only want to be obstructionist.
 * The only time I had encountered "verifiability" as a term, before getting involved in "wikipedia:" namespace, was in science philosophy course on Popper. And the connotation was "something to avoid" there (as opposed to its twin concept "falsification"). I tried to see clearer about that, for which I thought this page was the appropriate spot. I could never have started falsification properly without what I've learnt on this page. --Francis Schonken 14:45, 2 Mar 2005 (UTC)
 * I do not see anything that you have writtn on this page that involves improving the article. So far, you are just wasting time. Slrubenstein   |  Talk  16:35, 17 Feb 2005 (UTC)
 * Well, these last sentences are just unfalsifiable opinion, maybe better to ignore. I've no feeling I lost my time. Maybe you better speak for yourself. --Francis Schonken 14:45, 2 Mar 2005 (UTC)

Logic has more to do with validity than truth. I'm not sure what you're asking for. We're not proposing to insert into an article that falsfiiability is connected to logic, so I'm not clear why you want us to produce a reference. Sorry if I'm being dense, or perhaps I missed some of the discussion. I think our point was that a theory developed to describe scientific method is not applicable to the idea of verifiability as it's used here. I know what you're saying (I think): that editors shouldn't look for material that simply confirms their positions, but should be sure to include material that doesn't, and that's a good point to make on the NPOV page, but there's probably no need to introduce the idea of falsifiability, which would suggest we should actively hunt down material that contradicts material confirming a position. So in other words, if The Guardian reports that the former PM of Lebanon has been killed and that perhaps Syria was behind it, we shouldn't simply report that; instead we should actively hunt down sources showing that The Guardian report is wrong. But that raises two problems: (1) no editor is going to go to these lengths; and (2) we're not allowed to do original research and that comes dangerously close to it. I hope I haven't misunderstood what you were asking for. Best, SlimVirgin 08:41, Feb 17, 2005 (UTC)

Proposed amendment
After some discussion on Wikipedia talk:Confirm queried sources, I would propose the following to replace the "suggested procedure for verifying content". Thoughts? -- Beland 02:27, 11 Mar 2005 (UTC)

So you've found what looks like a dubious statement in a Wikipedia article. What should you do?

1.) Check existing documentation.


 * Check the bottom of the article for references.
 * Check the talk page.
 * Check other articles that the statement links to.
 * It would be embarrassing if you could have confirmed the fact you are disputing with a quick Google search, a peek in the dictionary, or similar trivial research.

If the existing documentation or references aren't clear, improve the article for the benefit of future fact checkers and readers. For more information on how to put citation footnotes in Wikipedia articles, see Cite sources.

2.) Notify others.

You have several options about how to do this. If you are short on time, it's more important that you let others know that you have found a problem than it is to pick the optimal method for doing so; just pick one of the actions listed below and do it. But if you can take a few moments, or you would like to improve your chances of getting along with other Wikipedians and making productive criticisms, you might want to consider the following.


 * Whether or not the article is controversial, and whether or not the change you are making is likely to stir up controversy. You might want to check the talk page and history to get a better idea.
 * Whether or not the majority of editors watching the article would agree with your input at first glance.
 * Whether or not the factual and personal motivation for the input is obvious.
 * Whether or not your input will cause the person(s) who put the statement there in the first place to disagree, or to be offended.
 * Whether or not you can cite a reliable source to justify your change.
 * Whether you want to make an incremental improvement (for example, making a dubious statement more accurate by making it more specific), have a specific claim that contradicts the existing text, or want to claim that a statement is false and should just be removed.
 * How certain you are that your claim is correct, or that the disputed claim is wrong. (Is it vaguely fishy, somewhat questionable, arguably wrong, or indisputably incorrect?)
 * Whether or not the original statement is documented, and whether or not the sources it references are reliable.
 * Whether or not you have any biases which may color your opinion.
 * Whether or not you can follow up in a few days.

Your options:

2A.) Make a note on the talk page. 2B.) Change the article itself.

It's recommended that you leave a note on the talk page if the rationale for your edit doesn't fit in the edit summary, or if it's likely that others will reply. The answers to the "Before you decide, consider.." questions above are also good talk page fodder.

If no one answers, it may be that the page is neglected, that participants haven't logged on to Wikipedia recently, or that no one has a substantive reply to offer.

It's always a good idea to cite specific parts of the article in any criticism. Which facts are of dubious accuracy? If you think something is biased, explain how.

Some situations clearly call for being bold and just fixing the articles (e.g. non-controversial improvements that you can cite a reliable source for); others clearly would benefit from being proposed on the talk page at least a day or so before being implemented (e.g. removal of fishy-sounding statements that you aren't too certain are actually wrong). There is a lot of gray area in between.

From one end of the spectrum to the other, you might:
 * Copy a statement to the talk page to ask for sources or note possible bias or inaccuracy.
 * Copy a statement to the talk page to ask for sources or note possible bias or inaccuracy, and in the article itself, tag it as dubious or disputed (see below).
 * Move a statement to the talk page and criticize it there.
 * Remove a statement without a note on the talk page. (Not recommended unless it's clearly a redundant statement.)

2C.) Tag the statement or page as dubious or disputed.

If you are reasonably certain that an article is factually inaccurate, biased, or both, you should tag it so that readers are aware of the question. See Cleanup resources for help with tags. Obviously, there's no need to put in such a tag if you fix the article yourself, and it is preferable to do so unless the changes should be discussed first. If you don't change the article, it's best to propose specific language on the talk page if you can, rather than make a vague suggestion that something should be improved. This will help show you are willing to help out and not just criticize, and will hopefully help move toward consensus and implementation more quickly.

