Wikipedia talk:Verifiability

Has Verifiability become the new Truth?

I think in some editors mind's Verifiability is truth and as a result anything presented on the talk page challenging that truth gets labeled as OR regardless of how reliable it is.--BruceGrubb (talk) 18:55, 29 September 2011 (UTC)

Yes, and particularly for some science articles that go into the techinical details this can be a real problem. This is why I wrote up WP:ESCA some time ago. Count Iblis (talk) 00:46, 30 September 2011 (UTC)

Bruce... perhaps you are not aware that OR is allowed to be discussed and explored on talk pages (especially if that discussion is relevant to the process of editing the article)... so it does not matter if something that is being presented on the talk page gets labeled as OR. WP:NOR applies to what is written in the article, it does not apply to what we discuss on the talk page. This is true for all of our content polices - in fact its why they are called content policies (they apply to the content of articles). Talk pages are governed by our behavioral policies, not our content policies. So, as long as we don't violate a behavioral policy, we can discuss just about anything on a talk page... even things that a content policy says should not be added to an article (as another example, there is nothing wrong with discussing unverifiable information on a talk page if need be... even though that unverifiable information should not be added to the article). Blueboar (talk) 01:36, 30 September 2011 (UTC)

I point you to the mess that started as Talk:Weston_Price/Archive_1#Weston_Price_and_Stephen_Barrett_in_their_own_words became Wikipedia:Reliable_sources/Noticeboard/Archive_79#Is_a_paper_.28possible_blog.29_by_a_psychiatrist_valid_regarding_old_claims_regarding_dentistry.3F and degenerated into the barrel of utter insanity that was Wikipedia:Administrators'_noticeboard/Incidents/Problem_on_BLP_noticeboard and Talk:Conspiracy_theory#The_first_recorded_use_of_the_phrase_.22conspiracy_theory.22_dates_from_1909.3F_WRONG.21 as example of editors that do NOT understand this.--BruceGrubb (talk) 04:04, 30 September 2011 (UTC)

Yes, there is a belief that wp:verifiability is truth, and that the purpose of Wikipedia is to report what reliable sources say, without concern for correctness.  I would note that this is a fringe theory for which no reliable source exists to support the view.  True believers in WP:V do not accept discussions of inaccuracy, unless a reliable source can be found that so says.  Unscintillating (talk) 02:13, 30 September 2011 (UTC)

Here's an excerpt from Wikipedia and the Meaning of Truth by Simson L. Garfinkel, Technology Review, MIT, November/December 2008

"So how do the Wikipedians decide what's true and what's not? On what is their epistemology based?

Unlike the laws of mathematics or science, wikitruth isn't based on principles such as consistency or observa bility. It's not even based on common sense or firsthand experience. Wikipedia has evolved a radically different set of epistemological standards--standards that aren't especially surprising given that the site is rooted in a Web-based community, but that should concern those of us who are interested in traditional notions of truth and accuracy. On Wikipedia, objective truth isn't all that important, actually. What makes a fact or statement fit for inclusion is that it appeared in some other publication--ideally, one that is in English and is available free online. "The threshold for inclusion in Wikipedia is verifiability, not truth," states Wikipedia's official policy on the subject."

At the end of the above excerpt, please note what its author has concluded from "verifiabiity, not truth", i.e. "On Wikipedia, objective truth isn't all that important". BTW note the wikilinks I gave above for the author etc. --Bob K31416 (talk) 03:01, 30 September 2011 (UTC)

Anyone who has watched James Burke's The Day the Universe Changed (especially the episode "Worlds Without End") knows there is no "objective" truth--not even in the hard sciences. The concept of an "objective" truth is a Newtonian view of how the universe worked and has been replaced by the relativity of Einstein and the probably of Mach.--BruceGrubb (talk) 07:10, 30 September 2011 (UTC)

Garfinkel's critique needs to be tweaked. Shame he was writing on a wiki. "Wiki truth" is based on "consistency or observability". We look at what can be consistently observed in reliable published sources. That's verifiability. We practice our "science", our pursuit, dare I say, of the truth, at least one step removed so to speak, because we aim to be an accurate tertiary source, not researchers in any normal sense of the term.--Andrew Lancaster (talk) 08:55, 30 September 2011 (UTC)

That's quite a good summary but rather idealistic - in many cases it is far from clear whether something is "consistently" observed in sources, or whether those sources are "reliable" (or even "published"). In making judgments on those questions, our desire to be accurate does (or at least should, and in practice probably normally does) lead us to think beyond just what words we can find on what page.--Kotniski (talk) 09:15, 30 September 2011 (UTC)

Of course it is far from clear. And our "tertiary researchers" are therefore constantly debating these things, as they should be, and policy writing can not or should not seek to stop this. This is what we want them to debate, as per our other content policies. But what WP:V is about is that we do not want editors debating as if conducting primary research concerning what is true, only about what is true about what is published (or otherwise verifiable) about what is true.--Andrew Lancaster (talk) 09:51, 30 September 2011 (UTC)

Doesn't "or otherwise verifiable" open the door to conducting primary research? (I'm not saying it shouldn't; just that I don't think we've succeeded in properly formulating this principle yet, assuming there is one to formulate.)--Kotniski (talk) 10:37, 30 September 2011 (UTC)

I think this has been discussed many times, and I think common sense tells us how to handle the definition of limits between something needing sourcing because it is original, and something which does not. I'm sure we've all seen people edit war based on a demand for proof of how the English language should be written. I do not think WP:V needs to be written in such a way that it implies that things like punctuation or word choice decisions can or should be tagged for sourcing. But even those things are verifiable in the end, one way or the other, so the principle holds always, even if demands for footnotes to explain the meaning of every word would not be justified.--Andrew Lancaster (talk) 12:13, 30 September 2011 (UTC)

(ec) Not if we retain NOR. And we do engage in original research in determining what to include and how to include it, but we do not use our OR in article space. So we think beyond what is on the page, and we discuss that, and we make judgments about how to phrase things so as to not include our conclusions or ideas in aritcle space, but rather to accurately reflect (as best we can) what the best sources (as far as we can determine) say (to the best of our ability to read them). Idealistic in this regard is a good thing--we set a high bar and we will pretty often fail to reach it, but the end results are better than setting a lower bar. Much like what Merleau-Ponty wrote about phenomenological reduction being both desirable and ultimately impossible, the goal sets a good course even if we do not actually reach it. Andrew Lancaster puts it very well in noting that we are about what true in terms of representing what others have determined the truth in a given context to be. I don't know that we'll ever be able to formulate it to everyone's satisfaction, and leaving wiggle room is probably a good idea--we could try to have a rule for every case, but then we need a rule to determine which rules apply, and so on, infinitely trying to refine rules instead of writing articles. --Nuujinn (talk) 12:25, 30 September 2011 (UTC)

Well, we could start by removing the first sentence of this policy (are people still discussing that? has anyone come up with a good alternative yet?), which implies something quite different from what you've just eloquently described.--Kotniski (talk) 13:25, 30 September 2011 (UTC)

Thank you for the compliment. I don't think, myself, that removing the first line outright would help, and would likely cause harm overall--I really like the first line in the lede the way it is. But I do think that Blueboar's proposal as it has been worked out is a reasonable compromise given the problems that others have with it, and I look forward to seeing what the community's reaction to it will be. Once that's done, if anyone is interested, I would be happy to work in a group to hash out an essay on how V, NPOV and NOR intersect with examples on what is considered reasonable and good practice in applying same. --Nuujinn (talk) 20:30, 30 September 2011 (UTC)

@Kotniski, I think I came up with a good alternative--take what we have and streamline it while getting rid of the whole "not truth" headache:

The threshold for the inclusion of information in Wikipedia mainspace (articles, lists, sections of articles, and captions) without exception is verifiability— ie being attributable to a reliable, published source appropriate for the content in question. This requires that all quotations and anything challenged or likely to be challenged be attributed in the form of an inline citation that directly supports the material.(See the discussion about sources in WP:NOR that describes summarizing materials in your own words, leaving nothing implied that goes beyond the sources.) For how to write citations, see Citing sources. Anything that requires but lacks a source may be removed, and unsourced contentious material about living persons must be removed immediately.

This is the guts of what we currently have but restructured so that you don't have the kind of wonky misinterpretations that occurred over on Talk:Conspiracy_theory#The_first_recorded_use_of_the_phrase_.22conspiracy_theory.22_dates_from_1909.3F_WRONG.21--BruceGrubb (talk) 05:41, 1 October 2011 (UTC)

Certainly seems an improvement over the present start to the policy, though I still don't like "the threshold" (which implies a single necessary, and hence sufficient, condition), nor do I like the wikijargon "mainspace" so early on. Then I think the rest of the paragraph gets a bit breathless, trying to mention everything without a proper logical structure. But if it's a choice between this and the present version, I'd go for this. (Fundamentally "verifiability" is still the wrong word if we mean specifically attributability; and this page should still be deforked by merger at least with WP:NOR.)--Kotniski (talk) 10:51, 1 October 2011 (UTC)

Hm, please do both of you keep in mind that in the various strawpolls and rfcs that have been held for the last several months here and elsewhere, about half the editors have wanted to keep the not truth "headache". Personally, I think the first line is neigh perfect rhetorically, but I'm not going to let my individual vision of the perfect interfere with community's desire for some changes, so I will oppose any compromise that removes "not truth" from a prominent position in the policy. Moving it down, as per Blueboar's proposal, is not desirable, but that, coupled with other changes intended to clarify what "not truth" means, seems a reasonable compromise to me. 12:22, 1 October 2011 (UTC)

Why not just say what "not truth" means, without using those words? (Because the words don't really mean what we are using them to mean - there's little point in introducing terms of art purely for the purpose of saying that they mean something quite other than you would expect them to. Same applies to "verifiability".)--Kotniski (talk) 12:52, 1 October 2011 (UTC)

Folks... could we please stop talking about "not truth" in quotes as if it were a term of art (or even a concept) that is discussed in the policy and needs to be defined... the policy is discussing the term (concept) "truth" (in its common everyday meaning)... not the term "not truth"... and says that truth is not a threshold for inclusion. The phrasing may be confusing, but what is clearly intended is: "The threshold for inclusion in Wikipedia is Verifiability - Truth is not the threshold for inclusion." Blueboar (talk) 15:00, 1 October 2011 (UTC)

Blueboar I think the position you are criticising deserves a bit more credit. Verifiability is one type of truth, a truth about what can be verified, and that is still the case. The word "truth" in the first sentence has qualifiers which are "understood". It is personal beliefs about what is true concerning the subjects of Wikipedia articles which is intended, right? If the sentence was expanded out to distinguish the two types of truth being distinguished, it would look more visually cumbersome.--Andrew Lancaster (talk) 15:58, 1 October 2011 (UTC)

If that really is the intention, then it's doubly wrong (or triply, or something - I can't keep track of the number of things that are wrong with that sentence) - anything that implies that there is one simple "threshold for inclusion" in Wikipedia is just wrong. And to pick "truth" out of all the things that it is not - if "truth" is to have its ordinary everyday meaning (as opposed to being some unattainable philosophical ideal) - is quite bizarre, since for the most part we would exclude something that appears not to be true; and if you're also claiming that "verifiable" has its ordinary everyday meaning, then that's wrong as well, since we certainly don't insist that people be able to verify the information (by doing experiments or their own synthesis of primary sources or whatever). We have to face up to the fact that these are terms of art; they are bad choices as terms of art; and we don't even need terms of art here (except for "reliable published sources", I guess). And there is plenty more wrong with the sentence. And the whole policy is a fork of (OK, I said that already). --Kotniski (talk) 17:54, 1 October 2011 (UTC)

I agree with Blueboar. I think it is important to make clear in this and related core content policies that WP does not claim that it contains the truth and that editors should not aruge over what is thr truth, as this is a waste of time. My understanding of "threshold" is that it is the basic or minimum. Not the only requirement or criteria for inclustion, just the starting point. And it does not make sense except in contrast to the "not truth" because of what we are verifying. We are verifying that others believe x to be true and the point is that this is quite different from actually saying that it is true. I suspect that the concern motivating this thread (V=T) is that many users interpret V simply to mean that one can find reliabl sources. This completely avoids the questions, hat is the value or use of these sources? For those editors who just skim V and go diretly to RS, many mistakenly believe that enough RS is "like the truth" or even "the truth." Reliable sources are a means of establishing that X really is a view that many hold. But the point is that what we mean by "verifiable" is that some pople believe x to be true. Anyone who comprehends this sentece cannot mistake it for saying it is the truth. Is this the only criteria for inclusion? No, weight thenb becomes a major question, obviously, but hey, one step at a time. Slrubenstein | Talk 19:25, 1 October 2011 (UTC)

Ack, Blueboar's statement is confounded by multiple meanings, for example, some editors will be looking at Blueboar's sentence and seeing:

• The threshold for inclusion in Wikipedia is Verifiability - Truth^TM is not the threshold for inclusion.

which is close to the reading of:

• The threshold for inclusion in Wikipedia is Verifiability - "Truth" is not the threshold for inclusion.

True believers will be looking at that sentence and seeing:

• The threshold for inclusion in Wikipedia is Verifiability - Truth is not the threshold for inclusion and we don't allow it to be a consideration for inclusion, only verifiability matters.

If there are those here claiming to be high priests with the inner knowledge of what the first sentence means, they need to broaden their theory so that it explains disparate viewpoints, and so that we can build consensus.  Unscintillating (talk) 20:07, 1 October 2011 (UTC)

You miss the point of my request. It was a plea to stop placing the two word phrase not truth in quotes. The two word phrase not truth (when placed in quotes) can be confused with the one word term untruth... and the point I was making is that the current policy says nothing about untruth. That omission is something I have tried to fix in my proposal. I think it is a flaw... but we should not imply that the policy currently discusses something that it doesn't. That's all I was asking. Blueboar (talk) 22:37, 1 October 2011 (UTC)

I won't deny that I am missing the point, meanwhile the point I am taking is that your request is ambiguous.  We need WP:V reform because "untruth" is going into the encyclopedia.  The phrase "not truth" is a figure of speech.  Figure of speech says, "A figure of speech is the use of a word or words diverging from its usual meaning...clarity may also suffer from their use, as any figure of speech introduces an ambiguity between literal and figurative interpretation.  I hope this doesn't sound harsh, but you have to agree that there is still a huge gap in our ability to build encyclopedia-wide consensus, so there may be more here than just a simple disagreement or simple misunderstanding.  Unscintillating (talk) 05:27, 2 October 2011 (UTC)

Blueboar's comments over at Wikipedia_talk:These_are_not_original_research#Accuracy_conflict shows the kind of misunderstanding we have regarding Verifiability and OR:

"If reliable sources say that the phrase was first used in 1980, and you (a Wikipedia editor) discover that the phrase was used before the date ... it certainly is OR to premiere your discovery in an article." (Blueboar)

Take a look at what OR actually states: "The term "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published source exists."

Source A presents as a fact that a phrase existed no earlier then 1920. Source B from 1910 demonstrates the fact that the phrase did indeed exist before 1920. But Blueboar is saying presenting the Verifiable 1910 source (ie a source that exists) is OR--HOW does THAT make any sense?!?

Discovery is NOT OR--if it was then looking for any source meeting Verifiability would be OR and that is just plain nuts. This is the same argument that made the Jesus myth theory a two year exercise in head banging and making the Weston Price article a similar exercise in frustration.

Here is an example of what we are actually dealing with:

"The first recorded use of the phrase "conspiracy theory" dates back to a history article from 1909." was made in a BBC Two article (Plots, paranoia and blame) written by Senior lecturer in American Studies, University of Manchester. RS by any reasonable standard.

"The fact, however, which makes the conspiracy theory completely illogical is that the political leaders in the slave states were not united in support of the southwest- ward movement, nor those in the free states against it." (Garrison, George Pierce (1906) Westward extension, 1841-1850 Edited by Albert Bushnell Hart LLD Professor in history in Harvard University pg 31 again RS by any reasonable standard.

"I must content myself with saying that the class conspiracy theory of economic development may generally be considered false,..." (The Economic review: Volume 1 Christian Social Union (Great Britain) Oxford University Branch 1891 Page 540) also RS by any reasonable standard.

The kicker to all this was in his earlier ABC-CLIO vetted book Knight stated on page 17 regarding conspiracy theory: The phrase first entered the supplement to the Oxford English Dictionary in 1997. (...) However, the entry suggests that the first recorded usage of the phrase was in an article in the American Historical Review in 1909." As good as BBC Two is I think we can all agree that is is not on the same academic quality level as ABC-CLIO. Also the "The first episode of The Conspiracy Files: How Diana Died was broadcast on Sunday, 10 December 2006 at 2100 GMT on BBC Two." that ends the BBC Two article makes one wonder if it was a reaction piece.--BruceGrubb (talk) 06:01, 2 October 2011 (UTC)

Bruce, you are focusing on what OR "is not." The policy page like everything else here has gone through many changes; it used to be clearer about what OR is but it is still clear to me reading the entire page that OR is forwarding one's own views in WP.

Editors will go around and around over how to use the word "fact" which is certainly a term of art - when Comte proposed that science is about facts he also made it clear that facts do not exist without theory. There are still serious debates among scientists and philosophers and sociologists of science as to what a "fact" is but they all distinguish "facts" from "the truth."

Moreover, "not truth" is not a figure of speech. The problem &ndahs; meaning, the source of much confusion and endless pointlss debate – is that "not truth" is shorthand. It is shorthand for "neither truth nor falsehood." The point is that the concept of truth depends on the concept of falsehood; although opposites, this pair of words go together because one only makes sense in relation to the other. When we say Wikipedia is about verifiable views, not truth, the full sentence "Whether or not we include something in Wikipedia depends on whether or not we can verify that it is a significant view, not on whether it is true or false" Editors who know little about science and philosophy often think that when we say WP is not about truth, we imply that it is therefore about falsehood. They are mistaking a categorical claim (it is not about "true versus false") for an empirical or descriptive claim ("That's not true!!"). Anyone reading a policy or principle should expect categorical claims rather than empirical ones. Alas, many of our readers do not understand the distinction, and this being a Wikipedia, these readers are also editors. It is the great paradox of Wikipedia, that it is so often the people who need to read more encyclopedia articles instead try writing them. Be that as it may, as long as we explain that whether or not we include something in Wikipedia depends on whether or not we can verify that it is a significant view, not on whether it is true or false" and that any shorter phrase, or slogan, is simply shorthand for this longer proposition, we would avoid a lot of this confusion. Slrubenstein | Talk 06:46, 2 October 2011 (UTC)

This goes to my contention of Verifiability vs believed truth. The statement 'Knight's BBC Two claim of "The first recorded use of the phrase "conspiracy theory" dates back to a history article from 1909." is wrong because the phrase "conspiracy theory" occurs on page 31 in the 1906 book Westward extension, 1841-185 edited by Albert Bushnell Hart LLD Professor in history in Harvard University , on page Page 540 of 1891 The Economic review: Volume 1 Oxford University Branch article, and on Page 67 of the The American: a national journal: Volumes 19-20 May 10, 1890' is a Verifiable truth. You can go to your local library and Verify that statement is true by checking the three pre-1909 works provided and seeing if they do indeed contain the phrase "conspiracy theory".

Obviously doing this is in the article space is not a good idea but it is still not a believed truth but a Verifiability truth. It certainly doesn't help that WP:NOTOR confuses things a bit regarding SYN via Verifiability: "Comparing and contrasting conflicting facts and opinion is not original research, as long as any characterization of the conflict is sourced to reliable sources. If reliable references cannot be found to explain the apparent discrepancy, editors should resist the temptation to add their own explanation."

The problem is when Verifiability is viewed as truth--the 'it is verifiable therefor it is true' mindset. This results in trips to WP:RSN and other misguided endeavors to try and keep inaccurate information out of article space via the 'if it is inaccurate it does NOT meet RS ie Verifiability' route which really don't help anyone.--BruceGrubb (talk) 07:43, 2 October 2011 (UTC)

The problem is you keep making claims about the truth. Neither you nor any other editor can ever prove to me that x is the earliest recorded use of the phrase "conspiracy theory." All we can say is that the phrase occurs in such and such a source in 1890. Bruce, all you are talking about is providing an accurate account of a particular source and what it says. That is not the same thing as saying that any claim is "true."

As for Kinight's BBC 2 claim, well, first of all we can discuss on the talk page whether any BBC 2 show is a reliable source on historical information. I personally think that no television program, not even Richard Starkey's or Simon Schama's, are reliable sources; if they say anything of importance it is in a book or article that went through a much more rigorous review process than any TV show. But, if I lose this particular argument, there is still no point in saying that Knight's claim is "false." This would be a clear violation of original research as you are making your on synthetic claim. All we need to do is say "Knight claims that the earliest usage is in 1909" and that "Y used the phrase in 1890." We are just providing accurate accounts of what sources say. It is still not for us to make rulings on the truth.

Would we keep a source like BBC 2's show out on RS grounds? First of all, I see absolutely nothing wrong with doing that. Discussing whether or not a source is reliable is one of our jobs as WP editors!!! But in some cases we will include a view that we believe to be demonstrably wrong because right or wrong it is nevertheless a significant view. We will always have to say that "many people today believe that in Columbus' day, people thought the world was flat" because it is a significant view. They key is not to judge whether it is true or false, but rather to attribute this view properly, and to provide other views. This is our system. It works. Any editor who introduces the concept of "truth" is unnecessarily screwing with a working system. The concept creates more problems, and solves none, because so far all the problems you bring up are easily managed through the proper application of current policy. Slrubenstein | Talk 08:14, 2 October 2011 (UTC)

"All we need to do is say "Knight claims that the earliest usage is in 1909" and that "Y used the phrase in 1890." We are just providing accurate accounts of what sources say." Well this example has been labeled as SYN OR by some editors:

"Peter Knight states "The first recorded use of the phrase "conspiracy theory" dates from 1909" (ref); however, the phrase "conspiracy theory" also appears in Garrison, George Pierce (1906) Westward extension, 1841-1850 Edited by Albert Bushnell Hart LLD Professor in history in Harvard University pg 31(ref) and The American: a national journal: Volumes 19-20 May 10, 1890 Page 67(ref)"

Mind explaining how the above is OR and your example isn't as they are basically the SAME EXACT THING!--BruceGrubb (talk) 10:09, 2 October 2011 (UTC)

You do not understand what the word synthesis means, and I am beginning to wonder about your understanding of NPOV and NOR. In my example, I simply observe that source a says one thing, and source b says another thing. This is not a synthetic statement. It is simply following NPOV, which demands that we include all significant views (for present purposes I am assuming that both sources are reliable and both views are significant. I am using the examples you provided simply because they are the examples you introduced. Obviously were we editors working collaboratively to decide that one of the views is fringe or one of the sources unreliable, we would not include it. But you were the one who raised the question of how to deal with contradictory claims. My point is only addressing the question of what to do when we have two reliable sources each presenting significant views that are contradictory or in some way irreconcilable.) NPOV requires us to include all significant views not "despite" their being irreconcilable; on the contrary, it is because two significant views are irreconcilable that we must include both (or all if more than two). This is what makes our article not only neutral, but so informative - we strive to include all significant views. Doing just so is never in violation of NOR.

