Wikipedia talk:Verifiability

Machine translation

The guidelines for non-English sources (WP:NONENG) include the following statement: "Translations published by reliable sources are preferred over translations by Wikipedians, but translations by Wikipedians are preferred over machine translations." I propose that this guideline should be revised. Translation is a complex cognitive task, and while machine translations may sometimes be accurate, assurance of quality requires a human agent. As the relevant article correctly states "Relying exclusively on unedited machine translation ignores the fact that communication in human language is context-embedded and that it takes a person to comprehend the context of the original text with a reasonable degree of probability. It is certainly true that even purely human-generated translations are prone to error. Therefore, to ensure that a machine-generated translation will be useful to a human being and that publishable-quality translation is achieved, such translations must be reviewed and edited by a human." The policy should state that fluently bilingual Wikipedians may use a machine translator, provided that they evaluate and correct the output, but that unassisted machine translations should never be permitted. I'd also suggest a new requirement for all translations to be independently checked by another editor. Rubywine . talk 02:23, 5 August 2011 (UTC) On second thoughts it's impractical to impose a proofreading requirement. Rubywine . talk 11:43, 6 August 2011 (UTC)

Unassisted machine translation pretty much has to be permitted, because we don't always have another good option, especially for less-common languages. A machine translation is more friendly to the readers who don't read the other language than presenting them solely with a completely unintelligible quotation in a foreign language. We have non-English quotations and phrases in all kinds of languages, and it is unreasonable for us to assume that any of our readers know anything other than English—much less that any single reader will be able to understand العربية and 한국어 and Македонски and 文言—and in the course of reading the English Wikipedia, they might easily encounter all of those languages at least once.

When and if we get a competent translator to that article, then the machine translation can and should be replaced by a good translation by a skilled human, but that doesn't mean that we have to leave readers in between wondering what it says.

Additionally, WP:Nobody reads the directions, so no matter how much you discourage it, it's actually going to happen. WhatamIdoing (talk) 03:08, 5 August 2011 (UTC)

I don't agree at all. There is no need to include machine translation into the article at all, in fact in doubt you would classify them as an unreliable source. The fact that we cannot expect readers to be versed in all those language is somewhat irrelevant, as all that the reader actually has to do is plugin the text into an automatic translator (such as Google). The notion that we have machine translation of foreign language quotes seems rather unappropriate to me.--Kmhkmh (talk) 06:04, 5 August 2011 (UTC)

Fluently bilingual (and multilingual) Wikipedians may use a machine translator, of course. Using a machine translator but evaluating and correcting the output is what I do when I'm translating, and I don't see a problem with it. As for the requirement that all translations should be independently checked—my question is, by whom? I certainly don't offer to spend my volunteering time checking other people's translations on a regular basis. I'll consider doing so on specific, respectfully-worded request on my talk page. No policy is going to require me to do it.

The basic problem with translations is the misapprehension that any editor is entitled to require other editors to help them check sources. I've seen editors demand that independent editors run around translating texts in languages they don't speak, on various occasions, and the answer is "bugger off".—S Marshall T/C 09:14, 5 August 2011 (UTC)

No policy requires you to do anything at all. Nobody here is entitled to demand anything from anybody. Regardless of how much translation work you volunteer to do, there's no reason for any policy to accommodate your personal disinclinations. There are thousands of bilingual and multilingual Wikipedia users; if translation work is only being undertaken by a very small group, that is undesirable and unnecessary. I am in favour of an active drive to engage a great many more people in this work. Rubywine . talk 11:11, 5 August 2011 (UTC)

I think nobody has an issue with machine translationen being used as tool. The issue here is merely that we should not include unassisted machine translation into articles (as quote or citation translations). Any unassisted machine translations shouldn't be a part of the article itsself due to potential unreliablity. That unsassisted machine translations are used by editors to perform a crude check of foreign language sources or on discussion pages to support an argument is perfectly fine of course. If somebody is using them for assisted translations that's probably fine, though personally I'm a bit skeptical there. However how an editor exactly produces a translation is out of our control anyway and probably not really our busines to begin with (as long as there is no legal issue like a copyright violation). The only thing that matters is the quality of the translations when he enters it into WP and that is responsible for it (rather than some program).--Kmhkmh (talk) 14:04, 6 August 2011 (UTC)

