Wikipedia talk:Verifiability

RfC: Should the threshold for inclusion in Wikipedia be changed from verifiability, not truth to not just truth, but verifiability?

Template:RFCpolicy

This is the change being discussed. Should the threshold for inclusion in Wikipedia be changed from verifiability, not truth to not just truth, but verifiability? I must ask: Whose truth? I'm totally opposed to this drastic change, which would amount to a repudiation of our core principles. Dlabtot (talk) 19:04, 6 April 2008 (UTC)

• PLEASE do not change the wording of this RfC, or any of my other talk page comments. If you wish to file a different RfC after this one closes, fine. Dlabtot (talk) 20:29, 6 April 2008 (UTC)
• I ASK AGAIN do not change the wording of this RfC, or any of my other talk page comments. If you wish to add commentary, fine. Do not delete mine. Dlabtot (talk) 20:39, 6 April 2008 (UTC)

A note - a third option, "not just accuracy, but verifiability" has also been proposed, and actually has, at this point, seemingly won over everybody who advocated "not just truth, but verifiability." Helpful comments will address both phrasings. Phil Sandifer (talk) 20:25, 6 April 2008 (UTC)

RfC Responses

• I do not think that an average reader will look at that sentence and see a debate about Platonic truth and the absolute nature of truth. Rather, I think most people will read "truth" there as meaning "accuracy." Anybody who disputes that we want to include only accurate information in an encyclopedia should probably find another project to contribute to. Accuracy is a core principle of any encyclopedia. Note that the new wording does not promote or allow the "But it's true!" argument to insert unverifiable information - it merely requires that we strive for both. Doing otherwise is obviously unacceptable. Phil Sandifer (talk) 19:09, 6 April 2008 (UTC)

Significant viewpoints about notable topics, that are published in reliable sources, deserve inclusion, even if in the opinion of some, those significant, published views are "not true". Dlabtot (talk) 19:22, 6 April 2008 (UTC)

Sure. But a claim in a reliable source that is not true should not be reported as a fact. Significant but untrue (or questionably true) claims can be reported with careful attribution so as to preserve both truth and verifiability. To utterly demote truth, as the old language did, invites sloppy work. The challenge - and it is at times a hard one - is to preserve both truth and verifiability. In cases where the truth is contested we use NPOV to present everybody's claims. But no article should ever contain inaccurate information. Phil Sandifer (talk) 19:25, 6 April 2008 (UTC)

• "Verifiability, not truth" has been a core principle of WP for a long time, and has a wide consensus. Diluting this crucial point, on which much of WP's edits hinge, would require a very wide consensus also. Crum375 (talk) 19:26, 6 April 2008 (UTC)
  • This isn't really an objection, to my mind. If there's something wrong with the new phrasing then I agree, it shouldn't be changed, but the argument "I oppose the change because it doesn't have consensus" doesn't seem to me to have a place in a discussion about the change's merits. It is also worth noting, it is inaccurate to say this weakens the phrase. Verifiability is still a threshold for inclusion. It just eliminates the idea that we should also ignore truth - in other words, it raises the bar. Describing it as a weakening is unfair. Phil Sandifer (talk) 19:34, 6 April 2008 (UTC)
    • It may seem to "raise the bar", but its effect is to lower it. See my reply to Steve McCluskey just below. The bar will be lowered if we aim for truth — we are not credentialed professionals, and we can only be trusted to correctly convey what published sources say about a topic. This can be verified (hence "verifiability") by any reader. If we assign to ourselves truth finding, we'll then produce material of dubious quality, that is unverifiable by our readers, with no author credentials. That is a drastic lowering of the quality bar. Crum375 (talk) 19:47, 6 April 2008 (UTC)
      • Again, though, you're treating the statement as putting truth above verifiability - it's not. The statement still clearly says verifiability is a must. (And indeed still stresses it above truth... well, accuracy now, as I think that sidesteps the big-T truth issue - verifiability is the one in boldface, clearly the emphasized point.) Phil Sandifer (talk) 19:50, 6 April 2008 (UTC)
• I strongly agree with this change and feel it should be kept. The quotation "verifiability not truth" is one of those silly catchphrases that can be so easily misinterpreted and satirized. The substance of the change is really minor -- little more than an editorial tweak; I do not see it at all as diluting a core principle of Wikipedia. We are, after all, writing an encyclopedia that is meant to be reliable and trustworthy. --SteveMcCluskey (talk) 19:31, 6 April 2008 (UTC)

I've come around to the view that "verifiability not truth", although paradoxical, is right on target. Since writing the above I've used the phrase to answer two different advocates of fringe points of view who believed that they were advancing the truth and that the works cited in opposition to it were wrong. --SteveMcCluskey (talk) 13:09, 18 April 2008 (UTC)

An encyclopedia is a tertiary source, not a textbook. Our goal is to present to our audience, just like a good librarian at a library, only what reliable sources have to say about a topic. We should ensure the sources we use are of good quality and present them neutrally, but we should not strive to find an "ultimate truth" (which rarely if ever exists anyway), only to do a good job presenting the published sources. This is the essence of our mission here. Crum375 (talk) 19:37, 6 April 2008 (UTC)

I think that the use of the word "truth" to refer to "ultimate truth" is limited to specialized philosophical settings of which this policy is not a part. Few readers, if any, will look at the slogan and realize that we mean Platonic big-T truth. That said, what if we changed it to the original language of the page - not just accuracy, but verifiability? Phil Sandifer (talk) 19:46, 6 April 2008 (UTC)

The word "truth" is a philosophical concept, strongly related to "ultimate truth", and does not belong here. Our starting point is verifiability, which means everything we write must be attributable to a reliable published source, that readers should be able to verify on their own. Accuracy is important to ensure that what we say has in fact been published exactly that way, but the accuracy is secondary to verifiability and is part of the overall quality, just like good writing. Crum375 (talk) 19:55, 6 April 2008 (UTC)

Though the word "truth" is a philosophical concept, I do not think that, in common usage, its philosophical meaning is the primary one. Regardless, I think "accuracy" does sidestep this problem. I have to say, however, I think you are wrong to suggest that it is secondary to verifiability - have a look at the "history of the term" section below. The statement that Wikipedia should be accurate is absolutely a core principle, and a longer-standing demand than the current phrasing. Phil Sandifer (talk) 20:03, 6 April 2008 (UTC)

• It's obviously the case that accuracy is a core principle of, uh, writing an encyclopedia, and that verifiability is useful as a way of getting to the accuracy of a given statement of a viewpoint. "Truth" as used in "verifiability, not truth" is directed at people who want certain POVs marked "right" or "wrong" rather than "widely held" or "not widely held" or "scientific consensus" or whatever. "Accuracy, but verify" also states the actual point perfectly well - David Gerard (talk) 20:12, 6 April 2008 (UTC)
• Just registering a generalized discontent with both options being debated here, with a grudging preference for the former iteration "verifiability, not truth," as it seems to me to convey the idea that verifiability is or may be a characteristic of "truth," but that the "truth" of a matter is not something WP is necessarily trying to convey, beyond reported facts and evidence, as it were. The phrase, "not only truth, but verifiability" seems to me to subordinate the concept of truth to the function or action of "verifiability," or capacity to be verified. "Accuracy" in this case can only mean conformity to facts established elsewhere, a characteristic best ascribed to "verifiability," although if something is accurate (to something else) it is often taken to be "true" of whatever that thing is. Amerique^dialectics 21:38, 6 April 2008 (UTC)
• I liked the original wording; it was useful. In my experience, the people who say "but its true, why are you removing it?" are precisely the people who need this policy to keep them in line. Plus, it means my Wikimania t-shirt of a lisping Ronald Reagan saying "truth but verify" won't sell. --Relata refero (disp.) 10:09, 8 April 2008 (UTC)
  • Are there really, you think, contributors who say "but it's true, why are you removing it" who would not also understand a policy that says "not just truth, but verifiability?" In both cases the statement is clear that truth is not in and of itself sufficient. It's just that one statement makes the ridiculous claim that we are truth-neutral, whereas the other makes the (quite accurate) claim that we ensure our accuracy through verification. Phil Sandifer (talk) 14:24, 8 April 2008 (UTC)
    • I don't think its ridiculous to imply that we should be "truth-neutral". I do think we shouldn't be here to put the truth on Wikipedia, but represent what other people think is the truth. (Hence my username.) Here for example is the sort of person who is drawn to Wikipedia who needs to be reminded that what's true doesn't really matter, because tertiary sources don't tell the truth about the world, only about secondary sources. --Relata refero (disp.) 13:33, 9 April 2008 (UTC)
• I strongly support the change. I worry that there are editors -- I believe that I have met some -- who justify the inclusion of some highly dubious material on the grounds that it has been published, and is therefore ipso facto "reliable" by virtue of being verifiable. I think that is is a particularly dangerous trend in BLP articles. I also question the inclusion of "mainstream newspapers" in the category of "most reliable" sources for the purposes of this policy statement. After seeing the "mainstream newspapers" in the U.S. and U.K. act as a megaphone for clearly bogus claims used to justify the Iraq war, I think that their claim to be "reliable sources" should be treated with some skepticism. --Terrawatt (talk) 06:51, 9 April 2008 (UTC)
• I support "not just truth, but verifiability." Philosophical arguments about the nature of "truth" are outside the scope of the proposal -- the proposed change is simply an attempt to bring our language into alignment with our intent. The proposed change is not about the nature of "truth", it's about correcting our current words, which say we want "not truth". Thirdbeach (talk) 18:23, 9 April 2008 (UTC)
• ...on second thought I'm still in favor of getting rid of "not truth" but favor the "accuracy" and "threshold" alternatives suggested elsewhere. Those alternate wordings get us past the whole set of concerns about the nature of truth. Thirdbeach (talk) 18:46, 10 April 2008 (UTC)
• I think it is fine just the way it is and has been since 2003, "The threshold for inclusion in Wikipedia is verifiability, not truth." _(1 == 2)^Until 18:57, 10 April 2008 (UTC)
  • Which 2003 version of the policy that says "not truth," exactly? (Clue - it wasn't any of the ones that said "Wikipedia strives to be accurate.") Phil Sandifer (talk) 19:00, 10 April 2008 (UTC)
    • Sorry, good catch. I was mislead by a comment lower on the page. It seem that it was indeed this edit in 2005 that introduced the concept of verifiability, not truth. My position remains the same. _(1 == 2)^Until 20:26, 10 April 2008 (UTC)

Original source of VNT: Thanks to Phil Sandifer and Until(1 == 2) I can add detail to Phil's observation (buried below) that the original coiner of "verifiability, not truth" (and the text inserted in 2005) was SlimVirgin on 8 Dec 2004 with this edit to a draft of NOR ("Wikipedia is about what is verifiable, not what is true"). The edit is explained here ("I may have labored the verifiability/truth distinction with my long example, but I did this because, time and again, editors don't seem to understand that their firm belief in something being true is not a reason to have it in Wikipedia") and again here ("They don't get that they're not allowed to insert something just because they personally know it to be true"). Slim's coinage based directly on Jimmy Wales's statement of 3 Dec which currently appears at footnote 1 of Wikipedia:NPOV, V and OR. JJB 14:35, 23 April 2008 (UTC)