You did something more than what I did. You wrote and I quote, "the statement 'Knight's BBC Two claim of "The first recorded use of the phrase "conspiracy theory" dates back to a history article from 1909." is wrong because the phrase "conspiracy theory" occurs on page 31 in the 1906 book Westward extension, 1841-185 edited by Albert Bushnell Hart LLD Professor in history in Harvard University ..." You wrote the words "is wrong because." You were not quoting another reliable source, you were forwarding you own argument based on your reading of different sources. This is "synthesis." A synthetic proposition is a proposition whose predicate concept is not contained in its subject concept. In your sentence, the subject refers to one source, and the predicate says it is wrong because of another source. The claim that it is wrong is not in the subject. Even if you rearranged it to say "Source A says x and source B says y" (a nominal phrase) and therefore source A is wrong (predicate) it is still a synthetic statement as long as source B does not itself say that source A is wrong.

I really do not know how you can say that you and I are saying the same thing. I am simply complying with NPOV by saying "source A says x and source B says y." I am providing all significant views in a neutral way. There is no synthetic proposition. You, however, are saying that "source A says x and source B says y therefore source A is wrong" This is the simplest, most basic version of a synthetic proposition. Since it is you who are making it, you are not writing in a neutral way and you are also violating NOR by introducing your own argument into the article.

As NOR explains, we can include synthetic propositions as long as they are themselves significant views found in a reliable source. If y="view x is wrong" (the = mans that y explicitly states) we editors of Wikipedia are not saying view x is wrong; we are simply reporting that there is a significant view in a reliable source stating that view x is wrong. This is wholly consistent with what I wrote; what I wrote is wholly consistent with NOR. What you wrote is not.

If we include a source y and y="view x is wrong" (the = mans that y explicitly states), then: "this is not truth" — we are not saying that this view is true or false, we are saying that it is significant; "this is verifiability" because we can verify that out there in the world – that is, not in my mind or your mind or any wikipedian's view – we canverify that someone important, or a large number of people, hold this view. This is the point of "not truth, verifiability." It's pretty simple. Slrubenstein | Talk 11:03, 2 October 2011 (UTC)

This should probably be discussed at WT:NOR or NORN... but I don't think we can comply with NPOV by saying "Source A says 1909 and source B says 1906" ... because in this case source B (Westward Expansion) does not actually say that the term 'conspiracy theory' was used in 1906, it simply uses the word. Westward Expansion is a primary source for claims about when the term was used. Contrasting these two sources with "A says x, and B says y" improperly compares secondary source opinion with primary source evidence. Opinion can be contrasted with counter-opinion, and evidence can be contrasted with evidence, but it is synthetic to contrast opinion with evidence. More importantly, no reliable source has noted this evidence before. Instead an editor has discovered it. To mention it in the article definitely crosses the line into OR.

However, the evidence does exist and it does "prove" that Knight (source A) is inaccurate. While it is improper to mention this "proof" in the article, it is not improper to discuss it on the article's talk page (we are allowed to discuss OR on talk pages). I think this "proof" is a very good argument for saying that we should omit Knight's opinion as inaccurate.

To relate all this to WP:V... I think this situation highlights that there is a distinction to be made between the single threshold for inclusion (verifiability), and the many thresholds for exclusion (or omission)... one of which is inaccuracy (or, if you will, truth). In other words, truth is not a consideration when determining whether something can be added, but it is a consideration when determining whether something should be omitted. I have tried to address that distinction in my proposal. Blueboar (talk) 13:30, 2 October 2011 (UTC)

As I have stated over at Wikipedia_talk:No_original_research many of these problems are in reality WP:NPOV issues and NOT OR ones. Slrubenstein stated "All we need to do is say "Knight claims that the earliest usage is in 1909" and that "Y used the phrase in 1890." and I asked why editors consider "Knight claims that the earliest usage is in 1909, however Y used the phrase in 1890." as OR and got nothing that addressed the issue I actually asked about.

"Sources should directly support the information as it is presented in an article, and should be appropriate to the claims made." (sic WP:RELIABLE and also presented with no bolding in WP:SOURCE).

"Peter Knight states "The first recorded use of the phrase "conspiracy theory" dates from 1909"" is directly supported via Knight, Peter. "Plots, paranoia and blame". BBC News 7 December 2006

"the phrase "conspiracy theory" also appears in Garrison, George Pierce (1906) Westward extension, 1841-1850 Edited by Albert Bushnell Hart LLD Professor in history in Harvard University pg 31(ref) and The American: a national journal: Volumes 19-20 May 10, 1890 Page 67(ref)" statement is also directly supported by the sources presented.

By Slrubenstein's own example everything is alright with the world but we use the word however to join these two directly supported statements and everything supposedly goes pearshaped into SYN-OR land. The use of the "wrong" conjunction suddenly messes everything up?!? Does this argument really make any sense?--BruceGrubb (talk) 14:08, 2 October 2011 (UTC)

@BruceGrubb, I wrote that your 07:43, 2 October 2011 included a statement that was SYNTH and violated OR. I was clear about hat I was referring to. And I was clear about how it is SYNTH and an OR violation.

Your argument is premised on the claim that Knight is a verifiable authority on this specific matter, the date of the first usage of the phrase. I do not agree with your premise. I will explain in my reply to Blueboar.

@Blueboar, I agree with the point you make about the 1906 case. However, I was deliberately shifting to abstract language (source A says view x) to make a point about what our policy is, as this is the policy page. That is, our talk here has to be directly towards improving the policy page e.g. through clarifying it - this really is not the right place to discuss other articles.

For what it is worth, I left a lengthy comment at the article on Conspiracy Theory talk page. That article should not use Knight, not because he is wrong per se but because he is not an expert on this specific issue. "Verifiability has long meant that we verify that x holds a certain view, and that we verify that x's view matters. Perhaps the policy, as it has changed over the years, no longer makes this clearly enough. Knight has a degree in literature and is also an expert on popular culture. His own research is on the interpretation of conspiracy theories, and for this reason his is a verifiably significant view on the meaning of conspiracy theories, especially those he has analyzed. But he is not a historian or philologist, and this mans that his is not a verifiably significant view on the dating of English language usage.

As an analogy: a biologist is not a historian of science, and while she may be an expert on the current understanding and uses of evolutionary theory, her views on the history of evolutionary thought may well not stand up by the standards of academic history. An expert tennis player may not know the etymology of "tennis." For editors to apply "verifiability" appropriately and effectively (the effect being a top-notch encyclopedia) it is not enough for them to find a quote in a book, even a quote in a book by a scholar on the topic. Academe, an the study of human beings both culturally and biologically, is now so specialized that someone with a PhD in biology may not be an appropriate source for views on all things biological.

A couple of years ago James Watson, who was one of those awarded the Nobel Prize in medicine for the discovery of the structure of DNA, made a comment about the innate intelligence of some group (I forget, women or blacks). Needless to say, some people thought that this was a significant scientific view to be included in the article on race and intelligence. The consensus was not to include it. Watson got the Nobel Prise because he trained in molecular biology and spent years researching the structure of DNA. That makes him an expert on the structure of DNA (and of course whatever else he has conducted scientific research on). Just because he made a comment on blacks and intelligence does not mean that he actually did research. His statement was a personal opinion and didn't rise to our standards. Hey, his Nobel Prize is in physiology and medicine - it does not mean he is an expert on all fields of medicine!!

We really need to make clear to users that sometimes, you actually need to know more than the title of the book or the academic title, to verify that the book or person in question is a verifiably significant view. All I had to do was go to Knight's website to see that his research is not on the history of the concept or phrase "conspiracy theory" so I would not use his book as a reliable source for the first recorded use. Academic books and articles typically make a range of claims some of which are central to the author's argument and based on considerable research and others of which are practically "throw away" lines. Editors need to be able to recognize these distinctions to use sources properly. Slrubenstein | Talk 14:31, 2 October 2011 (UTC)

BruceGrubb should know how academic research works, and how someone with a degree in x, or the editor of a volume on x, really has much narrower expertise. And if we want to write a great encyclopedia, verifiability must mean verifying not only that someone has a view, but that their view matters. Knight's views on the cultural meanings of conspiracy theories matter, but his views on the history of the English language do not. Slrubenstein | Talk 14:33, 2 October 2011 (UTC)

What is surprising is that instead of arguing that we have no definitive source on the first usage of the term (unless the OED provides it), BruceGrubb wants to insert Original Research into Wikipedia. Bruce, the reason we have an NOR policy is precisely because it is inevitable that some WP editors will want to put original research into wikipedia. Aren't the reasons obvious? First, in the whole wide world of knowledge, most things actually have NOT been the object of good research. We need a lot more original research! Second, many Wikipedian's are good researchers and thus tempted to do original research. It is because people will want to do it that it is proscribed; there is no need for a proscription against something no one wants to do. But our NOR policy is about a simple commitment: that all that OR that needs to be done should be done through other venues. If you want to do original research you are welcome to do it, you just cannot use Wikipedia as a venue for self-publication. WP is for publishing established research. It is as simple as that. Slrubenstein | Talk 14:38, 2 October 2011 (UTC)

@Slrubinstein, Ironically you claim to know the truth that "not truth" is not a figure of speech.  Your belief system does not make this true or correct or accurate.  Nor does my opinion have standing as being "correct" given that you/others disagree.  What we have is evidence, which I hope you will consider.  To repeat, [Figure of speech] says, "A figure of speech is the use of a word or words diverging from its usual meaning".

Some editors report a "jolt" from reading the first sentence of WP:V.  This jolt does not come from the literal reading of the words, the jolt comes from seeing a meaning diverging from the usual meaning.  The jolt is the contrast between an expectation that an encyclopedia should be truthful to the extent possible, with (what can be read as) a stark denial that Wikipedia is intended to strive for truth.  This is the key element that makes "not truth" a figure of speech.  But a figure of speech also has a literal meaning.  Continuing to map this out, we find that the literal meaning is itself multiply ambiguous.  (1) Some readers see the words "not truth" and read it as "not Truth^TM", meaning that the sentence is targeted toward blocking crackpot theories from being discussed.  (2) Other readers see "verifiability, without regard to truth or the lack of truth" as a staging point from which they can use WP:IAR to backdoor truth into the encyclopedia.  (3) Yet other readers take the literal reading on its face value that truth is not a consideration at Wikipedia.  This last, IMO, is a destructive view that has been propogated into other policies by removing discussion of truth or accuracy therein.  Unscintillating (talk) 16:25, 2 October 2011 (UTC)

Unscintillating, if figure of speech means just what you say, then I can only conclude that what "truth" usually means to you and what it usually means to me are quite different things. Be that as it may, I never claimed "to know the truth that "not truth" is not a figure of speech. " I have only claimed to have an internally consistent understanding of Wikipedia policy that has matched my experience editing here over the past ten years. I am not surprised that readers of WP who wish to edit Wikipedia feel "a jolt" when they read our policies. I think that they would feel a jolt if they stood graduate level courses on philosophy or in the sciences, where such bracketing of truth claims is common (and in the sciences, standard). I also think that they would feel a jolt if they ever tried to write a scholarly article for publication in a major peer-reviewed journal. Most people consume information without really examining what they mean by knowledge, and most people read newspaper accounts of scholarly research or encyclopedia articles, without having any comprehension of the work that produced the knowledge and researchers' own understandings of their work. This sometimes has some really contentious consequences, when you see people in the US and many other parts of the world who do not believe in evolution or global warming, or people who believe in it but don't really understand how it happens - dangerous controversies when these same people vote for those who make policy. Other times, it is utterly inconsequential. But here at Wikipedia we want to manage an encyclopedia that anyone, including people who do not have a solid grasp of the sciences or humanities, and people who have not gone through the long training usually required to, say, write an encyclopedia article. And that requires policies that work. It is not at all surprising that anyone who has never written an encyclopedia article would feel a jolt, or that people who are used to writing peer-reviewed journal articles and have a good grasp of the conventions of academic writing might also feel a jolt when they have to collaborate with people who have no knowledge at all about these conventions. Wikipedia is still a relatively new idea. It certainly is unique. One thing a scientist knows when encountering something new is to be open-minded and expect something new. Why would people who arrive at Wikipedia fir the first time, and have no experience in writing wikipedia articles not expect new groundrules? Frankly, this is what gives me a jolt. If I saw you playing a game I had never played and I asked if I could join in, I would expect the rules to be different from football, baseball, or basketball. That anyone might come here thinking that they might not have to learn entirely new groundrules really surprises and baffles me. This is a very particular project. Whatever rules you use when you hang out at a coffee shop or pub or at home with guests, talking about whatever you believe about God or whether people should be forced to go to school or why are some people gay — why on earth would anyone expect that the same rules apply here? So strange. Slrubenstein | Talk 17:34, 2 October 2011 (UTC)

Yes. Citation to reliable references (verifiability) is not an ordinary element of writing, nor is it an ordinary element of truth. Moreover, it requires some effort, not generally expected of people (even those people who know the truth). Yet, we require it. Alanscottwalker (talk) 18:24, 2 October 2011 (UTC)

The problem here is as Slrubenstein shown with his "Knight is a lecturer in American Studies and his expertise is literature and popular culture. His expertise is not history or philology." comment is that editors are not always clear what a expertise in certain field entails. Expertise in literature and popular culture would logically include the history of these two fields (ie when a word became popular or first appeared). By Slrubenstein's logic only historians can write reliable histories--the reality is most histories of fields are written by non historians.

By Slrubenstein's "logic" Pallasch, Thomas J. DDS; MS, and Michael J. Wahl, DDS (2000) "The Focal Infection Theory: Appraisal and Reappraisal", Journal of the California Dental Association is totally unreliable for the history of focal infection theory because Pallasch is NOT a historian. That is just plain nuts.--BruceGrubb (talk) 18:11, 2 October 2011 (UTC)

First, it is you who has been insisting that Knight's statement about the history of the phrase is unreliable. So, which is it: is Knight right or is he wrong? If he is wrong, you are just proving my point. What is nuts is that you will say Knight is wrong then you will say he is right, whichever one at the time you think will somehow justify your desire to use WP as a venue for self-publishing your original research.

Second, please do not put words into my moth. Please do not fabricate fictions. Pallasch and Wahl make claims about a theory. is their account of the history good history? I have no way of knowing. If someone found evidence that they were wrong, as you found regarding Knight, I would know. Or, if a real historian wrote a critique of their work, or presented a contrasting history, I would know. I await empirical evidence. Then I apply the policies in a reasonable fashion. But at this point it is clear that you just don't like WP policies. Slrubenstein | Talk 19:33, 2 October 2011 (UTC)

Random break 1

To put it as blunt as possible--is it verifiable that the phrase "conspiracy theory" appeared before 1909--yes or no? We don't want to hear anything else but if the statement "the phrase "conspiracy theory" appeared before 1909" is verifiable. Yes or no.--BruceGrubb (talk) 03:50, 3 October 2011 (UTC)

At the risk of jumping into the middle of this conversation, I think the issue is that we can agree on the Discussion page, based on the evidence and as per WP:Inaccuracy, that there is a strong likelihood that "conspiracy theory" appeared before 1909, but we can't make this statement on the Article page because we have no source for such a statement, i.e., the statement is not wp:verifiable.  Unscintillating (talk) 14:56, 3 October 2011 (UTC)

As I understand the issue as described, we do have a source for such a statement, namely the very source that uses the phrase before 1909. Sometimes (and this is yet another thing wrong with the opening statement of the policy) we can make statements not because the statements themselves have appeared in a reliable published source, but simply because we can see they're true by inspection of a reliable source (we do this every time we say "X has written Y" with a citation to X). --Kotniski (talk) 18:45, 3 October 2011 (UTC)

Eh, yes, we expect editors to be able to read.  You've reminded me of something I wrote at Pandora Reef, "The name Pandora Reef dates back to at least 1925.^[1]".  This edit may contradict my previous post here.  Comments?  Unscintillating (talk) 19:03, 3 October 2011 (UTC)

If so, then I firmly believe that your edit was right, and your previous post (and, once again, the first sentence of the policy) has it wrong.--Kotniski (talk) 19:08, 3 October 2011 (UTC)

This all goes back to how OR itself is defined: "The term "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published source exists."

Source A presents as a fact that a phrase existed no earlier then 1920. Source B from 1910 demonstrates the fact that the phrase did indeed exist before 1920.

Regarding the Pandora Reef example above I popped "Pandora Reef" (with quotes) into Google books and found these pre-1925 references to it:

"Pandora reef: 5 1/2 miles West of Fly islet, is a dangerous coral reef, nearly three-quarters of a mile long east and west, and a quarter of a mile broad." (1920) Australia Pilot United States. Hydrographic Office Page 414

"Pandora reef, 5 1/2 miles West of Fly islet, is a dangerous coral reef, nearly three-quarters of a mile long east and west, and a quarter of a mile broad." (1889) Australia directory, Volume 2 Great Britain. Hydrographic Dept pg 300--BruceGrubb (talk) 20:20, 3 October 2011 (UTC)

Unscintillating, I think your edit at Pandora reef was fine, and I would not object to it being changed to say "at least" 1889 based on the source that Bruce Grubb presented above. But this is not a parallel example to the one from the Conspiracy Theory article, where the issue is SYNTH. There's no SYNTH in your example. We do have primary sources that Bruce Grubb presented that establish clearly that the term was used prior to 1909, there's no debate about that and that's why we removed the date from the article. We may put it back, but we have to figure out how to do that without violating OR, and Hans Adler made a good suggestion as to how to approach it. Or we could simple use a statement similar in structure like "the term was in use as early as X", but I personally find that less interesting since the connotations associated with the term have changed over time. So I don't see a V issue in regard to the discussion Bruce Grubb is presenting as a problem. Indeed, given that the matter was settled in a few days, I don't see much of an issue at all, this largely seems to be a tempest in a teapot.

I think the fundamental problem in these discussions is that Bruce Grubb fails to grasp the essence of SYNTH. For example, he said "Actually, per WP:NOTOR (comparing and contrasting conflicting facts and opinion is not original research, as long as any characterization of the conflict is sourced to reliable sources) the second WOULD be allowable but only if it was done like this: Peter David states 'The first recorded use of the phrase 'conspiracy theory' dates from 1909' (ref); however, the phrase 'conspiracy theory' also appears in Garrison, George Pierce (1906) Westward extension, 1841-1850 Edited by Albert Bushnell Hart LLD Professor in history in Harvard University pg 31(ref) and The American: a national journal: Volumes 19-20 May 10, 1890 Page 67(ref)." Leaving aside the NOTOR is an essay rather than policy, that kind of phrasing, strictly speaking and in my opinion, violates SYNTH, in that the juxtaposition makes an implicit judgement of David's work and is based on Bruce Grubb's research into primary sources, and both are things NOR enjoins us not to do. There have been extensive discussions of this issue at the Conspiracy Theory article, and at the talk page for WP:NOTOR following Bruce Grubb's edits of that page, in case any one is interested. For what it's worth. But I do think it's pretty clear that the issue has nothing to do really with V. --Nuujinn (talk) 22:41, 3 October 2011 (UTC)

When NOTOR talks about "as long as any characterization of the conflict is sourced to reliable sources" it is talking about something like this:

• "As Joe Expert notes: Peter David says the first use of the phrase appears in 1909, however the phrase appeared in Westward Expansion which was published in 1906.<cite Joe Expert noting both facts together>"

In this example, we are merely reporting a comparison that Joe Expert made... Joe Expert discovered the 1906 date and uses it to correct the 1909 date ... not a Wikipedia editor. The comparison does not originate on Wikipedia, thus it is not OR for us to note that Joe made it... but, without a source like Joe acting as an intermediary between us and the facts, the comparison is WP:SYNTH, and thus it is OR. Blueboar (talk) 23:35, 3 October 2011 (UTC)

Yes, I think that is correct, and I think if we had just "Joe Expert says the first use of the term appeared in 1705", we could contrast these views by saying something like "The date of the first use of the term is variously given as 1909 and 1705", citing both sources. But unfortunately, we have no such sources. What we do have are the primary sources uncovered by Bruce Grubb, and we have used those to verify that David's assertion is incorrect, and removed the inaccuracy. Would that that were all there was to it. --Nuujinn (talk) 23:46, 3 October 2011 (UTC)

Is that not all there is to it? What are people still arguing about?--Kotniski (talk) 08:40, 4 October 2011 (UTC)

I wouldn't really say most people are arguing at this point, most of the discussions are really about interpreting SYNTH, which is relatively subtle, and on which I readily admit I hold a narrow view. Bruce Grubb has proposed a new guideline, Wikipedia:Inaccuracy (which is a pretty good essay) and has been seeking to alter WP:NOTOR to support his view, and continues to push his view here, at Wikipedia talk:Verifiability/First sentence and at Talk:Conspiracy theory. From his contributions, I take it that this issue seems very important to him, but I think he's misrepresented the nature of the initial mutual misunderstandings of positions and subsequent discussions at Talk:Conspiracy Theory, and spun this up more than necessary or beneficial to the project, characterizing the discussions there as a long battle and a disaster and expressing undue emotion when he believes that other editors disagree with him. I leave it to others as to whether any of our actions rise to the level of tendentious editing, but I just do not see a real issues from Talk:Conspiracy Theory beyond WP:SYNTH and WP:DUE. Certainly I do not see any issues related to verifiability. I have not read through the entire discussion at Talk:Jesus_myth_theory, however, so there may be some meat there, but mostly I see very little chicken --Nuujinn (talk) 11:18, 4 October 2011 (UTC)

Kotniski summed up my position perfectly--"We can make statements not because the statements themselves have appeared in a reliable published source, but simply because we can see they're true by inspection of a reliable source (we do this every time we say "X has written Y" with a citation to X)"

No matter the handwaving done the statement "the phrase "conspiracy theory" appeared before 1909" is verifiable

As for claiming I don't have a firm grasp of SYN I suggest editors read the archives of Talk:Jesus_myth_theory especially Talk:Jesus_myth_theory/definition#Defining_.22Christ_Myth_theory.22_via_the_reliable_sources and try to continue to make that claim. The "directly states" plea was used by several editors to ignore the fact there were conflicts in the very definition of the term which was acknowledged to have several synonyms. That is anther reason that the article was a two year headbutting migraine--BruceGrubb (talk) 07:28, 4 October 2011 (UTC)

Random break 2

This seems to be going back and forwards between a discussion of policy and a content discussion for a particular article. For an encyclopedia editable by anyone, it's important to have verifiability be absolutely, categorically, without exception, a criteria for inclusion. Most folks intend "Not truth" to reinforce this, by specifically addressing and ruling out the most often claimed exception to verifiability. ("But it's true!") IMHO it is well intended, and it does some good, but also misfired into various wrong interpretations in other areas. IMHO the proposed change keeps the "does some good" part, keeps "not truth" with a context explanation, while reducing the "misfires". Sincerely, North8000 (talk) 20:43, 2 October 2011 (UTC)

"The problem is "But it's true!" can be "But it's true, because this source says so!" and that is where the whole "Not truth" argument falls apart. One of the main reasons the Jesus myth theory article was a two year exercise in head banging was that editors would use one source to determine if another source was talking about the Christ-Jesus myth theory. in addition to the the exact phrase nonsense (Talk:Jesus_myth_theory/Archive_33#Wells.2C_Price.2C_and_Doherty) you had editors using other sources and interpreting passages to say what people thought saying rather than what they actually said (talk:Jesus_myth_theory/Archive_34#Background_and_definition_section.