I agree with Kmhkmh in every respect. WhatamIdoing, you seem to be proceeding from an underlying set of assumptions which I think are wrong, and inconsistent with policy. Wikipedia: Translation specifically states: "Translation takes work. Machine translation often produces very low quality results. Wikipedia consensus is that an unedited machine translation, left as a Wikipedia article, is worse than nothing." I see no reason to take the view that unedited machine translations of smaller quotes and phrases are any more acceptable. The worst possible outcome is for readers to be misled and misinformed. Rubywine . talk 11:30, 5 August 2011 (UTC)

Machine translations are vaguely accurate but can introduce problems. At the very least the grammar is usually very poor; at worst the translation is misleading and just plain wrong. It really depends on their usage - if it's copy/pasted without checking then no, it shouldn't happen. However I and I'm sure many others have used a machine translation to inform and suggest content which is then written/copyedited into prose. That is fine. violet/riga _^[talk] 14:34, 5 August 2011 (UTC)

It sounds like you you've all missed the context of this question. Machine translation is only mentioned once in this entire policy, and it is mentioned solely in the context of supplying information about non-English reliable sources. This is not about translating whole articles: This is not about the fact that Bacterial pneumonia is way better over at the Spanish Wikipedia and I'd like to have some competent person translate it for us, or any situation even remotely like that. For that purpose, which has nothing to do with this entire policy, I agree that machine translation would be inadequate.

This recommendation appears in WP:NONENG and is specifically about direct quotations from non-English sources. There are only two reasons why you might be typing a direct quotation from a non-English source into Wikipedia:

1. You've been asked for a quotation from a non-English source, and the person who asked you doesn't speak <fill in the blank>, so you've been asked to provide a translation into English.
2. The article needs to include a famous quotation, book title, or similar short phrase for some purpose.

The policy says that in such cases, you need to supply the best translation you can, which means (in order):

1. A proper translation by a reliable source is best.
2. A translation by one of our competent editors is second-best.
3. If all else fails, a machine translation is better than nothing.

This means, to give a practical example, that at Thomas Corneille#Notes, we don't stop with just "In a letter to her father, shortly before her execution, Charlotte Corday quotes Thomas Corneille: "Le Crime fait la honte, et non pas l’échafaud!"" and leave the 97% of the world that can't read French wondering what it means.

Instead, we add "(The crime causes the shame, and not the scaffold!)", which is a decent translation (since we have one) immediately afterwards. If we didn't have this decent translation, then we would add the worse-but-not-too-awful machine translation, "(Crime is a disgrace, not the scaffold)". What we don't do is leave all the people who don't read French scratching their heads over why that sentence is in the article and what it means. [ip address redacted] User:Fred Bauder Talk 19:41, 6 August 2011 (UTC)

Thanks for your thoughts and comments, and also for pointing us to the relevant policy - that's most helpful. However, you're wrong to think that we've all missed the context. Yesterday, after quoting Wikipedia: Translation: "Wikipedia consensus is that an unedited machine translation, left as a Wikipedia article, is worse than nothing", I said "I see no reason to take the view that unedited machine translations of smaller quotes and phrases are any more acceptable."

Your argument has actually consolidated my viewpoint. Direct quotations are very likely to be idiomatic, so I disagree with WP:NONENG re machine translation. The example you've given is not bad at all, but I'm confident there'll be worse ones. I'd rather provide readers with a foreign-language quotation than with a hit-or-miss, potentially very bad translation of an idiomatic phrase. I think it's ok to leave the readers scratching their heads. They can look elsewhere for a good translation, or, if we really want to be helpful to them, we can source one for them ourselves. Rubywine . talk 18:18, 6 August 2011 (UTC)