If you do add such a tag, you'll almost certainly want to post a note on the talk page giving specifics.

2D.) Write a personal note to the author of a specific edit or the primary author of the page (if any).

You might do this to draw attention to a question you've asked on a talk page, to ask a question you think only this person will know the answer to (for example, to clarify a statement they made, or to ask if they were drawing on a specific source), to ask for help, or to make a comment that's not for public consumption.

Keep in mind that not everyone logs on to Wikipedia as often or as infrequently as you do, so you might not get an immediate response.

2E.) List the edit on RC patrol.

Only for recent changes. See the page itself for inclusion criteria.

2F.) Tag the page for cleanup. 2G.) Tag the page for expert attention. 2H.) Tag the page for merger with another page. 2H.) Tag the page for deletion.

See Cleanup resources for help with tags.

Keep in mind that tagging for deletion is a strong statement. You will definitely want to check the article's talk page and history. Consider whether or not there should eventually be an article with this name, even if it needs to be re-written from scratch. Especially if you are new to Wikipedia, read Votes for deletion and related policies to get a feel for site conventions on deletion. Be extra careful to explain your reasoning.

3.) Cope with the aftermath.

If you've edited the article directly and your change is immediately reverted, then it's time to stop and discuss the situation. Instead of re-implementing the change, tag the statement or page as disputed (so readers aren't misled), make a polite post to the talk page, and wait a while to see what happens.

Whether you are making an article change, posting a note to the talk page, or adding a tag, someone may ask you a question in reply. If you care about getting your proposal implemented, you may need to participate in the ensuing discussion. To facilitate this, you might want to add the page(s) involved to your watchlist when you make changes, nominations, or comments. Then you can just check your watchlist on subsequent visits to see if anything has happened.

(A related policy discussion is ongoing at Confirm queried sources.)

Requests for arbitration/RFC
An important policy discussion has started concerning ways in which our content-related polices, such as NPOV, No original research and Verifiability could be better enforced. I've made a proposal to give the Arbitration Committee the ability to consult Wikipedia users who are knowledgeable in  subject-areas that apply to cases before them. Such consultation is needed due to the fact that the ArbCom does not by itself have the requisite knowledge to easily tell what is NPOV, original research, or a fringe idea in every field. Please read my proposal at Requests for arbitration/RFC and comment. Thank you! --mav 02:50, 7 Jun 2005 (UTC)

Changes to obscure section
I've tightened this to remove a few weasel words (Some say, Other [s] say, etc.) and off-topic talk about what should be included. brenneman (t) (c) 02:28, 17 August 2005 (UTC)

Verifiability
I'm sure that the section noted above is good, but is it necessary? I'm for "shorter is better". brenneman (t) (c) 23:40, 25 August 2005 (UTC)

Mind-reading
Please forgive me if this is already addressed somewhere, but I'd like to see it stated somewhere that Wikipedia is not an exercise in mind-reading. I'm sure that for most people, this already goes without saying, but I've seen some articles that have included unsubstantiated assertions about the mental processes of broadly defined groups of people. Thanks. Doctor Whom 18:50, 12 September 2005 (UTC)
 * Just cite the combination of this policy and no original research. That locks it up pretty tight. If someone can't cite a prominent published opinion from a quality source, the speculation should be removed. - Taxman Talk 19:47, September 12, 2005 (UTC)

Hawking example
is the line about tape recording a conversation with Stephen Hawking meant to be a joke? if so it's in pretty poor taste. Nateji77 06:32, 15 September 2005 (UTC)
 * I'm not sure when exactly it was added(you can check the history if you want), but I agree, it is in poor taste, and it's not necessary. I've reworded it to make the physisist, like the theory, unspecified.  Thanks for bringing it up. JesseW, the juggling janitor 07:13, 15 September 2005 (UTC)
 * Because I'm crazy, I took the time to track down the insertion of the example. It was moved by Uncle G from WP:NOR, where it was added by Slrubenstein from the draft version of that page.  On the draft version it was (originally, as far as I can tell) added by SlimVirgin in this edit at 17:36, 8 December 2004.  Interesting.  And people claim you can't tell where the text on Wikipedia comes from.  Sigh. JesseW, the juggling janitor 07:40, 15 September 2005 (UTC)

Obscure topics
Hi Jesse, I removed the topic because it's potentially confusing. It says: "If an article covers a subject which has never been written about in published sources, or which has only been written about in sources of doubtful credibility, it is difficult to verify the information." Well, that means it must not be written about &mdash; so why are we even mentioning it? SlimVirgin (talk) 07:21, 29 September 2005 (UTC)
 * That makes sense. Let me try and rephrase it.  How 'bout this version?  I've slimmed it down, and made it more clear that subjects with no reputable, published sources should not be included in Wikipedia. The part I value is the description of obscure topics: "never been written about in published sources, or which have only been written about in sources of doubtful credibility" and the connection being made to NOR. JesseW, the juggling janitor 07:45, 29 September 2005 (UTC)


 * Looks good to me. &mdash; mark &#9998; 09:21, 29 September 2005 (UTC)


 * Me too. SlimVirgin (talk) 17:51, 30 September 2005 (UTC)

Addition of "sources should be in English" - strongly object
SlimVirgin added that sources should be in English. I strongly object to that change.


 * many sources are not available in English
 * the best sources are often not in English
 * it is better to use the original source where available.

This is a particularly strong way to introduce anglophone bias and would seriously reduce the value of Wikipedia.