• I support the change, or the weaker phrasing "not just accuracy, but verifiability". — Carl (CBM · talk) 20:48, 10 April 2008 (UTC)
• Oppose. The proposal is to change from something we can legitimately promise -- that our product reflects the major opinions of reliable sources -- to something that we cannot promise -- that our product reflects the truth. Honest doctors and honest stockbrokers have the humility not to promise health or wealth. They promise only to do what is possible for doctors and stockbrokers to perform, and know that actual results depend on factors beyond their capacity to control and cannot be guaranteed. If we are willing to be similarly honest and similarly humble, we will recognize that while verification lies within our capacity to guarantee, truth does not. Best, --Shirahadasha (talk) 21:06, 10 April 2008 (UTC)
• Strongly oppose "Not just truth ..." is a complete reversal of the actual policy. If anything, it should be, "Not truth, nor accuracy, nor facts, nor correctness, but verifiability." To propose anything other than verifiability of notable points of view from reliable sources as a criteria for inclusion (e.g. that it is true, or accurate, or factually correct) defeats the whole purpose of this policy and NPOV which is that different editors have different views of what is true (or factually correct, or accurate, etc.). If "truth" or "accuracy" or "Correctness" were the criteria, we would have at least two articles on Jesus, and both of them would violate NPOV. "Verifiability, not truth" is the principle that givs NPOV its force. Slrubenstein | Talk 21:19, 10 April 2008 (UTC)

We have even more than that.. Besides Jesus, we have: Biblical Jesus, Religious perspectives on Jesus, encompassing separate subarticles on Christian views of Jesus, Jesus in Islam etc. etc. ; Historical Jesus, Nontrinitariansism, Jesus myth hypothesis, Jesus and comparative mythology ; not to mention Cultural depictions of Jesus . Genealogy of Jesus and about 30 other special topics and aspects. We do similar in other subjects, usually using slightly different titles to disguise the fact that we actually are accommodating a POV split. Personally, I feel that we ought to make it explicitly permitted. As now, each article would still be required to be fair and to at least refer to disagreements--but the full exposition of the other sides could be in other articles,if there is enough material. We already have an article Positions on Jerusalem with sections on Israeli Position, etc. No reason why those sections shouldn't be separate cross referenced articles. DGG (talk) 01:17, 12 April 2008 (UTC)

• Strong support for either new version "not just {truth|accuracy}, but verifiability". Other terms that would work are "source accuracy" or "faithfulness to source". Have felt for many moons that the "accepted" version was mangled and stunted, even after wrapping my mind around its accompanying rationalizations. Have skimmed Phil and am empathetic to most of what he says. One clear reason is that it breaks WP:LEAD: "truth" is not mentioned in the policy (except by Jimbo!), but "accuracy" is! The ease of satire is another negative. And no matter what people think truth is, it is not hard to define accuracy or verifiability; better to not even raise the issue. Precluding the "But it's true!" inclusionist objection happens in both versions, but the old is a poor compromise, while the new is sufficient but does not have the weaknesses of the old. JJB 23:08, 13 April 2008 (UTC)

• If the point is that verifiable sources should be accurately represented, I don't think anyone would disagree - but this demand is entirely consistent with "Not truth, but verifiability" Slrubenstein | Talk 23:21, 13 April 2008 (UTC)
  • Thank you, but even in your view, the language should still be changed (due to ambiguity) to something like "not all that is true, but all that is verifiable", which also resolves some of the wording problems-- though I favor the proposal version. I was going to add, philosophically I see this particular debate as a battle between camps called perhaps "mainstream" and "alternative", or less charitably the "guild" and the "fringe", each of which fears the undue influence of the other. The socks object to the cabal and vice versa. This debate, like its close relative deletionism vs. inclusionism, is not likely to be solved easily by consensus between the two. The present language favors the mainstream too much, and definitely needs change; the proposed change counterweights toward the alternative stream (but perhaps too much), and is open to variations; but at any rate the language should not favor either camp. Both camps should rally around the fact that if all assume good faith, pursue improvement, and do not cheat, the project will ultimately succeed, whichever version is favored here. That is where to address the fears that either side will abuse verifiability for their own POV ends. In this kind of wording debate we can only hope not to throw licenses for anyone to abuse either inclusion or deletion. JJB 00:57, 14 April 2008 (UTC)

• In my view, there are two kinds of encyclopedias, and people are free to work on which ever one suits their fancy. One kind is the conventional encyclopedia with a editorial board that establishes and enforces certain criteria for quality and, to the point, veracity; if you care about truth, work for an enecylopedia that cares about truth and submit yourself to their editorial control. The other Kind is wikipedia, in which anyone can edit at ay time and there is no editorial control and no standards for veracity or quality established or enforced by an editorial board. Every editor has their own ideas about truth, veracity, accuracy. NPOV is the sine qua non that enables all these people to work together. It is because we reject truth as a criteria that we need the NPOV policy, and it is really the NPOV policy that enables us to dispense with truth as a criteria. Since NPOV goes along with dispensing with "truth" as a criteria, we need some other minimum standard for inclusion. One is included in the NPOV policy - the POV must be notable. The other is enshrined in this policy, that the view is verifiable. I have no problem with people rejecting this framework - indeed, virtually all other encyclopedias do not use it. And like I said, people are free to try to work for them if they wish. I think we are the only encyclopedia that has this alternate framework ... out of so many others ... and yet there are people that wish to get rid of this framework, which makes us distinctive? That would be a real loss. Might as well just go back to EB. Not that it is a bad encyclopedia. I just think it is good that out of the hundreds of encyclopedias that do exist, there be one, just one, based on the NPOV/V framework and not the "truth" criteria of all the others. Slrubenstein | Talk 01:13, 14 April 2008 (UTC)

• I'm afraid you have been misinformed about this project. We are striving to make an encyclopedia - that means we want our content to be equally accurate (or true, if you prefer) as that in other encyclopedias. NPOV and verifiability are methods we use to help reach the goal of accuracy; they aren't a rejection of it. — Carl (CBM · talk) 01:18, 14 April 2008 (UTC)

• I am afraid it is you who are misinformed. You just do not seem to get it. Perhaps the idea is too coplex or subtle for you. Too bad. Slrubenstein | Talk 01:35, 14 April 2008 (UTC)

Rubenstein, your POV is fascinating but contains a flawed argument. The short version is that you make WP ripe for hoaxers. With a little work, creating a reliable-looking, independent-seeming website farm and publishing otherwise noncontroversial articles with occasional completely false statements, any small group could manufacture a verifiable notability for, say, a nonexistent rock band or church. Separating verifiability from accuracy is a very messy divorce. The potential (meaning probable) inclusion of "verifiable" and "notable" false information is not encyclopedic and I must politely disagree with your view that it should be a WP distinctive. But would you agree with the above suggestion that your inclusion standard is fairly described as "not all that is true, but all that is verifiable"? JJB 13:47, 15 April 2008 (UTC)

No, I still adhere to "Verifiability, not truth." The cases you mention are already handled by other provisions of our NPOV policy, that we include all notable views, and do not give undue weight. See also our policy on fringe views. NPOV and its various adjunct policies covers this concern. Slrubenstein | Talk 13:58, 15 April 2008 (UTC)

Sorry, the philosophical flaw remains. Your implication that WP does not care about truth is fundamentally contrary to WP:AGF. I do not wish to give the long version of the argument now. Do you believe "verifiability, not truth" is equivalent to WP:ATT: "whether material is attributable to a reliable published source, not whether we think it is true"? I do not think they are equivalent. If you think they are not equivalent, you should comment on how to fix the problem that the long WP:ATT version is presented as a summary of the short WP:V version. If you think they are equivalent, you have no reason not to accommodate my view that the longer statement is better here. JJB 16:56, 15 April 2008 (UTC)

• Support. I've long complained about this problem. People seldom listened. Ken Arromdee (talk) 15:46, 15 April 2008 (UTC)

• "The threshold for inclusion in Wikipedia is verifiability, not truth." The statement is a contradiction, for the process of verifying information by reference is itself a guarantee about what the reference actually said/did(in truth to the best of our knowledge). The two are not mutually exclusive, also for the reason that referencing a "reliable" source is to get closer to the truth of any particular scenario. --Jweston2 (talk) 08:29, 20 April 2008 (UTC)
• Strong, strong oppose. The revised wording makes verifiability seem secondary to truth. Verifiability comes first. Otherwise the Wiki process becomes a game of my word vs. yours. - Chardish (talk) 06:38, 25 April 2008 (UTC)
• Conditional support The problem with truth is that it's hard to define or measure. In articles on physical sciences the "truth" is usually clear enough, but in social sciences "truth" is often a matter of opinion because the systems they describe are too complex, models are to inaccurate, and empirical data is too unreliable. So I suggest the following: "the threshold for inclusion is verifiability and falsifiability, not truth". This still would prevent Wikipedia from presenting falsifiable theories (such as Young Earth Creationism) from notable, yet flawed, sources as "knowledge". Falsifiability is easier to establish than "truth". Cambrasa confab 21:36, 27 April 2008 (UTC)

History of the term

The earliest phrasings of the policy had the following language: "The goal of Wikipedia is to become a complete and accurate encyclopedia." Indeed, the earliest form, from 2003, explicitly linked accuracy to verifiability: [1]. The change away from that was made by SlimVirgin in 2005, in this edit: [2]. There, accuracy was changed to "reliability." This is a subtle change, but I do not see it as one that moves us away from truth - the assumption is still that one relies on an encyclopedia to be true. In fact,t he purpose of the change seems to be one of parallelism - because we want to be reliable, we use reliable sources.

The "Verifiability, not truth" statement was imported a few days later from NOR in the following edit: [3]. The larger context of the statement makes the usage of the word "truth" clear - it's one about the absolute nature of truth - what I earlier referred to as big-T Truth. The statement used in the example, about a theory of Hawking's, would be stated as an attributed theory if it was being done right - that is, "In Article X, Hawking said...". Thus nothing that would fail the truth test, in this example, would be included - it's just that the unverifiable but true information of Hawking's later denunciation of the theory would not be included.