Now consider the other situation: Someone asked you for a quotation to show that the source really does support the text. You provide a quotation, and the other editor says, "But I don't know what that means; I don't read ____". And you say, "Well, that's too bad, because I'm no good at rendering it in English, and Rubywine refuses to let me give you a machine translation." Is providing no translation, on the grounds that you can't provide a really good one, going to help that editor at all? Would that not be exactly the sort of rule that is described as "a rule that prevents you from improving or maintaining Wikipedia"? WhatamIdoing (talk) 18:55, 6 August 2011 (UTC)

If the machine translation of the quotation is merely clumsy, then it's probably better than nothing. However, there's every chance that a machine translation of an idiomatic expression will be nonsense, or far worse, completely misleading. A machine translation can convey the precise opposite of the original text. Providing misinformation is not improving and maintaining Wikipedia, it is damaging it. And that is why I believe we should either go to the effort of sourcing a good translation, or we should provide none. It is much better to say "I don't know" than to tell a lie. What's more, I think it is unlikely that we will often be in a situation where a good translation of a direct quotation cannot be externally sourced. Throwing in an unedited machine translation amounts to laziness. Rubywine . talk 19:12, 6 August 2011 (UTC)

It depends entirely on what class of direct quotation you're talking about. Professional translations are very frequently available for famous proverbs (which is convenient, because aphorisms are hard). Good translations are often available for titles of major books and plays.

However, they are almost never available for sentences in non-English sources that we are using to support article content. There are thousands of academic journal articles and millions of newspaper articles that are not written in English, and there are almost never professional translations available for them. If someone asks for a quote and a translation, and the best we can do is a machine translation, then we definitely should provide that.

On the general question, WP:There is no deadline. Translations do not have to be perfect on the first try. If providing a poor translation (whether due to machine translation or a lack of skill in a human translation) is the best we can do, then perhaps the existence of the poor translation will inspire some more capable editor to improve it. WhatamIdoing (talk) 04:05, 7 August 2011 (UTC)

No we shouldn't! Why should we offer unreliable ,questionable translations? Editors should never include content they can't really vouch for and that's exactly what an unassisted machine translation is. If a reader wants to roughly verify the accuracy of some citation or quote he can copy it into a translation machine (such as google) himself, there is no need for us to include such "rule of thumb" reasoning into the article itself. If you're afraid the reader is unable to use or find a translation engine, provide him a link to some online translation machine like google in the footnotes, but do not include that translation into the article or worse create the possible impression to readers the translation was actually produced by somebody who knew what he was talking about.--Kmhkmh (talk) 11:16, 7 August 2011 (UTC)

One answer is that we can offer "unreliable questionable" translations when they ipso facto support the relevant point.  See [Jaakko Pöyry], in which the mostly-incomprehensible machine translation of the name of the book is a strong indication of the notability of the subject.  English-only readers entirely get both the point that this is a machine translation of idiomatic Finnish, and that Jaakko Poyry is a man who has come from a hinterland to a world-wide stage.  Unscintillating (talk) 13:06, 7 August 2011 (UTC)

I would take such a translation to be evidence that an assertion is verifiable, but not verified, and would suggest that they would be useful in AFD, or on the talk page of the article to help others help find a decent translation, but I would be very hesitate to accept such as cited references in the article. Of course, it would depend on how bad the translation is, but comes a point that an editor who does not have the language in question cannot evaluate the quality of the translation. Here I speak from the experience of rereading works in various languages after having improved my skills--my early understanding from 2-3 years of instruction of such were very poor. --Nuujinn (talk) 14:36, 7 August 2011 (UTC)

I certainly agree that it's irresponsible to add material to the encyclopaedia based on a machine translation of a language you do not speak.—S Marshall T/C 17:24, 7 August 2011 (UTC) I don't think machine translation can ever be considered a reliable source. Indeed, the meaning can easily be the opposite of what the original said. If the person providing a machine translation is in a position to check that the translation is adequate, it effectively becomes his/her translation, and the fact that it is (based on) a machine translation needs only be mentioned for the purposes of intellectual integrity. If he/she is not able to verify the correctness of the translation, it should not be used. The editor introducing the translation must assume responsibility for its accuracy, preferably explicitly. Perhaps we could change the text to something like

Translations published by reliable sources are preferred over translations by Wikipedians. If the translation is not the editor's, it must be attributed. If the translation provided by the editor is substantially the result of machine translation, this should also be mentioned, but an editor should never provide a machine translation unless he/she is in a position to confirm its accuracy.