For example, I pointed out that Bromiley stated "This view states that the story of Jesus is a piece of mythology, possessing no more substantial claims to historical fact than the old Greek or Norse stories of gods and heroes,..." and that saying a story of a person is a piece of mythology is NOT the same as saying the person himself didn't exist but some editors were so hung up on what other sources had said that they literally forced Bromiley into saying something he was not actually saying.--BruceGrubb (talk) 04:18, 3 October 2011 (UTC)

Yes, I mean, this is the fundamental point - a fact does not become verifiable because a source says it, any more than it becomes true because a Wikipedian thinks it. So "verifiability", and especially "verifiability, not truth", is just an abuse of language, that doesn't mean what it's intended to mean. Put it together with "the threshold", implying that there's a single condition for inclusion of information (or as we wrongly call it, "material"), and we have a complete mess of a sentence, that no-one can even remotely defend except by saying "there is something behvind it that it's supposed to mean", "we like it", or "it's been here for a long time, so ha ha, we're going to revert anyone who changes it". To me it is ludicrous that one of our core policies, which is being read by the outside world as well as those on the inside, continues to be headed with a sentence so manifestly absurd. --Kotniski (talk) 08:47, 3 October 2011 (UTC)

Well, the proposal to change it, developed over many months, is nearing completion. North8000 (talk) 11:53, 3 October 2011 (UTC)

If you view wp:ver as simply stating a requirement for inclusion, NOTHING more, it is fine with respect to all of the above. The first sentence has caused problems by contributing to lots of other claims and chants that it doesn't say. For example:

• that an editor can use verifiability/sourcing as a mandate to force inclusion of the material.
• that potential falseness of sourced material can't be discussed when discussing possible exclusion/leaving out of material (recognizing that in special cases wp:npov trumps this discussion and provides a mandate for inclusion)
• That meeting verifiability means that it is true.
• That wp:verifiability weighs in on the side of inclusion of material rather than just setting a condition for it.
• That accuracy is not an objective of Wikipedia

North8000 (talk) 12:00, 3 October 2011 (UTC)

If it really is just stating a requirement (not a sufficient condition) for inclusion, why are we not allowed to make that clear, by changing "the threshold" to something unambiguous like "a minimum requirement"? Does anyone actively disagree with doing at least that (I mean have any real arguments against doing it, other than "this sentence is under discussion so it can't be touched")?--Kotniski (talk) 12:49, 3 October 2011 (UTC)

I agree with your logic and reasoning. I think the main venue would be at the talk page of the first sentence page where it is up for final tweaks. My thought is that logically "Verifiability is a requirement for inclusion" says it precisely and totally. But it doesn't provide sufficient impact. Everything else is for additional impact and impression. Anything that does this without doing harm is OK, (e.g. "initial", "fundamental", "threshold", "absolute") although superfluous or imprecise from a logical standpoint. But also folks who said "let's deal with that later, not try to to change too much at once", and I sort of agree with that. North8000 (talk) 13:57, 3 October 2011 (UTC)

But "verifiability is a requirement for inclusion" is still wrong, because the requirement is not "verifiability" in any normal meaning of the word, but (something closer to) "sourceability" or "attributability". However I agree that small change in the right direction is better than no change at all (though things have come to a pretty pass when the rewriting or removal of just one sentence is held up as too big a step to take in one go).--Kotniski (talk) 18:41, 3 October 2011 (UTC)

I don't have a problem with "verifiability is a requirement for inclusion" as long as we explain what verifiability actually means. In my own field of anthropology the term "culture" must have an provided context to understand how it is being used--that is why I suggested this restructuring:

The threshold for the inclusion of information in Wikipedia mainspace (articles, lists, sections of articles, and captions) without exception is verifiability— ie being attributable to a reliable, published source appropriate for the content in question. This requires that all quotations and anything challenged or likely to be challenged be attributed in the form of an inline citation that directly supports the material.(See the discussion about sources in WP:NOR that describes summarizing materials in your own words, leaving nothing implied that goes beyond the sources.) For how to write citations, see Citing sources. Anything that requires but lacks a source may be removed, and unsourced contentious material about living persons must be removed immediately.

Take a good look a how that first sentence in the above paragraph is structured:

1) It establishes verifiability as a threshold requirement

2) It establishes where verifiability actually applies--Wikipedia mainspace (articles, lists, sections of articles, and captions)

3) Finally, it spells out what verifiability actually is--being attributable to a reliable, published source appropriate for the content in question.

This is all there in the current lead but it is broken up into disjointed segments that don't flow together into a unified whole. As a result you have some editors going for "Verifiability soundbites" where one disjointed segment is presented while other equally valid ones are ignored. This restructuring should hopefully minimize that--BruceGrubb (talk) 20:03, 3 October 2011 (UTC)

All of which can be summed up in one simple sentence... "Verifiability--whether readers can check that material in Wikipedia has already been published by a reliable source." Hmmm... I think I have read that somewhere before. Blueboar (talk) 22:35, 3 October 2011 (UTC)

That is one of the "Verifiability soundbites" I was talking about. Note what is missing:

1. It says nothing about verifiability as a threshold requirement

2. It does not state where verifiability actually applies--allowing the it applies to talk page nonsense presented.

This don't address the actual problems raised here.--BruceGrubb (talk) 04:24, 4 October 2011 (UTC)

What do you mean by a "threshold requirement"? Why not just say "a minimum requirement" or simply "a requirement"? But certainly the article should be "a" and not "the" - as pointed out ad nauseam, if we say "the" then the phrase is naturally interpreted as meaning "the only", and is therefore liable to be read as referring to a sufficient condition, which is Wrong.--Kotniski (talk) 08:46, 4 October 2011 (UTC)

The online Oxford advanced learners dictionary defines threshold as the "the level at which something starts to happen"

The general Oxford online dictionary defines threshold as "the magnitude or intensity that must be exceeded for a certain reaction, phenomenon, result, or condition to occur or be manifested"

It is clear that passing a threshold is only the start of the process but given the wonky way editors have used this we only need to add one word to shut the kind of nonsense that you suggest could happen:

The first threshold for the inclusion of information in Wikipedia mainspace (articles, lists, sections of articles, and captions) without exception is verifiability— ie being attributable to a reliable, published source appropriate for the content in question. This requires that all quotations and anything challenged or likely to be challenged be attributed in the form of an inline citation that directly supports the material.(See the discussion about sources in WP:NOR that describes summarizing materials in your own words, leaving nothing implied that goes beyond the sources.) For how to write citations, see Citing sources. Anything that requires but lacks a source may be removed, and unsourced contentious material about living persons must be removed immediately.

That should address everything as far as WP:V is concerned as this

1) It establishes verifiability as the first threshold requirement (implying there are others)

2) It establishes where verifiability actually applies--Wikipedia mainspace (articles, lists, sections of articles, and captions)

3) Finally, it spells out what verifiability actually is--being attributable to a reliable, published source appropriate for the content in question.

There would be the rest of the article explaining all this in detail but this boils verifiability down to its key components: the first threshold, where in Wikipedia it applies (mainspace and NOT talk page), and finally formally defines it.--BruceGrubb (talk) 09:22, 4 October 2011 (UTC)

I still don't see why you want to use such an unnatural phrase as "threshold" or "first threshold" when we can just say "a requirement". --Kotniski (talk) 10:49, 4 October 2011 (UTC)

If you take a hard look at WP:CCPOL you will see that WP:V and WP:NOR are in essence mirror images of each other.

Note the way No original research is defined on that page: Wikipedia does not publish original thought: all material in Wikipedia must be attributable to a reliable, published source. This makes it clear that verifiability is the "first threshold" for inclusion but a work meeting verifiability may not get in due to it failing other policies or guidelines. Let me use an example I presented long ago:

"There is not a shred of evidence that a historical character Jesus lived, to give an example, and Christianity is based on narrative fiction of high literary and cathartic quality." (abstract) and "It is not possible to compare the above with what we have, namely, that there is not a shred of evidence that a historical character Jesus lived." (main text body) (Fischer, Roland (1994) "On The Story-Telling Imperative That We Have In Mind" Anthropology of Consciousness American Anthropological Association Volume 5. Issue 4. December 1994 Pg 16 - 18)

A peer reviewed statement in a recognized professional journal published by the American Anthropological Association--clearly Verifiability...and it fails to pass WP:WEIGHT muster. No matter how Verifiability the above is--it remains an obscure one time hiccup against the mountain of other references that were found. But to even argue for possible inclusion you had to first meet the Verifiability requirement.

This is why verifiability is the "first threshold"--it is the first hurdle an editor must clear to even have a chance of the view being presented to be taken seriously. It, in theory, makes editors realize that they have to produce high quality sources rather than use questionable things like tabloids or self published works.

This why I try to use the highest quality sources I can for my arguments--it addressed the WP:V issue right from the beginning so one can go to WP:NPOV issues.--BruceGrubb (talk) 14:59, 4 October 2011 (UTC)

*comment The reason that "not truth" is important is that it stresses that wikipedia does not make any judgements or evaluations of what is true. We only represent other people's ideas of what is true, and we represent them weighted in accordance with the degree of agreement that it is true. Verifiability means that in order to include statement X we have a source that shows that it is true that some person publishing in an appropriate forum has stated that X is the case. If there is no such source then the statement X is not verifiable. Whether the statement X is true or false is not relevant for whether it should be included in the encyclopedia. False statements can be included, and indeed they have to be included if they are notable, i.e. held by a significant group of people. E.g. wikipedia does not judge whether it is true that evolution produced the diversity of life on earth. Because what matters isn't whether its true, but that most scientists believe it is true. Truth value is simply not a part of the calculation in figuring out what to include. That is what "Verifiability not truth" is supposed to mean, and why it is importance.

Now of course we shouldn't include rumours of dubitable veracity in articles about living persons, but that is not because they are likely to be untrue, but because they are likely to be non-notable and could create potential legal liabilities of a court decides they are not true. The fact that this is the case can be seen by the fact that we can and do include rumours of doubtful veracity as long as they are 1. attributable to sources that can be considered accountable for the statement and 2. have received sufficient coverage to be notable. The reason we want sources of high quality is not that statements in them are more likely to be true, but that they are more likely to be significant, and in BLP cases because the source is more likely to be legally accountable. E.g. if a peer reviewed article writes that celebrity X was running around naked in downtown London we can include that whether or not the statement is in fact true because the journal that published the claim is the one that will be legally liable if it turns out to be untrue, and not wikipedia who merely repeated their statement. ·ʍaunus·snunɐw· 18:54, 5 October 2011 (UTC)

Misinterpretation of "verifiability, not truth" in progress

Certain editors on this page have repeatedly said that they cannot find examples of "verifiability, not truth" being misused. As myself and others have pointed out, it has been misused on a daily basis for years, so they must not be looking very hard. For a current example of this misuse, I would invite interested parties to review the ongoing discussion over at Talk:Karen_Gillan#No Mention of the "Lady Godiva" Incident, where it is being argued that a risque, trivial story about a celebrity found wandering naked and "disoriented" in a hotel must be added to a BLP because 1) it was reported by the Daily Mail , and 2) no RS has denied the story, therefore 3) we must include it. Viriditas (talk) 00:19, 4 October 2011 (UTC)

That example appears to be more of a straight BLP issue, which takes into serious account all three core content policies. It certainly doesn't have to be included simply because it is verifiable. As it says in BLP: "Wikipedia is an encyclopedia, not a tabloid: it is not Wikipedia's job to be sensationalist, or to be the primary vehicle for the spread of titillating claims about people's lives, and the possibility of harm to living subjects must always be considered when exercising editorial judgment." But I truly don't see how "verifiability, not truth" specifically is being "misused" there: that phrase isn't even mentioned by anyone in that discussion. No one's touting it around incorrectly there. If it is being misused, whomever is misusing it is not taking into account the other policies that affect BLPs. Jus' sayin'. Doc talk 00:46, 4 October 2011 (UTC)

You must not have read the same thread. As I explained above, the argument for inclusion in that thread amounts to 1) we can verify a tabloid claim in a purported RS 2) it doesn't matter if it is true or not, and no RS has disputed or refuted it 3) therefore, we must include it. If you do not see this as an explicit example of "vefifiability, not truth" being distorted, then we have a problem. Viriditas (talk) 01:13, 4 October 2011 (UTC)

No, I see it as WP:UNDUE and controversial material (that is also not backed by at least three reliable sources - this covers the "tabloid" aspect of this as well) trying to be inserted in an article about a living celebrity. There's three strikes against including that stuff right there. And, again, I don't see that particular phrase appear even once in all those comments. Yet outside analysis shows that it is somehow being misinterpreted and misused with out anyone there pointing to it and incorrectly attributing it to justify anything that could be considered troublesome. Doc talk 01:25, 4 October 2011 (UTC)

I don't know if there is some kind of institutional policy blindness going here or what, but are you actually saying that you don't see how the argument for inclusion is based on the type of misinterpretation of "verifiability, not truth" that has been discussed here? Because it sounds to me like you recognize it, but are in denial about it. The proponent of inclusion in the thread in question has cited the verifiability policy as a reason to include the tabloid material, and has argued that it does not matter if it is true or not and that no other source has disputed it, therfore, it should be included. Do you see how this is an example of how the current wording of the policy is continually misinterpreted on a daily basis, especially in regards to BLP articles, and how the current wording results in many disputes that could otherwise be avoided simply by changing the wording? I'm having great difficulty understanding how you could not see this. Viriditas (talk) 01:40, 4 October 2011 (UTC)

That argument does not appear in this discussion: a search shows that the word "true" is not even there at all, and the only mention of the word "truth" by anyone there is, "It is not within your wheelhouse to determine if the reliable source has told the truth or verified its sources." Nobody there says "It doesn't matter if it's true or not" - maybe you interpreted it that way, but no one actually said that. I also see no links from anyone to WP:V, and the two mentions of the word "verifiable" are accompanied by "reliable" (and WP:RS is linked) Now, if he specifically pointed to the "verifiability, not truth" phrase and said, "See? My understanding of this phrase allows me to include this and trumps any other considerations." maybe you'd have one decent example here. Instead, you appear to be lumping multiple arguments coming from a user and interpreting (for him) that his reading of the phrase "verifiability, not truth" is the #1 argument for inclusion. Now, one of the many others who is "policy-blind" and "in denial" will have to take over this one; because you are correct that I cannot see this example as an accurate representation of what you are saying. Doc talk 02:32, 4 October 2011 (UTC)

You appear to be in denial. The entire agument for inclusion is based on a misunderstanding or what "verifiability, not truth" means, and is illustrative of dozens, perhaps hundreds of unnecessary disputes that could be avoided by removing this poor wording from the policy. Viriditas (talk) 02:39, 4 October 2011 (UTC)

This is related to a 'no Wikipedia policy or guideline is an island' problem I commented on: too many times a Wikipedia policy or guideline is taken on its own rather than how it should be viewed--as part of a whole.--BruceGrubb (talk) 04:36, 4 October 2011 (UTC)

The examples have been provided again and again and again. North8000 (talk) 01:02, 4 October 2011 (UTC)

I agree with Doc that this "example" doesn't seem to have much to do with WP:V. I also think that shouting that editors "appear to be in denial" because they don't agree that this particular WP:RS and WP:BLP issue demonstrates a crisis with WP:V hurts the case of those demanding a change in the wording of the verifiability policy. Finally I think that overreaching claims that examples have been provided "again and again and again" demonstrates a disconnect with what has actually happened in the course of this seemingly interminable discussion. North8000 has been arguing this for a long time—how many examples has he provided? Quale (talk) 07:16, 4 October 2011 (UTC)

I don't know about examples North8000 himself may have presented but I have presented some. I might add that thanks to the way WP:V was worded BLP itself has been used as a magical censorship hammer as demonstrated by the nonsense that started as Talk:Weston_Price/Archive_1#Weston_Price_and_Stephen_Barrett_in_their_own_words became Wikipedia:Reliable_sources/Noticeboard/Archive_79#Is_a_paper_.28possible_blog.29_by_a_psychiatrist_valid_regarding_old_claims_regarding_dentistry.3F finally escalated into Wikipedia:Administrators'_noticeboard/Incidents/Problem_on_BLP_noticeboard. BLP was being used to prevent any meaningful discussion on the talk pages if Stephen Barrett was reliable regarding the information he provided on Weston Price. It was totally insane how BLP was being used there.--BruceGrubb (talk) 07:43, 4 October 2011 (UTC)

Quale, you must be reading a different thread. In this thread, the example I provided specifically and explicitly deals with the "verifiability, not truth" problem we have been discussing for several years—a problem you and a number of other editors are in denial about. This does not require any interpretation. When a single editor argues that a risque, trivial story about a celebrity found wandering naked and "disoriented" in a hotel must be added to a BLP because it can be verified in the Daily Mail , we are dealing with verifiability. And, when that same editor claims that no RS has denied the story and it is not up to editors to decide if it is factual or not, we are dealing with not truth. Finally, when the same editor says that because of verifiability, not truth, we must include this content regardless of its accuracy or appropriateness, we are dealing directly with a misinterpretation of "verifiability, not truth". You can continue to claim that we aren't, but it is clear, beyond question, that we are. The wording doesn't work and needs to be changed. Viriditas (talk) 08:30, 4 October 2011 (UTC)

I know that I have provided several just in the last few months, and I'm just one of several people who have. North8000 (talk) 10:07, 4 October 2011 (UTC)

Let's be clear here. We're talking about examples where the phrase "the threshold for inclusion is verifiability, not truth" is causing a problem, but changing the phrase to "the threshold for inclusion is verifiability", would make the problem disappear. And you say you have provided several examples "just in the last few months", and several others have as well. In that case, it should be easy to list well over 10 cases where it's clear that removing the words "not truth", would have stopped an edit war, ended disagreements on a talk page, or caused a troll to give up and leave, etc. As a favour, could you point us to some talk page discussions where removing the words "not truth" would have caused this happy outcome? LK (talk) 10:40, 4 October 2011 (UTC)

It's really not necessary to show real cases where a change of wording has had an identifiable effect like this. If the current wording is wrong (and we have seen through countless arguments, including real-world cases, that it is) then we just change it, we don't require evidence that its wrongness has had a harmful effect. Just as when we find demonstrably wrong information in a WP article we change it forthwith; we don't require real-world instances of someone having used this information and suffered as a result. The extreme extent to which it's being attempted to ring-fence this sentence just confirms how much it's become an article of faith for some people, rather than a topic for rational discussion and seeking of a best solution.--Kotniski (talk) 10:46, 4 October 2011 (UTC)

Responding to Lawrencekhoo the examples were where the situation was made significantly worse by "not truth" and thus would be improved by removal or mitigation of that phrase. If you ever see any example of anything that meets the utopia slam-dunk perfection, including proof of what the outcome would-have-been-if-it-were-changed new standard that you just wrote, you have seen a simpler world than I ever have in Wikipedia. :-) Sincerely, North8000 (talk) 10:51, 4 October 2011 (UTC)

I do not think that because "not truth" is misunderstood my some editors means we should get rid of it. What it does mean is that we should explain it better. I think that goes for this whole policy. I do not think "verifiability" simply means that one can find a source. NPOV is our core policy and V and NOR derive from it, which is why I always start with NPOV. NPOV demands that we include all significant views. The "not truth" means that we WP editors are not assumed to be able to distinguish between what is true or false, and thus it is not our responsibility (you may think this is absurd when it comes to a question like, when is the earliest recorded use of the phrase "conspiracy theory" - but most scientists would say that even this simple empirical claim is not "true" because five or fifty years from now someone may discover an even earlier recorded instance. When it comes to "the meaining of the Heisenberg uncertainty Principle," what I just said should be evident to everyone). However, we can verify whether something is a significant view or not. The meaning of "verifiability" derives from what we call neutral editing. We do not claim that such and such is true, we claim that some people hold the view that it is true. But NPOV does not insist we include all views, only the significant ones. Also, the very idea of a reliable source cannot be giving a particular source (say, a tabloid) a blank check for information. The reliable source is the right source for verifying that this is the actual view we wish to include in the encyclopedia, and that it is significant. So "verifiability" must mean verifying all these things: verifying that this is a view actually held by some people; verifying that this is a significant view, and verifying that the source we are using provides us with a reliable account of what this view is. For excample, the New York Times is (I think) a pretty reliable source for the views of politicians and celebrities in the news. It may be highly reliable. But I would not assume that it is the most reliable source on what a particular philosophical school, say, the Epicurians, believed. But Bertrand Russell's A History of Western Philosophy is (I think) a reliable source for that view.

In short, we are verifying that the requirements of NPOV are met. We may not be able to agree on what "the truth" is, but we should be able to agree on what kind of documentation is appropriate and adequate to document that (1) someone or group of people hold the view that x is true and (2) this view is significant and not fringe.