At the same time it is not needed since non-English sources are clearly verifiable. It is possible to get a person who reads the language or, if needed a translator can be found to translate them.

I propose instead


 * Where available, sources in foreign languages should be backed up with references to English sources, e.g. translations of the non English source.

This should perhaps be moved to cite sources in any case since it should be a best practice rather than a policy. Mozzerati 22:15, 29 September 2005 (UTC)
 * I don't see why it shouldn't be a policy. This is the English Wikipedia; if you want articles (and sources) in another language, see the Wikipedia for that language. The original source is useless to people who cannot speak the language. A note about the original source could be included, but if possible there should also be an English version. I've changed the policy to include the original source, even if non-English... feel free to object/discuss here... ··gracefool |&#9786; 22:28, 29 September 2005 (UTC)


 * I, too strongly object to the idea of citing only English sources. I support Mozzerati's amendment, which makes more sense to me. Take my own field, African linguistics. Many good sources, especially on languages of countries that were under French colonial rule, are in French and in French only. I can see no good reason to avoid those sources in writing articles for the English Wikipedia. Imagine the next step: omitting subjects because only the French have written about it. &mdash; mark &#9998; 22:34, 29 September 2005 (UTC)
 * Mazerrati's amendment is exactly what I was hoping someone would mention in relation to this issue, for bias and quality reasons. The only thing I would add is that basically we do prefer references in the native language (English in this case) if they are high quality enough, but if not use the highest quality source available and also make sure to include English supporting references or translations if possible. A foreign language source of course places a higher burden on the person adding it to justify the source covers what they claim. I do see a potential problem in a dispute of someone claiming a Swedish source says X and that being hard to disprove. - Taxman Talk 23:14, 29 September 2005 (UTC)
 * Adding a requirement that sources must be in English is a fundamental change in policy and should be discussed on the village pump and at WP:BIAS not just here. Kappa 12:36, 30 September 2005 (UTC)
 * I agree that this section is something which ought to be moved to the WP:CITE section. When I add references (and I know that I am often a bit sloppy with that...) I will usually provide an English reference if possible, but sometimes such a reference is simply not available. I think it should be reworded to something like "Since this is the English Wikipedia, sources should be in English if possible". Sjakkalle (Check!)  14:24, 30 September 2005 (UTC)


 * Sjakkalle, I added "whenever possible," as you suggested. But I'm not sure what this sentence means: "If the original source is in a language other than English, it should be included, as well as a later/translated source, if possible." SlimVirgin (talk) 17:05, 30 September 2005 (UTC)


 * Mozzerati has asked me to discuss a "change to policy" here. Not sure what s/he means. That sources for the English-language Wikipedia should (whenever possible) be in English is not a change in policy. The whole point of NOR and Verifiability is that readers should be able to check that what we claim the sources say really is what the sources say. They can't do that if the sources are in Italian. SlimVirgin (talk) 17:08, 30 September 2005 (UTC)
 * I disagree: they can. It just takes a little more time. Which is why we prefer English sources. But if there are none, or if the best sources are non-English, we should certainly use them and cite them. &mdash; mark &#9998; 17:19, 30 September 2005 (UTC)


 * That's what it says now. SlimVirgin (talk) 17:41, 30 September 2005 (UTC)


 * Sources shouldn't use hard math. The whole point is that readers should be able to verify claims.  They can't do that if they're written in calculus  ( :-)  :-) :-) for the humor impaired).
 * more Wikipedia readers can read Italian than can understand tensors (and I have a survey to prove it, but unfortunately it's written in statistics :-).
 * now the difference between "sources MUST be in English" and "whenever possible an English version of a source SHOULD be provided" is immense and I can see almost no problem with the second formulation. The only things I can see are that it isn't the main point of this document and so shouldn't be the first secton and that I think that it's then a guideline and not a policy, and so probably should be moved to a guideline document such as cite sources.  Does that seem reasonable to you? Mozzerati 17:25, 30 September 2005 (UTC)


 * There's no reason it shouldn't be here. This page, NOR, and NPOV are our editorial policies, and this is an important editorial point. Also, editors do as a matter of fact already practise this as policy, not just as a guideline. If someone tries to submit a source that's only in, say, Hebrew, they're asked to look for an English-language source instead. SlimVirgin (talk) 17:41, 30 September 2005 (UTC)

I support putting a requirement for an english version of a source in this page, as a policy. If there are no english language sources for a claim, that claim should not be included into wikipedia until there are such sources. Those who want to include them should be strongly encouraged to go to the proper non-Wikipedia channels to get such translations or new works published that make such claims, then the claim can be included. Otherwise, the material cannot be verified by most Wikipedians. JesseW, the juggling janitor 21:44, 30 September 2005 (UTC)


 * I don't agree with this sentence: "Note that, as a last resort, foreign language sources are normally verifiable through translation and so can be acceptable even where no translation currently exists." The burden of proof regarding sources and verification is on the editor who makes the edit, not on other editors or readers. If it's reasonable to expect readers to arrange for a translation, then it's also reasonable to expect the original editor to provide one in the first place. SlimVirgin (talk) 22:07, 30 September 2005 (UTC)

Strongest possible objection. I believe there should be no policy, or even guideline on this. Whatever good effects it could have go without saying and would and do occur without it. Assume good faith works. It's like having a policy THAT WE SHOULD NOT WRITE IN ALL CAPS. Is it part of a project to increase systematic bias? - or one to insult those who may benefit the encyclopedia by their maybe hard - to -find knowledge? - something I thought was the opposite of the spirit of the encyclopedia. Frankly, it makes Wikipedia look ridiculous and amateurish in a bad sense. Slim, I think you are generally a sensible sort, but this idea is madness - and if you remember I opposed it when I pointed out that there (rightly) was no such policy. It is an attempt to cure a nonexistent problem, which is far worse than the problem could ever be. The only good thing is that it is a certainty that the "policy" will be ignored. As a matter of fact editors do not practice this policy. If people submit Hebrew or French or whatever only sources, people accept it. They might be asked for something in English, but if it hard to find, or doesn't exist, and is necessary to the article, as it is in many, many, many cases, they stay.