--Boson (talk) 19:25, 7 August 2011 (UTC)

Machine translations, even between major languages, can come up with just the opposite of original meaning of a text. It's better to leave a text untranslated, because a machine translation conveys the appearance of accuracy, and it's even worse if the machine translation is being smoothed by someone who does not know the original language. We may look at the issue in five years again, and maybe machine translators have made further progress until then. In the meantime, I suggest that we advise against using machine translations, and that an editor must take personal responsibility for all translations. An editor who can confirm the accuracy of a machine translation will also be able to translate the text without the help of a machine translator. I also don't think it would be helpful to indicate whether an editor used a machine translation, as it would also not be helpful to indicate whether the editor used a dictionary or not.  Cs32en Talk to me  11:33, 8 August 2011 (UTC)

It depends on the language in question. I notice, without any surprise, that every single person opposed to machine translations, and who identifies a language on his or her user page, speaks German. I've seen some truly lousy machine translations from German (double negatives seem to be a real problem). On the other hand, I've never seen a really poor machine translation from Spanish. It's often a bit awkward, but it's never the opposite of the original. You're basically proposing that nobody be allowed to use a tool merely because it doesn't work so well for your particular language.

And, again, I think you're overlooking the context. This is primarily about providing a translation when you're using the |quote= parameter in the {{Citation}} templates. You'd be supplying both the original language and the English translation, and any person who read the language could correct or improve the translation at any time. If someone cited a source as supporting a statement, based on a machine translation or a plain old misreading, then we need to know this. "Showing their work" by providing the mistaken translation helps us find and correct errors. WhatamIdoing (talk) 16:45, 9 August 2011 (UTC)

Google translate does fairly well with Romance languages. It's not to be trusted for others.—S Marshall T/C 16:53, 9 August 2011 (UTC)

And so you would issue a universal ban. That does not seem appropriate to me, nor does it seem like it reflects the actual practice of the community. WhatamIdoing (talk) 22:33, 10 August 2011 (UTC)

What I'm saying is that from my personal experience, I would trust google translate (but not babelfish) to turn French into tolerably accurate English. I would not trust google translate with German or Dutch, and I definitely wouldn't trust it with something highly inflected like Finnish. If in doubt, I would have to reject the machine translation unless confirmed by a human speaker.—S Marshall T/C 22:51, 10 August 2011 (UTC)

I am German, but I almost never have used translations into German. Translations to German are so bad that I - at least for the last two years - have always used translations from various languages (Spanish, Italian, Japanese) to English. My comments are based on my experience with machine translations into English and have nothing to do with the German language.  Cs32en Talk to me  23:41, 10 August 2011 (UTC)

Under NONENG, all translations are to English, not from English. There is never any excuse to translate something into any language except English under NONENG.

WP:V and NONENG are not the general translation guide. NONENG is about what to do when somebody cites a Polish paper to support some claim in an article, and nobody knows for sure if the Polish paper says anything even remotely like what the original editor claims it says. NONENG says that in such cases, it's best to get a professional translation, second-best to get a Wikipedian to translate, and if all else fails, you can ask Google translate to have a go at the title—and that having that machine translation is better than leaving a citation in a state that is unintelligible to 99% of the world's population.

NONENG always moves to English. You would never translate anything into German under NONENG. WhatamIdoing (talk) 19:51, 12 August 2011 (UTC)

Of course, we are talking about translations to English. If "nobody knows for sure if the Polish paper says anything even remotely like what the original editor claims it says", then, with a machine translation only, still nobody knows whether the Polish paper says exactly what the editor claims it says. It is not sufficient to know whether the papers says something remotely similar to the content of our article, and therefore a machine translation is not helpful in such cases.  Cs32en Talk to me  20:04, 12 August 2011 (UTC)

A proposed rewording of WP:NONENG

Thank you everyone. Maybe it's impossible to reach consensus but I'd like to try. I have the impression that most people favour a change. I don't feel there's much point in trying to formulate fine-tuned rules about the translation process; it's really about whether the editor has the language skills to take full personal responsibility for the translation. Based mainly on Boson's suggestion, but with some alterations, this is a proposal which I think is in line with the majority view:

• Translations published by reliable sources are preferred over translations by Wikipedians.
• A Wikipedian may only use a machine translation if he/she speaks the source language and is able to confirm the accuracy of the translation.
• All translations must be attributed to the translator by name.