The overly long example, however, was rightly trimmed out, and we were left with the slogan, which, coupled with the (in hindsight regrettable) removal of the word "accurate" took us to the current position where we appear to be truth neutral. However, I see no evidence that the truth-neutral position was ever intended, and, if we are talking about core principles, it is clear that truth-neutrality is not a core principle - the original core principle did demand that Wikipedia be accurate. Phil Sandifer (talk) 19:44, 6 April 2008 (UTC)

How many times are you going to reply to this RfC? The point is not to provide you with a forum for a filibuster, but to get comments from a wide variety of editors. Dlabtot (talk) 20:00, 6 April 2008 (UTC)

Oddly enough, he does this for a living and counts as an actual, no foolin', expert on the subject; you don't. But I'm sure you're not letting that stop you either - David Gerard (talk) 20:12, 6 April 2008 (UTC)

What are you talking about? Dlabtot (talk) 20:15, 6 April 2008 (UTC)

That "LOL UR TALKIN 2 MCH" contravenes long-established practice on Wikipedia talk pages, and doesn't make you look too good either - David Gerard (talk) 20:17, 6 April 2008 (UTC)

I guess somehow that explains "he does this for a living and counts as an actual, no foolin', expert on the subject". Then again, maybe not. Dlabtot (talk) 20:21, 6 April 2008 (UTC)

I assume he's referring to the fact that I teach college composition and research skills. Which is less germane to this particular debate (as I don't teach "fundamental nature of an encyclopedia"), but does still heavily inform my perspective. Certainly I am disinclined to give my students a good grade when they cite a source and then totally screw up the whole "getting it right" thing. :) Phil Sandifer (talk) 20:21, 6 April 2008 (UTC)

Right, what we are talking about here, is writing an open source encyclopedia, which is not something you do for a living. As to whether you are an expert on the subject, I'll leave that for consensus to decide. Dlabtot (talk) 20:24, 6 April 2008 (UTC)

Fight the power! - David Gerard (talk) 20:30, 6 April 2008 (UTC)

I don't think the open source nature of the project changes the fact that it ought be a reputable and acceptable presentation of researched information. There seems to be a general consensus on the part of the states of Florida and Illinois that I am a sufficient expert on the presentation of researched information to be allowed to teach it to college students. Phil Sandifer (talk) 20:27, 6 April 2008 (UTC)

As I noted above, "accuracy" is part of the overall quality of the article, similar to good writing. Verifiability, OTOH, is a key concept, which underpins our effort as a tertiary source. Crum375 (talk) 20:02, 6 April 2008 (UTC)

I think accuracy is the point of verifiability - it should be possible to verify our articles, yes. But verify what? The answer, at least in the original policy, is clearly that one should be able to verify our accuracy. And I see no evidence that anybody ever intended the meaning to drift from this - the word "accuracy" was removed in favor of reliability, but that does not seem to me to have been intended as a seismic shift in our goals. (Though, SlimVirgin - can you explain your thoughts on that change?) Accuracy has always been the fundamental goal underlying our demand for verifiability. Phil Sandifer (talk) 20:06, 6 April 2008 (UTC)

Accuracy is not the point of verifiability — verifiability is our ultimate goal. We'd like our readers to be able to verify that our writing is properly sourced. "Accuracy" is not well defined: it could mean accuracy relative to some hypothetical "truth", or relative to the sources (i.e. that we correctly copied or paraphrased them). Since we can't address ultimate truths (as we lack credentials and it would violate WP:NOR), we can only strive for the second kind of accuracy, which is similar to ensuring proper spelling and overall good writing quality, and is secondary to verifiability as a key concept. Crum375 (talk) 20:32, 6 April 2008 (UTC)

Accuracy in this context - on any topic, we want to show the various significant viewpoints that have been cited in reliable sources. Enterprising readers or prospective editors need to be able to verify that those citations are accurate and additionally are provided the ability to judge our characterizations of those sources, and the overall topic, for themselves. Truth doesn't really enter into it. Dlabtot (talk) 20:36, 6 April 2008 (UTC)

For the most part, truth and accuracy are treated by people as synonyms. But you're right - "truth" does have philosophical connotations that, while secondary to its everyday usage, could be confusing in their own right. "Not just accuracy, but verifiability" captures the point perfectly without making the ludicrous (and embarrassing) statement that we don't care if we get it right, we just care that we're not the only ones who are wrong. Phil Sandifer (talk) 20:40, 6 April 2008 (UTC)

"Don't just be accurate - be verifiably accurate." Indeed. "Verifiability not truth" is a phrase that only makes sense as Wikipedia jargon. In terms of relations with the outside world, it's nearly as problematic in my experience as the Wikipedia jargon usage of "notable" - David Gerard (talk) 20:50, 6 April 2008 (UTC)

With all due respect, that's nuts. Accuracy is clearly our ultimate goal. It always has been. It was in the original policy, and no consensus for the sort of radical subjectivism you're proposing has ever, as far as I can tell, been garnered - the idea of it came from a general drift of the language as documented above. The idea that Wikipedia should not pursue accuracy has no consensus. Which is unsurprising - the idea that an encyclopedia would be indifferent to its own accuracy is anathema to any sort of reputability, and aims towards a standard that is wholly unacceptable by any generally regarded method of assessing the presentation of information. Phil Sandifer (talk) 20:38, 6 April 2008 (UTC)

Accuracy is not unimportant — of course it is. But it is a derivative of verifiability. If we demand that a statement X be verifiable to source Y, every reader must be able to verify that Y said X (directly or paraphrased). If there is inaccuracy there, e.g. Y really said Z, and X != Z, then we have violated verifiability. Thus accuracy, as a measure of adherence to the source material, is not an independent parameter. Crum375 (talk) 20:55, 6 April 2008 (UTC)

I agree wholeheartedly - accuracy and verifiability should pretty much never come into conflict, and when they do it's usually a quick phrasing change to explicitly attribute claims that fixes it. The statement really is, in practice, deeply uncontroversial and trivial, serving mainly to close up a silly-sounding slogan that is too easily read as something we don't mean. (That is, we really, really do not mean "We don't care if our information is true," we mean "We're uninterested in discussions of TEH SEKRIT TRUTH TEH GOVERNEMNT DOESNT WANT U TO HEAR." And while those of us who are prone to getting into lengthy discussions of the policy understand that we don't mean "we don't care about small-t truth," we can probably avoid the appearance of visible idiocy with a slightly tighter phrasing. Phil Sandifer (talk) 21:01, 6 April 2008 (UTC)

I would support the change back to accuracy, as it makes the whole issue about neutrality preventing us from the determining "the truth" irrelevant and simply returns us to "please keep the encyclopedia accurate", as we should strive to do whether it's policy or not. Verifiability is a key concept, to be sure, but it's just a way to ensure that we're as accurate as possible. —Scott5114↗ [EXACT CHANGE ONLY] 20:07, 6 April 2008 (UTC)

The change to "reliability" appears to be part of the push to canonised "reliable sources" for all uses (not where they make sense - canonicalisation of authority, rather than reliability per fact). This too is something we're needing to pull back from as the attempt to mechanise the process of thought produces blithering absurdities - David Gerard (talk) 20:15, 6 April 2008 (UTC)

The apparent trouble with "verifiability, not truth" is that it presents the two as if they are in tension with one another, whereas people would normally consider the set of all verifiable claims to be a subset of all true claims. But in fact there are claims that are verifiable but untrue: if (for instance) today the scientific consensus is that there's a black hole at the center of the galaxy but next year it's discovered to be a (say) superdense wibblecluster instead, the previously verifiable claim will be found to have been untrue all along. But when writing today we would say that it is a black hole. --FOo (talk) 22:37, 6 April 2008 (UTC)

Actually, no, the Milky Way article does not say that it is a black hole. What it actually says, which is much more proper, is: The galactic center harbors a compact object of very large mass (named Sagittarius A*), strongly suspected to be a supermassive black hole. Most galaxies are believed to have a supermassive black hole at their center. Dlabtot (talk) 23:46, 6 April 2008 (UTC)

Beg to differ. the Milky Way article (I am assuming good faith here) accurately and verifiably reports that that assertion is made by Blandford, R.D. (1999). "Origin and evolution of massive black holes in galactic nuclei". Galaxy Dynamics, proceedings of a conference held at Rutgers University, 8–12 Aug 1998,ASP Conference Series vol. 182. {{cite conference}}: Cite has empty unknown parameter: |coauthors= (help); Unknown parameter |booktitle= ignored (|book-title= suggested) (help) In another article (e.g., Young Earth creationism) that assertion and the cited supporting source might not be as readily accepted as it is in the Milky Way article. -- Boracay Bill (talk) 03:48, 7 April 2008 (UTC)

I don't know how you arrived at your misunderstanding of my comment. I quoted verbatim from the article. It does not say that there is a black hole at the center of our galaxy. It says it is strongly suspected. And uses the citation you've reproduced here in support. Dlabtot (talk) 15:38, 7 April 2008 (UTC)

Again, you're hitting on big-T vs small-t truth. Yes, we cannot hope to aspire to absolute universal truth, and aspirations to that have no place in the project. But for the most part, when we say that there is a black hole in the center of the galaxy, we are not implicitly appending the statement "and this is ontologically and absolutely true" to the end of it. Instead, we implicitly append "to the best of our scientific knowledge" to the end of it. Big-T Truth is a philosophical concept. But when we tell people "Wikipedia does not care about truth," anyone who is not a policy wonk will assume we mean small-t truth, which amounts to "what, to the best of current human knowledge, is known to be accurate." And that's a problem. But as I've advocated, what about "accuracy," which seems to sidestep the philosophical implications of "truth?" Phil Sandifer (talk) 22:42, 6 April 2008 (UTC)

Why not just eliminate the unnecessary mention of "truth" in the first place, and open with "The threshold for inclusion in Wikipedia is verifiability"? This would save us the problem of trying to develop an overly complicated or simplistic explanation of what WP's relationship to, or stance on, "truth" is, and then we could keep this page focused on discussions of verifiability in terms of accuracy. Amerique^dialectics 00:34, 7 April 2008 (UTC)

Boo-yah. It seems to me that Amerique nailed it precisely (and note: it didn't take a lot of words!) But I'd like to give people a chance to say if the "not truth" phrase has somehow made their job easier. - Dan Dank55 (talk) 02:32, 7 April 2008 (UTC)

That is a good question. I think the reason for this specific wording is to make it crystal clear that we don't strive to find "correct" information, by anyone's standard, but we instead only try to find what the published sources say about the subject. Many editors are unfortunately confused about this issue, and see WP as a big textbook written by the Internet community, which it isn't. Saying right up front that we don't aim for "truth", whatever that may mean, but to simply present reliably published information that is verifiable, hits that nail on the head. Removing it would cause this crucial point to be shoved lower down the page, and make it less clear than it is today. Crum375 (talk) 01:20, 7 April 2008 (UTC)

Yeah, but that seems to me an over-correction. Which is why I'm inclined to demand both explicitly. Verifiability does not replace accuracy - it was always conceived as the way we ensure accuracy. I mean, that is, factually and historically, what the core principle was. "Verifiability, not truth" is a catchy slogan, but it is misleading - in its drive to boldly state the importance of verifiability, it makes a transparently silly statement. Phil Sandifer (talk) 02:32, 7 April 2008 (UTC)

More history. The term, as I said, was imported from WP:NOR. The original phrasing of WP:NOR was straight-up Jimbomancy: [4]. The key phrase here is "If your viewpoint is held by an extremely small minority, then whether it is true or not, whether you can prove it or not, it doesn't belong in Wikipedia." The key thing here is that *this is a restatement of NPOV*. In other words, what's being said here is that instead of making true statements about science, we make true statements about what important people say about science. i.e. not "For every action there is an equal and opposite reaction," but "Newton said that for every action..." Thus the initial rejection of truth is, in fact, a restatement of the fringe theories clause of NPOV.