My aim is not to comment on the particular case because the purpose of this page is to improve the policy page. I completely believe that every day many people misunderstand this policy. But the policy makes sense, and what we should be doing is using this talk page to find clearer ways to express it. This is what I am trying to do. We have to accept the fact that some people will always misinterpret our policies - this is the nature of any writing, it is vulnerable to misinterpretation. But we can always try to make our writing clearer. it is worth trying. Slrubenstein | Talk 10:54, 4 October 2011 (UTC)

Yes, I think most of us agree that the wording of this policy needs to be made much clearer, that's what we're trying to do (except for a few serial reverters who seem determined not to allow it). But about your analysis: what do you mean by "we are verifying that this is a significant view" and that "[this is a reliable source]"? Surely we can no more "verify" these things than we can verify the "truth" of the statement itself?--Kotniski (talk) 11:02, 4 October 2011 (UTC)

Reply to Slrubenstein. I think that the main proposal makes "not truth" clearer as you advocate. North8000 (talk) 11:28, 4 October 2011 (UTC)

@Kotniski and others: "most of us agree" – I think this is true, if "most of us" means most of the people still interested in discussing this complaint after nearly a year of scant progress. (I think I saw some dissent about "not truth" on this page in November 2010.) I think most editors who don't see this as The Single Most Critical Issue Facing Wikipedia Today have become burned out by the monomania and (quite sensibly in my opinion) dropped out weeks ago. If you drag this out a few more months you will probably be able to truthfully claim "all of us agree" as everyone else will be gone. Eventually you will come up with a concrete proposal that will be voted on. You will either get what you want or you won't. Either way, I hope that that puts at least a temporary end to the siege of this talk page, but I'm concerned that some people here have come to enjoy trench warfare too much. Quale (talk) 04:34, 5 October 2011 (UTC)

@Quale, didn't occur to you that the so called monomania is because there is a problem? The fact I have just experienced something akin to this in 2011 shows to me at least there is a problem that same editors are in denial regarding.--BruceGrubb (talk) 05:57, 5 October 2011 (UTC)

@LK, Re "it should be easy to list well over 10 cases where it's clear that removing the words 'not truth'..." - Could you list well over 10 cases where the words "not truth" specifically were essential. Please note that just doing a search for "verifiability, not truth" only shows that "V not T" has been used, not that the words "not truth" were essential or even relevant to a particular discussion. Furthermore, doing the wiki-search, noting the number of hits, and doing a simple arithmetic calculation, would show that "verifiability, not truth" is used less than 3 times a day on Wikipedia on the average. Note that there are more than 30,000 active editors. Also note that there are around 4,000 views of WP:V a day, so it's more important to have a clear policy page than to have a questionable slogan like "verifiability,not truth" that is used less than 3 times a day on the average. --Bob K31416 (talk) 13:37, 5 October 2011 (UTC)

I found rather quickly in that list an example of an editor using the phrase the way it's been understood for many years. "Verifiability, not truth. It doesn't matter if what you're saying is correct. You can't verify it with, "someone told me so."[1] It really should be no more complicated than that. Doc talk 13:59, 5 October 2011 (UTC)

That's stating the part that everyone agrees on, and was the intended meaning of "verifiability, not truth". The problem is the other unintended meanings that have been invented based on that term. The new proposal keeps the term (moved down and explained) and reinforces the intended meaning, while reducing the issue of unintended meaning. 14:10, 5 October 2011 (UTC)

An old idea revisited

A long time ago I suggest the use of this chart which tried to address varies issues then present in RS:

Source	Reliability	Possible non WP:RS issues
Peer reviewed journal (in relevant field)	Highest reliable with subrankings from specific field (highest) to general (lowest)	Outside journal's field of expertise ie anthropological journal making medical claim WP:OR issue
University Press (Accredited)	Second highest reliability (Even University Presses can publish vanity books)
Academic division of reliable publisher	Third highest reliability (large reliable publishers may not clearly separate their academic and non academic divisions)	unclear if book is from academic division and division in question can be shown to do vanity books Wikipedia:Verifiability
News organizations (including online versions)	Fourth or lower reliability depending on quality (ex: Washington Post (United States), Times (Britain), and Associated Press are high quality)	involves a living person
Self-published sources	Only usable if by an established expert on the topic of the article whose work in the relevant field has previously been published or favorably sited by reliable third-party publications (per above)	Same as News organizations. See WP:SPS for additional requirements.
personal/unknown quality websites, webforums, blogs, Wikipedia itself	Unusable

It was a rough stab a very complex issue but it tried to address the issue that not all Verifiability sources are created equal.--BruceGrubb (talk) 15:10, 4 October 2011 (UTC)

I think the idea of a table is a very good idea, and I appreciate the effort Bruce put into setting up this table. We need a good starting point, like this! But I see a couple of problems that can be improved. First, I do not agree with a ranking of "reliability." Different sources are reliable in different ways. A newspaper is more reliable than an academic journal for reporting the news. Mein Kampf is a very reliable source for Hitler's views; it is a very unreliable source for, say, information about Jews. Sources are the concrete means for verifying that a certain view exists, and is significant. What is the most reliable source depends on which/what kind of view we are trying to verify.

I think that the category of problematic issues is a good idea. But the way it is expressed here would confuse many readers, because it doesn't really match our policy. For example, using an anthropology journal as a source for a view about medicine is not a violation of NOR; the major NOR problem when using sources is not using the wrong source, but misusing the right source - most NOR violations are the result of violating SYNTH and using two reliable sources to make a claim niether source makes. Second, anthropology journals can be very highly reliable sources for some topics concerning medicine. Medical journals like JAMA and the NEJM are probablyu the most reliable sources for the "Western" or alleopathic medical view of disease and treatment. Mut the fields of medical anthropology and the sociology of medicine provide veryu important scholarly analyses of medical beliefs in practice in modern (industrialized, bourgeois) culture and of course in other cultures. the journals Medical Anthropology Quarterly and Social Science and Medicine might well be highly reliable sources for what MD's believe about illness and healing, if that is the aim of a particular article. Otherwise, it is the most reliable source for how social scientists interpret and analyze medical doctor's claims about illness and healing. The key point is, JAMA is a reliable source for the views of MDs, and NAQ is a reliable source for anthropologists' views of illness and healing.

I would also go further and say that any educated reader of a journal article has to be able to distinguish between the authors central argument based on her own research, her accounts of the research of others, and tangential comments. All three kinds of content are found in most academic journal articles. The article is a highly reliable source for verifying the view of a (discipline x) on the topic that is the main object of the author's research and argument. The article is a highly reliable source on the view of a researcher (from discipline x) of the research of others (this may sound obscure. I will give an example. I rfecently read an article by an anthropologistg in immigrants. In the couirse of the analysis, the author reviewed certain theoretical literature, including concepts of Sartre. In my view, the author's views on Sartre are plain wrong. I would not consider an anthropologist to be the best source for Sartre's views - I would turn to work by other philosophers or intellectual historians as reliable sources on Sartre. But I do think that this article is a reliable source on how many anthropologists interpret Sartre). Finally, I do not think that there is any grounds for conmsidering the article a reliable source on tangential statements in the article.

In short, the degree of reliability of the source depends on whose view it is being used to verify. It is not simnply a matter of more versus less reliable (I agree with Grubb in general that when it comes to any topic that is an object of academic research, established peer-reviewed jouirnals and books published by acadsemic presses are the most reliable source). It is also a case of what is the most appropriate source to verify a particular view.

In short, OR occurs when an editor presents his or her own view. Even if the editor cites several highly reliable sources, if she is deriving a view from combining or synthesizing information from the sources, but this view is not expressed clearly in any one of the sources, it is OR. OR usually occurs not because one uses the wrong source, but because one uses sources wrongly. Slrubenstein | Talk 08:19, 5 October 2011 (UTC)

Wikipedia_talk:Reliable_sources/Noticeboard/Archive_2#Decision_Memory was where I first produced this chart and it was also an attempt to separate relevant from reliable. However regarding the ranking I would argue that an academic journal analyzing Mein Kampf would be more reliable then the book itself especially if you try to relate the work to what happened later.

One of the problems I have over at the Weston Price biography is that nearly all of the material regarding root canals is based on how people currently use Price's self published 1923 work Dental Infections, Oral and Systemic rather then what Price himself wrote but getting Price's own words (using exact quotes--see [[2]] for what I want to put in the article) into that article even when they come from a 1925 JAMA article and 1939 book published by Paul B. Hoeber, Inc; Medical Book Department of Harper & Brothers is impossible as we have one editor going "it's primary and therefore cannot be used" and nearly every other editor seems to have left the article in disgust.--BruceGrubb (talk) 09:26, 5 October 2011 (UTC)

In my opinion we should recognize the three main metrics of reliability, plus the one "metric of the metrics"

• Expertise with respect to the item which cited it
• Objectivity with respect to the item which cited it
• Raw (easily mis-used) data vs. digested coverage (primary vs. secondary)

And the metric of the above metrics which is:

• Source credentials/attributes (e.g. newspaper vs. blog)

Of these four, wp:ver really only acknowledges two. It skips the first two.

I've been (long term) trying to work on something regarding this at: Wikipedia:Strategic issues with core policies#WP:ver and wp:nor need additional source metrics, and a way to apply them North8000 (talk) 11:17, 5 October 2011 (UTC)

I don't think it is so much wp:ver skips things but that it fails to explain things well. Cleaning up the table above into something more usable produces this:

Source	Issues relevant to WP:V	Relevant factors outside WP:V
Peer reviewed journal (specific field (highest) to general (lowest))	Primary, secondary, and tertiary information, WP:MEDRS	WP:WEIGHT, WP:NPOV, WP:MEDASSESS
University Press (Accredited)	Primary, secondary, and tertiary information, WP:MEDRS
Reliable publisher	identifiable as academic and not vanity division
News organizations (including online versions)	Highest quality such as Washington Post (United States), Times (Britain), and Associated Press preferred. Tabloids are generally viewed as low quality.	involves a living person
Self-published sources	Only usable if by an established expert on the topic of the article whose work in the relevant field has previously been published or favorably sited by reliable third-party publications (per above)	Same as News organizations. See WP:SPS for additional requirements.
personal/unknown quality websites, webforums, blogs, Wikipedia itself	Unusable	Unusable

Again this is a rough stab so feel free to suggest modifications to clean it up a bit.--BruceGrubb (talk) 12:29, 5 October 2011 (UTC)

With all due respect, I have problems with North8000's criteria. What makes a source reliable is that we have confidence that it provides an accurate account of a significant view. But according to our NPOV policy, we include all views that are significant. Unfortunately, many views are not objective and not based on expertise, but are significant - and we need to include them. If your criteria for identifying a reliable source is focussing solely on those criteria that allow us to have confidence in the accuracy of the account of the view presented, well, then, I agree completely. but we need to be clear to distinguish between the accuracy of the source and the significance of the view the source represents. For example, with regards to Jews, hitler was no expert and certainly not objective, but his view is significant. And Mein Kampf is a reliable source for his views. Certainly, Hitler was an expert in what Hitler thought. But did Hitler really have an "objective" understanding of his own views? Maybe. But I do not believe that most historians and biographers of Hitler are convinced that even Hitler understood the real sources of his anti-Semitism. Perhaps you would then argue that we should not use mein kampf as a source for Hitler's views. you might say that we really should favor historians and biographers, who have not only studied mein Kampf carefully but have analyzed it in the context of his life and times. (This is the value of your distinction between primary and secondary sources). I am not sure what the answer is. I do hope you see how, when talking about certain significant views, and certain sources, objectivity may not be useful and in fact may not even be a desirable criteria. Slrubenstein | Talk 17:11, 5 October 2011 (UTC)

Slrubenstein, my idea/framework fully incorporates what you are saying. (and I agree with what you are saying). Actually, I think that my idea incorporates what you are saying even more than the current policy. Since your comment covers a lot of ground, I'll just reinforce this on a couple of points:

Mein Kampf is certainly an excellent source on what Hitlers views are. Under my idea, (rating 0-10) on that topic "expertise" would be a "10", objectivity (since he's the type to say what he really thinks) a "9", on raw data vs, digested by others (primary/secondary) a "0", and on credentials for what it is, a "10". Under my idea you combine all of those and say that it's a pretty good source for that. Under current wp:ver, it flunks primary/secondary, which is a categorical/stand alone criteria, and so it's use would be very limited. Now, if there was an article on the intelligence level of Jews, (not Hitler's opinion of such) Mein Kampf would rate much lower under my idea. North8000 (talk) 18:15, 5 October 2011 (UTC)

I am glad we agree. I understand that the table is meant to be concise and schematic and I really do see the value of that. I do not want to quibble over words - your point is that however differently I might understand these words from you, we agree on how they should be used. So I think that it is critical that if we add anything like this to the policy, these terms be operationalized, and perhaps a few very different kinds of examples b provided. An example of a highly contentious article where multiple views must be provided given due weight: "race" (as applied to human beings). Some academics (including some biologists) believe that race is a social construction and can be analyzed only sociologically (by which I do not mean "by sociologists" but rather "by social scientists"). Some academics believe that races are biologically real and can be studied biologically, but their definition of race and how they designate different races is quite different from the popular understanding. Then, many non-academics (i.e. ordinary people) in the US and UK believe races are biologically real, but they do not use the word or think of different races the way actual biologists do. And in many parts of Latin America, races are very important but popularly understood to be cultural, thus, a person's race can change based on wht language they speak, how they dress, how well-educated they are etc. I bet that you could use your three criteria to identify what you and I would agree are the most reliable sources. But, in fact over the years there has often ben a good deal of contention over what are the reliable sources and how much weight to give them (I assume your ranking sytem is meant to help determine due weight) - I can easily imagine diferent editors using your criteria and yet assigning each variable very different scores. This is why I think that whatever words are used, people will need guidance for you to use them or how to settle arguments over its use.

Another good example is the Jesus article. It is no surprise that many of the people who, in college and later graduate school, are drawn to studying the Gospels and Jesus are Christians. Some of these scholars are religious Catholics, yet write books that non-Christians assign in university courses on the history of religion. Some of these people remains Christian, but reject the Bible as a reliable historical source and reject the divinity of Jesus (thus placing them in a minority group among Christians) and some abandon Christianity altogether. For reasons that I think are obvious, very few universities have several professors who are fluent in Koine Greek, Aramaic, and Hebrew - languages which one must know if one wishes to study 1st century Judea and Galilee as a professional historian. If one wishes to get a PhD. on Jewish-Christian history in the first centuries CE, one would want professors who are fluent in these languages and the primary sources. You can find such departments at top universities like Oxford and Duke, but some of the best departments are at Seminaries. Some Seminaries only train priests, ministers, or rabbis, but some have secular graduate schools (i.e. students need not be religious, professors may not be religious, the methods used are the same as those used by any other historian studying any other time and place. Sorry to go into so much detail, but my point is this: the point of view of graduates of these programs, or professors at these programs, is not necessarily "Christian." And the problem is, this would be obvious to anyone who has studied history at a graduate level and who actually reads the books, or someone who took courses in the history of religion or history or Near Eastern Studies or (as it is sometimes called) Biblical Studies, at a university like Oxford, Duke, or Harvard, and was assigned some of these books. But: many editors do not have advanced training in history and of these, many do not actually read the books. But by looking at a web-page, they learn that professor X got a debree in Biblical Studies, or got a degree at X Seminary, and on this basis they insist that any book or article these historians have written express a Christian point of view.

In other words, to use your system appropriately, in many cases one must start out being generally well-informed about academic conventions, and one must thus do considerably more research than looking at the author's homepage. This is not a criticism of your table or terminology. I guess my point is, the simpler the scheme or table, the more important it is to give people clear guidelines about how to use it. Sometims, the simplest guideline may demand that the editor do a great deal of research. You know how uneven are the research skills and the efforts editors put into research. I am sure we agree, we would want people to use your table as a tool, but not as a crutch. I do not think they should be incorporated into the policy until we have worked out clear guidelines for their use, what kind of data one would need in order to measure any of the criteria, etc. Slrubenstein | Talk 19:44, 5 October 2011 (UTC)

I've always been of the opinion that it is much better to provide a sound, effective framework, and let folks struggle with filling in the blanks than expecting folks to try to make a system work that does not have as sound of a framework and does a ham-handed job of trying to dictate the details. I consider that wp:ver, having only two separate metrics, (RS criteria and primary/secondary) with the former often unrealistic as written, and no assessment of knowledgeably and objectivity with respect to the item that cited it, and no scaling in accordance with how controversial the cited state to be the latter case to a certain extent. North8000 (talk) 20:33, 5 October 2011 (UTC)

Stealth canvassing

From the lead of WP:Canvassing

"In general, it is perfectly acceptable to notify other editors of ongoing discussions, provided that it is done with the intent to improve the quality of the discussion by broadening participation to more fully achieve consensus.

However canvassing which is done with the intention of influencing the outcome of a discussion in a particular way is considered inappropriate. This is because it compromises the normal consensus decision-making process, and therefore is generally considered disruptive behaviour."

With the upcoming consensus poll re the first sentence of WP:V, editors should particularly be careful not to participate in the unethical activity described in the section on stealth canvassing.

"Because it is less transparent than on-wiki notifications, the use of email or other off-wiki communication to notify editors is discouraged unless there is a significant reason for not using talk page notifications. Depending on the specific circumstances, sending a notification to a group of editors by email may be looked at more negatively than sending the same message to the same group of people on their talk pages."

--Bob K31416 (talk) 12:30, 5 October 2011 (UTC)

A caution against stealth canvassing. This RfC is in serious trouble already, I'd wager. Doc talk 12:34, 5 October 2011 (UTC)

Bob, please WP:Assume good faith... there is no indication that anyone has engaged in stealth canvasing or that anyone intends to do so. Blueboar (talk) 13:26, 5 October 2011 (UTC)

For those of us not in the loop, could Bob and Doc please clarify this? Is there any evidence behind this allegation? --Tryptofish (talk) 00:17, 6 October 2011 (UTC)

That's pretty doubtful, but you'd have to ask Bob. I've never seen someone issue a warning on stealth canvassing before a RfC. Doc talk 00:21, 6 October 2011 (UTC)

Looks to me like Bob is just telling people to be on their best behavior (rather than alleging anything). It would be a shame to lose a lot of people's hard work because of anyone being a bad actor in this. -- Avanu (talk) 04:49, 6 October 2011 (UTC)

Yeah. How about another form of "inappropriate canvassing": almost the definition of votestacking. Only the most influential editor on the project is notified when it is already known that he wants "not truth" removed from the lead. Bravo. Doc talk 05:34, 6 October 2011 (UTC)

I don't think that is a valid objection. It is never inappropriate to notify me of anything. Never. I really don't like it when people try to make out like talking to me is some kind of policy violation. It never is.--Jimbo Wales (talk) 06:33, 6 October 2011 (UTC)

I was merely pointing out that you're the only one that was notified. I don't truly think any policy was violated, but perhaps more people could have been notified from both sides of the issue if anyone was going to be notified. Sorry if my comment irritated you, Jimbo. I'd still like to know what this "stealth canvassing" is about: was something seen on an outside website? Doc talk 06:56, 6 October 2011 (UTC)

I have left a notice at the Village Pump, and at both WT:NPOV and WT:NOR. So, hopefully, word should spread that this RFC is going on. The key is to cast our net widely. Remember that an RFC is not just a head count (although a simple vote tally can be useful information)... The actual comments are important (one insightful comment can, and should, carry more weight than several pile-on "me too" votes). I also think we should pay attention to who is commenting (I am not talking about individuals here, I mean "who" in a broad sense ... we should pay close attention to the comments that come from editors who have not been part of our previous debates.) Blueboar (talk) 14:34, 6 October 2011 (UTC)

RFC - Compromise proposal re first sentence

The proposal is in two parts...

• 1) change the opening paragraph:

From: The threshold for inclusion in Wikipedia is verifiability, not truth—whether readers can check that material in Wikipedia has already been published by a reliable source, not whether editors think it is true. To:     The initial threshold for inclusion in Wikipedia is verifiability—whether readers can check that material in Wikipedia has already been published by a reliable source. While verifiability is a requirement for inclusion, it is not a guarantee of inclusion. Wikipedia has other policies and guidelines that affect inclusion (especially whether specific material is included in a specific article).

The other paragraphs in the lede will not change.

• 2) Insert a new section (as the first section after the lede, following the index box) to deal with the issue of truth/untruth...as follows:

==Assertions of truth and untruth== An editor's assertion that something is true is not enough for inclusion in Wikipedia. It does not matter how convinced you are that some bit of information is true; if the material is unverifiable, do not add it. In this context, Wikipedia requires "verifiability, not truth". Assertions of untruth (i.e., an editor's assertion that some bit of information is untrue) are a more complicated issue. If the dubious information is not supported by a source, it should be challenged; but the question of how to challenge (whether to tag the information as needing a citation or to remove it immediately) depends on the nature of the information (see: WP:Burden, below). If the dubious information is supported by a reliable source, the problem should be discussed on the article talk page, with reference to policy concepts such as maintaining a neutral point of view (and especially the sub-concept of due weight). Often rewording to present the information as opinion rather than fact can resolve issues of verifiable but potentially untrue information.