Do you agree or disagree? Rubywine . talk 20:53, 9 August 2011 (UTC)

I disagree with everything except the first point. People add information based on machine translations for languages they don't understand (well enough) every single day. If they paste in their machine translations, we increase the likelihood of finding their errors. We also use machine translation for purposes far outside the subject of this policy, e.g., to decide whether the person who says "this five-page magazine article in Chinese is entirely about this company" is telling the truth. That's a perfectly legitimate use of machine translation, since even the most busted machine translation is going to reliably identify whether the name of the company appears repeatedly throughout the five pages. This sloppy statement would ban any and all use, not just uses that are higher risk.

Also, the names of the editors who did the translation have no business appearing in the article under any circumstances. WhatamIdoing (talk) 21:00, 9 August 2011 (UTC)

• Contributor names are always recorded in the usual Wikipedian way, of course. If the translator was a reliable source, then it's not just reasonable but strongly encouraged to attribute them, since not attributing them is potentially plagiarism or breach of copyright. As for your second point, who will police this and who will enforce it?—S Marshall T/C 22:29, 9 August 2011 (UTC)

• Yes of course. On my second point, who can stop dissenting translators, determined to foist unassisted machine translations upon the unsuspecting world? I don't know. The existing WP:NONENG guidelines can't be policed or enforced. Most other guidelines can't. The system assumes that the majority respect consensus. Rubywine . talk 01:03, 10 August 2011 (UTC)

• The existing policy is enforceable by individual editors on talk pages (with backup from uninvolved administrators if necessary). The proposed version would require editors to know which language other editors can speak, before being enforceable.—S Marshall T/C 07:14, 10 August 2011 (UTC)

• No it doesn't, it only requires that other people can speak the source language. If an editor's submissions were found to match unassisted machine translations, and were deemed to be incompetent by a consensus of competent speakers, then they could be asked to stop. That's all. It is unlikely that this scenario would arise. Rubywine . talk 12:18, 10 August 2011 (UTC)

• That means that we have to decide who's "competent" to pass judgment. It would be far simpler for these competent people to simply fix bad translations when they find them, in exactly the same way that I fix bad grammar when I find it.
  Given how few pages I see with any non-English content, I find it hard to believe that we have a widespread problem with seriously misleading machine translations in our articles. This issue may not be worth the time we're spending on it. WhatamIdoing (talk) 17:48, 10 August 2011 (UTC)

• We don't need to decide who is competent. Firstly, because consensus about correct use of language is implicit in every discussion we have on Wikipedia, and there's no real difference here. If anything, it's easier than usual. Any human being can recognise a bad machine translation of a language they speak; there's nothing subtle about it. Secondly, because the enforcement scenario is a red herring. The point of this guideline is not to enforce and police, the point is to guide. There have been instances of bad translations being submitted, which have been mentioned by translators, and that is avoidable and regrettable. It's clear that most people in this discussion say there is no place for unassisted machine translations on Wikipedia. We just need to reach a consensus on the wording of guidance to that effect. Regarding your point that there are few pages with any non-English content - I think that is likely to change. There is a vast quantity of non-English material which has been requested for translation. This issue will be increasingly relevant in the future. Rubywine . talk 22:06, 10 August 2011 (UTC)

No, if you're going to have "a consensus of competent speakers", then you have to decide who qualifies as a "competent speaker".