The term "verifiability, not truth" came from a drafted revision to NOR from late 2004/early 2005. It was added by SlimVirgin in this edit: [5]. The Stephen Hawking example discussed above is present in the earliest version. I can find no discussion of that phrasing in the discussion around the draft, nor anywhere else. As far as I can tell, in fact, there has never been a serious discussion about either the removal of the word "accuracy" or the statement "verifiability, not truth" (which, as I discussed earlier, was not originally phrased in opposition to small-t truth). I see no evidence that there is or ever has been a consensus against the statement "Wikipedia strives to be accurate." Phil Sandifer (talk) 16:15, 7 April 2008 (UTC)

A problem if you are suggesting "Wikipedia strives to be accurate" for the lead sentence is that the policy itself is called "verifiability," not "accuracy." I would support the inclusion of a statement on accuracy but I don't see how it becomes an "either/or" proposition with "v≠t". I agree with you that the latter statement would seem like a silly contradiction to people not accustomed to philosophic subtleties, i.e. most of the general reading public, but I think it could be solved by eliminating the extraneous suggestion of a binary opposition or dichotomy that the policy doesn't try to clarify anyway. Making reference to any kind of "truth" is a bad thing, I think, if the policy does not explain what WP means by that, and trying to do so, I think, would create more problems than it would solve. Leaving "truth" in gets us in needless debates about what the referent of that concept is, which is a distraction, I think, from what this policy is trying to do, which is explain sourcing practices. Otherwise, I see no issue with including a specific, well-worded statement on accuracy in the lead section, but don't immediately see how the core ideas of either of these propositions negate each other. (While the latter one does negate itself, as is.) Amerique^dialectics 17:45, 7 April 2008 (UTC)

I don't think, in practice, that it's either-or - quite the contrary, I think that, if we are limiting ourselves to NPOV statements, there will be no verifiable statements that are inaccurate. The problem is largely one of self-presentation - saying flat out that we are not interested in truth appears to set a tone that is simply wrong for an encyclopedia, is not a principle that this page (or any other) was ever founded on, and that flies in the face of what people expect from an encyclopedia. Nobody, I don't think, seriously believes that anybody loads up Wikipedia and looks up information without the expectation that there it is put there because it is supposed to be accurate. Thus the phrase "not truth" becomes deeply misleading in a way that, I think, puts the whole page off on the wrong foot, because it makes the whole page seem like something that is presenting an alternative to truth. It's not - it's presenting a way of guaranteeing our accuracy. "See, we're accurate. Check for yourself." Even if we make no comment on accuracy - simply saying the threshhold is verifiability, that is preferable to the "not truth" phrase, which is actively wrong (as many people here have pointed out). Phil Sandifer (talk) 21:20, 7 April 2008 (UTC)

Mmmm... Accuracy is not the term either. The only accuracy we can assert is that what you find in Wikipedia is accurate as it relates to the verifiability of the sources we use. That is, we strive to describe accurately what reputable sources say about a subject, without asserting that the claims and viewpoints made in these sources are necessarily accurate. ≈ jossi ≈ (talk) 21:28, 7 April 2008 (UTC)

That's just NPOV, though. I mean, in essence, the three content policies fit together thusly. V notes that we ensure accuracy by having information be checkable in sources. NPOV notes that we restrict the sorts of things we say to statements about what other people say - that is, not "Every action has an equal and opposite reaction," but "Newton's third law of motion is..." NOR notes that we further restrict the people whose perspectives we offer to those who have mainstream, reputable views. Together we get the overall picture - Wikipedia provides an accurate and verifiable collection of what the major viewpoints ona given topic are. Phil Sandifer (talk) 21:36, 7 April 2008 (UTC)

Phil, the reason the current wording ("verifiability, not truth") is important is that Wikipedia (and Wikipedians) aren't really equipped to discover with "the Truth" is; I doubt anyone really is, and we shouldn't focus our efforts on that. All we can do is report what reliable sources have to say, so we need to guide editors to do exactly that, rather than trying to figure out what the "real truth" is about a matter, and writing an article to suit that view. Unless we strongly discourage people who are attempting to write "The TRUTH", we will end up with articles that start "Sun Myung Moon is the Second Coming of Christ, the "Savior", "returning Lord", and "True Parent"." After all, according to hundreds of thousands of people, that is "The TRUTH". Jayjg ^(talk) 04:55, 8 April 2008 (UTC)

Oops, Jay, I didn't see this comment before I deleted "not truth" just now. My sense was that we had consensus, although the discussion has died down so I'm really not sure. I felt two arguments were persuasive: everything that Amerique said, and also my point, which is: it's perfectly okay to say "truth", but don't do it on a core content page. If something on a core content page is unclear, it will cause arguments, and arguments will lead to changes, and lack of stability in a core content page leads to FUD. Let's let people argue what "truth" is somewhere else, and follow Amerique's advice not to distract the reader from a discussion of what verifiability is and isn't. However, I concede that, if my deletion of "not truth" survives, and now that we've taken Jimbo's big quote out of Burden of evidence, there's a legitimate criticism that the page may not do a good enough job of making our position perfectly clear to new editors. Can you think of language that doesn't have "truth" in it that could get the job done? -Dan Dank55 (talk) 09:46, 8 April 2008 (UTC)

I didn't see this comment before I posted above in unknowing agreement. I think Jay is quite right. --Relata refero (disp.) 10:11, 8 April 2008 (UTC)

First, none of the proposed wordings remove verifiability as a requirement - that is, there is nothing that is currently forbidden that is now allowed under the new wording. Second, this idea that people will read "verifiability, not truth," "not just truth but verifiability" or "not just accuracy but verifiability" as meaning, as JayJG so elegantly puts it, The TRUTH seems to me unfounded. Most people, I think, do not take the word to mean the sort of Platonic TRUTH you mean - the term, I think, is well-understood among Wikipedia policy wonks, but to someone who is not a policy wonk the term sounds like we don't care about accuracy. Which is not at all something we want to present. I mean, there are other ways we could fix it - we could reintroduce the long-standing sentence "Wikipedia strives for accuracy," then note that the threshhold for inclusion remains verifiability. But it is manifestly silly of us to have a policy page that openly and in boldface suggests that Wikipedia is unconcerned with accuracy. Phil Sandifer (talk) 14:22, 8 April 2008 (UTC)

Jayjg is right. Obviously. This is a pointless discussion. Slrubenstein | Talk 14:17, 8 April 2008 (UTC)

I think that the number of people who have expressed skepticism about the formulation, the fact that this wording *never* got serious discussion when it was proposed, and the fact that when it was proposed a lengthy explanation accompanied it which is long since gone suggests that, actually, there is nothing obvious about the statement that "verifiability, not truth" in its current form has or ever had consensus. Furthermore, there seem to be sensible objections being raised, not the least of which is that it is profoundly silly for an encyclopedia to openly declare itself uninterested in whether information within it is true. None of which is to say that we ought change our policy on requiring verifiability. But to declare the discussion pointless is utterly unhelpful to it. The discussion is, actually, pretty good. Please consider contributing to it with more substantive statements. Phil Sandifer (talk) 14:22, 8 April 2008 (UTC)

hmmm... I have always read the "Verifiability not Truth" phrase a bit differently than most of you seem to be doing. I have always understood it to mean that you can not include something just because you know/think it is true... you have to show that someone (a reliable source) other than you has reached that conclusion. In other words, the phrase in question is really a restatement of WP:NOR at its most basic level. Blueboar (talk) 14:34, 8 April 2008 (UTC)

Would the phrase "not just truth, but verifiability" thus seem equivalent to it in your mind? Phil Sandifer (talk) 14:44, 8 April 2008 (UTC)

To me, the phrase "not just truth, but verifiablity" implies that the editor knows the truth, and is avoiding claims that can't be verified because the editor realizes that readers don't trust editors. But the reality is the editor may not know the truth, is aware of his ignorance, and just puts in a balanced set of verifiable claims so the readers can draw their own conclusions. --Gerry Ashton (talk) 15:22, 8 April 2008 (UTC)

Either way seems to say the same thing to me. I think the real problem is that we don't fully explain what the phrase means (thus the confusion, and potential for abuse by wikilawyers). A brief sentence or two needs to be added. something like:

• "The threshold for inclusion in Wikipedia is verifiability, not truth. While Wikipedia strives to provide accurate information on every topic it covers, editors should not add material simply because they know or think it is "true". The information also needs to be verifiable. "Verifiability" in this context means that readers should be able to check that material added to Wikipedia has already been published by a reliable source. ...." Blueboar (talk) 15:40, 8 April 2008 (UTC)

(edit conflicts) The problem as I see it is, the first line of this policy is being intended for widely different audiences: POV pushers who might need such an admonition re: "truth," and everyone else, I would say the vast majority of users, of varying degrees of sophistication, who will read "v≠t" and either think "smart" or "what BS!" To take care of one problem caused by some users, "v≠t," while appeasing some, offends others who basically see it as a logical contradiction, as evidenced by numerous debates over it located here:

• Wikipedia_talk:Verifiability/Archive_6#Verifiability_not_Truth

• Wikipedia talk:Verifiability/truth

• Wikipedia_talk:Verifiability/Archive_7#The_role_of_truth_in_WP_policy

• Wikipedia_talk:Verifiability/Archive_7#Why_negation_of_truth.3F

• Wikipedia_talk:Verifiability/Archive_8#.27Verifiable_truth.27

• Wikipedia_talk:Verifiability/Archive_8#Definition_of_term

• Wikipedia_talk:Verifiability/Archive_8#Terms

• Wikipedia_talk:Verifiability/Archive_8#WP_not_into_truth_.28yet.29

• Wikipedia_talk:Verifiability/Archive_9#Introduction.2C_defining_verifiability

• Wikipedia_talk:Verifiability/Archive_10#Removal_of_truth

• Wikipedia_talk:Verifiability/Archive_10#Justify_this

• Wikipedia_talk:Verifiability/Archive_11#Wikipedia.3DNot_Encyclopedia

I could not read more past archive#11, but the point that the line "v≠t" has been a constant cause of controversy from its inception has been made. By not explaining what WP means by truth, the policy leaves an opening for these debates to continue. Removing "not truth" removes this source of controversy. The "not truth" concept can perhaps be more effectively pursued at WP:NOT. Amerique^dialectics 16:09, 8 April 2008 (UTC)

←Thanks for the research, Amerique. Correct me if I'm wrong, but I have re-read every statement about the word truth, and I think we could get very wide consensus around keeping it in, but as the word of the people we're talking to, not as our word. So we couldn't keep it in the first sentence, but we could put it in the lead. Many suitable formulations, it seems to me, have already been proposed in this discussion. - Dan Dank55 (talk) 19:15, 8 April 2008 (UTC)

P.S. I like most of Blueboar's words just above, but I think Amerique's research demonstrates what happens when we even imply that we know what "not truth" means. I also agree with many others that, even though "not truth" has served a useful purpose on Wikipedia, it is a major turn-off for academics, policymakers, professionals and journalists. There's also a general principle here that we should keep hard-to-define and contentious words such as "truth" out of core content policies, no matter how useful they are. But I have no problem with Blueboar's "truth" in quotation marks. - Dan Dank55 (talk) 20:32, 8 April 2008 (UTC)

As before, it's a rather important point for editors to understand that although millions think they know "the truth" (and these truths are often diametrically opposed), neither we, nor really anyone else, is qualified to decide what "the truth" is. Jayjg ^(talk) 01:23, 9 April 2008 (UTC)

So you've said. And I, as well as others, have responded with concerns that the phrasing is misleading, that many people will not and do not distinguish between the sort of big-T philosophical truth you're discussing, and that the sentence looks like we're saying we don't care about accuracy. The "verifiability, not truth" phrasing does not seem to have consensus now, nor am I convinced that it ever has. Please propose an alternative. Phil Sandifer (talk) 01:27, 9 April 2008 (UTC)

Is there any reason why 'accuracy' is not open to identical problems of dual significance? --Relata refero (disp.) 14:38, 9 April 2008 (UTC)

It is, at least, not used in philosophical discussions about epistemology very much. So it's harder to bring that meaning into it. Phil Sandifer (talk) 14:42, 9 April 2008 (UTC)

Proposed logjam break

I will take the liberty of WP:BOLDly importing the following clause from WP:ATT, after an em dash: "whether material is attributable to a reliable published source, not whether we think it is true". This represents ongoing consensus on that page, which is a summary of this page. The only opposition might come from any who wish to baldly state that unelucidated "VNT" is superior to all other formulations, but per my argument above, either VNT is identical to the longer version and the question is moot, or there is a problem with ATT's "summary", which can be solved by an edit like this one. JJB 03:43, 23 April 2008 (UTC)

Having read the final I will note that the first time I read this phrase was when someone defended a speculative negative opinion about what two people might think about a third person at a RS blog. Does there need to be a clear distinction between fact and opinion?? Carol Moore 16:21, 23 April 2008 (UTC)Carolmooredc {talk}

Thanks Carol! So VNT has overlapped into personal attacks being defensible because truth is not our threshold? That's a consequence foreseen in this little essay I just wrote: JJB 17:37, 23 April 2008 (UTC)

Please reference Original source of VNT subhead above. In Jimbo's statement at Wikipedia:NPOV, V and OR he wisely finesses the sometimes hard question "Is X true (credible)?" by answering two easier questions instead: "Who says X is true (credible)?" and "Is that source reliable (reputable)?" But note that, while changing the question, he retains the individual editor's responsibility to use good judgment (determination) while answering. For Jimbo, the standard is still that edits (even attributions) must be true: untruths of the form "Y says Z" can be reverted immediately on either ground, either that Y does not say Z, or that Y is not reliable. In short, this seems strongly to support versions like "verifiability, not just truth".