Rationale

Introduction The first sentence of the policy currently reads: "The threshold for inclusion in Wikipedia is verifiability, not truth—whether readers can check that material in Wikipedia has already been published by a reliable source, not whether editors think it is true." There are concerns about this sentence, and particularly about the two words "not truth". In RfCs held in April and June, about 50% of editors responding supported change and about 50% opposed. After further discussion, with neither viewpoint gaining a solid WP:CONSENSUS, a working group formed to examine the concerns of those on both sides of the debate. The working group's deliberations can be found primarily at Wikipedia talk:Verifiability/First sentence and its project page Wikipedia:Verifiability/First sentence; although a few threads continued at Wikipedia_talk:Verifiability. Out of this working group has emerged the proposal above. It is seen as a compromise - one that addresses the core concerns of both sides. The proposal keeps the "verifiability,not truth" phrase in the policy, but moves it to its own section and clarifies it. It is hoped that both those who are content with the current wording and those who advocate change, to whatever degree, will support this measure as a compromise. Main rationale presentation Background: The concept that truth is not the threshold for inclusion in Wikipedia was added for a specific reason - to support WP:NOR in saying that material should not be included unless there is a source that directly supports it. At that time, we had a persistent problem with editors wishing to add unverifiable material purely because "it's true" (a rationale commonly used by editors trying to "prove" their pet fringe theory). However, as WP:V has changed over time, the sentence has been moved earlier and earlier in the policy, and it has lost some of its original context. It has taken on meanings that were never part of its original intent. Concern: The sentence can be misconstrued to mean that any material that appears in a source must be included...simply because it is verifiable. This misinterpretation is in conflict with several other policy and guideline statements (especially the WP:Undue weight section of WP:NPOV), but examples of this misinterpretation happening in practice have been provided. How the proposal resolves this concern: The proposal adds an explanation that "while verifiability is a requirement for inclusion, it is not a guarantee of inclusion", and it notes that other policies and guidelines can affect inclusion. Concern: The sentence can be misconstrued to mean that we may not discuss the possibility that verifiable information is untrue (i.e., that the source may contain an error). Specifically, this reading says that editors need not discuss the exclusion of material on the grounds of being not true, in the same way that editors need not discuss the inclusion of material on the grounds of being true. That is, we must accept what a published source says simply because it is published, even if we have reason to believe that the material is inaccurate or that the source is less than reliable. Counter concern: This was never the intent. We often need to make judgment calls about the reliability of specific sources when it comes to specific information. A source may be reliable for one statement, but unreliable for some other statement. We also have to make judgment calls about the relative reliability of one source when compared to others. As Jimbo Wales puts it, "We are not transcription monkeys." We do want the information we present in Wikipedia to be accurate, as far as possible. Further, as NPOV notes, we cannot omit significant viewpoints just because we disagree with them (or even because most sources disagree with them). Sometimes we should discuss facts and opinions that may be untrue, because doing so gives the reader a complete picture of disagreement among the sources. How the proposal resolves these concerns: The proposed language acknowledges that inclusion of potentially untrue information depends on context. We cannot make a firm one-size-fits-all rule on this. The proposal points out that the question of whether to include controversial and potentially untrue material is a complex one, that involves applying editorial judgment. It points the reader to other policies and guidelines that may help. Concern: Introducing the concept of "truth" in the lede is distracting and confusing, particularly for new editors. The lede should focus purely on explaining what Wikipedia means by Verifiability without introducing secondary concepts. To the extent that it is relevant for the policy to discuss the issue of truth/untruth, this belongs in the body of the policy. How the proposal resolves the concern: The issue of truth is moved out of the lede and into its own separate section. Conclusions: The goal of this proposal is not to change the meaning of the policy, but to clarify it and reduce the potential for real or feigned misunderstanding. The concepts behind the phrase "Verifiability, not truth" should remain part of the policy. But they are complex concepts that need to be better explained.

Notice: A change was made to two sentences of the Introduction to the Rationale on the first day of the RfC. This change is documented here. Also note that there were three intermediate versions of these two sentences on the first day of the RfC.

Comments

Support

1. I obviously support this proposal... I was, after all, its primary author. The rational that is presented with the proposal explains most of my thinking on it. Not only is it a good (and, more importantly, a workable) compromise between the various positions... I think it actually improves the policy by making what I have always understood the intent of the current language clearer. Blueboar (talk) 23:41, 5 October 2011 (UTC)
2. Support This is a compromise crafted over many months to resolve an ongoing, otherwise unresolvable debate that has gone on intensely for almost a year. It retains all of the intended meanings from the original AND reduces the unintended meanings and effects. I hope and request that folks from both sides of the debate....including those those advocating more or less change...support it as such.North8000 (talk) 00:04, 6 October 2011 (UTC)

   This proposal deals with the largest, most discussed issue. It does not preclude tweaks in other areas can then be discussed later. North8000 (talk) 15:46, 6 October 2011 (UTC)

3. Support as a reasonable compromise and a well thought out approach to address a wide variety of concerns. --Nuujinn (talk) 00:03, 6 October 2011 (UTC)
4. Support - HOWEVER, I think the initial paragraph needs to remove the word 'material' and replace it with 'information' (or similar). The use of the word 'material' can lead a person to think that we only accept words that are verbatim from other sources, which is actually against policy, yet it is FAR too common in Wikipedia. I've been involved in too many debates where people argue that "no, the source said 'rough' not 'coarse'" and people spend endless hours debating whether changing a word constitutes original research. But during this RFC, we're supposed to be focused on the "truth" part of the intro, so I guess despite my misgivings, I'll say 'SUPPORT'. -- Avanu (talk) 03:06, 6 October 2011 (UTC)
5. Support - This version addresses many of the problems presented though I agree with the information rather than material reword.--BruceGrubb (talk) 03:30, 6 October 2011 (UTC)
6. Support  Balances many tradeoffs, not all of which I agree with, but overall a huge step forward.  Unscintillating (talk) 05:09, 6 October 2011 (UTC)
7. Support, there is still much that I would change (including "material" to "information"), but this is undoubtedly an improvement.--Kotniski (talk) 07:09, 6 October 2011 (UTC)
8. Support as a major step forward for this policy. This proposal will reduce the potential for inexperienced or tendentious editors to misinterpret or pretend to misinterpret what it says.—S Marshall T/C 07:35, 6 October 2011 (UTC)
9. Support Wifione ^Message 08:10, 6 October 2011 (UTC)
10. Support. I had no problems with the current version either, but this is a compromise I can accept. Additional changes (material vs information, threshold vs. fundamental requirement, ...) should be discussed afterwards (but, assuming that this rfC gets support, should stay away from the whole "truth" thing for a long time). Fram (talk) 08:30, 6 October 2011 (UTC)
11. Support the current wording is simply misinterpreted too often and this is an important step forward. Yoenit (talk) 08:37, 6 October 2011 (UTC)
12. Support Sir Armbrust ^{Talk to me} _Contribs 10:23, 6 October 2011 (UTC)
13. Support part 2. I particularly like the part on "Assertions of untruth". The new sentences in part 1 read like a camel, but the intent can be understood. --SmokeyJoe (talk) 10:28, 6 October 2011 (UTC)
14. Support. I've sometimes found myself asserting that although WP:V prohibits us from adding unverifiable information, it does not compel us to add information that is verifiable but untrue. This codifies the need for editorial discretion in those instances. I also prefer "information" over "material". Adrian J. Hunter^{(talk•contribs)} 11:44, 6 October 2011 (UTC)
15. Verifiability and truth should go hand-in-hand. In the rare cases that they do not, we should not be actively encouraging people to post untrue information. I therefore support most of the rewording. —WFC— 12:43, 6 October 2011 (UTC)
    • However I oppose the bracketed bit as unnecessary. —WFC— 18:29, 6 October 2011 (UTC)
16. Support Even though I understood what was meant by "verifiability not truth", it was clear just glancing at past discussions that editors were misreading and/or wikilawyering that counter to the policy's intent. Clarification is always a good thing. --MASEM (t) 12:49, 6 October 2011 (UTC)

    The problem is, people thought "not truth" = "false" when what we really meant was "neither true nor false." I know some people think we must make truth a criterion to improve the quality of articles, but the problem is this will lead to a conflict with NPOV. NPOV says we must include all significant views. What if a significant view is, we believe, false? I think that the cause for the low quality of many articles is many people confuse "V" for "cite sources." It is easy to "verify" a view by finding a source. The hard thing is to verify that the viw is significant. This is what I think we need to work on! Slrubenstein | Talk 19:47, 6 October 2011 (UTC)

17. Support Was tempted to oppose because it retains the Toxic Triad (albeit farther down the page). But it's a start, and in fact quite a good start. It addresses head-on the old "it's been printed in a newspaper so it's gotta go in" argument that we hear so often. Short Brigade Harvester Boris (talk) 12:56, 6 October 2011 (UTC)
18. Support. Not perfect, but it is a change in the right direction. Count Iblis (talk) 14:55, 6 October 2011 (UTC)
19. Support as a good compromise, with the hope that it can be tweaked over time (without giving preeminence back to the too often misused and misapplied 'Toxic Triad' that Boris mentions). First Light (talk) 15:27, 6 October 2011 (UTC)
20. Support - The "verifiability not truth" mantra has always been the single most idiotic thing about Wikipedia. One of the PILLARS of Wikipedia should be VERACITY — and here we have a very explicit, up front, unambiguous declaration that PUBLISHED FALSEHOOD is perfectly fine. The threshold should be VERIFIABILITY and VERACITY. Period. Carrite (talk) 16:22, 6 October 2011 (UTC)

    The problem is, people thought "not truth" = "false" when what we really meant was "neither true nor false but something else." I know some people think we must make truth a criterion to improve the quality of articles, but the problem is this will lead to a conflict with NPOV. NPOV says we must include all significant views. What if a significant view is, we believe, false? I think that the cause for the low quality of many articles is many people confuse "V" for "cite sources." It is easy to "verify" a view by finding a source. The hard thing is to verify that the view is significant. This is what I think we need to work on! If we really took the time to make sure that we were accurately representing sources in context (which belongs with "cite sources") and that we are verifying that the views included are all of the significant views, the quality of many articles will increase tremendously. But bring in truth and NPOV goes out the window. People holding equally significant but contradictory views may think they are equally true and the other's false!Slrubenstein | Talk 19:52, 6 October 2011 (UTC)

    If I may comment: Intended meaning: Verifiability rules as a requirement for inclusion. Truth or claimed truth is not a substitute. But potential inaccuracy of a statement should not be excluded as a consideration when editors are discussing possibly leaving something out. (Recognizing that in some particular situations, wp:npov trumps the discussion and dictates that it stay in.) Sincerely, North8000 (talk) 20:00, 6 October 2011 (UTC)

    I admire your desire to support compromise. I would support Blueboar's proposal if it were modified. If you are using "accuracy" simply as a synonym for "truth" I fear we just disagree. If you mean that we must ensure that we are accurately representing the views we include, and that we are accurate in our claims that some views are significant and others fringe, I would agree with you, entirely (e.g., it is accurate to say that Darwin's theory of evolution is a significant view, and it is accurate to say that Darwin's view was that evolution occurred through natural selection). But if you mean that we must judge whether an established, significant view is or is not accurate (e.g. "is the theory of evolution by natural selection accurate" or "is the theory of relativity accurate"), then I think you are simply contradicting NPOV - and perhaps misrepresenting how science progresses. Slrubenstein | Talk 14:06, 7 October 2011 (UTC)

21. Support - A revision like this has been needed for a long time; this will end a whole lot of senseless talk page wikilawyering. --Ludwigs2 16:39, 6 October 2011 (UTC)
    • P.s. I'm not thrilled about the parenthetical statement at the end - it doesn't seem to make any sense, and I'm not certain what it is intended for. That should be removed. --Ludwigs2 16:50, 6 October 2011 (UTC)
22. Support - Sounds reasonable. --Iantresman (talk) 16:58, 6 October 2011 (UTC)
23. Support seems like an improvement. -- Taku (talk) 17:44, 6 October 2011 (UTC)
24. Tepid support. I'm supporting because I think that the clearer explanation of "not truth" is a net plus. I also think that Blueboar deserves enthusiastic applause for devising such thoughtful wording for the sentence in which it appears, and for working productively with the diversity of editor opinions that went into the drafting. I am also taking to heart the suggestion that editors adopt a spirit of compromise. However, my support is only tepid, and I actually agree with a lot of the comments that I read in the oppose section. I see no good reason to move the discussion of truth out of the lead, and I worry that it has been buried in tl;dr. I also think that the proposal has been weighed down with additional changes that go beyond the central issue of truth. I believe this policy has long served Wikipedia very well in its present form, but that this proposal is, net, a very small improvement. --Tryptofish (talk) 18:27, 6 October 2011 (UTC)
25. Support - Mainly because of the inclusion of "While verifiability is a requirement for inclusion, it is not a guarantee of inclusion." in the lede, though I think the whole thing is beneficial. — TransporterMan (TALK) 18:32, 6 October 2011 (UTC)
26. Support in principle What we have needs to be changed, and the three things that need to be made clear are the three things presented in the proposal; however, the proposed text is far too verbose—a little more work and we'll be there. Uniplex (talk) 19:24, 6 October 2011 (UTC)

    This addresses just the main topic of concern/debate. Other items can then be tweaked after that. North8000 (talk) 19:37, 6 October 2011 (UTC)

    Yes, the verbosity was my main concern also, and that can be improved over time. First Light (talk) 19:43, 6 October 2011 (UTC)

27. Qualified support The addition to the lede has the correct intent but is too wordy; it needs to be tightened up considerably. The additional section still fails to address the issue of simple factual error in sources, and, with WP:BURDEN, still encourages the creation of fake controversies by discouraging the rejection of manifest factual errors in otherwise "reliable" (that is, conforming the Wikip's formal standards) sources. Mangoe (talk) 20:05, 6 October 2011 (UTC)
28. Support Although I agreed with the philosophy behind the original wording, its actual expression caused far too much difficulty. This proposal is an improvement. I shall not suggest any changes to the proposed wording and I suggest other "supporters" do not do either. Refinements can be made after any implementation is done. Thincat (talk) 21:01, 6 October 2011 (UTC)
29. Weak Support. The examples show that something about the current formulation of this policy is problematic. It's like the burden is always on those arguing for the removal of well-cited material to prove that it is irrelevant, unreliable, or otherwise inappropriate. However, it is not at all clear where this well-cited is holy attitude is coming from; in particular, I do not see how one can be so sure that "not truth" is to blame. Still, I largely agree with Tryptofish, this is a net plus, and my view is clearly closer to those wishing to remove "not truth" than to those who deny the problem altogether. Vesal (talk) 22:21, 6 October 2011 (UTC)
30. Support per Blueboar. --JN466 22:25, 6 October 2011 (UTC)
31. Support While "verifiability, not truth" is a great slogan, it's confusing enough that more explanation is warranted. I appreciate the further explanation of truth wrt verifiability and, in the strange circumstance that someone in a dispute actually read policy, would help. Just lop off the parenthetical statement at the end of the first change. On a different note, I'm seriously annoyed that this was set up as a vote by the proposer. When did we leave behind that other great slogan, "Voting is Evil"? 2009?Danger (talk) 22:33, 6 October 2011 (UTC)

    I don't think that is fair. Blueboar is not so much that author as he is a de facto manager of a very long and comlicated process involving many people on some subpages. I think that the subpages were a serious exercise in good faith consensus building, and this poll should be seen more as a ratification or reality-check for the subpage process. --SmokeyJoe (talk) 23:11, 6 October 2011 (UTC)

    Also, Blueboar just put it up as an RFC. Others added the format and structure for the feedback. Not that I disagree. North8000 (talk) 00:12, 7 October 2011 (UTC)

    Actually, Tryptofish is evil. I made the edit creating support and oppose sections, so please blame me, not Blueboar. And I think that it has proven to be a useful edit. Please note that there are hardly any "me too" type comments. Rather, editors are generally taking the time to explain their thinking. --Tryptofish (talk) 16:47, 7 October 2011 (UTC)

    I, for one, was glad that Blueboar was willing to do the primary work to create the RfC, and I appreciate his role in getting this going.  It was a decision of the working group to proceed with the RfC.  Unscintillating (talk) 21:24, 7 October 2011 (UTC)

32. Support the compromise. Not the most ideal solution in my view but better than the "not truth" abomination. violet/riga _^[talk] 23:31, 6 October 2011 (UTC)
33. Support When I read paragraph 1, I was ready to jump flaming away onto the oppose side. Luckily, I kept reading and thus avoided making an idiot of myself. I think the compromise of keeping the wording but moving it out of the lead is one that will help clarify the point for new readers, while still leaving the concept available to this of us to whom the old wording makes sense. Qwyrxian (talk) 23:43, 6 October 2011 (UTC)
34. Support. The existing language does more harm than good. — Carl (CBM · talk) 01:54, 7 October 2011 (UTC)
35. Support As one of the party working on making the policy clearer and with less leeway for genuine misunderstanding or "creative misunderstanding", this is a very good start. We've spent a long, long (and at times tiring!) time working towards an improvement here, and finally come up with something which we all agreed could be put up for a "vote" by the wider community. This baby has been months in gestation - it's time for it to poke its head out into the world now. Pesky (talk …stalk!) 02:47, 7 October 2011 (UTC)
36. Support. We are concerned about truth - it's just that that concern is secondary to our concern about verifiability. We first make sure our statements are backed by reliable sources, and then and only then consider whether they are true. But that doesn't mean we don't consider truth at all, which WP:V could be mistaken to mean right now. -- King of ♥ ♦ ♣ ♠ 03:10, 7 October 2011 (UTC)
37. support not truth helps those who defame wp, dont believe what you read in wp, anyone can edit it and it doesnt have to be true. a clear example of wp being more concerned with the process than the result. it reminds me of a romantic scene where the troops know the are marching into certain death and continue because it is their duty. i also think it is stupid  :) Darkstar1st (talk) 04:59, 7 October 2011 (UTC)

    People defame WP because most of its editors are too lazy to take the time to go to a library and read several books in order to represent the current research on a topic. This is what academics do, and the result is not claims about "truth" but actual well-informed research. "Truth" is a metaphysical concept that people throw around when the believe in something that they do not understand, or use as a shortcut that saves them the trouble of investigating the complexity of just about anything and everything actual scholars study. No reform of WP will be effective unless we distinguish between two very different kinds of critics of WP: POV-pushers who hate WP because it includes views they think are false - this is a matter of faith (even if we are talking about secular claims), and people who recognize quality research. Quality research does not depend on claiming that something is true; it does show that one has read the most recent and well-reviewed books published by academic presses, and peer-reviewed journal articles, and, just as important, explain the views found in these sources in context. I have read lots of WP articles where people cherry-pick quotes from academic sources to promote a view that many people believe is "true" but that distorts the scientific research. As long as this happens, university professors will rejct any student research based on WP. Slrubenstein | Talk 12:55, 7 October 2011 (UTC)

38. Support. The "verifiability, not truth" is catchy, but the literal interpretation is not entirely accurate. It tries to make "verifiability" and "truth" opposing concepts, when in actuality they are supportive of each other. We should never knowingly add false information to an article just because a "reliable source" is erroneous. (Indeed if the "reliable source" is erroneous on a certain statement, then the source isn't reliable on that statement, even if it is reliable otherwise.) Sjakkalle (Check!) 05:29, 7 October 2011 (UTC)

    Not quite. We must include all significant views, it is not for us to determine whether they are true or not (especially when "truth" is a word most scientists avoid and philosophers question). The most we can say is "it is true that this is a significant view," if this is what you mean about mutually supportive. But what we are verifying is that it really is a significant view, and that we are providing an accurate account of it. This is a far cry from saying that th view is true! Do you truly believe that we are supposed to verify which of the following interpretations if quantum mechanics: The Copenhagen interpretation; Many worlds; Consistent histories; Ensemble interpretation, or statistical interpretation; de Broglie–Bohm theory; Relational quantum mechanics; Transactional interpretation; Stochastic mechanics; Objective collapse theories; von Neumann/Wigner interpretation: consciousness causes the collapse; Many minds, is "true?" If one of them is true, the others must be false. Scientists acknowledge that they cannot all be correct. But you think our policy should be to report the true one? Slrubenstein | Talk 12:46, 7 October 2011 (UTC)

    Of course there are cases where there are genuinely competing views in high-quality sources about what the truth of a matter is. But that is addressed in WP:NPOV policy. It's not a matter for WP:V. --JN466 13:12, 7 October 2011 (UTC)

    I agree with you that this is covered in NPOV. But my position starts with NPOV as axiomatic and asks what then follows? My point is (1) V must be consistent with NPOV, and (2) given NPOV we have to ask ourselves, what is it that we are verifying? For me, if we accept (1) the answer to (2) cannot be "the truth." If we accept (1), then what we must verify is that "x is a significant view." It may be a view of the truth (as is often the case with religion, and never the case with science). Some people who are polling "support" are motivated by a perception (accurate in my experience) that most academics do not respect WP. They are also motivated by the belief that the cause of this perception is our use of "verifiably significant as opposed to fringe views" as our criterion for inclusion, rather than "true as opposed to false." I vehemently disagree with this.