Slim, as you most certainly did not "make sure that changes you make to this policy really do reflect consensus, before you make them." - in fact when it came up before, in the old Israeli terrorism article, iirc the only person for it was you, and there were at least 2 against - me and one other (who might have been a troll though, I forget), and because it most emphatically is a "change in policy" - I am moving it entirely to the talk page. If I had seen it before I would have the minute you put it up. Here, I hope you notice that the comment is more negative than positive, especially for it as a policy. In its current state it the English is hard to understand and adds an even more absurd and unintelligibly longwinded requirement of originals and translations.

The I-hope-never-becomes-policy in any way, shape or form, IMHO completely useless and entirely destructive:

Sources should be in English

As this is the English Wikipedia, English language sources should be given whenever possible. If there is no original source in English, but only a translation or restatement in English, then the original source in the foreign languages must be given alongside the English source. If a non-English language source is translated into English for the purposes of a quotation, the original-language quote must be given alongside it, so that readers can check the accuracy of the translation. Note that, as a last resort, foreign language sources are normally verifiable through translation and so can be acceptable even where no translation currently exists.

John Z 22:17, 30 September 2005 (UTC)


 * John, first, there are more positive comments about it on this page than negative, particular for the edited version (that English-language sources should be given wherever possible, rather than must be given). Secondly, could you show me a few examples of edits backed up only by foreign-language sources? I ask for this, by the way, only because I'd like to review them, not because I intend to delete them. Third, I don't recall there being a discussion about this on Talk:Israeli terrorism, though I do recall editors asking for English-language sources. I would say there's unquestionably a consensus throughout Wikipedia that, in the English WP, sources should be in English wherever possible, and in the German WP, in German wherever possible, and so on. SlimVirgin (talk) 22:29, 30 September 2005 (UTC)


 * We do have a problem with unverified, and quite possibly unverifable, information in en.wikipedia. I don't see how requiring the sources need to be in the language readable by those reading the encyclopedia.  While it may be good to start a project to translate a lot of material that has not been translated - Wikipedia has a policy against original research - that sort of work should be done outside of this project and under more supervision and peer review than we can do here.  JesseW, the juggling janitor 22:50, 30 September 2005 (UTC)

Slim, there's a bit of discussion on that page, section 4, sources, the guy does not seem to be a troll there. I can't find my comments, perhaps they were on another page, maybe Zionist Terrorism. You were the only one who said that sources should be in English, and you yourself only say "I don't know whether there's an explicit policy (I'll check Wikipedia:Cite sources), but I'd say it's in the spirit of that guideline." - rather far from a belief in an "unquestionable consensus."


 * John, I've been editing for awhile now, and I can tell you there is a consensus without any doubt on the pages I have edited that sources should be in English, wherever possible. And I'm very big on asking for sources, so this is an issue I pay close attention to. Please don't delete the passage again. I'd say there's agreement on this page to have it in some form. The objection to start with is it said "must" not "should," and didn't say "whenever possible." It has now been qualified to include those terms. SlimVirgin <sup style="color:purple;">(talk) 00:27, 1 October 2005 (UTC)


 * I pointed to an example of a page you edited where there was no such consensus, and you were the only one favoring such a rule. When I find where I opposed you earlier, you will have another one.  By your own words you seem to have doubt yourself there! John Z 01:33, 1 October 2005 (UTC)

The unquestionable consensus is that common sense should be used, and common sense is quite enough in this case.


 * Sadly, it isn't. I've had editors try to produce foreign-language sources that they've translated to mean something quite different to what the sources said. If common sense were enough, we wouldn't need any of these policy pages to start with. SlimVirgin <sup style="color:purple;">(talk) 00:27, 1 October 2005 (UTC)


 * I saw this happen once, with Jayjg and somebody's French source a while back, it got resolved quickly no muss, no fuss, no need for a new policy, common sense being enough.John Z 01:33, 1 October 2005 (UTC)

Anything beyond it is entirely destructive, and makes Wikipedia look utterly parochial, and dedicated to having ridiculous unfollowable policies.


 * The hyperbole probably isn't helpful, John. The section reflects what is currently the case at Wikipedia, and it hasn't fallen apart yet. SlimVirgin <sup style="color:purple;">(talk)  00:27, 1 October 2005 (UTC)


 * In my opinion, it is not hyperbole I genuinely feel it is a ridiculous parochial unfollowable destructive (or unnecessary) policy; I think the hyperbole is in saying there is a problem for which this rule is needed to cure. I hope we each consider the other a reasonable person who however has gone completely insane on this point.   In my opinion, it does not reflect what is currently the case at Wikipedia..  The part about translations certainly doesn't. To the extent it does, it is because so much is not sourced at all.  In my opinion, this rule will due more harm than good, the fact that in your original form it had "must" is frankly incredible - and that JesseW still apparently believes something like it too is dumbfounding - there are plenty of things in the sciences where the English sources are inferior or utterly inadequate.  This policy would put a straitjacket on growth.  The only good thing is that it will be ignored.  John Z 01:33, 1 October 2005 (UTC)