Requests for translations of articles from other Wikipedias has nothing to do with this policy. NONENG is not about whole articles. NONENG is about what to do when somebody cites a Polish paper to support some claim, and nobody knows for sure if the Polish paper actually says what the original editor claims it says. What you need to do is explain why omitting the translation of the paper's title, on the grounds that we don't know how it was translated, would improve Wikipedia. WhatamIdoing (talk) 22:31, 10 August 2011 (UTC)

• Fair enough re NONENG. But I'm not going to be dragged into further discussion of a red herring, or repeat myself. Rubywine . talk 22:38, 10 August 2011 (UTC)

I suggest two changes:
(a) Change "is able to confirm the accuracy of the translation" to "confirms the accuracy of the translation". Maybe a comment can be added that this can be done either on the talk page or in the edit summary.
(b) Translations should not be attributed to the name of editors, but they should be attributed to real world sources if such sources have been used. If not, as with all other content, the attribution is in the history of the page, plus the confirmation on the talk page (or talk page archive) and/or the edit summary.  Cs32en Talk to me  23:48, 10 August 2011 (UTC)

Thanks, I agree with your suggestions and I'm happy to adopt them. In the absence of an explicit RS attribution it would be implicitly understood that the editor was the translator. I'll leave this a bit longer in case anyone else wants to comment. Rubywine . talk 22:13, 11 August 2011 (UTC)

A proposed rewording of WP:NONENG #2

The revised proposal is to replace this sentence in WP:NONENG :

• Translations published by reliable sources are preferred over translations by Wikipedians, but translations by Wikipedians are preferred over machine translations.

with the following :

• Translations published by reliable sources are preferred over translations by Wikipedians, and should always be attributed. A machine translation may be used only if the Wikipedian speaks the source language and confirms the accuracy of the translation.

Footnote: Attributions and confirmations may be provided on the talk page or in the edit summary.

Rubywine . talk 16:41, 14 August 2011 (UTC)

When we say that something "should always be attributed", we normally mean that it requires WP:INTEXT attribution. Your first sentence is therefore going to confuse a lot of people. Also, the name of the translator is commonly unknown and therefore cannot be "always" provided in any form. Furthermore—who cares who the translator is, except for the translator himself (who presumably would like to build his career)? If the translation is published by a reliable source, then we don't need to say "Translation provided by Bob Smith, who was hired by The London Times to provide a translation for this phrase in their article."

I firmly disagree with the second sentence. A machine translation may be used for many purposes that have nothing to do with verifiability, and your proposal fails to acknowledge that. Furthermore, as admitted above, machine translations aren't always bad, especially for Romance languages. Finally, since WP:Nobody reads the directions, and this rule is directly counter the community's actual practice, it's both ineffective and inappropriate. Written policy must describe the actual policy, and the actual policy is determined by what the community actually does, not by what a couple of people wish they would do instead. (It would be appropriate to say "Be cautious in using machine translations, which are lousy for many languages".) WhatamIdoing (talk) 23:18, 14 August 2011 (UTC)

"Be cautious in using machine translations, which are lousy for many languages" would be a step forward. The rest is ground we've been over before. WP:Nobody reads the directions is a weak argument. People try to read guidelines; long pages of tortuous prose get ignored because they're taxing and not understood. WP:NONENG is not a remotely arduous read, so instruction creep doesn't really apply here. If guidelines were so uninfluential, so universally ignored, you wouldn't spend so much time arguing over them! Rubywine . talk 03:05, 15 August 2011 (UTC)

Holy cow do I have a lot to say on this subject... I'll try to be concise. It touches on several different translation issues, actually. Some exclusive to Wikipedia, some not.