    I am convinced that the credibility problem with WP is cased by the fact that very few of our editors take the time to do enough research - which often times must take place in a library - to distinguish between fringe and significant views, and to be able to provide an adequate context when presenting significant views. This is a skill that any PhD must have when writing a literature review, but one does not need a PhD to have this skill (and I have always opposed making one's academic credentials or lack of them relevant to editing WP). One does however need to make an effort, and in my experience a great many editors do not. And academics can see this. Moreover, I am convinced that making "truth" a criterion for inclusion will degrade the quality of the encyclopedia either by encouraging OR or by leading to overly-simplistic coverage of complex topics. We would be much better off if we took more time to provide guidance about what kinds of research are required to verify whether a view is significant or fringe, and to be able to provide sufficient context to understand why people in good faith might have competing views (e.g. why different US Supreme Court Justices hold different views on the constitutionality of abortion, gun control, or torture. Or why equally respected scientists have mutually exclusive interpretations of quantum mechanics. I am not criticizing these articles, just using them to illustrate my point) This, and not the absence of "truth" is in my experience why so many scholars are dismissive of WP. Slrubenstein | Talk 14:00, 7 October 2011 (UTC)

    The problem here is significant and fringe are not exclusive things. Regardless of how you define it the Jesus myth theory is significant but because the definitions for it are all over the map it is next to impossible to tell if the entire idea is fringe.--BruceGrubb (talk) 14:37, 7 October 2011 (UTC)

    I would rather say that Jesus Myth Theory is significant in one context and not in another - just as I would say that Creationism is significant in one context but not in another. As many scientists have pointed out (Gregory Bateson most notable) context is crucial. Bruce, I think you make a valid point. I just do not think that making truth, fact, or accuracy a criterion is the solution. I think that clearer criteria for significant and fringe views, criteria that make clear the importance of context, will not only address the specific problems people here are concerned with but would moreover improve the encyclopedia in many other ways. Slrubenstein | Talk 16:16, 7 October 2011 (UTC)

    Several of the examples Slrubenstein mentions are cases where the truth is controversial or unknown. In many cases the fallacy of a statement in a reliable source is fairly clear cut and objective. The latest one I ran into was Talk:2011 Norway attacks/Archive 2#Transportation where a "reliable source" erroneously said that the (hideous) Oslo bombing in July halted all public transportation. Would it have been right to propagate that statement? (In this case the issue was resolved by finding an even more reliable source, but even if not, I would support eliminating any sentences that are in clear and objective error.) Sjakkalle (Check!) 19:20, 7 October 2011 (UTC)

    True and in the case of the Jesus Myth Theory just what the idea even is is up for grabs. For example, take this definition from Bromley, International Standard Bible Encyclopedia: E-J Page 1034: "This view states that the story of Jesus is a piece of mythology, possessing no more substantial claims to historical fact than the old Greek or Norse stories of gods and heroes,...". Even if you ignore the whole story of issue (George Washington and the Cherry Tree is a mythical story of George Washington but doesn't mean there wasn't a George Washington) you still have the huge range of Greek and Norse mythology. Hercules was regarded by Eusebius in Preparation of the Gospel as a flesh and blood person who was later deified and as late as 1919 it was stated "Osiris, Attis, Adonis were men. They died as men; they rose as gods." So Bromley's definition depends how well you know Greek and Norse mythology and is only regarding story of Jesus rather than the man himself--not very helpful.--BruceGrubb (talk) 03:34, 8 October 2011 (UTC)

39. Support. The proposed wording is a welcome clarification on what to do with material that is verifiable but not necessarily true. I think that it is a lot clearer on the subtleties of this issue for someone who is reading the policy for the first time. — Mr. Stradivarius ^♫ 08:35, 7 October 2011 (UTC)
40. Support. Useful clarification. Murray Langton (talk) 19:12, 7 October 2011 (UTC)
41. Support I have previously opposed any change on this guideline, which has served us well. But the proposed wording is simpler and clearer, and doesn't confront new users right up front with a basic rule which at first glance appears to be a logical contradiction-- or perhaps the nearest analogy might be a koan. Koans are very attractive as ways of expressing some deeper truth that is otherwise difficult to verbalize, but it's not necessary to have one here. The basic threshold is, after all, quite simply verifiability in the ordinary meaning of the word. If we can not find any information about a subject, even as a concept, we cannot write about it. Everything else is a subsequent step. We need not, for example, consider whether something would be notable if it existed, or whether a possibly insulting paragraph about an made-up person is a BLP violation. For content also: we cannot find some evidence for a quotation, we can not use it. If there's no source for a birth date, we can't give one. And so on, as the basis for WP:OR and much of WP:NOT , and many other fundamental policies and guidelines. DGG ( talk ) 04:06, 8 October 2011 (UTC)

    But DGG this change doesn't suggest that if there is no source we can't give a birthdate - it suggests that a sourced birthdate can be removed if one editor says "I know that birthdate is incorrect because they guy used to be my neighbor".·ʍaunus·snunɐw· 12:17, 8 October 2011 (UTC)

    I do not think it can be read that way. --Andrew Lancaster (talk) 19:36, 8 October 2011 (UTC)

    Several editors voting support are clearly stating that that is the way they read it.·ʍaunus·snunɐw· 22:45, 8 October 2011 (UTC)

    I don't think so.  Unscintillating (talk) 03:38, 9 October 2011 (UTC)

42. Support The proposed wording is more precise and the additional section is accurate with regards to community consensus. ThemFromSpace 05:08, 8 October 2011 (UTC)
43. Support per TransporterMan. I like how it reminds editors to get down-and-dirty in article-talkpages and discuss the verifiable sources - so they can form a consensus as to whether such-n-such is really appropriate/reliable.--Brianann MacAmhlaidh (talk) 05:24, 8 October 2011 (UTC)
44. Support While I appreciate the pithiness and power of the "verifiability, not truth" aphorism, it really doesn't belong in the beginning of the lead of this policy page, where the proper nuanced reading of it may be lost, and it is too easily open to misinterpretation. Insofar as the "not truth" part distracts from the intent of this policy page, it should be removed as described, and the addition providing context is also nice. --Jayron32 05:37, 8 October 2011 (UTC)
45. Support. I think the link to the essay "verifiability, not truth", even the whole sentence including it, is unneeded; but this is still a better version. Congratulations to the many editors that discussed this (I've read once in a while over the months) and then brought it to a RfC. I am learning from you, thanks. - Nabla (talk) 08:49, 8 October 2011 (UTC)
46. Support. This wording is better. I am amongst those who feel that a further improvement could be achieved by tweaking the first two words, "The initial...". Initial implies a chronological set of steps in editing Wikipedia which do not exist, and the use of "The" instead of "A" makes it sounds like it is even a strict one. During discussion of this draft many people clearly preferred something like "A fundamental threshold..." and I do not believe that consensus has been tested upon this point yet.--Andrew Lancaster (talk) 12:58, 8 October 2011 (UTC)
47. Support.Drift chambers (talk) 17:47, 8 October 2011 (UTC)

Oppose

1. Oppose per my statement above[3]. I am not wedded to having the slogan in the first sentence of the lead, but I don't like the proposed section on truth since it doesn't explain that wikipedia doesn't care about truth at all, only about significant and verifiable views. It is not the wikipedia editors job to evaluate whether claims published in otherwise reliable sources are true. That is the job of researchers and scientists who make science progress by critiquing the claims of other researchers - that is not what we do in wikipedia. If an editor finds that a claim in an otherwise reliable source conflicts with his view of reality then what she needs to do is publish a research article about the topic, not bring wikipedia in line with their own view of reality. Certain editors are already arguing here that certain kinds of OR is ok and that this policy should back that notion. In my view this argument is 100 times more damaging to wikipedia than the odd pov pusher citing WP:NOTTRUTH in order to include a fringe view. We would be turning wikipedia into a publisher of original thought. Here is the citation I go by in my relation to truth here on the encyclopedia[4] "Perhaps the easiest way to make your writing more encyclopedic is to write about what people believe, rather than what is so. If this strikes you as somehow subjectivist or collectivist or imperialist, then ask me about it, because I think that you are just mistaken. What people believe is a matter of objective fact, and we can present that quite easily from the neutral point of view." Jim Wales. ·ʍaunus·snunɐw· 16:33, 7 October 2011 (UTC) ·ʍaunus·snunɐw· 00:26, 6 October 2011 (UTC)
2. Oppose. I highly commend Blueboar in particular for this well-reasoned proposal, but I feel that the current wording is fine and widely understood by the majority of editors: especially if they read beyond that sentence. I still think the only change that might help clarify the strength and meaning of that first phrase might be to put "truth" in quotes, but that was shot down. Doc talk 03:27, 6 October 2011 (UTC)
3. Opposes:.

   adding "initial". Verifiability is the starting point of inclusion - it is the threshold. After the starting point we have other process that material may encounter in order to remain on Wikipedia, but they are not thresholds, because the starting point of verifiability has been passed. "The threshold for inclusion in Wikipedia is verifiability" is a simple, clear statement that loses effectiveness when qualified with the tautological "initial". Verifiability is basic. It is the essential and major inclusion criteria. It all starts from that point. We need to drum home that unverifiable material should not be added. People can quibble about the value of the material later - but let's let people know that inclusion or formatting or editing discussions shouldn't even take place without the threshold of verifiability being passed.

   adding "not a guarantee of inclusion" / "other policies and guidelines". The essence of this policy is that material must be attributable - discussions about how the material is handled, edited, formatted, questioned, etc, do not belong here and simply cloud the issue. It is inappropriate to try to cram the whole of Wikipedia into one policy. This policy page is about verifiability, not notability, which is a different page. Awareness that material may be subject to further scrutiny is covered by the elegant "threshold".

   removing "not truth". The phrase "The threshold for inclusion in Wikipedia is verifiability, not truth" is simple, elegant, and powerful. As far as we are concerned, it doesn't matter that your local band exists, and that you are staring right at them: if there is no means to verify their existence, we don't have them on Wikipedia. While I agree that a section expanding on "truth" is worthwhile (and I support that section), removing "not truth" from the opening sentence removes a simple clarification, and removes a powerful phrase. We can change "truth" to "existence" or any other such synonym, if people are uncomfortable with the word "truth", but that distinction needs to be there, and it needs to be in the first sentence. We define things by what they are not as well as by what they are. We need to make clear that truth/existence by itself is not verifiability. SilkTork ^{✔Tea time} 10:06, 6 October 2011 (UTC)

   I believe that the new proposal "agrees" with and furthers all of the principles which you just described. Sincerely, North8000 (talk) 10:12, 6 October 2011 (UTC)

4. Oppose per User:SilkTork. Roger (talk) 10:58, 6 October 2011 (UTC)

   This proposal is trying to force this policy to do the job of WP:RS (and a bunch of other policies and guidelines). The supporters seem to be under the impression that this is the be all and end all of content policy. Roger (talk) 12:16, 7 October 2011 (UTC)

5. Oppose per Maunus: "I am not wedded to having the slogan in the first sentence of the lead, but I don't like the proposed section on truth since it doesn't explain that wikipedia doesn't care about truth at all, only about significant and verifiable views" Moreover I object to the wording of the first paragraph - what we are verifying is not that there is material documentation, we are verifying that it is a "significant view." Reliable sources (which are typically some kind of publication) are a means for documenting that it is a significant view, but what we are verifying is that it is a significant view. Note I appreciate the hard work Blueboar did and I think that these objections can be resolved with some relatively minor rewording of the text Blueboar wrote. If we can modify the text in lines with these objections - which I think affect just a few sentences - I would support it. Slrubenstein | Talk 11:49, 6 October 2011 (UTC)

   Slrubenstein, you are confusing a Neutral Point of View (WP:NPOV, WP:DUE) with Verifiability. These are not the same thing. -- Avanu (talk) 14:55, 6 October 2011 (UTC)

   No, they are not the same thing. But V derives from NPOV and they are linked, which by the way is true for NOR as well. I think you are confusing V with Cite Sources. These are not the same. I stand by what I wrote. It expresses the principal that has guided my editing since we first agreed on a V policy. Slrubenstein | Talk 17:00, 6 October 2011 (UTC)

   First line of Encyclopedia: An encyclopedia (also spelled encyclopaedia or encyclopædia) is a type of reference work, a compendium holding a summary of information from either all branches of knowledge or a particular branch of knowledge.

   It's a sad state of affairs when apparently many editors seem to think that an encyclopedia is nothing more than a collection of opinions, biographies, views on historic events, politics and popular fiction.

   While I can understand the need for rules focussing on the areas where most problems arise, when "truth" becomes a dirty word for many editors, maybe some change is in order. I was going to comment on the absurdity of "verifiability, not truth" when applied to articles about mathematics, but I'll save that for another time. DS Belgium (talk) 01:51, 7 October 2011 (UTC)

   This is sophistry. First of all, the line you quote does not mention truth. No real scientist claims that they know the truth or teach the truth. Second, everyone knows that if they want to read a great encyclopedia written the conventional way, by actual experts on the topics, they would be better off reading Encyclopedia Britannica. When WP was created, the idea was clearly not to emulate EB but to try a whole new model. As a "wiki"pedia there is no requirement and thus no guarantee that the editors be experts. WP:NPOV was the framework that would not only enable editors with conflicting beliefs to work together, but the principle that would distinguish WP from other encyclopedias. And that principle is neutrality, not truth. NPOV demands that we include all significant points of view. period. Even if we think one view is false, we include it if it is significant. Even if one view says another view is false, we include the other view if it is significant. This is how we achieve neutrality. And if we keep our NPOV policy, then we cannot take it upon ourselves to verify that the view is the truth, and most of us are not qualified even to judge one view as better than another anyway. The only "truth" we can verify is that "it is true that someone holds this view." This is what we are verifying, not that a view is true, but that the view is significant and accurately represented. If you do not like this you should have sought a job at EB rather than volunteer here. And if you want to change this, then we have to get rid of NPOV too. We would have to change it from "neutral point of view" to "truth point of view." Slrubenstein | Talk 11:53, 7 October 2011 (UTC)

   Seems I was wrong, thinking you and the others were defending "Verifiability, not truth" because it was useful in keeping the wackos and fringe theories out and trying to make WP a trustworthy source. Looking at your user page, I realise you want the opposite, every opinion that is notable should be represented, because you believe in "great truths". And that's why you say "no real scientist" claims they know the truth or teach the truth, yet in the next sentence you acknowledge that "actual experts" would use such judgement to decide the content of an encyclopaedia! Since I don't think works of fiction and arts express some great truth, you see me as an idiot ("I think of those people who would answer "no" to my question whenever I read this passage from Umberto Eco's Foucault's Pendulum: Idiot... How could you fail to kneel down before this altar of certitude?")who doesn't understand a thing ("If the person answers no, I know that they understand nothing."), so there's no use in talking to you any further. I just hope you stay away from science topics, especially maths, if you think formal proofs are based on opinions, and the truth or falsity of conjectures can never be determined. DS Belgium ٩(͡๏̯͡๏)۶ 18:37, 7 October 2011 (UTC)

   I certainly am defending "Verifiability, not truth" in order to make WP a better source of knowledge. Moreover, one reason I think we should keep some kind of "not truth" formula is precisely because I think the real solution lies in clarifying and elaborating on what we mean by verifiability, the criteria for reliable sources (which must include other criterion besides publication or even publication in a respected venue) - better guidance concerning fringe views is precisely one of the ways I think we can do this. Yes, I do believe in "the truth." But for reasons that Jimbo and Larry laid out a long time ago, and that are still central to our NPOV policy, I do not believe that my own view of the truth should ever influence my editing. If you think that is the meaning of my user page. I will not discuss my user page here as it is not the appropriate venue but if you actually cared about what I think you can ask on my talk page. Suffice to make three key points: first, I never called you an idiot and that you feel a need to put words into my mouth says more about you than about me. Second, you clearly do not understand the quote from Eco — the capacity for so many WP editors to believe that they understand what a quote means when taken out of context is in my view one of the major problems with Wikipedia (eliminating "not truth" will not address this problem and I think it will actually make the problem worse). Third, I never said formal proofs are based on opinions, although I do know that they are based on axioms and I appreciate the implications of that fact. Would I ever argue against including in an article a statement like "Mathematicians agree that Euclidean geometry provides a proof for the Pythagorean theorum, or a proof that the sum of angles a, b, and c of a triangle is 180 degrees? No. But – and this is the key point – "verifiability, not truth" in no way can support removing such a statement from our article, and, indeed, our policies in their present form help us write great articles that include just such statements. So I do not think your comment is rational (and no, I am not calling you an idiot), and it certainly is not constructive. Slrubenstein | Talk 18:30, 8 October 2011 (UTC)

6. Oppose – to me, the proposed version seems more cumbersome than the current version, and it seems like we're trying to make this more complicated than it should be. –MuZemike 14:41, 6 October 2011 (UTC)
7. Oppose per Slrubenstein, Maunus, MuZemike. Obfuscates rather than clarifies, and places undue prominence on whether sourcing exists, not what is verified by that sourcing, if you follow. I would support a modified version of the 1st paragraph second suggested edit, but strongly oppose the change to the nutshell verbiage opening paragraph, which has lasted for many years for excellent reasons - because it is clear and concise. I see no reason to change it so that it is neither clear nor concise. . KillerChihuahua^?!?Advice 15:20, 6 October 2011 (UTC)

   Question for KillerChihuahua... the proposal does not call for any changes to the nutshell... but perhaps you were referring to something else... could you clarify your comment? Blueboar (talk) 15:26, 6 October 2011 (UTC)

   This is what happens when I'm editing in three windows while on the phone. Apologies, I have corrected my statement. KillerChihuahua^?!?Advice 15:32, 6 October 2011 (UTC)

   Thanks... your views are much clearer now. Blueboar (talk) 15:36, 6 October 2011 (UTC)

   They would have made sense before, if I'd actually typed what I was thinking. Sometimes I have a brain=keyboard disconnect. Thanks for asking for clarifying so politely. :-/ KillerChihuahua^?!?Advice 15:41, 6 October 2011 (UTC)

8. Oppose, and indeed absolutely, utterly oppose. Avoiding the statement that "Wikipedia is not about truth" will open the door to untold of headaches and debates on many, many pages. My real concern is that the Wikipedia community is by and large unaware of a scary phenomenon that is creeping up on us: "we may soon run short of editors who are generally familiar with the policies". In most systems, there are usually underlying processes that drive the system and Wikipedia is no exception. Some time ago I came across a paper based on a thesis by a student about Wikipedia and he had done some studies that made sense. As I recently searched again, I only found a summary of the thesis here, but his main argument, supported by various graphs was that as the number of Wikipages and Wikiusers increase, reality will catch up with us, and he predicted that there will be:

   an untenable trend towards progressive increase of the effort spent by the most active authors, as time passes by. This trend may eventually cause that these authors will reach their upper limit in the number of revisions they can perform each month, thus starting a decreasing trend in the number of monthly revisions, and an overall recession of the content creation and reviewing process in Wikipedia.

   So as more and more IPs require comments, the level of effort to support them may become a burden. If those IPs feel that "they know what truth is" and try to do good by setting the record straight in Wikipedia, the effort to explain things to them will be tremendous.

   It is essential that the millions of new IPs coming in be told upfront that what they consider to be "truth" will probably differ from what someone 3 blocks away from them considers "truth", let alone someone three continents away. It is essential that the idea that "your truth may be different from the next guy's concept of truth" be stated upfront to save us the effort of repeating t again and again to new IPs. I personally feel like a broken record player repeating it again and again.

   And I would go further and point out to the new readers/editors that in some fields "there is no truth". Period. A suitable case is monetary economics where truth is all but elusive. Nobel laureate A says X and Nobel laureate B says Y and usually X and Y are not equal: there really is no "concept of truth" in monetary economics, just ideas and references. And Wikipedia can not even begin to pretend there is truth therein. This must be explained to the new editors upfront.

   We must remind users upfront that Wikipedia is not about truth but verifiability, to save the explanation efforts again and again. That effort is really needed elsewhere: improving content.

   I would, however, also keep something like the 2nd paragraph to repeat the same.History2007 (talk) 17:27, 6 October 2011 (UTC)

9. Oppose does more harm than good. There is wikilawyering where people use sources to push known untruths. But a much more prevalent problem is editors bickering back and forth about what they know to be true, and never checking what a reliable authority would say on the matter. You're focusing on the small problem to the detriment of the bigger problem. You saw the dam leaking, so you decided to tear it out and replace it with a towel. If there's a problem with the occasional wikilawyer, put an explanation in the body of the policy. Don't obfuscate the intro to the policy which has worked well for years. Dzlife (talk) 17:38, 6 October 2011 (UTC)
10. Oppose Adds more verbiage contrary to WP:CREEP. The policy should be simplified, not bloated. Warden (talk) 18:18, 6 October 2011 (UTC)
11. The creepier our policies, the harder it is to edit. The proposed changes do not make it any easier for any inexperienced user to edit, and only add verbiage to be parsed by axe-grinders. No. -- Y not? 18:36, 6 October 2011 (UTC)
12. Oppose. The possible benefit of this particular addition is definitely outweighed by the negative effect of the instruction creep it introduces. Shorter is better. —Mark Dominus (talk) 21:26, 6 October 2011 (UTC)
13. Oppose the new section "Assertions of truth and untruth". Too much emphasis on the term truth that would be better referred to as majority or minority viewpoints. No need to open up a remote possibility it will be misinterpreted to squash multiply-sourced, minority viewpoints on the grounds that it is not true. WP:UNDUE already does a wonderful job with only one instance of the word truth. I am OK with proposal in the first part for changes to the opening paragraph. Willing to reconsider opposition if truth is removed or de-emphasized.—Bagumba (talk) 22:17, 6 October 2011 (UTC)
14. Oppose as unnecessary. The best proposal I've seen was to link "verifiability, not truth" to the essay that explained the topic quite adequately. Jclemens (talk) 23:57, 6 October 2011 (UTC)
15. Oppose. I concede that the "not truth" maxim can be taken the wrong way. So can anything if one does not want to take the point, but the point here is clear enough. My real objection is that the proposed new section is a CREEPy digression to wade through before getting to the nuts and bolts of sections on when a citation is required and what sources are reliable. To the extent that issues in the proposed new section need to be raised at all, they would belong in the "other principles" section at the end of the page. ~ Ningauble (talk) 11:21, 7 October 2011 (UTC)
16. Oppose. Verifiability, not truth may be a slap in the face when reading it the first time – it certainly was to me because it gave me the impression that the Wikipedia community does not care about the (objective) truth but only about verifiability. But you quickly start to understand the rationale behind this, and the meme becomes a powerful and appealing one. While I understand that not truth provides futile grounds for wikilawyering and a more clarifying version is therefore welcome, I am missing what our current meme delivers so clearly, that "truth" by itself is not an acceptable criterion for inclusion. I am willing to support a version that addresses the issue about truth in its lead paragraph, but not this one. Nageh (talk) 13:28, 7 October 2011 (UTC)
17. Oppose I think the new wording is less clear and actually moves us in the wrong direction with this policy. This is not the place to say that we don't include everything in every article. And I really prefer the idea that just because you think it isn't the truth you still need to source it. I think the current wording does that better. Hobit (talk) 23:32, 7 October 2011 (UTC)
18. Oppose. The discussion below (at my previous neutral comment) makes it clear that this would be a major change in policy, wherein a statement could be challenged or removed not because it was inadequately sourced, and not because any reliable source contradicts it, but simply because an editor claims it to be "untrue". This is a complete reversal of the way our policy has worked for many years. I'm shocked that we would trust some Wikipedian's assertion of "truth" as a reason to censor or exclude sourced content, and more shocked that so many Wikipedians would approved of such a change. The Italian Wikipedia recently shut itself down rather than submit to a proposed law that would mandate removal of material based on an unsupported claim that a statement is detrimental to someone's image. But when it comes to the English Wikipedia, do we really want an unsupported claim of inaccuracy to be a reason for removal, as S Marshall clearly supports below? – Quadell ^(talk) 12:09, 8 October 2011 (UTC)

    I did not say that an unsupported claim of inaccuracy is a reason for removal, and that is not my view.—S Marshall T/C 12:14, 8 October 2011 (UTC)

19. Oppose I think you are right Quadel. I think the great danger to Wikipedia’ usefulness is unfaithfulness to reliable sources because we know the truth. This leads to two blights on the project unsourced statements in articles, and misrepresentation (and ignoring) of sources in articles. At least where an untrue statement is sourced, the reader has the tools to review reliability for themselves. Sourcing requires uncommon effort; anything that denigrates that effort should be rejected. “Verifiability, not truth” is a strong reminder that Wikipedians, for all their ego, still need humility. We are not tellers of truth; we are faithful recorders of what others "out there" have studied -- we merely claim to understand it enough, so that we have recorded their (those out there) truth faithfully.