I believe an unintentional slippage from Jimbo's statement has occurred. He refers only to when it's "quite difficult for us to make any valid judgment", but that implicitly omits easy valid judgments, such as "1 + 1 = 2" (tagged as "refimprove" in 1 (number)). Millions of unsourced, noncontentious statements are retained where "Is X true?" is easy. In fact, it's somewhat harder to judge "Is there some source which reliably says 1 + 1 = 2 in the right context?" than it is to judge "Does 1 + 1 = 2?" (Insert random speculative 'I heard it somewhere from notorious paradoxer Bertrand Russell' pseudo information here.) The same omission is also implied in the current WP:OR as the material not "likely to be challenged" or not "open to interpretation".

This implication seems to slip away gradually and unconsciously in the comments I quoted above (not to pick on Slim, as the same slippage happens psychologically to us all at times, and is evident in several others' comments, but Slim's are handy right now). I repeat them in order, supplying qualifiers in italics where necessary to preserve Jimbo's implication: "what is verifiable, not only what is true"; "firm belief in something being true is not by itself a reason to have it in Wikipedia"; "not allowed to insert something contentious or not obvious just because they personally know it to be true". We slip into thinking that the reason we include "1 + 1 = 2" is not at all because we judge it independently obvious, but only because we judge it obvious that some reliable source must've said it by now. When we fail to allow for Jimbo's implication, we mistakenly think we must never judge statements of the form "X is true", but only those of the form "Y says Z" and "Y is reliable". The philosophical error is to forget that "Y says Z" and "Y is reliable" are also statements in the form "X is true". If we are presumably capable of judging reliability, it is only because we are capable of (often) judging truth itself. The converse philosophical error is to forget that if we are presumably incapable of judging truth itself, we are (always) incapable of judging reliability.

Now, every dispute on WP comes down to a matter of editors asserting X and not-X, and very often in cases other than attribution and reliability. But strict "VNT" would be compelled to say: WP cannot judge X, so it is inappropriate to take any action toward resolving a dispute over X. "VNJT" would say: WP can judge X, so it is appropriate to take action toward resolving those cases where I can best contribute-- and in fact that's how most people behave, even though they profess VNT over VNJT. Strict VNT is in conflict with our duty (IAR) that our behavior improves WP, because strict VNT encourages inaction in unresolved disputes, which worsens WP. On the other hand, the way most people actually behave in dispute resolution presumes that DR is fruitful, we can actually make independent truth-based judgments, and therefore the effective policy really is "not just truth". (Arguments like "we can judge truth in talk but not in mainspace" fail to split the difference, because truth claims have the same nature on both sides.)

In sum, I have demonstrated that we actually have slipped from VNJT to VNT; that this slippage leads to erratic claims about our ability to judge the validity of ordinary statements; and that this error in judgment, logically followed, contradicts the goal of improving WP. Having finally taken the time to write this out for you, I will be happy to answer questions and criticisms. JJB 17:37, 23 April 2008 (UTC)

RfC: Should WP:SELFPUB be reorganized?

Template:RFCpolicy I attempted to reorganize SELFPUB, but was reverted here. I explained above why I feel there is a categorical difference between restrictions #1-5, #6, and #7, and why not making that distinction is potentially misleading. To sum it up, [what are now] #1-5 would seem to apply to statements specifically referenced in wikipedia which originate from "questionable" sources, whereas #6 would exclude an entire source from being used at all, while #7 limits how questionable sources can be used in an article overall. I don't feel this is a substantive change, simply a reorganization. PSWG1920 (talk) 01:57, 7 April 2008 (UTC)

I do not think that a re-organization of self-publication is what is needed. The position of the deck chairs is not why the Titanic is sinking. Phil Sandifer (talk) 02:35, 7 April 2008 (UTC)

Rumors that the Titanic is sinking are greatly exaggerated. - Dan Dank55 (talk) 04:25, 7 April 2008 (UTC)

If the deck chairs were positioned differently, then maybe trouble could be seen more clearly. In other words, reorganizing this could help illuminate other problems. PSWG1920 (talk) 05:10, 8 April 2008 (UTC)

Huh, I don't even see a ship. _(1 == 2)^Until 05:13, 8 April 2008 (UTC)

• WP:SELFPUB needs tightening more than it needs loosening, so of course I oppose this change. This re-write is an attempt to make websites promoting Perfect Sight Without Glasses more palatable as sources. Dlabtot (talk) 02:54, 9 April 2008 (UTC)

At this point, I am not suggesting that SELFPUB be loosened, simply reorganized. Other issues can be discussed later. PSWG1920 (talk) 03:58, 9 April 2008 (UTC)

Can we more clearly define what does it mean "self-published"? Then discussions like in the next section below (books by Tregubova) would not occur.Biophys (talk) 16:13, 9 April 2008 (UTC)

I agree, Dlabtot, SELFPUB does need to be tightened, or at least the full extent of the restrictions should be emphasized. This edit seems very reasonable. Or perhaps questionable sources should be excluded completely, so these issues wouldn't come up. 88.164.4.15 (talk) 20:34, 20 April 2008 (UTC)

Support because the two versions have the same logical content, but PSWG believes that the old version can be misunderstood, and there is no downside to the new version. If of course some misunderstanding can be possibly predicated upon the new version, what it might be has not been made clear by the objectors to the new version. 67.191.48.90 (talk) 02:44, 14 April 2008 (UTC)

Another burden of proof proposal

See diff, which I immediately self-reverted. I proposed something similar some months ago as a newbie here, but was drowned out because I improvidentially made the suggestion hot on the heels of a real case (now long stale), which you can read about there.

The problem I outlined, which others have now brought up independently above, is the extant possibility of thoughtless challenging or reversion of good-faith edits as "unreliably sourced". My solution is that the claim that source X is reliable or not is a verifiable claim like any claim in mainspace (it just happens to be a metaclaim), and the editor who first makes such a claim should carry the burden of proof. Either the one who challenges a good-faith source as unreliable, or the one who preemptively inserts a source with simultaneous claim of reliability, should be the first to provide the goods; and both editors should be encouraged to discussion of reliability as a verifiable fact like any other.

Because there is the now-ever-so-tenuous linkage between my proposal and a potential conflict of interest, I urge you to read the whole prior discussion at this talk and ask any questions. While my interest in the policy arose from a particular case, I hope it is clear by now that I still believe in good faith this is a policy gap that needs addressing and would like to discuss the merits of the proposed change. JJB 04:06, 14 April 2008 (UTC)

We don't require other sources to "verify" that a source is reliable or not. We decide that by discussion on the talk page. I suppose if you wanted the article itself to claim the source X was unreliable you would need a source for that, but editorial discussion on the talk page does not require sources (although they may be beneficial in making your arguments of course). — Carl (CBM · talk) 04:26, 14 April 2008 (UTC)

OK, granted that the verification process for metaclaims is not exactly "like any other", but the need for verification of reliability through discussion is still underweighted. Some editors have a practice of deleting sourced material by claiming unreliability and relying on the current WP:V to leave the burden of evidence on the inserter as to whether the source is reliable. When a source appears reliable prima facie, the initial burden should be carried by the one who claims unreliability. Otherwise new content can be held hostage through "show me a rock" arguments where each new insertion is challenged reflexively; my edit here would discourage this problem, an admitted concern in other discussions above. JJB 14:38, 15 April 2008 (UTC) The essay WP:WHEN provides an excellent example of this concern as well in the topic "Challenges should not be frivolous" in relation to mere tags on unsourced material; how much more so with deletions of sourced material. JJB 14:45, 15 April 2008 (UTC)

Nobody else? Looks like I'm primed to try a more permanent version of this edit taking Carl and WP:WHEN into account. Any issues? JJB 14:16, 17 April 2008 (UTC)

Normally, when two editors disagree on whether a piece of information is cited to a reliable source, the information is removed at least once and a discussion takes place on the talk page as to the reliability of the source with WP:RSN called upon if an outside voice is needed. Your proposed change would shift the conversation from reliability to whether or not the source is reliable on its face, which just adds unneeded drama and calls to WP:ANI for "failure to AGF". Burzmali (talk) 16:13, 17 April 2008 (UTC)

Huh? Shifting from reliability to whether it's reliable? If you mean shifting from "reliable for the proposed edit" to "reliable in general", I don't intend that. The purpose of my proposal is to emphasize the responsibility to back up claims that the source is (un)reliable to support the edit sourced. But perhaps you mean "An editor who reasonably disagrees with another about level of source reliability should invoke a consensus-building method such as the bold, revert, discuss cycle or the reliable sources noticeboard."? JJB 16:22, 17 April 2008 (UTC)

Prima facie meaning on its face in English. You are proposing that if a source is reliable on its face, it must be challenged before it can be deleted. However, what I think is reliable on its face and what you think is reliable on its face are clearly different, so when I remove your source, we start fighting over whether or not the source is reliable on its face and whether or not I acted in good faith without even considering if the source is actually reliable. For example, let's say someone is sued by the government, and the government asks for summary judgment. If the respondent can't show that he has a Prima facie case, the government wins without even considering a trial. The last thing I want to see is Wikipedia become more like a government bureaucracy. Burzmali (talk) 18:10, 17 April 2008 (UTC)

Perhaps my summary should add what was implied: "When a source appears reliable prima facie to the inserter, the initial burden should be carried by the one who claims actual unreliability." If the reverter thinks the edit is not even prima facie reliable, why wouldn't that reverter go and win the argument by arguing from the stronger ground of "not actually reliable"? Also, doesn't a good-faith revert of a sourced insert always need to state a good-faith challenge before (i.e., while) deleting? Also, how can I be preventing the challenger from reverting, if I'm inviting BRD? Are you afraid something in my text (as proposed at the diff) encourages an inserter to sit on a prima facie source rather than "invoke a consensus-building method" to resolve the actual disagreement? JJB 20:14, 17 April 2008 (UTC)