What is unfollowable is the longwinded part about translations. Be serious. You really think anybody is going to do that? The problem on Wikipedia is not enough sources. Anything that keeps them off is to be discouraged. Again, it is for curing a non-existent problem - basically all the people who are putting in extraordinary claims in unrelated articles based on their documents published only in Tupi-guarani. For your question, following this policy could be fantastically destructive in the sciences. For example, Algebraic geometry is an integral part of modern mathematics, many would say central. Any algebraic geometer worth his salt would say that if you had to take all of the work on it and divide it into two piles, one in never translated French and one in English, then maybe the second pile would be larger nowadays, but if one had to cast one into the fire and erase it from human consciousness, there would be no question about keeping the French stuff. If the rule is watered down to the point of non-destructiveness, it is utterly meaningless, it is saying nothing more than common sense and should not be on the page. If it just says sources in English are preferable, the answer is "Well, duh."


 * It says what it says: English language sources should be used whenever possible. If you feel that is "well, duh," why not just leave it? There's quite a lot on this page of that ilk. Nevertheless, these points serve to emphasize certain things that, yes, should be a matter of common sense, but often aren't. SlimVirgin <sup style="color:purple;">(talk) 00:27, 1 October 2005 (UTC)


 * The part about translations should be eliminated; nobody will ever follow it.John Z 01:33, 1 October 2005 (UTC)

The way I count, there are only 3 people who support this policy - Gracefool, yourself, and now JesseW. Mozzerati, Mark Dingemanse, Taxman, Sjakkalle, Kappa and myself make comments with various degrees of negativity, often saying that it should not be policy, not on this page which requires consensus, maybe a guideline or best practice. To be blunt, I think that some of these people tried to water it down to something innocuous, and were not as bold as you were in creating this entirely new policy - in its original form, taken seriously, it could wreck Wikipedia. So there are only 3 who really want this policy, on the other hand I am unalterably against it in any form. 3-1 is not a consensus and therefore it does not belong on this page. Again, many of these 5 others have said they do not think it should be here, so my second count is extremely prejudicial toward inclusion, and it still fails.

I think you have seen me enough to know that I am levelheaded and very respectful of others. I cannot underscore enough how bad an idea this is. Before this rule had you ever seen anyone complain that there should be one ?


 * Yes, often. As I said above, requesting and checking sources is something I do a lot, and so I have seen the most ridiculous things be offered as sources, often untranslated, sometimes translated badly, and on a couple of occasions, translated to mean the opposite of what they actually said (I believe deliberately). I've also seen a request for English-language sources completely stop edit wars between, for example, Japanese and Chinese nationalists over some tiny point of little consequence. I see it as a good and sensible policy. How would the German Wikipedians feel if we all traipsed over there with our English-language sources? SlimVirgin <sup style="color:purple;">(talk) 00:27, 1 October 2005 (UTC)


 * Well, if that worked without this policy, why is it necessary? The only person who I have ever seen make this request is you. Again, imho it is either destructive or unnecessary.  If we invaded Germany that way, I hope that they would look at what the sources said and eliminate them if they were superfluous or replaceable, but keep them if not. I give an example of a peaceful Gernman invasion of English Wiki below - No complaints, no fuss.John Z 01:33, 1 October 2005 (UTC)

What is the point, except to insult people's intelligence with absurd rules? So I am doing something I have never done before on Wikipedia, and will revert it back. Also, Gracefool does not seem to have seen my comments here before reverting, as she asked for them when they were already there. Please think again - why on earth do we need such a policy? Where is the harm that it would prevent? - give me an example. I think it is your burden to show why there should be a new rule. Please don't claim consensus or that it is not a new rule. Look at the talk page - it looks like a new rule to many more than me. Look at the title of this section.John Z 00:10, 1 October 2005 (UTC)


 * John, could you give me some examples of the pages you mentioned earlier &mdash; where edits are supported only by non-English language sources? SlimVirgin <sup style="color:purple;">(talk) 00:27, 1 October 2005 (UTC)


 * Anything really serious on Algebraic geometry or Alexander Grothendieck would have to be. On the latter, David Madore just added something to the SGA page that mentioned how the unavailability of his books in any language, never translated, has made serious obstacles for researchers. In my opinion, the implied insult to non-native English speakers of this rule would greatly outweigh the mythical benefits. Others in the above discussion have mentioned cases where other languages are necessary.  For another, Nahum Goldmann has or had mainly German sources (I personally replaced some with English where I could find and be sure about the existence of a translation, some of which I have somewhere probably but can't find right now.)  It was based on the German version of the article.  This rule is an obstacle to translating clearly superior articles from another language - I would bet that that happens English --> other langs a lot more.  There is so much in English on everything, that that is the only reason that such a rule is conceivable - I doubt that any other Wiki has a similar rule.John Z 01:33, 1 October 2005 (UTC)

No matter how strongly you feel about this issue, Slim, Wikipedia rules require policy to be formed by consensus. You completely ignore the fact that many/most objectors, while trying to modify it into something reasonable, said or implied that it even if it is kept, it should not be on this page,. Do you have more than three people who definitely say it should be? Is 3 to 1 a consensus? I don't understand your counting. You are the only person who has explicitly said that this is not a new policy; many have disagreed. Regards John Z 01:33, 1 October 2005 (UTC)

It's been a while. There is nothing new in talk. I will not revert now, but I will tomorrow if there is no discussion. I have been asked not to, but equally well I can ask that new policies not be promulgated when contrary to policy, there was no pre-existing consensus, arguably, based on their own words, even in the opinion of the person who inserted the policy, and the discussion page hardly gives evidence that there is now. The argument that an edit war between Japanese and Chinese nationalists was stopped by a request for English sources is illogical. For an argument that this policy is necessary, what should have happened was the edit war was not stopped. This incident argues for "If it ain't broke, don't fix it."