Verifiability is the ability to verify -- in the current policy, the ability to verify that a reliable source has said something. Truth is not the ability to verify, it has no ability in it - it's an assertion, an ipsa dixit. In the scheme of what editors are doing, it makes sense that we first require them to come to agreement that a reliable source has said something, and agree what that something is (no misrepresentation or mistake because, we know the truth). This is the humility required of us -- we must first take the source on its own terms, even when (especially when) we agree or disagree with it because we know the truth, because in most (all) cases we don't in fact know the truth, and we should in any case assume we and our readers do not, when first approaching the source. Moreover, in practice, and by the dictates of this policy, one editor cannot convince another editor that they have the truth (don't tell me the truth, that's irrelevant, convince me with reliable sources). Only then can editors proceed to agree on correct representation of the source, relevance to the topic, POV and other considerations for putting it in or leaving it out of an article. Unfortunately, the present proposal does not promote these values, it is increased license for editors to promote and propagate unfaithfulness (as if such license were needed) to reporting on reliable sources. Alanscottwalker (talk) 15:56, 8 October 2011 (UTC)

Neutral

• Neutral (Switched to oppose, above.) I agree with some of this. I agree that verifiability is an "initial threshold for inclusion" (while sometimes verifiable information may not meet other thresholds). I agree that verification is not a guarantee for inclusion (though that fact is mentioned elsewhere, so I'm neutral about its inclusion here). My concern is that the proposed wording intends to de-emphasize the fact that truth is not an acceptable criterion for inclusion -- but it could effectively remove that criterion by basing it only on an essay that will no longer be accurate. See my question below. I might switch my comment to "support" if my concerns are adequately resolved, or to "oppose" if they are not. – Quadell ^(talk) 12:36, 7 October 2011 (UTC)

I can't speak for others, but as the primary author of the proposal, I can assure you that my intent isn't to "de-emphasize the fact that truth is not an acceptable criterion for inclusion" (I firmly agree that it is not), my intent is to address the opposite side of the coin: whether untruth is an acceptable criterion for exclusion.

These are two sides of the same coin... The "unverifiable truth" side of the coin has a black and white answer... if the material isn't verifiable, we shouldn't include it, no matter how true it may be. But there isn't a clear black and white answer to the "verifiable untruth" side of the coin... sometimes we should exclude verifiable material that is untrue, and sometimes we shouldn't. It's a case by case determination... and the determination is (in most cases) based primarily on policy concepts other than verifiability. The problem is that the current policy doesn't mention this. It only discusses one side of the coin. The point of the proposal is to address both. It may not do a perfect job of doing so, but at least we try to address it. Does this resolve your concerns? Blueboar (talk) 13:58, 7 October 2011 (UTC)

I apologize if I misrepresented your intentions. I want to be clear here... it sounds like you're saying that there are times that verifiable statements might still need to be deleted based on the claim that they are untrue, and that this wording change is designed to facilitate that. Am I correct in that? – Quadell ^(talk) 16:57, 7 October 2011 (UTC)

Yes, that's clearly the case. An example that the working group came up with, in discussion, was the statement that "Pluto is a planet". That statement is verifiable and I can prove it by reference to otherwise-reliable sources, but it's also been false since 2006 when the definition of "planet" changed. The proposed wording allows for that kind of situation, and more controversial ones too.—S Marshall T/C 18:20, 7 October 2011 (UTC)

I don't believe the truth of that statement matters, in terms of whether it should be included. It's a matter of the weight of the sources. Certainly statements should be challenged if more recent sources contradict it, if corrections are issued, etc., but not merely based on a claim of untruth. To make a simplistic example, if I claim "Cromulent widgets tend to be blue", and I provide a reliable source, and Joe Blow claims the statement is untrue (without providing contradictory sources), his claims are null here. The current wording makes this clear. The proposed wording, while having some advantages, seems to open the door to the possibility that Joe's claim of untruth has weight. If so, I can't support it. – Quadell ^(talk) 19:10, 7 October 2011 (UTC)

It's up to you whether you can support it, but I want to say that I completely disagree with you. Jimbo expressed my view very well when he said: "We are not transcription monkeys, merely writing down what sources say. We want to only write true things in Wikipedia, and we want to verify them." My position is that wikipedians are educators, and there's something sinister about an educator who doesn't care about the truth. We absolutely can, do, and must, make judgments about what's true. Oh, sure, we can put it in terms like "weight of the sources" and try to judge which source is "most reliable", but in fact what we mean when we say that a source is "reliable" or "deserves weight" is that that source is "likely to be true". Thus we put a semantic layer in between article content and truth. But when our judgment about what to include depends on which source is the most likely to be accurate, then the difference between that and making judgments about "truth" is semantics and nothing but. And my position is that we should be intolerant of those who wish to introduce lies into encyclopaedia articles.—S Marshall T/C 19:36, 7 October 2011 (UTC)

I disagree. When we say that a source is reliable and has weight we do not make a judgement about its presumed veracity, but about the degree to which it represents current consensus in the relevant field. I must say that I am surprised at Mr. Wales in my view rather naive understanding of what truth is. To make judgments about what is true is the job of specialist researchers - not wikipedian editors. It is implicit in the word "editor" that we make editorial decisions - of what to include and what not. But editors do not overrule professional researchers conclusions even if they disagree with them, they leave that to the scientific community. You are trying to give wikipedia a function of knowledge creation in addition to its role of knowledge transmission. That could be fine given that that is what a majority of wikipedians feel it should do, but it is not what I signed up for. ·ʍaunus·snunɐw· 20:05, 7 October 2011 (UTC)

Yes, where there's a controversy in any field, Wikipedia prefers the mainstream academic consensus. It's right that we do that. We present the mainstream academic consensus in the simple indicative ("Evolution has led to the diversification of all living organisms from a common ancestor...", quoted from Evolution) and the alternative views as reported speech ("Some advocates believe that major differences in the appearance and behavior of two organisms indicates (sic) lack of common ancestry", quoted from Baraminology). From my point of view, the reason to use reported speech is because the addition of "some advocates believe" turns what would be a false statement in the simple indicative ("The major differences... indicate lack of common ancestry") into a true one. And from my point of view, the reason to present the mainstream academic consensus in the simple indicative is because we default to believing that the mainstream academic consensus is the correct view. But if you take away the value judgments about "truth" from this process, then actually there's no reason to treat the mainstream academic consensus about evolution any differently from baraminology, so either both belong as reported speech, or else both belong in the simple indicative (presumably the latter, since if we don't care about truth, the simpler construction is to be preferred).

This is why I believe the what is truth? view is logically inconsistent and fails to document good practice as well as a truth-based view.—S Marshall T/C 20:31, 7 October 2011 (UTC)

The idea that "verifiability" does not include a brief review of whether something cited by a source is accurate or not, is and always has been an erroneous POV. Verifiability implies much more than confirming that a statement can be found in a reliable source. It also implies that such a statement can be evaluated based upon other, corresponding criteria, such as the authority of the author, the reliability of the publisher, and the relevancy and currency of the statement. In the Pluto example, such a statement fails the currency criteria. We can verify it, but it is no longer current, and this means, it is no longer accurate. This is very simple to understand, so I am unable to grasp its opposition. Viriditas (talk) 09:34, 8 October 2011 (UTC)

Questions

Linked essay bases itself on the current wording

• The proposed wording for this policy links to the essay Wikipedia:Verifiability, not truth in order to explain that truth is not enough for inclusion (a clearly important concept). But that essay bases itself on the current wording here; it's initial sentence is "Wikipedia's core sourcing policy, Wikipedia:Verifiability, defines the threshold for inclusion in Wikipedia as verifiability, not truth." How will this discrepancy be resolved? – Quadell ^(talk) 12:36, 7 October 2011 (UTC)
  • Well, clearly, by editing the essay. We need to see what the consensus wording for the policy will be before we can fix the essay, though.—S Marshall T/C 12:53, 7 October 2011 (UTC)
    • Does it make sense to propose a change to a policy's wording that would both link to an essay and make that essay inaccurate at the same time? – Quadell ^(talk) 13:37, 7 October 2011 (UTC)
      • Yes, it makes perfect sense. Essays are not rules... however, one of the main purposes of essays is to explain the thinking and intent behind the rules. The essay in question is a very good explanation of the the thinking and intent behind the phrase: "Verifiability, not truth". Linking to it will help readers gain an deeper understanding of what we mean by that phrase. Now, we will need to amended the essay slightly if this proposal is passed, but that does not mean we should not link to the essay at all. Blueboar (talk) 14:19, 7 October 2011 (UTC)
        • Although this isn't exactly responding to Quadell's question, I'd like to suggest that, assuming the proposal passes, we have a shortcut that goes directly to the new section about truth. Currently, WP:NOTTRUTH directs to the top of WP:V. I'd very much like to see it target, instead, the proposed new section. --Tryptofish (talk) 16:56, 7 October 2011 (UTC)

The case where a reliable source is wrong in their statement

• How do you propose to resolve the case where a reliable source is wrong in their statement? I remember a case where a statement was made about Wikipedia by a reliable secondary source that was directly contradicted by the primary source of the page histories. (I'd give more details, but I don't feel like digging through archives until I remember). Easily verifiable, but in contradiction to the way we're supposed to be working here. I don't think the old version addressed the issue any better, but the new version hammers more on the published RS side of things.--SarekOfVulcan (talk) 15:35, 7 October 2011 (UTC)

If a source is demonstrably wrong it is ipso facto not a "reliable source" - that is explained in WP:RS. I really wish people would stop trying to force this policy to do the work of a different policy. Roger (talk) 15:43, 7 October 2011 (UTC)

Where in RS is there any guidelines on that? I think RS does a very good job in explaining who to judge a book by its covers, and that's what is needed most of the time, but nowhere in Wikipedia, even not in this proposal, is there particularly helpful advice on what to do when a secondary source is contradicted by primary or raw data sources. Vesal (talk) 15:51, 7 October 2011 (UTC)

Yes actually have a clear policy on that: What we need to do if we find that a peer reviewed publication is contradicted by primary sources is that we write an article about our interpretation of the sources arguing that previous interpretations are wrong and submit it to a peer reviewed journal.·ʍaunus·snunɐw· 16:29, 7 October 2011 (UTC)

And how is checking a primary source to see if it aligns with a secondary sources description of it not Original Research? Are we not in effect saying that our skills as researchers trump the ones of peer reviewed professionals. ·ʍaunus·snunɐw· 16:09, 7 October 2011 (UTC)

I think the problem everyone here is grappling with is the misuse of reliable sources. The simple fact that something is published in a book or journal article is not always sufficient to establish that this is an appropriate source or that the view it expresses is significant. I think everyone would agree about this. But whatever our solution is, it cannot be one that promotes original research, I agree fully with Maunus. I think the problem is that people have turned publication, which was a minimum criterion for verifiability, into the maximum criterion. The problem is that we need more information to know whether the source is appropriate or the view is significant. This requires more research, but not original research as Maunus describes. It means reading enough secondary sources to understand what exactly makes any given book or article important and to whom, perhaps. Slrubenstein | Talk 16:20, 7 October 2011 (UTC)

We are not talking about conflicting views expressed in secondary sources. Sarek's question is about when a reliable source is contradicted by a primary or raw data source. To simplify the example, assume The New York Times, and a bunch of outlets repeating it, state that Jimbo Wales edited the entry of president Obama to say Obama is a really nasty piece of work, and you check the contribution logs and see this edit was made by an impostor, such as User:Jim Wales. What would you do? Wait until New York Times retract their statement? No other views are published in secondary sources, so it is up to you to decide whether you want to repeat untrue information. What kind of research, based on secondary sources, would help us here? What kind of NPOV considerations? Vesal (talk) 16:48, 7 October 2011 (UTC)

We have an editorial discussion and decide whether the statement in the Times is significant enough to be notable- If we decide that it is then we write "On february 10th the New York times published a piece accusing Jimbo Wales of dissing Obama". And we do it because that is a fact. When Wales then decides to retort via the press then we include his statement as well. Any other approach would be Original Research. (you could for example write your finding in a letter to the editorial staff of the times - thus making them retract the statement)·ʍaunus·snunɐw· 16:52, 7 October 2011 (UTC)

Okay, that is a pretty good answer, although perhaps my example was too harmless and does not adequately reflect the situation that Sarek was referring to. Still, I would not say any other approach is original research: there is no attempt to include something in the article, and one should try to avoid repeating libelous information, even attributed, until of course the situation reaches such degree of notability that it can't be avoided, but instead Wikipedia is one of the first places that reports, and escalates, such controversies. Vesal (talk) 17:36, 7 October 2011 (UTC)

I would argue that we handle it this way. The Wikipedia log is a primary source, but in this hypothetical case its reliability trumps the reliability of the secondary source The New York Times. We should always use the best and most reliable source when we try to verify something, and if that source refutes the allegation that Jimbo Wales vandalized the Barack Obama article, then that is the version we should go by. The next questions are "Is the NYT allegation notable?" and "Is it notable that someone impersonated Jimbo Wales?". If the answer to these questions is "no", then I would probably exclude that content altogether. Sjakkalle (Check!) 19:39, 7 October 2011 (UTC)

The proposal attempts to deal with all of this in two ways... first, by noting that Verifiability is a requirement for, but not a guarantee of inclusion... and by noting that we should look to other polices and guidelines to make the determination as to whether to include a specific bit of verifiable information. Blueboar (talk) 17:42, 7 October 2011 (UTC)

But the policy needs to state specifically that untruth is not in itself a valid reason for exclusion, we can include untrue statements if they are notable and attributed to their specific source, and weighted with any contradictory statements according to significance. For example we can include in the article about Monical Lewinsky that "President Clinton stated I did not have sex with that woman", and no one can remove that statement because it is a lie - because it is vrifiable fact that he made the statement.·ʍaunus·snunɐw· 18:07, 7 October 2011 (UTC)

It's a true statement. President Clinton did state "I did not have sex with that woman". Therefore the policy supports its inclusion.—S Marshall T/C 18:15, 7 October 2011 (UTC)

Exactly. It is also true to write "the national inquirer wrote that Leonardo di Caprio has been analprobed by aliens". The reason we don't include that is not that it is untrue, but that it is insignificant. For that reason the question of "truth" is not relevant - only the question of verifiability and attribution. User BruceGrubb is for example arguing that we have to remove a statement by a recognized professional making a claim about the first usage of the word, because he himself has found an earlier usage. I say if the statement is significant we include it attributed to its source, regardless of whether BruceGrubb's or another editors original research suggest that the statement may be factually incorrect. Similarly we don't remove Clinton's statement because an editor happens to have a photo of a sexual act betwen Clinton and Lewinsky, falsifying Clinton's statement. ·ʍaunus·snunɐw· 18:23, 7 October 2011 (UTC)

This kind of example got discussed to death by the working group over a period of about eight months. The situation is that if an editor writes, "Clinton stated, 'I did not have sex with that woman'," then the statement is true; but if an editor writes, "Clinton did not have sex with Monica Lewinsky" then that statement is very arguably false and certainly needs to be refactored. Even though it's sourced to the President of the United States himself. The proposed policy wording clarifies how to deal with such situations.—S Marshall T/C 18:28, 7 October 2011 (UTC)

I don't think it does. On the contrary it suggests that wikipedians can remove statements that they "know" to be untrue regardless of whether there are any reliable sources supporting that conclusion. This undermines both WP:V and WP:NOR.·ʍaunus·snunɐw· 19:10, 7 October 2011 (UTC)

WP:Inaccuracy is intended to be a response to this type of discussion.  This essay currently says, "Ultimately, with allowing for due weight considerations in how the material is presented, and notwithstanding copyright violations, the only reason to exclude verifiable material from the encyclopedia is because it is insignificant."  And

* Levels of exclusion regarding potentially inaccurate material

• We don't use Wikipedia's voice to say it, instead we use inline attribution.
  
• We mention the anomaly in a footnote.
  
• The potentially inaccurate material has so little prominence (WP:DUE), that we don't mention it at all.

Unscintillating (talk) 19:27, 7 October 2011 (UTC)

This policy with which I was not familiar should be integrated much more prominently in the proposed text of WP:V then.·ʍaunus·snunɐw· 19:55, 7 October 2011 (UTC)

This essay is purposed as a guideline to be linked from WP:V and WP:NPOV.  It is new and needs more eyes.  Unscintillating (talk) 20:11, 7 October 2011 (UTC)

Elevating this would go a long way towards alleviating the problems that are driving this discussion. Mangoe (talk) 22:12, 7 October 2011 (UTC)

Need explicit wording

1. The casual reader needs some help to make the distinction between two definitions of verifiability:

• common: The ability to determine if a claim corresponds to truth.
• Wikipedia's: The ability to determine if a summary of a claim corresponds to its appearance in a published reliable source. patsw (talk) 15:27, 8 October 2011 (UTC)

2. I would also like to see an explicit denial that as a collaboration we are indifferent to truth. My own contribution along these lines is here: Wikipedia:Verifiability, not truth#Editors are summarizers not truth finders patsw (talk) 15:27, 8 October 2011 (UTC)

Example of Primary/Secondary source conflict

I will try to neutrally present the Phil Knight example for editors that haven't followed the discussions. A reliable secondary source, Phil Knight, wrote that the term conspiracy theory dates back to an article published in 1909. An editor, User:BruceGrubb, showed that a number or earlier works contained this term. The simple truth here is that the earliest usage noted by the Oxford English Dictionary is an article published in 1909, and the OED does not claim to be noting first usages, only important early usages, but this is not what our source said. Current consensus seems to be:

• It is okay in a case like this, when everyone agrees that the secondary source is wrong, to simply remove it.
• It is not okay to go for an NPOV-based solution in the article: Peter Knight states the first recorded use of the phrase "conspiracy theory" dates from 1909" (ref); however, the phrase "conspiracy theory" also appears in Garrison, George Pierce (1906), ..." This is a synthesis implying that Knight is wrong.

If this is an unfair summary, feel free to correct me. Vesal (talk) 17:01, 7 October 2011 (UTC)

• Proving Chris Knight wrong in that way is Original Research, yes. In this case it is also difficult to know whether the data actually contradicts him. Two words occurring together is not a term, it is only a term when it takes on a fixed meaning. We don't know whether Knight was saying that the first time "conspiracy theory" was used in its modern meaning was 1909, or whether he meant it as an absolute statement of the first time the words occurred together as a phrase. In any case the right thing to do would be to write a professional article about the term "conspiracy theory" using the editor's knowledge of earlier ysages to suggest that there are earlier significant usages than the ones in the OED. Wikipedia does not publish original research. It is of course always OK to make an editorial decision that a claim in a certain source should not be included. It is also fully acceptable to include a statement "Peter Knight states the first recorded use of the phrase "conspiracy theory" dates from 1909" because he did state that, and the claim coincides with the earliest usage given in the OED. Nothing untrue about that statement. ·ʍaunus·snunɐw· 17:32, 7 October 2011 (UTC)

  I mostly agree with your analysis, but I would rather not discuss so much what is the right thing to do in this example. Rather, please think more in terms of whether you think current policy gives adequate guidance on cases like this, and whether you think the proposal would improve the situation, as Blueboar says above, and if not, what could be said to provide some guidance about these kind of cases. Thanks, Vesal (talk) 17:50, 7 October 2011 (UTC)

The policy needs to make it clear that wikipedia editors are not supposed to judge what is true and what isn't - we only judge which views about the true are significant and which aren't. That is why I think it is important to have a policy that does not suggest that we can remove sourced statements if an editor finds them to be "untrue" or nt in line with their view of reality. I think that it can be made clear that in cases where there is any reasonable doubt about the truth value of an otherwise significant statement we need to attribute the view to its specific source and never state it as fact in wikipedias own voice. ·ʍaunus·snunɐw· 18:01, 7 October 2011 (UTC)

I do not think this is an acceptable way to write an accurate encyclopedia. The contrary argument, the one that also helps keep reliability going as a hot topic, is that this theory substitutes form for accuracy, and that real editing requires evaluating sources against each other as to their real accuracy as well as their conformity to are preferred forms for sources. Here's an example: an interviewer incorrectly stated that Erin Burnett attended St. Andrew's Episcopal School (Maryland), presumably because her hometown is Maryland. However, it is simply physically impossible for her to have attended that school, as her hometown and the school are on opposite sides of the Chesapeake Bay, and the school does not board students. She actually attended St. Andrew's School (Delaware), which is a boarding school much closer to her hometown, and I was eventually able to prove this inarguably by finding a copy of her high school yearbook. In the meantime, however, I was met with a great deal of resistance in removing the incorrect statement. Since it could be cited to a news outlet, people insisted that need to remain in the article even though nearly everyone in the discussion agreed that it was extremely unlikely to be true.

That is why "truth is not an acceptable criterion for inclusion" is, on its own, a problematic statement. The confidence is that verified statements can be assumed to be true, and for the most part we are willing to let the possibility of verification stand as a proxy for actual checking. That's why we talk about "reliable sources": reliability ultimately means being trusted to convey accurate information. The implication through all of this is at truth is a necessary criterion for inclusion, but that one of our assurances of truth is that statements can be verified (and in a thorough world, would be verified). But the possibility of verification does not always come through on its promise to deliver accuracy, and when it fails, then verifiability is not of itself a sufficient criterion for inclusion. Mangoe (talk) 22:35, 7 October 2011 (UTC)

In my view you are undermining the foundations of wikipedia by suggesting that using primary sources to trump secondary ones should ever be allowed. The case about the schooling of Erin Burnett in my view should have been settled by including the sourced information and then upon evidence to the contrary simply remove the statement - without including the statement from the primary source. In this case it matters little since it is a trivial issue - but it in other more controversial cases it is of the top most importance not to encourage any kind of original research, but to rely solely on published interpretations of primary sources. Verifiability in tandem with notability/significance and weight will always be sufficient for inclusion regardless of what any wikipedian thinks is the truth-value of a claim.·ʍaunus·snunɐw· 22:54, 7 October 2011 (UTC)

We disagree about the reference to the yearbook; it is a secondary source. That is not, however, germane to the point: that people insisted on including the erroneous information simply because it could be cited to a source which fit the formal standard for reliability. That is what you advocate in your final sentence, so your defense of what I wished to do before I found some other source, primary or not, is in contradiction to your own principles. Your principle says that the article was obligated to continue repeating the mistake potentially indefinitely, as there was never a guarantee that someone would find a formally acceptable source. Mangoe (talk) 02:29, 8 October 2011 (UTC)

No my principle doesn't contain any such obligation - it is an editorial decision whether to include or not. ·ʍaunus·snunɐw· 03:44, 8 October 2011 (UTC)

No, that's not what you said. You said it was a "sufficient" condition, meaning that no other argument is needed to justify inclusion of the material. But it's actually a necessary condition; other considerations may result in the discarding of the material. This is really the crux of the matter: in the Real World, accuracy is what is necessary, and verification is one of the means to that. But saying that a source is all that is sufficient keeps leading people to the conclusion that, since a source is sufficient, the accuracy of the source cannot be taken into account. That's how we get the insistence that patently erroneous sources must nonetheless be reflected in the article.