←That pretty much reduces the change to "All sources should be considered reliable unless proven otherwise", since I should assume good faith on the behalf of the editor who inserted the source. The model we have now basically assumes that some sources are almost always reliable (NyT, WSJ, peer-reviewed journal), some are reliable within a reasonable scope (local newspaper, trade journals) and some are reliable in only specific cases (blogs, self-pub). If you want to add a source to support something outside its scope, you should either mention on the talk page or not be offended when someone remove it with a cite to WP:RS. You change suggests that all sources should be handled like the NYT until proven otherwise, and I contest that such a system is a open license to add and defend poor sources to push a WP:FRINGE agenda. Burzmali (talk) 15:32, 18 April 2008 (UTC)

You're getting warm. You describe the current model correctly, but my proposal is upholding that model, not creating a model where all sources become like the NYT (often unreliable BTW). The problem is not that I want anyone "to add a source to support something outside its scope"-- it's that people don't know, and disagree on, when something is outside the scope in the first place! If you imagine anyone (in good faith) would willingly intend to add outside the source's scope, you have prejudged the reliability question. If one suspects the add might be considered outside the scope, yes, one should proceed at talk or with say WP:BRD. And if someone does unwittingly add obvious fringe sources, by the definition of "fringe" the clear consensus will be against the add. But my proposal is dealing instead with the issue where there is not consensus about reliability, in which case the deletionists should not have the upper hand automatically. I am talking about what happens after the inserter has met the initial burden of proof as to sourceability-- I think that then anyone opposed to the add should meet an initial burden of proving unreliability, rather than just stating it and demanding that the inserter also meet the burden of proving reliability. JJB 20:44, 19 April 2008 (UTC)

Conflicting sources Sport event records

I didn't see, in this article, any orientation about how to deal with conflicting sources. For example, there is a sport competition own by an organization. That organization says that competition replaced another one that was not own by them. They also clearly say that the records of the previous competition are NOT part of the records of the new one. But, some third part sources, consider the previous competition's records as part of the new competition's records. So, how we suppose to deal in this situation. Use the records by the owner of the competition (with or without a comment that some sources say different)? Or use the records by third parts? Thanks. --ClaudioMB (talk) 01:21, 15 April 2008 (UTC)
It's a bit complicated, I'm going to reformulate. There is 2 sport competitions: A and B. Competition B replaces competition A. Competition B is own by Owner B that doesn't own competition A. Owner B say clearly that records of Competition B does NOT include records from competition A. Differently, some other sources consider records of competition A as part of records of competition B. So, should Wikipedia use records from Owner B or from other sources? My, PoV is that since that is a formal competition, with formal regulations and records, Wikipedia should use the source from the owner of the competition. If the other sources are considered by a large number of people, it could be mention that "sources A, B and C consider as one record". Maybe, this article could have a section explain such situations. Thanks. --ClaudioMB (talk) 03:12, 15 April 2008 (UTC)

In general, we should present all sides fairly and without bias when there are differing and/or conflicting views on a topic. That is largely what is meant by "neutral point of view". Cheers! Vassyana (talk) 03:23, 15 April 2008 (UTC)

Not just all sides ... the keyword is significant viewpoints. ≈ jossi ≈ (talk) 15:45, 15 April 2008 (UTC)

I understand that Wikipedia should present significant viewpoints on a subject. But, Wikipedia cannot change the subject. I mean, I consider that a sport event owned by an organization has only one record, the record defined by the owner. There is no "neutral point of view" on this. This is not something abstract like "who is better". This is just a fact defined by someone. If the owner of a tournament says that "A" is champion and "B" is not, that's it. Anyone could disagree with it, but no one could redefine it. So, Wikipedia must present the official records, but could mention that some sources disagree with them. Also, I still believe this subject should be mention in this article in order to guide editors and avoid long disputes. Regards. --ClaudioMB (talk) 04:57, 16 April 2008 (UTC)

The issue here is whether the two competitions are connected or not. The owner says no, but another source says yes. Thus, we have two viewpoints. If both viewpoints are significant, both should be discussed in accordance with WP:NPOV. Blueboar (talk) 14:39, 16 April 2008 (UTC)

I agree with you. Both significant viewpoints should be presented. But, how should they be presented? When Wikipedia presents the records, it should use only the owner's source. Any other significant viewpoint should be present to readers to know that there are some people that believe differently. Could we work in this way?--ClaudioMB (talk) 22:22, 16 April 2008 (UTC)

I am not sure it is correct to present only the owner's records. Perhaps a chart with different collumns for different record sources would work? Blueboar (talk) 01:17, 17 April 2008 (UTC)

Initially, I thought do exactly that, 2 columns. But, I realize that, on our society, the ownership gives all the power to the owner. Below, I'll explain better.--ClaudioMB (talk) 02:06, 17 April 2008 (UTC)

I've changed the title of this discussion because the former title was not correct to the problem. Now, I realize this is not a conflicting source problem, this is a source reliability problem on sport event records. What I mean is that any organized sport event has an organization behind it, like IOC or NBA. Those organizations own the rights of their sport events. They have all the powers an ownership gives to them to dictate everything about it: who was the winner, when the competition start to happen, anything. Anyone could disagree with them, but no one can change what they define.

For example: when, for any reason, IOC disqualify an athlete after he/she won an Olympic gold medal, anyone could disagree with the disqualification, but no one can keep saying that he/she own that gold medal. IOC cannot deny that that athlete was the winner at the end of the competition, but they can deny to him/her what belongs to IOC, an Olympic medal and the title of Olympic gold medalist. In the end, IOC owns the records.

So, my point is very simple, when Wikipedia presents sport event records, it must use only the owner's records (official records). Any other significant viewpoint that disagree with that (unofficial records), should not be consider as records, but mention in order to show readers that some people disagree with the official one.

I'm sorry for the initial confusion, but, sometimes, that's the way discussions and ideas develop. --ClaudioMB (talk) 02:06, 17 April 2008 (UTC)

Wikipedia as source

I have changed the 'permissive not':

"Articles and posts on Wikipedia may not be used as sources."

to the imperative form:

"Articles and posts on Wikipedia must not and shall not be used as sources."

If anyone objects, please feel free to change it and explain your rationale User:Pedant (talk) 07:04, 15 April 2008 (UTC)

←I reverted Pedant because of the second sentence in the first infobox, which says "Before editing this page, please make sure that your revision reflects consensus." A few relevant points that have come up in the past:

• There is some confusion over what the words "source" and "cite" mean. Sometimes they mean "used as a wikilink", sometimes they mean "mentioned by someone somewhere other than Wikipedia", and sometimes they mean "formally cited in an academic or policy paper".
• Don't forget the general principle behind WP:UNDUE. Is that sentence really the most important thing in all of WP:V? If not, then the proposed language was too dramatic.
• See the 5 pages listed in WT:Verifiability/Archive 26#Academics and journalists, which show that Wikipedia is increasingly being used as a valid source by journalists and academics. One of the articles at WP:GAN, 2006 Atlantic hurricane season, was recently cited both by academics and by the National Climatic Data Center.

On another point, Relata reverted Pedant to "cannot" rather than the original "may not". (Was that in response to my "may vs. can" thread above? If so, thanks for listening :) But the case is weaker for "may not" than for "may"; what does everyone think? - Dan (talk) 12:10, 15 April 2008 (UTC)

"may not" is not permissive, anyway, at least in terms of how it usually used in such contexts. "You may not steal cars" is not the same thing as "you might not steal cars"; "you may" is a grant of permission, "you may not" is the negation of this, and indication of the lack of permission. SamBC(talk) 12:15, 15 April 2008 (UTC)

On the other point, this page only governs what may be used as source on wikipedia, for purposes of verifiability; it can't limit any use of wikipedia elsewhere. Further, the idea that it would prevent wikilinking is contradicted simply by common sense, and anyone trying to claim that it prevents that is probably trolling. SamBC(talk) 12:15, 15 April 2008 (UTC)

I think "may not" is fine. "Cannot" is incorrect if read literally. "Must not and shall not" is overly polemical. — Carl (CBM · talk) 12:21, 15 April 2008 (UTC)

Agreed. "You may not do such-and-such" is not going to be misread in the way that "editors may" might be misread (discussed above). - Dan (talk) 12:31, 15 April 2008 (UTC)

Indeed, You shall not is the future, not the imperative; you may not is the imperative. Slrubenstein | Talk 13:20, 15 April 2008 (UTC)

I did indeed replace cannot in response to the thread; while it is incorrect if read literally, I doubt anyone would actually read it like that. --Relata refero (disp.) 18:04, 15 April 2008 (UTC)

Why not ditch the passive voice? "Do not use articles and posts on Wikipedia as sources." (Though I never understood why we wouldn't allow them as primary sources when relevant... if Jimbo had posted a detailed comment on the Essjay controversy on a talk page I am hard pressed to see why we would not treat that as a notable part of a notable figure's response to the issue. But that's a separate matter - in terms of this phrasing, ditch the passive voice and lose the ambiguity.) Phil Sandifer (talk) 18:08, 15 April 2008 (UTC)

Phil, we certainly could use Wikipedia as a primary source when appropriate... But there are very limited situations where that would be the case. About the only situation I can think of would be if a particular Wikipedia article became a source of controversy and comment in the real world, and we were to write another article about that controversy... In that limited situation it would be appropriate to cite the controvercial article (using an archived version to show what the article looked like at the time of the controversy) to back statements of fact about what that article said. However, I don't think we need to spell out this very limited exception. In any case... for once I agree with your rephrasing :>) ... active voice is clear and better wording. Blueboar (talk) 18:36, 15 April 2008 (UTC)

"I did indeed replace cannot in response to the thread"...oh the irony, the one person who's giving me support in the "may vs. can" debate, and I revert you :) I was really reverting Pedant, you were just collateral damage. After reading the last 2000 or so messages, it seems to me that edits in policy and guidelines pages without consensus do seem to fan the flames; whoever put "don't edit without consensus" at the top of every policy and guidelines page had exactly the right idea, I think. I don't know why it has become so common not to revert undiscussed changes on sight on policy and style pages; can anyone tell me how that developed? Is it a good thing? Should we remove the top infobox, or should we start reverting undiscussed changes on sight? - Dan (talk) 18:53, 15 April 2008 (UTC)

A revert is also an edit. A revert of a good faith edit *never* has consensus. The rule is actually "Before editing this page, please make sure that your revision reflects consensus.". O:-) --Kim Bruning (talk) 19:00, 15 April 2008 (UTC) Ps. the person who wrote that was me. O:-)

Okay, that makes sense. Suppose that, every time I see a recent (like, in the last 24 hours) edit on a policy or guidelines page where I have seen no discussion on that or similar pages, I immediately make a boilerplate post on the talk page saying something like "there was an undiscussed edit; I plan to revert per the instructions in the top infobox; editors are requested to respond within x hours (you pick the x) if you're aware that there was consensus for that change". (And obviously, I would encourage everyone else to do the same, but I can only speak for myself.) I'm saying "boilerplate" because I don't want it to look like I was targeting any edit in particular, only that it's standard operating procedure to revert edits for which there was no consensus. Might even use some small "warning"-type graphic just so people know I'm not picking on them. Would that be a good thing or a bad thing? - Dan (talk) 19:10, 15 April 2008 (UTC)

(ec) P.S. Don't get me wrong ... I'm not saying that what we've been doing is wrong, per se, or that it has to happen this way. I'm suggesting that keeping up with style guidelines pages (where I usually hang out) and other guidelines and policy pages has become difficult, and instability creates FUD, and perhaps it would be better to get a little more routinized in the way we respond to edits that do not reflect consensus, so that we don't all wind up passing the buck and letting bad things stay in place. Obviously, this would need discussion at WP:VPP, at the least. - Dan (talk) 19:18, 15 April 2008 (UTC)

I think an undiscussed reversion is appropriate if you are confident that the new version doesn't reflect consensus, or you at least have significant doubt. It's a sliding scale, really... just use judgement. SamBC(talk) 19:16, 15 April 2008 (UTC)

The top infobox does not instruct you to revert. It is not standard operating procedure to "revert edits that do not have consensus" (As reverts never have consensus, this used to lead to infinite revert wars, hence WP:3RR. )

An undiscussed reversion is never ever appropriate in the case of good faith changes.