In addition, in its present form it is partly contradictory with the end of "when adding information" which says all that should be said along these lines already.

Unless one is referring to a source in Linear A or another undeciphered language, this has nothing to do with verifiability, just ease of verifiability. One could equally well restrict sources to those available on the net, or in every public library in the US. In any case, even as a guideline "whenever possible" is far too strong. "Is preferable" is better - who decides what is possible - does somebody have to make an exhaustive search for alternative sources before they insert something? The overwhelming problem with wikipedia is no sources at all; compared to that the problems addressed are microscopic. John Z 04:17, 1 October 2005 (UTC)


 * The discussion is getting long and I don't have too much time on my hands right now. I just want to point out that I was mostly OK with the page as it was before John Z's objection and that SlimVirgin's changes adressed my point. &mdash; mark &#9998; 08:42, 1 October 2005 (UTC)


 * OK, one against me. But do you agree that this should be on this prominent page? Taken as a whole, do you feel that the stuff about translating sources together with the stuff in  "when adding information" makes sense and will be followed? John Z 09:20, 1 October 2005 (UTC)

John Z - let me see if I can summarize your specific objections to this version of the policy. Please correct me if I mis-state your views.


 * "English language sources should be given whenever possible." - you consider this obvious and not worth stating, especially in a policy, as it is just a general guideline.


 * I also stated this is much too strong. Who decides "whenever possible"?  The implied deprecation of non-English sources on this prominent page - I believe I have already seen examples of people saying they won't make edits because all they have are non-English sources - is going to damage the encyclopedia.  Putting it here is saying Wikipedia is happy to have a policy of having blinkers on.  At the most there should be some style guideline that English sources are  preferable.  This is so obvious it is hardly worth stating.  With so many sources on the web, and the possibility of however crude online machine translation, fraud from many languages' sources is a small and decreasing problem.John Z 00:30, 2 October 2005 (UTC)


 * "If there is no original source in English, but only a translation or restatement in English, then the original source in the foreign languages must be given alongside the English source." - I am not sure why you object to this, as it does not seem to add any more work to people adding sources; i.e. if they have a non-English source, they need to enter it, in the original language, rather than just entering a translation. Where's the extra work?  Reading this over, I am concerned that the sentence is not clear enough about the difference between a non-English source, and a English source that is based on another, non-English source; although I suppose this can be explained based on NOR - if the translation has been published, it's an English source, otherwise, it's not.  But this should be clarified.


 * Well, the way it is written, I found it incomprehensible and probably asking for bizarre behavior. Now that you explain it, I see that it is perhaps nothing more than what was written at the end of the first section, and so entirely superfluous. I still don't understand the words, and I have read and reread it. A person has something to say, in English of course.  He (tries to) back it up by a non-English language source, cite and page number. Should be end of story.  It is in any normal scholarly environment. What does "entering a non-English source" mean?   Is it talking about quotations?  What does "giving the original source in the foreign language alongside the English source" mean?  If this is talking about quotations, which doesn't seem to be the case as there is a second clause, then are we demanding that non-English sources be used  only for direct quotes - unlike English sources?  Is it talking about titles of sources?  Who restates titles?  Policies should be in intelligible English. Is it talking about translated works - who makes up phony translations?  If a work is translated in toto, refer to either the original or the translation, the way normal people do, not caring, unless there is some textual problem.  It is an irrational excess burden on users of translated works then. Suppose it is a translation of something in Chinese, (or Arabic or another not easily transliterated language), do we insist on Pinyin, Wade-Giles, or what for the original title?John Z 00:30, 2 October 2005 (UTC)


 * "If a non-English language source is translated into English for the purposes of a quotation, the original-language quote must be given alongside it, so that readers can check the accuracy of the translation." - slightly expands on and restates the previous sentence; these two sentences should be merged into one.
 * "Note that, as a last resort, foreign language sources are normally verifiable through translation and so can be acceptable even where no translation currently exists." - expands the caveat in the first sentence - I don't expect that you object to this.


 * Minor objection to the deprecation implied in "last resort." Of course the first is a best practice, but eliminating cited stuff from a reputable editor, say one who may have left the project, because it does not follow this "must" is not a good idea.John Z 00:30, 2 October 2005 (UTC)

So, that's how I break down your objections to this specific wording of the policy. You have also said that it is very bad, with great vehemence, but I did not see any specific objections other than the ones mentioned above. Again, please correct me if I left out or mis-stated your position. JesseW, the juggling janitor 22:33, 1 October 2005 (UTC)


 * I also object to the way it was proposed. There was no consensus,  when I had seen Slim propose it in its pernicious earlier form in various places, I do not recall seeing a single person supporting it. All I saw was opposition, not just from me.  I just don't think that many people - Slim in particular - understand just how bad that earlier form was, what enormous gaping holes there would be in the Encyclopedia if it were - and it of course never would be - followed --  and thus how bad anything that even remotely resembles it is, especially on this page, where it is entirely out of place.  The main effect will be to keep people from adding valid material to the encyclopedia, it will increase systematic bias,  and this will be entirely invisible.  It is a cure for a nonexistent problem.  It makes the encyclopedia look ridiculous.  I don't think "not really so bad", or "no particular objection" should be the criterion for making something policy.  I think it should be that people should definitely agree that it is a positive good. I see zero benefit from this bright idea, and a real possiblity of harm and abuse, however slight, from other people who are not as well-meaning and scrupulously sensible as Slim. I still point out that there are at least 2, maybe 3, other people who, based on their words above, do not want this on this page, which is my strongest objection.  I don't see anywhere near 12 people who definitely say this should be a policy here.John Z 00:30, 2 October 2005 (UTC)