• Machine translation: I do make use of Google's app, which does make readily recognizable glitches. At times producing a translation that is actually the complete opposite of what the original text is saying. Yes, I am fluent enough to recognize this, and I still find the translation apps useful. If I look up a single word, and it doesn't suit the context of the original text, I know enough to do more searching with other tools, look in a thesaurus, etc.
• Professional translation is preferred: Did someone in WP:NONENG say that a professional translation is preferable? I do translation work in my capacity as a professional with the federal gov't of Canada. I've been tested and assigned a formal evaluation of my abilities. Do I qualify as a 'professional translator' by Wikipedia standards? (This is in my second language, French, so I do consider myself qualified to do French -> English translations, and vice versa, up to a level that includes technical language in a handful of fields. I've done editing for an academic journal in both languages, years ago.)
• A specific example: I just did a translation of a biography article (Anita Aubspurg) that appears in the French, Italian, and German Wikipedias. The original article appears to be the German version, all the references are to German books/Websites. The Italian version just copied and pasted the German references, and that's what I did too. No, I don't know German. My Italian isn't as good as my French, but I can read it pretty well. (The French version was all but empty, they didn't translate the Italian or German article in full at all). I used a translation app on both the Italian and German versions of the article, compared them side by side, and edited the article so that it would be coherent in English.
• Citing sources: I've been reading quite a few French Wikipedia articles, and glancing through some Italian and German, and guess what? Articles using inline cites to references are in a small minority. Who's enforcing the Wikipedia reference rules for the non-English Wikipedias? If the articles are acceptable in those Wikis without inline citations, why are they not acceptable in the English one? Should the articles in the non-English Wikis be purged if they do not meet that standard?
• Citing sources, part 2: No, I can't verify the original source texts used by the original article author. But then, I can't do that for an English article if it's citing books, journals, etc. that I don't have access to. I could read the original French books if I had access to them, but if all the copies are in France, or Senegal, or Cote d'Ivoire, or Tunisia, what should I do then? What would you do if you were editor for an English language article that was citing documents in the library of Oxford University, but you live in California? Does the article get purged? If the original article is on a subject for which information only exists in publications in a language for which English translators are rarely available, is the consensus that it not be translated at all? FWIW, I do try to look for some equivalent English sources online to add to articles I've translated, but for many subjects, there's just no information in English. (Doesn't mean the topic isn't noteworthy.)
• My two cents' worth -- I think it comes down to a question of reasonable due diligence. When you read an article in Wikipedia, or the New York Times, or a paper encyclopedia, you do take a certain leap of faith that the article isn't complete fiction with imaginary sources. But I think if an article contains inaccuracies, the open nature of Wikipedia is that it will be found out and either removed or corrected by someone more knowledgeable.

OttawaAC (talk) 01:24, 15 August 2011 (UTC)

OttawaAC, thanks for your comments. There's an awful lot of context here that you'll need to get a handle on. Please read through the previous parts of this discussion topic. This is just a quick reply to your point number 2. In this context, reliable source doesn't mean professional translation; a Wikipedian can't be recognised as a professional translator. A reliable source on Wikipedia means a published, third party, reliable source. Of course you could cite your own externally published work, but only in the same way that you could cite someone else's. Rubywine . talk 02:17, 15 August 2011 (UTC)

OttawaAC, in theory, NONENG has nothing at all to do with interwiki translations... but you wouldn't know that from Rubywine's proposal. Rubywine's proposal would ban the use of machine translation for any and all purposes, even for figuring out if a long article that is alleged to demonstrate notability in an AFD discussion actually talks about the subject. I do not know whether this is what Rubywine intended, but it is what s/he actually hopes to put in this policy. WhatamIdoing (talk) 02:44, 15 August 2011 (UTC)

WhatamIdoing, this is ridiculous. You are misrepresenting my intentions and the scope of the policy. Tiresome. Rubywine . talk 03:14, 15 August 2011 (UTC)

Yes, the proposal as written is ridiculous. Writing policies is hard. If you say "no machine translations unless you understand the language", then people will believe that you actually mean "no machine translations unless you understand the language", full stop. They will not magically know that you mean "no machine translations unless you understand the language, and, by the way, this only applies to direct quotations in the text of the article, because discussing sources on the talk page, translating whole articles from other Wikipedias, and figuring out whether a non-English source indicates WP:Notability in a deletion discussion are outside the scope of this policy, and we know that Wikipedia:Policies and guidelines#Content demands that we maintain the scope of the policy without wandering into off-topic stuff, like what's okay at AFD", which is what you probably mean to say.