If you want to "revert an edit because it does not reflect consensus", then consider that perhaps consensus isn't what you thought it was. After all, at least one person disagreed enough to make an edit (as per consensus can change, and documentation on how silence and consensus are related). While people certainly mean well, reverting such edits is actually borderline disruptive, because they disrupt the normal consensus process.

It is true that more people have started reverting with that particular rationale in recent months. It is not a good practice, so I've started explaining that to people everywhere I can.

The template actually states that people should feel free to document consensus (but that they must be sure that it *is* the current community consensus). As per WP:BRD, one way to find out for sure that something has consensus is to make the edit, and then see if someone reverts. If you revert on the basis of "I didn't see the discussion anywhere", as opposed to "I disagree"; then that method doesn't work anymore, and it becomes impossible to discover actual consensus, or who is a "most interested party" in that consensus--Kim Bruning (talk) 19:31, 15 April 2008 (UTC)

That sounds reasonable. I still get a sense that, on style guidelines pages, in practice, it has sometimes tipped the scales in favor of some people feeling a lack of responsibility to take action, with the result that bad edits flourish and bad people feel rewarded. I'm pretty sure that the way it plays out in practice on style guidelines pages is quite different than on, say, policy pages, because ... despite the last 1000 edits on this talk page ... policy is straightforward compared with style. Not that policy is easy, but it's incredibly easy to ask a question on a style guidelines page that no one knows the "right" answer to. It's often easy to spit out "quick" answers, but coming up with an answer that correctly respects variations in culture, writing style, etc, and balances the goals of looking nice to professional and "persuasive" writers vs. being sufficiently tolerant of diversity so that we don't lose any editors as a result of meddling ... well, it's really hard. I guess the more I write, the more I think that some kind of different community standards would be appropriate and not hard to get approved on style guidelines pages. We've got a lot of open questions at WT:MoS at the moment, but for the benefit of people reading this who want the link, I'll go ahead and create it now: WT:MoS#Future discussion on reverting edits to style guidelines. - Dan (talk) 21:46, 15 April 2008 (UTC)

Well, people on Wikipedia talk:Consensus have been having a surprising amount of trouble thinking up any kind of discuss first or revert-if-not-discussed kind of workflow, to the point where they simply gave up. Each time, they ended up with potential situations where people simply got stuck, and new consensus could never be found anymore. So while your proposed change *sounds* simple, it turns out to be very hard in practice. (or worse, it might seem to work for a while, and then fail with lots of wikidrama involved). I guess tweaking the actual wiki-model for editing and gaining consensus is very tricky, it's no wonder it's one of the foundation issues. :-)

I understand that you may want a bit of stability in MoS, for diverse reasons? Just be careful that it doesn't get entirely stuck, alright?

--Kim Bruning (talk) 22:21, 15 April 2008 (UTC) Initially I thought that "discuss first before using the wiki" would turn out to be merely inefficient, but it kept coming up as unworkable in almost every analysis. That did spook me a little.

Absolutely, and you'll get an engraved invitation at the trial-and-error stage. - Dan (talk) 22:51, 15 April 2008 (UTC)

Kim, I disagree with your statement "A revert of a good faith edit *never* has consensus." and hence your understanding of what a Wikipedia consensus is (No surprise in that for you ;-) ). A Wikipedia consensus is not the same as Cabinet Government consensus, just because one person does not agree does not mean that there is not a Wikipedia consensus on a particular issue. --Philip Baird Shearer (talk) 18:12, 16 April 2008 (UTC)

That's a persistent meme, and it's true in certain time-limited situations where we have no choice (several oft visited pages/systems), but it just doesn't work that way in none-time-limited situations (which is the entire rest of the wiki).

Page content is determined by WP:SILENCE, whether we want it to or not, and you can see it in action every day.

Exactly there is a sentence in an article someone comes along and changes it. It gets reverted, the first editor reverts, a third editor reverts to the consensus version ... . It takes more than one swallow to make a summer and more than one edit to indicate that a consensus no longer exists. Equally if no one objects to an edit then the edit has a consensus (even if it is a consensus of one). But I dont agree with your statement that "A revert of a good faith edit *never* has consensus." --Philip Baird Shearer (talk) 09:47, 17 April 2008 (UTC)

Well, as long as you don't actually try to revert back to the mythical "consensus version", as the only definition for that is at WP:WRONG ;-) . Reverting a good faith edit could yet gain consensus, but you're going to have to talk with the editor. Not doing so would be rather rude (not to mention you would be assuming that the author was acting against consensus, which is not a good idea). As you can see, a lot of policies and guidelines and essays all intersect at this point, and all explain the same thing over and over from different angles; so that hopefully people get it, one way or the other. --Kim Bruning (talk) 15:41, 17 April 2008 (UTC)

With "Reverting a good faith edit could yet gain consensus", you have just made a very similar statement to your previous one (and repeating it does not make it true) "A Wikipedia consensus is not the same as Cabinet Government consensus, just because one person does not agree does not mean that there is not a Wikipedia consensus on a particular issue." :-) --Philip Baird Shearer (talk) 18:55, 17 April 2008 (UTC)

I have provided a large number of references that all imply or explicitly state that what is so is so. I have provided multiple different reasonings (either directly or per link) that all come to the same conclusion. Even if those references did not exist, I can even reason from scratch. If you start with 3 people in a discussion, you might get a different consensus than if you start with 4. And it might differ somewhat yet again if you start with 5. (Hopefully, 5 people know more than 3, so the consensus among 5 people is "superior"). Optimally, if people come along at a later date, the result will be the same as if you started with x number of people in the first place. And that's some of the basis behind WP:CCC in its current form.

Note that so far you've been repeated a bald (and opposite of bold ;-)) position twice, but have provided neither references nor reasoning to support that position. Would you care to do so now? --Kim Bruning (talk) 15:59, 18 April 2008 (UTC)

I do not that that anything you have written in this thread give credence to the initial statement that you made: "A revert of a good faith edit *never* has consensus." --Philip Baird Shearer (talk) 18:35, 18 April 2008 (UTC)

Oh well. Since you do not present an argument, I guess this conversation is over? --Kim Bruning (talk) 22:23, 18 April 2008 (UTC) a revert not having consensus is a trivial consequence of the above. A revert is an edit, but the edit you revert already shows that the version you are reverting to does not have consensus. (Example: v1: "Peter pier picked a peck of pickled peppers"-> v2. "Peter piper picked a peck of pickled peppers." v3 {rv to r1}. Which version has consensus, or is most likely to gain it: v1, v2, or v3?)

Conversely, in a recent example at WP:NFCC, a group of 7 people thought they'd gained rough consensus on a policy. However, when they tried to implement their agreement, they simply got reverted, the situation got escalated, and they completely failed to get their preferred change pushed through.

--Kim Bruning (talk) 00:14, 17 April 2008 (UTC)

Need: "Widely acknowledged as extremist by reliable sources"

We're having this discussions in Wikipedia_talk:Reliable_sources#extremist_sources. This article says: Questionable sources are those with a poor reputation for fact-checking. Such sources include websites and publications that express views that are widely acknowledged as extremist, are promotional in nature, or rely heavily on rumors and personal opinions. We are discussing that this should be revised to say either widely acknowledged as extremist by reliable sources or widely acknowledged as extremist by a number of reliable sources or even widely acknowledged as extremist by a preponderance of reliable sources. This is because since the definition of who or what views are extremist can be so subjective and politicized. I think I'd like to put in the middle one - "number of reliable sources" - so if you object, pony up in next few days. :-) Carol Moore 23:31, 17 April 2008 (UTC)Carolmooredc {talk}

People must agree cause haven't commented, but i'll give it another day before put in the article. Carol Moore 15:34, 21 April 2008 (UTC)Carolmooredc {talk}

Didn't see this. Strong agree with "number of". But "preponderance of" can be defended, because if it's truly extreme then a majority would consider it to be so. If it appears a minority of "reliable" sources describing an "extreme" source X, it might also be the case that it's a minority of extreme sources describing a reliable source. Majority does not decide reliability, but it does decide center/edge concepts relatively well, while minority doesn't even do that very certainly. I like how this edit is making much more clear that claims of "questionable source" also arise from sources that, thereby, might possibly be somewhat questionable themselves. Remember when Drudge listed how he heard every allegation of source unreliability from another source, in a long chain of allegation? JJB 15:11, 24 April 2008 (UTC)

Exceptional claims and conspiracies.

The present text includes among exceptional claims:

• claims that are contradicted by the prevailing view within the relevant community, or which would significantly alter mainstream assumptions, especially in science, medicine, history, politics, and BLPs. Be particularly careful when proponents of such claims say there is a conspiracy to silence them.

When I first encountered it I was puzzled by the intended meaning of "be particularly careful"; did it mean be especially cautious about triggering the editor's sense of persecution or be particularly alert since belief in conspiracy is a sign of an exceptional claim?

I checked the phrase back to it's origins in WP:RS and found that it originally said:

• Claims which are not supported, or which are contradicted, by the prevailing view in the scientific community. This is especially true of claims whose proponents consider that there is a conspiracy of "official science" to silence them.

Although the phrase "be particularly careful" has been around since 2006 and reflects well-established consensus, I think a change to "be particularly alert" would be somewhat less ambiguous and better reflects the original sense. --SteveMcCluskey (talk) 13:52, 18 April 2008 (UTC)

I agree it is clearer. DGG (talk) 15:45, 18 April 2008 (UTC)

Agree. JJB 15:12, 24 April 2008 (UTC)

Where are the teeth?

In the past the nutshell stated clarly "or [unreferenced claims] can be removed by any editor". It's gone now... why? --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 22:37, 18 April 2008 (UTC)

It is still in the main body of the Policy: "Editors should provide a reliable source for quotations and for any material that is challenged or is likely to be challenged, or it may be removed." Blueboar (talk) 12:10, 20 April 2008 (UTC)

Personal Web Pages

NOT BLOGS which tend to fall into the rant category. No the web sites created by individuals where they disseminate their knowledge on their interests. I consider such sites to be as reliable as those officially sanctioned by wikipedia.

Only when my contribution [Retr0 Retr0] was bounced did I really find out about this. Why? Because included in the references is my own site siltec which has been up and running since early 1991.

As 'reliable' sources includes newspapers I had to question the rest. Surely it comes down to the researchers equivalent of caveat emptor - let the researcher beware - of any source.

If the source looks reasonable, seems reasonable and does not break rules of good taste then why should a personal web page, or even blogs, be considered no more reliable than the so called verifiable sources.

An academic paper might only ever have been read by two or three people. Web pages by hundreds, if not thousands.

It seems to me that this approach is going against the open common sense concept of wikipedia. OK, it might be a site for the self promotion of an original concept. But is NOT within wikipedia so wikipedia should be able to reference that site. It might just be that it is the information that someone is seeking.