 * You're managing to persuade me a little on this issue, particularly because of the example you left on my talk page. However, could you address two points: (1) Do you have any examples of edits that are currently supported by only non-English sources? And (2) what do you say to the issue of readers needing to be able to check the sources we use, which is the point of giving sources in the first place? We can't expect readers to commission their own translations of obscure academic books. And if we can expect it of the reader, we can also expect it of the editor who uses that source in the first place. I see the burden of proof for an edit as lying always with the editor: that is, he must produce the source if challenged, he has to convince others than the source is reputable, and he has to make sure it can be understood by others (either by producing an English-language source, or by arranging a translation). I've quoted from German sources a few times, but I've always translated the quotes into English, and then I've placed the German directly underneath so that people can easily check my translation. What in your view is wrong with asking editors to do that? SlimVirgin <sup style="color:purple;">(talk)  05:31, 2 October 2005 (UTC)


 * Well, no good examples I can think of this moment, but  It is more that it is a policy restricting growth that I oppose.  On the burden, I completely agree with you for the first two points, but not entirely the third  (If it were followed to the letter, writing about Hegel would be illegal)  There is nothing wrong in asking, even requiring editors to "make sure it can be understood by others", especially when there is genuine doubt about veracity, and it is not just obstructionism, but insisting on it ab initio is too restrictive, and I do not believe will or should happen.  Often people might just refer to a page number in a German book, and I don't think we should require more unless challenged, which should not be more automatic than with English.  Asking for a few sentences in German then machine translated would rule out much fraud, unless the guy is just making up the sentences.  With google.print and google scholar, checking total fraud with sources should be becoming easier, no?  I still don't understand exactly what is being proposed, and I guess I can't see an objection in the direct quote case - here the presence of the German would be more important than the English actually.  I can see people with reading knowledge of English but whose writing is so hard to understand the foreign language would be easier to understand.  There's a guy who is putting up a lot of material on Japanese and Chinese history I think machine translated from Spanish, and I think it would be easier to just leave it in Spanish and translate than read the machine stuff. John Z 15:18, 2 October 2005 (UTC)

I'm not sure what to make of this, but About-Picard_law provides some good examples for this discussion. I'm also beginning to lean more towards the anti-policy view. Although I still think there is some critical element here that should be expressed... JesseW, the juggling janitor 08:20, 2 October 2005 (UTC)


 * Me too. But what is it? ;-D SlimVirgin <sup style="color:purple;">(talk) 08:30, 2 October 2005 (UTC)

John, regarding your statement I saw this happen once, with Jayjg and somebody's French source a while back, it got resolved quickly no muss, no fuss, no need for a new policy, common sense being enough, as I recall that situation, the person in question insisted on using French sources, and also insisted on translations which disagreed with existing English language translations. I tried using on-line French translators, which were less than satisfactory, and also to get other French speakers involved to verify, with middling results. Overall I wasn't happy with the results, even if the matter was settled without much "fuss". Verifiability is difficult enough, but when you have only foreign language sources, it can be almost impossible. Jayjg <small style="color:darkgreen;">(talk) 08:43, 2 October 2005 (UTC)


 * Well, I have too much confidence in your and Slim's etc. abilities to intelligently use the existing policies to want to modify them. You guys should go to rogue admin school, then maybe I will change my mind. :-)  I'd have to look at the case again, I don't remember what happened.  If the translations (his own?) disagreed, then I think it is clear that without very strong support for his view (confirmation that he was a professor of French say, or pointing out an obvious one word mistake) that the existing translations should be accepted as correct and his as unacceptable (not reputable, OR).  From my knowledge of French, what he was saying seemed wrong IIRC -- and though I didn't say anything, I favored being tough in that case, tougher than you IIRC.


 * I feel that the burden should be on new policy proposals, particularly why it is necessary, just where would it prevent damage that just can't be prevented by rigorously, perhaps more rigorously than before, enforcing existing policies? - and then one should carefully looking hidden costs and undue burdens that don't assume good faith.  The sysop's view may be jaundiced, as they are constantly looking at the violators, rather than the harm (maybe minute in each particular case, maybe infrequent) to the good faith majority.


 * A major point is that there is a very real chance that we already use fewer non-English cites than Britannica, say. If the consensus here departs from existing scholarly consensus, the Wiki consensus may already be too strong and invisibly discouraging good material, and the policy seems unwise.  On the other hand, if we are using more, then I begin to support it as a guideline.  John Z 15:18, 2 October 2005 (UTC)


 * I am very glad that I have not totally destroyed the case with my crazed vehemence, that I have swayed people a little with my farragoes. I was beginning to think I may have been going nuts myself.  Regards to all.John Z 15:18, 2 October 2005 (UTC)


 * They were very intelligent farragoes, and therefore most welcome. ;-D SlimVirgin <sup style="color:purple;">(talk) 16:00, 2 October 2005 (UTC)