Having a source is usually sufficient as well as necessary, first because as a practical matter we are deficient in fact-checking, but second, because most formally reliable sources turn out to be actually reliable, and thus there's no real reason to approach them too skeptically. But this sufficiency must give way in the fact of actual verification; having a source is NOT always sufficient. Mangoe (talk) 13:12, 8 October 2011 (UTC)

You are taking the word "sufficient" out of context I said verifiability in tandem with notability/significance and weight (= editorial decision). Please don't twist my worth to something that suits your argument better.·ʍaunus·snunɐw· 13:17, 8 October 2011 (UTC)

Editorial decision renders all that goes before it as necessary rather than sufficient.Mangoe (talk) 14:02, 8 October 2011 (UTC)

(ec)Here is analysis using WP:Inaccuracy for Erin Burnett.  By inductive reasoning that can be done by a high school student, I induce that the interview had potentially inaccurate material.  This means that I reduce to some degree the due weight.  For a complex of reasons, I conclude that there is no reason to think that readers want to know about this material with inline attribution, or in a footnote, so I conclude that the material is insignificant.  The high school classbook is reliable to say that Erin Burnett attended that school.  So we now have a secondary source that is insignificant, and a reliable primary source.  Without the classbook, the article says nothing about the high school, consistent with WP:Editing policy, "an absence of information is preferred to false information".  Unscintillating (talk) 03:04, 8 October 2011 (UTC)

Phil Knight's exact Verifiability statement was "The first recorded use of the phrase "conspiracy theory" dates back to a history article from 1909". The statement "the phrase "conspiracy theory" also appears in Garrison, George Pierce (1906)" is also Verifiability.

WP:NOTOR (which many editors like to ignore as it shoots their argument to blazes) expressly states "Comparing and contrasting conflicting facts and opinion is not original research, as long as any characterization of the conflict is sourced to reliable sources." The "characterization" of the conflict in this case is that the phrase "conspiracy theory" does indeed appear before 1909 and per WP:NOTOR showing this is NOT OR--QED.--BruceGrubb (talk) 03:52, 8 October 2011 (UTC)

You are wrong in that interpretation - the characterization of the conflict would be if there was a published source noting that Chris Knight was wrong when he said that. Since there isn't saying so is 100% classical OR.·ʍaunus·snunɐw· 12:21, 8 October 2011 (UTC)

By the way it is clear from the edit history that you like that essay a lot. It seems you have in fact inserted the example about the occurence of a word yourself and editwarred to keep it there. In any case the paragraph you quote says "Comparing and contrasting conflicting facts and opinion is not original research, as long as any characterization of the conflict is sourced to reliable sources. If reliable references cannot be found to explain the apparent discrepancy, editors should resist the temptation to add their own explanation. Present the material within the context contained in reliable sources, but avoid presenting the information in a way that "begs the question". An unpublished synthesis or analysis should not be presented for the readers' "benefit". Let the readers draw their own conclusions after seeing related facts in juxtaposition."·ʍaunus·snunɐw· 12:27, 8 October 2011 (UTC)

Maunus, I am 100% with you on the WP:OR / WP:SYN point. However, I don't see Blueboar's wording as encouraging synthesis. Synthesis is a perennial problem—that is why we have WP:SYN, and I don't see the proposal as weakening it. When it's a question of whether Oslo's entire transportation system came to a halt or not (both reported in reliable sources), or whether the Italian Wikipedia was locked down for two days or three (both reported in reliable sources), then as editors we are sometimes in a position to be able to decide which of the sources was better researched and has it right. That is something for discussion and agreement on the talk page, and that is all the new wording is saying. To insist on giving the factually wrong version as well as the right version in such a case does not serve the reader. --JN466 11:14, 9 October 2011 (UTC)

This policy and the core policies

If, as certain editors are clearly arguing, this proposed change means that wikipedia editors can now challenge claims in secondary based on editor's own interpretation of primary sources, or perhaps even on unpublished "evidence", then this will have profound effects on our other policies e.g. WP:OR and WP:SYNTH (because doing so clearly falls under the definition of original research by synthesising two sources into a new conclusion not contained in either of them) and WP:NOT (because it will turn wikipedia into a publisher of original thought, in effect competing with venues for the publication of research). In my view therefore this forum is too small to make this decision and a broader community consensus should be sought. We are playing with the very foundations of wikipedia here, and doing it on a single policy talkpage. Doesn't seem right.·ʍaunus·snunɐw· 13:09, 8 October 2011 (UTC)

• Let's imagine that Michael Green makes a breakthrough at the bleeding edge of string theory. His paper is published in Physics Letters B (primary source). It is reported by the BBC (secondary source) as front page news. But there's a trivial difference between the equation as published in Physics Letters B and the equation as published by the BBC. Which do we trust?

  In my opinion we would be right to prefer what's said in Physics Letters B. I think the view that we should never challenge secondary sources based on primary sources is very extreme indeed, and it would not be accepted by the majority of reasonable editors. I agree that we should not normally do so lightly.

  I object to moving the discussion to a wider venue. I understand why someone whose view is not in accordance with an emerging consensus would want a fresh start in a different venue, but I want to draw your attention to WP:FORUMSHOP.—S Marshall T/C 13:20, 8 October 2011 (UTC)

That is not what I am arguing. Some editors here are arguing that we could challenge Michael Green's letter B article by making our own calculations. In the case you give Letter B is a more reliable source than BBC news and any editorial group would automatically reject BBC's claims as unreliable compared with his own statement. It feels like you are twisting my words, you are certainly not reading them in a very collaborative way.·ʍaunus·snunɐw· 13:26, 8 October 2011 (UTC)

Also cut it out with the accusations of Forumshop - forumshop would be if I took it to a different forum. Which I probably should since you are making a decision here in a small closed forum that will have a much wider impact. It is completely customarily to take such decisions to a venue where the entire community who will be impacted can be involved. ·ʍaunus·snunɐw· 13:28, 8 October 2011 (UTC)

Maunus, most of what you say makes sense to me, but you are seriously confusing what some of us might have said in these discussions with what the proposal is actually saying. The proposal does not say that assertions of untruth, even if they be based on primary source evidence, would suffice for removal. It really does not say that. Sarek asked how this proposal would deal with such situations. Why ask this if the proposal actually said something one way or another about such cases? All the proposal is saying is that such cases are complicated, and it actually defers judgment to WP:DUE. Do you really find that problematic? Vesal (talk) 13:37, 8 October 2011 (UTC)

I realize that the proposal is not saying this directly - it is however already being interpreted as if it is saying those things by several editors. The rationale for the proposal is that some editors misuse "not truth" - now we have some editors misusing this proposal before it is even instated. The very least we can do then is modify the proposal to specifically not support any instances of original research.·ʍaunus·snunɐw· 13:56, 8 October 2011 (UTC)

I'm not seeing how the concerns that Maunus states lead to the conclusions.  If "wikipedia editors can now challenge claims in secondary based on editor's own interpretation of primary sources, or perhaps even on unpublished 'evidence' ", this does not mean that any such discussion leads to inclusion. Or, as Maunus states it, "turn wikipedia into a publisher of original thought".  I'm only aware of three end results for such a discussion: the questionable material is stated with inline attribution, the questionable material is noted in a footnote, or editors agree that the material is sufficiently insignificant that readers don't want to know about it.  In none of these cases does there exist publication of original thought.  Since the word "initial" in the proposal confounds the idea of inclusion, maybe I'm missing something, but if so, I think that Maunus needs to better explain the problem, preferably with an example or examples.  Unscintillating (talk) 14:31, 8 October 2011 (UTC)

Indeed in the examples we hear of, all or most people on different sides of the argument were still following the principle of trying to avoid OR. There just was not agreement that there was OR?--Andrew Lancaster (talk) 14:37, 8 October 2011 (UTC)

• OK example: If we accept that BruceGrubb's philological research discovering that Chris Knight mistakenly referred to the OED as providing a reference to the first usage of the word "conspiracy theory" when in fact there are earlier usages is not to be considered Original Research, then we have to ask what happens when we apply the same criteria in another, less trivial venue.? Let us say that the established consensus in the field of history is that Bartolome de Las Casas was born in 1474 and many biographies and documents repeat this date. Now BruceGrubb or another very thorough and dedicated wikipedian decides that this is weird since we know that Las Casas died in 1566 and that that would mean that he would have lived the surprisingly long life of 94 years. He travels to Seville gets a research permit and conducts an investigation in the Archive of the Indies fiunding out that there are records showing that his actual birth year was 1484 and not 1474. He now comes to wikipedia and produces a scan of the original document suggesting that we remove the claim that Las Casas was born in 1474 although all available published sources say so. What do we do? Do we assume that BruceGrubb or whoever did the research is competent enough in that field of research to correctly interpret the document? Do we remove the claim supported by multiple sources, because they are now possibly incorrect? If we follow the way that BruceGrubb interprets the proposed policy we should. This would make wikipedia the first media to publish the change of Las Casas birthdate with ten years, a considerable change to the written history of history Spain and colonial Latin America. Wikipedia would also be in direct contradiction to all professionally published research in the area. Does this not look like a problem to you all? It does to me, since it basically means that we have to rewrite WP:NOT, crossing out "not a publisher of original thought" or at least adding "not a publisher of original thought, unless it is true". That's one example of a possible consequence of this interpretation. I can make up more. The examples are fully parallel except that BruceGrubb only conducted a google search to find an earlier instance and not actual archival search, but I see no good reason why these should be considered different circumstances policywise. (BTW: Helen Rand Parish found the document that changed Las Casas birthdate and published an article about it in the The Hispanic American Historical Review in 1976 - that is what I would encourage the wouldbe BruceGrubb to do instead of publishing it here first. It would also enhance his own prestige as a scholar more.). ·ʍaunus·snunɐw· 18:09, 8 October 2011 (UTC)

• In that situation, what we do is tell the truth. The truth being: "Almost all sources state that Las Casas was born in 1474."—S Marshall T/C 18:28, 8 October 2011 (UTC)

Then why shouldn't we handle the other situation by stating the "truth" that "the earliest use of conspiracy theory recorded in the OED is 1909"? (supposing that Chris Knight is simply relying on the OED when he said 1909, and accepting that an interview with a scholar who is not a philologist is not a more reliable source than the OED)·ʍaunus·snunɐw· 18:33, 8 October 2011 (UTC)

• There's no reason why that sentence shouldn't form part of what we say.—S Marshall T/C 18:45, 8 October 2011 (UTC)

What do you mean exactly? That you agree with me that BruceGrubb's argument for removing the claim because he found an older source is OR? If that is the case then why do you not think that the proposed version should make it clear that policy does not support that kind of approach?·ʍaunus·snunɐw· 18:47, 8 October 2011 (UTC)

• I think this discussion is getting seriously hung up on this one particular case, and I want to point out that hard cases make bad law. My opinion is that it would be reasonable, in that particular case, to say: Peter Knight states the first recorded use of the phrase "conspiracy theory" dates from 1909,(ref) but the phrase "conspiracy theory" also appears in Garrison, George Pierce (1906). I feel that it can't possibly be original research to identify what's said in a published work with citations.—S Marshall T/C 18:56, 8 October 2011 (UTC)

YouThey asked me to provide an example, then you say that hard cases make bad laws. ... Bad laws are laws that don't consider the implication they may have before hand. Here we have a case where several editors are expressing that they are already interpreting the proposed "law" to mean that they can now freely do types of edits that would previously be rejected as synthesis. One would have to be very optimistic to pass the law without making sure that doesn't happen first.·ʍaunus·snunɐw· 22:37, 8 October 2011 (UTC)

I don't recall asking you for an example. Have you confused me with someone else? I also think you're confusing my opinion on this one particular article with the general consequences of this policy. I would have said what I said above irrespective of how WP:V read. I have never agreed that it's appropriate to edit without regard for the truth, and I have never had any patience with those who think adherence to our policies means it's acceptable to allow lies in an encyclopaedia. But, and this is the key point, my personal view on this particular article is not what Blueboar's proposed edit to WP:V says. Blueboar's proposed edit says that if there is a dispute about whether reliably-sourced material may be untrue, then editorial judgment is required and the correct response is to seek talk-page consensus.—S Marshall T/C 22:55, 8 October 2011 (UTC)

I think you are confused about the difference between the meanings of the words "lie" and "inaccuracy".·ʍaunus·snunɐw· 23:00, 8 October 2011 (UTC)

Really? I think that an inaccuracy might be a mistake. I think that once there's a good faith suspicion that material might be inaccurate, but we continue to say it anyway, it's become a lie. Could you point out where I'm confused about that, please?—S Marshall T/C 00:05, 9 October 2011 (UTC)

It does not because we are not saying that it is the case, we are saying that a particular authority says it is the case. Because Wikipedia does not produce knowledge only transmit it. The average wikipedian is not, and should not be expected to, be able to be qualified to judge whether a statement published by a reputable researcher in a reputable venue is right or wrong - that should always be left to other experts publishing in the same venue.·ʍaunus·snunɐw· 00:40, 9 October 2011 (UTC)

You say that as if it was uncontroversial—as if everyone agreed that Wikipedians really were just transcription monkeys, reporting what the sources say without evaluating whether the sources are telling the truth. I'm really quite surprised that an encyclopaedia-writer can think like that. I don't agree at all that Wikipedians aren't qualified to judge the truth. I know that there are fields, and levels of complexity, where I'm personally quite capable of deciding what's true and what isn't; and I'm sure that there are fields, and levels of complexity, where you're just as capable. If there's a disagreement then we discuss it on talk. I mean, yes, I agree that our job is to summarise what the sources say, but I think we should evaluate the sources before we use them. The evaluation process is to decide which sources are the most "reliable", by which we mean, which sources are the most likely to be true. The point is to make sure that what we write is reliably-sourced, trustworthy, accurate, and other synonyms for "true".

As I said, I don't understand why you think Wikipedians aren't capable of deciding what's true and what isn't, but as you can see, the consensus is against you. An emerging but clear majority prefer Blueboar's wording, which requires sources but also requires that those sources tell the truth. I hope this doesn't make you too unhappy.—S Marshall T/C 01:08, 9 October 2011 (UTC)

• Yet many editors have argued this very example is OR via SYN because it implies Knight is wrong. Much this same "reasoning" is being used to keep two reliable sourced direct quotes out of the Weston Price article because they imply something regarding Price. Never mind they would fix the serious NPOV issue that article has because much of it is based on what others currently are doing with his earlier work rather than what Price himself actual said and did. In otherwords OR via SYN is being used to effectively break NPOV in that article--wonderful. Regarding the Hard cases make bad law that argument falls flat on its face the moment you realize that that hard cases like the Winterbottom v Wright situation eventually resulted in consumer protection laws reversing the president set forth in that case.--BruceGrubb (talk) 19:07, 8 October 2011 (UTC)

Exactly it implies a synthetic by begging the question - therefore it is synthesis.·ʍaunus·snunɐw· 22:39, 8 October 2011 (UTC)

• As I've already said, hard cases make bad law. My position in this example is that we can't say the 1909 date without qualifying it, because it's a crime against the basic purpose of an encyclopaedia to publish known error. Exactly how we qualify it is a matter for editorial judgment and local consensus. But my position is also that we need to consider the phrasing of the first sentence in the light of the many other examples discussed on WT:V over the last eight months, and to follow this example so closely without considering the others we've exhaustively discussed is inadvisable.—S Marshall T/C 19:25, 8 October 2011 (UTC)

• The problem is that hard cases make bad law doesn't always apply--the Triangle Shirtwaist Factory fire comes to mind as does the public outcry regarding food quality triggered by Sinclair's The Jungle. Sometimes the hard case is the straw that brakes the camel's back or would you prefer a world where you have a good chance of getting real Swede in your Swedish meetballs because some worker fell in the meat grinder?--BruceGrubb (talk) 19:45, 8 October 2011 (UTC)

BruceGrubb, you yourself said that your intent was not to add the statement to the article, and you've brought this same issue up in multiple venues, so this appears to be a bit of a crusade. If there were some controversy covered in reliable sources about the first usage of the phrase, it might be appropriate to include the statement as S Marshall suggests. If we had a reliable source that said that Knight was incorrect, it might be reasonable to use that sources to show that Knight was wrong. As it is, we have an instance in which Knight made an incorrect statement about the first usage of the term--he is a reliable source in general for the topic, but his speciality is American studies, not etymology or philology, so as a source for first usage he is weak, there's been no evidence presented that there is any controversy amoung experts on the subject of conspiracy theories about when the first usage occurred, and Knight appears to only states the first usage date to contrast the sparse usage in the early part of the century with the increased use of the phrase in its current meaning during the 1960s. The only apparent purpose inclusion of that statement in the article serves is to imply that Knight is wrong, and in that sense it is, in my opinion, a violation of SYNTH. That appears to have been the general consensus amoung those directly involved in that discussion.

That all being said, S Marshall is entirely correct that hard cases make bad law. How we deal with an incorrect source is not just a question of V, but also of DUE, NPOV, and NOR, and very likely other policies and guidelines as well. In the hypothetical case of Las Casas birthdate, I concur with Maunus that even if we believe something is incorrect, if reliable sources agree about it, we stick with the reliable sources, and not publish our views. Let us say that one travels to Seville gets a research permit and conducts an investigation in the Archive of the Indies finding out that there are records showing that his actual birth year was 1484 and not 1474, and produces a pdf. Who are we to judge whether or not those records were the right ones to consult, that they are not forgeries or mistakes in themselves, that the pdf presented is what the presenter claims to be? Those are tasks for experts in the field, not us. The situations are parallel--we are not experts, we should not attempt to determine whether the use of "conspiracy theory" by primary sources in the 1800s is the same as current usage, that the comparison is apt, any more that we are in a position to judge in what sense a particular instance of the phrase "National Socialism" is used in a primary source, see [5]. This is the reason we prefer to rely on secondary sources for most material in articles and advise caution in using primary sources. BG's investigation has shown that a secondary source was incorrect about a minor detail, but it can show us neither what that inaccuracy means nor the implications of the earlier usages in the context of the subject of the article. To try to do either would violate NOR. In the particular case at Conspiracy Theory, I think we took the best course, by consensus, of simply removing the material, but I do not think we can draw a significant conclusion in general from that particular situation. Each situation will be different, we can come up with rules to help us deal with cases, but not a rule to tell us what rules apply to which cases--at some point we must rely on our ability to reach consensus as to our general intent. I think that's why IAR is a pillar, to give us the ability to ignore rules when it becomes clear that applying them is not a good choice when doing so violates our goal. --Nuujinn (talk) 20:12, 8 October 2011 (UTC)

Reply to ·ʍaunus·snunɐw· 18:09, 8 October 2011 regarding the Las Casas example:
WP:OR states, "The prohibition against OR means that all material added to articles must be..."  So to have an example of WP:OR we need an example of material that has been added.

The example considers a "document" wherein it is implied that this document is not considered as a Wikipedia reliable source, perhaps because it is not accessible publicly. There is a disconnect in the example, a leap that removing bad material in an article implies adding unsourced material to that article. The example states, "Do we remove the claim supported by multiple sources, because they are now possibly incorrect?...This would make wikipedia the first media to publish the change of Las Casas birthdate with ten years..."  No, a removal would not be followed by adding anything, it is as stated in WP:Editing policy, "on Wikipedia a lack of information is better than misleading or false information".  Since nothing would have been added to the article, there is no WP:OR to consider.  Unscintillating (talk) 20:25, 8 October 2011 (UTC)

So you are suggesting that the article shouldn't have the information presented in all literature because one wikipedia suggests it may be incorrect? That is quite similar to the situation the italian wikipedia is striking against - the power of one person to veto the inclusion of wellsourced information that they believe to be factually incorrect.·ʍaunus·snunɐw· 22:43, 8 October 2011 (UTC)

No, Unscintillating is suggesting that one editor can raise his/her concern that the information is incorrect... even if this concern is based on OR, there is nothing wrong with discussing the concern on the article talk page... and after that discussion, a WP:Consensus of editors might agree to omit the information. We do this all the time. We do not violate any policy if we choose to not discuss something.

Nothing says we have to give the date of someone's birth in an article. If our sources conflict, or if we have convincing evidence (even OR evidence) that the source we got the birth date from is inaccurate... we always have an option of simply not mentioning any birth date. Blueboar (talk) 02:06, 9 October 2011 (UTC)

There is also nothing wrong with an editor voicing his opinion on the talkpage of Elvis Presley stating that he knows for a fact that he is still alive because he saw him in Walmart in Reno Nevada yesterday and has a photo to prove it. But there is no reason to give him extra backing in the policy to suggest that his claim should influence what goes in the article on Elvis Presley or not. We don't have to mention death dates either.·ʍaunus·snunɐw· 02:33, 9 October 2011 (UTC)

The issue presented is that an editor has a recent photo of Elvis taken in Reno, Nevada, and the editor asserts that Wikipedia has a potential inaccuracy because it says that Elvis is dead.  What changes with the proposal, IMO, is the degree to which we consider the assertion.  Before, we could say that the claim is not verifiable and that we don't want to spend any time on it.  With the proposal, we analyze the assertion, because if we ignore assertions of potential inaccuracy, the effect is to leave potential inaccuracy in the encyclopedia.  At this point, further analysis seems not to be useful since the example is a hypothetical, but the idea that this particular assertion could or would "influence what goes in the article" is IMO not warranted.  Unscintillating (talk) 04:42, 9 October 2011 (UTC)

I don't see how anything changes. The hypothetical example you offer goes on all the time, every day in fact. And we don't take it seriously unless there are RS saying, "yes, Elvis may have faked his death, and the authorities are looking for him", or if the topic of the article is the Elvis Presley phenomenon and there are RS supporting the Reno sighting. Viriditas (talk) 09:09, 9 October 2011 (UTC)