As wikipedia itself is not regarded as a reliable source by many within academia, and for the very good reason that it is open to change by anyone, to condem sites with information harder to change than on wikipedia seems a bit daft.

After initially condemning wikipedia is academia now trying to hi-jack it into the traditional ways of an academic elite under attack from an open world. —Preceding unsigned comment added by Wahiba (talk • contribs)

Personal webpages are definitely not reliable sources. The reason is simple: we have no way to know how accurate and reliable the material on any given personal webpage is... It might be full of well researched, scholarly material, etc. Or it might be someone's ill informed speculation. Your personal web page might be an example of the first type... but what about the personal web page of someone who disagrees with you? How are we to judge? The fact is, we can't. The best we could do is state that the website reflects the website owner's opinion... but then we would have to ask if that opinion is notable. If website owner were an acknowledged experts in a given field, writing about that field, an exception could be made... but otherwise, no. Blueboar (talk) 12:04, 20 April 2008 (UTC)

Verifiability not truth in the press

Found via this digg discussion: "Wikipedia's zealots"

Looks like public opinion so far is that this makes wikipedia less reliable?

So we either need to explain what's going on, or think harder. :-)

--Kim Bruning (talk) 15:35, 21 April 2008 (UTC) any obscure references to Feynman entirely intentional

Will Bulten's edit adding the language from WP:ATT help? ("whether material is attributable to a reliable published source, not whether we think it is true") My guess is it will probably help, and certainly won't hurt, and it leaves the "verifiability, not truth" slogan that a lot of people like. I'm really busy with wikistuff at the moment but I'll check back in a few days; I hope this survives. - Dan Dank55 (talk)(mistakes) 20:54, 23 April 2008 (UTC)

The last line of the "news"paper story reveals the reality is based in a WP:COI. The editor is a staffer for a lobbying organization, misusing WP. LeadSongDog (talk) 12:53, 24 April 2008 (UTC)

Specifying Measures of Advocacy Group Reliability

One of the most frequent debates here is whether actual or defacto advocacy groups and publications and web pages are reliable sources. For example currnet (and often repeated discussions on Wikipedia:Reliable_sources/Noticeboard: Antiwar.com;Political Research Associates (and often Chip Berlet); CAMERA; the Nation; CounterPunch; Monthly Review; as well as FrontPage Magazine; Daniel Pipes; ADL; LewRockwell.com; Jewish Virtual Library, etc.

Is it possible to set up some objective measures by which to judge their reliability to avoid constand debates on talk pages, noticeboards, PRAs, etc? Carol Moore 15:48, 21 April 2008 (UTC)Carolmooredc {talk}

I really don't think we can set up objective measures beyond what we now have. The reliability of advocacy group publications often depends on the specific group and publication in question. Some have a reputation for factual (if biased) accuracy, others for playing fast and loose with the facts. I can certainly agree that anything cited to an advocacy group should be attributed so that readers know who is stating what. Beyond that, I think we simply have to hash it out at an article by article and citation by citation level. Even the worst of them are going to be OK as an SPS in articles about the group. Blueboar (talk) 16:40, 21 April 2008 (UTC)

Such groups are going to be judged by academics and other assessors. In general it's a problem to use them as authorities to back up controversial claims; however, if they are the origin of statements the truth of which is accepted/conceded by all parties, then they are acceptable as a source for that. For example the research of the Alan Guttmacher Institute is generally accepted even though they are an arm of the blatantly political Planned Parenthood. If such a group is used as a secondary source, their citation should be checked; and if they don't have a citation, the claim shouldn't be used. Mangoe (talk) 13:44, 24 April 2008 (UTC)

I suppose this comes down to whether we are using a given advocacy source as a primary source for a statement about the advocacy group's claims (ie a statement of their opinion - as in: "The position of X Advocacy Group on Y issue is that 'Y is bad' <cite to advocacy group publication>") or whether we are using the source as a secondary source to back a statement of fact (as in: "it has been shown that Y is bad "). The first is probably OK under this policy (although it might have NPOV issues)... the second is not. Blueboar (talk) 15:14, 24 April 2008 (UTC)

But I presume it would be fine to say "It has been shown that Y causes warts in a dozen clinical trials ", if there are no reliable objections to X's research? If there is a reliable objection, of course, that takes priority over X Advocacy Group and it must be framed as "Study by X says" with the reliable source mentioned separately. JJB 15:27, 24 April 2008 (UTC)

I would recommend against that, the information is only semi-reliable because of the source, not because it is refuted. If LewRockwell.com claims that Ron Paul is winning 50% of the vote among left-handed Hindus it should still be attributed as no one is likely to bother refuting it. Burzmali (talk) 16:50, 24 April 2008 (UTC)

Right, thanks, meaning-- "if there are no reliable objections and the research is significant enough that the nonadvocacy community can be expected to have commented if it were objectionable". Actually, WhoWouldTheWorldElect.com puts him at only 37% in India. JJB 18:11, 24 April 2008 (UTC)

No, the nonadvocacy community may choose not to comment because they often ignore fringe claims. Silence isn't agreement. It's the same as sending your credit card company a letter saying "If you don't respond to this letter in 10 days, you must forgive my debt", you are creating a false dilemma. Burzmali (talk) 18:36, 24 April 2008 (UTC)

Thanks for responding. Is there some way to put that all into two or three sentence paragraph in this article? (I'll try myself in couple days after recover from head cold.)

Also, this reinforces idea of how useful it would be to have a List of previously discussed sources -- which I have proposed with sample list at the link. Carol Moore 00:22, 25 April 2008 (UTC)Carolmooredc {talk}

Someone (other than Burzmali or me) should probably add the graf to WP:RS instead: easier entry. (He and I disagree in that he thinks we disagree and I think we don't.) JJB 04:04, 25 April 2008 (UTC)

I look forward to returning to this subject after some form of WP:Flagged revisions comes in. Any data-mining that helps identify vandals could also, in theory, help identify sources that are often reverted; thus we would bypass the argument over whether a source "is reliable" and go straight to whether it's accepted by Wikipedians, with big piles of data either way. - Dan Dank55 (talk)(mistakes) 14:32, 26 April 2008 (UTC)

Verifiability of fiction

Don't you think, that when it comes to fiction verifiability shouldn't be taken so seriously? Strong, independen sources are something that fiction often just doesn't have. Where it hurts when Hermione Granger's article contains some not 100% accurate information? Pseudohuman (talk) 18:14, 21 April 2008 (UTC)

The entire purpose of a reference work is to present accurate information.... Articles on works of fiction can include content that is cited to primary sources but information in articles should always be accurate and verifiable, regardless of the topic. Black Falcon ^(Talk) 19:41, 21 April 2008 (UTC)

I agree. I was just confused by some discussions, and thought that primary sources are not good enough even in case of fictitious subject. Now i'm wiser. Pseudohuman (talk) 20:16, 21 April 2008 (UTC)

They are good enough for basic statements. If you are going to include something larger, such an an analysis of the freudian psychology of the character, you need to have a secondary source that discusses it. Blueboar (talk) 20:26, 21 April 2008 (UTC)

There seems to be a common habit on wikipedia of people insisting on secondary sources for everything; they are not required. They are better for most things, worse for a few things. We also generally require independent (third party) sources for a lot of things, but not for the things that primary sources are okay for. I've also known some people to not understand the difference between the primary/secondary distinction and the first-party/third-party distinction, which tends to confuse things more. SamBC(talk) 20:29, 21 April 2008 (UTC)

Needed clarification

I believe this addition is advisable, because WP:SELFPUB is so often misapplied in an attempt to include questionable material which it actually excludes. And this doesn't really change the policy, it only emphasizes what is already there. 202.109.116.3 (talk) 19:13, 22 April 2008 (UTC)

Should WP:ATT be marked as historical?

We are (once again) debating whether WP:Attribution should be marked as "historical". The page states that it is not a policy itself, but is a summary of two policy pages (WP:NOR and WP:V). The question has been raised as to whether it accurately does this. Please pop over to WT:ATT and share your thoughts. Blueboar (talk) 11:29, 24 April 2008 (UTC)

Frieda Harris

The article is a direct, verbaitim copy of this webpage. This is apparently not a copyvio as the talk page states that permission was granted to copy it. But I am still bothered by having an article that is simply a cut and paste copy of another webpage. There is only one source listed for this article (the webpage that is copied)... and that source does not list its sources. This makes me wonder if it is reliable (I have raised that issue at RSN). The question for this page is: Does this article really meet our Verifiability standards? Blueboar (talk) 12:59, 25 April 2008 (UTC)

I read "© Supertarot 2008" at the bottom of that web page. Unless there is a verifiable release and free-use license, the article should be stubified. I see an OTRS ticket with the release. So, what is the problem? ≈ jossi ≈ (talk) 00:36, 26 April 2008 (UTC)

It looks to me as if the deprecated {{confirmation}} template was incorrectly added to this article without the required but unsupplied an OTRS ticket number in this edit. -- Boracay Bill (talk) 01:17, 26 April 2008 (UTC)

I assume good faith, and thus will assume that the material has been released and that the ticket info is simply incorrectly formatted. It simply bothers me to see an article on Wikipedia that is essentially just a cut and paste job from another website, and I am wondering how to deal with situations like this. I don't think it is appropriate for any article to take so much of its text from another source (quotes are perfectly OK... even large quotes... but not an entire article). I raise the issue here simply because I can't think of another talk page/noticeboard that might apply. Blueboar (talk) 16:04, 26 April 2008 (UTC)

I've left a note at User_talk:Splash#OTRS ticket number needed (he placed the {{confirmation}} template). I see that he has been only moderately active lately, so I guess that it may be a few days before he reacts. -- Boracay Bill (talk) 01:11, 27 April 2008 (UTC)

I'm not sure it matters--like many copies of material from websites with a nonencyclopedic purpose, the article is essentially unverified and untenable in its present form.DGG (talk) 03:16, 27 April 2008 (UTC)

Note that I placed that tag 2½ years ago, when the tag had a different set of instructions and indeed, had only just been created and was neither deprecated nor used incorrectly. In particular, the procedure at the time was to obtain permission, forward the permission-granting email to permissionsatwikimediadotorg and then simply place the tag oneself. (OTRS was brand-new at the time). I dealt with several in this way at the time, but I no longer have access to the email facility where I did so from. I presume that OTRS has archives and that a request to an appropriate person will turn up the magic number that is required (I note that Jossi above says he see an OTRS ticket, but it's not clear whether he might mean 'tag' instead). I should say that it is also possible that I viewed the email copy on the talk page as sufficient in its own right, and tagged it on that basis; I have absolutely no recall of this article so long after the fact. Splash - tk 22:16, 27 April 2008 (UTC)

So how should we proceed? Take the article back to a stub?... leave it as it is?... add citation requests on questionable statements? Blueboar (talk) 23:56, 27 April 2008 (UTC)

Example of acceptable use of questionable sources

It would be helpful to include an example of when and where questionable sources can and can't be cited. This seems like a good one. Thoughts? 88.112.32.5 (talk) 19:01, 27 April 2008 (UTC)

Not particularly. Its fairly obvious to most readers without examples, and whatever example you choose is likely to be: based on a particular point of view, and attract lots more examples. The one you chose is a particularly extreme point of view and gives undue weight to that POV. Gwernol 19:03, 27 April 2008 (UTC)

Indeed, that is a good example of how use of questionable sources are is not acceptable. _(1 == 2)^Until 19:10, 27 April 2008 (UTC)