Just my little addition. Having sources in English as a requiement would be wrong. It should be encouraged to use English sources when available, but if it's not any other language would do. Having it as a requirement would introduce a systematic bias not to mention that not everything is available in English. // Liftarn
 * As some of us seem to have moved closer to each other's opinions, is the version currently on the page all right for everyone? It makes clear that English sources are preferable, but that sometimes non-English language sources will be used, and that where quotes are translated into English, the original-language version has to be supplied too so readers can check that the translation is correct. We should probably add something about the English needing to be acceptable, as I've seen translations offered where the English was completely unreadable. Any thoughts? SlimVirgin <sup style="color:purple;">(talk) 09:16, 3 October 2005 (UTC)

The heading of the section is misleading to the point of being downright wrong, and a lot of the ground covered above is not reflected in the text. I've completely rewritten the section. I've tried to keep the text neutral on how sources should be cited (which is a matter for style guidelines) concentrating solely upon what sources should be cited and why. I've tried to prevent any implication whatsoever that Wikipedia should be biased against non-English sources, which is an idea that is counter to our goals. And I've tried to include enough rationale to make it clear what the verifiability and accessibility issues actually are that underpin the requirements. Uncle G 13:34, 15 October 2005 (UTC)


 * I feel the new wording is dangerous as it almost seems to be encouraging creating derivative translations in breach of copyright. It also seems to have dropped the requirement to cite the original language source when citing recognized translations, even though it has already said that even those may be unreliable.  (I suspect wars have been fought over the translations of the Bible!)


 * For material that is not directly quoted it is not clear whether it rejects the use of the material entirely if there are no independent translations (one danger of this is that it will make articles about certain countries anti-governemnt because the government's own case is only in the native language), or whether it is encouraging the inclusion of copyright breaching translations --David Woolley 14:22, 20 November 2005 (UTC)

Dutch scholarship and English sources
I object to such strict guidelines when using sources in major European languages, like German and French. Citing sources from those languages is acceptable in Dutch scholarship without such strict guidelines. I do not see why Wikipedia should deviate from these generally accepted practices. Of course, it is a different matter when citing minor languages and I think that those strict guidelines do apply here. Andries 15:31, 15 October 2005 (UTC)
 * The sources we use, and especially any we quote, have to be accessible to the readers, and because this is the English Wikipedia, we assume they are mostly English speakers. We don't have the same readership as Dutch scholarship. It's not that we oppose non-English speaking sources, as Uncle G I think understood. It's just that we need what they say to be in the English language, or else the whole concept of verifiability by readers and other editors goes out the window. SlimVirgin <sup style="color:purple;">(talk) 15:50, 15 October 2005 (UTC)
 * Here in Wikipedia we try to copy scholary and encyclopedic standards so it is important what Dutch scholars do. Scholarly articles in Dutch have to be verifiable too and they are considered verifiable without such strict guidelines. So I do not see the difference. (I understand this is a clash of different scholarly traditions, though.) I request the addition of a statement that the "smaller" the language, the more is expected from the person using the non-English source. Andries 16:26, 15 October 2005 (UTC)
 * You have not actually stated what the difference is between Dutch scholarship and our policy. Uncle G 23:33, 15 October 2005 (UTC)
 * The difference between Dutch scholarship and the current Wikipedia policy is that in Dutch scholarship it is acceptable to cite untranslated sources from other major European languages. Andries 23:47, 15 October 2005 (UTC)


 * I'm confused about what you see as the relationship between Dutch scholarship and Wikipedia. And whatever languages are spoken in the Netherlands are what you're calling minor languages anyway. Could you clarify? SlimVirgin <sup style="color:purple;">(talk) 01:31, 16 October 2005 (UTC)

I'll state my thoughts in the style of a draft proposal, if I may, jguk 18:11, 15 October 2005 (UTC) :
 * On the English-language wikipedia, we have a strong preference for English-language sources. A problem arises where the information being added does not have an English-language source, or where the English-language source is significantly inferior to a foreign-language source. Where possible provide English-language sources, and where the English-language sources are significantly inferior to foreign-language sources, cite both.


 * Jguk, your position elsewhere is that we shouldn't adhere to academic standards, but should be accessible to ordinary readers. Are you going to be consistent? SlimVirgin <sup style="color:purple;">(talk) 01:31, 16 October 2005 (UTC)


 * Here is my counter-draft to user:Jguk. It is quite a difference if you use a source of a major language like German, or of a minor one, say a Slovak source. This is generally accepted in Dutch scholarship. Andries 19:40, 15 October 2005 (UTC)


 * But whatever is accepted in Dutch scholarship has nothing to do with us. SlimVirgin <sup style="color:purple;">(talk) 01:31, 16 October 2005 (UTC)


 * On the English-language wikipedia, we have a strong preference for English-language sources. A problem arises where the information being added does not have an English-language source, or where the English-language source is significantly inferior to other-language sources. Where possible provide English-language sources, and where the English-language sources are significantly inferior to other language sources, cite both. In general, major international non-English languages like German, French are preferred over minor ones. Andries 19:40, 15 October 2005 (UTC)

This discussion of languages that are "major" is the wrong direction to take this policy. Neither proposal encompasses all of the issues brought up in the above discussion, moreover. Uncle G 23:33, 15 October 2005 (UTC)


 * I'm not sure what the difference is between a major and a minor language. SlimVirgin <sup style="color:purple;">(talk) 01:31, 16 October 2005 (UTC)

I concur with Uncle G, and I think we are opening a can of worms here with the major/minor point. I agree with SlimVirgin that Wikipedia's readership is not the same as the readers of Dutch scholarship. &mdash; mark &#9998; 07:23, 16 October 2005 (UTC)