Writing policies is much harder than you think it is. WhatamIdoing (talk) 20:26, 15 August 2011 (UTC)

I have amended the text of the proposal to rule out this plethora of unlikely misunderstandings and threats to personal liberty. Rubywine . talk 03:43, 16 August 2011 (UTC)

I actually don't think that the amendment is necessary, because I don't think editors will fear getting blocked for posting a machine-based translation on a talk page. I support both the amended and the original version of the proposal.  Cs32en Talk to me  19:23, 16 August 2011 (UTC)

(original version=#2, amended version=#3)

For the sake of continuity I have restored the original text of #2, have moved the amendment to #3, and have copied Cs32en's comment below. Rubywine . talk 22:11, 16 August 2011 (UTC)

A proposed rewording of WP:NONENG #3

Do you support the proposal to amend the guidance in WP:NONENG regarding the use of machine translations, as given at WP:V:talk?

Please note that the scope of WP:NONENG is limited to the translation of non-English sources for use in English Wikipedia.

The proposal is to replace this sentence in WP:NONENG :

• Translations published by reliable sources are preferred over translations by Wikipedians, but translations by Wikipedians are preferred over machine translations.

with the following :

• Translations published by reliable sources are preferred over translations by Wikipedians, and should always be attributed. A machine translation may be used in the text of the article only if the Wikipedian speaks the source language and confirms the accuracy of the translation.

Footnote: Attributions and confirmations should be provided on the talk page or in the edit summary.

Rubywine . talk 03:56, 18 August 2011 (UTC)

I actually don't think that the amendment [to your previous proposal] is necessary, because I don't think editors will fear getting blocked for posting a machine-based translation on a talk page. I support both the amended and the original version of the proposal.  Cs32en Talk to me  19:23, 16 August 2011 (UTC)

• (where the original version=#2 without the bolded text, and the amended version=#3) Rubywine . talk 22:56, 16 August 2011 (UTC)

No, they won't fear getting blocked, because violating WP:V almost never results in blocks. But we would have some tendentious editors who use such unqualified statements for obstruction. This change dramatically reduces my concerns about including this sentence.

Rubywine, are you prepared to explain what "should always be attributed" means now? WhatamIdoing (talk) 23:43, 16 August 2011 (UTC)

I'd like to remind you that this is an open discussion, and not a 1 on 1. The idea came from Boson. It was strongly supported by S Marshall, and then phrased as an amendment by Cs32en. I have reworded it more concisely, and nobody else has objected. It is clear and unambiguous, since the footnote rules out the possibility of in-text attribution. It won't confuse anybody. Rubywine . talk 01:12, 17 August 2011 (UTC)

The footnote does not rule out anything at all. It rules in additional options. WhatamIdoing (talk) 16:33, 17 August 2011 (UTC)

I have never seen an inline attribution on a reference, i.e. a nested attribution, and I doubt that it is possible to create one. Even so, I have changed the wording so that the only options are talk page or edit summary. Rubywine . talk 17:11, 17 August 2011 (UTC)

We use the word "attributed" in Wikipedia's content and sourcing policies to mean two specific things: as a synonym for "supported by an inline citation", and to refer to WP:INTEXT attribution. I think introducing a third meaning is a very bad idea that will inevitably lead to confusion. I suggest that you drop the word "attributed" entirely and merge the footnote into the text: "* Translations published by reliable sources are preferred over translations by Wikipedians. Please leave a note in the edit summary or on the talk page to tell other editors who did the translation." WhatamIdoing (talk) 16:11, 19 August 2011 (UTC)

Ok, I'll do that. It's totally clear and I cannot imagine that anyone would change their vote to oppose it. Rubywine . talk 00:43, 20 August 2011 (UTC)

A proposed rewording of WP:NONENG #4

The problem this RfC seeks to resolve is that WP:NONENG allows people to put unassisted (therefore totally unreliable) machine translations of sources into the text of articles. This may be done in good faith, and in the belief that any translation is better than no translation, but we don't agree. An unassisted machine translation is worse than no translation, because it may be highly inaccurate. Its usage in article text is contrary to the principle of upholding quality and reliability.

Please note that the scope of WP:NONENG is limited to the translation of non-English sources for use in English Wikipedia.

The proposal under discussion is to replace this sentence in WP:NONENG :

• Translations published by reliable sources are preferred over translations by Wikipedians, but translations by Wikipedians are preferred over machine translations.

with the following :

• Translations published by reliable sources are preferred over translations by Wikipedians. Please leave a note in the edit summary or on the talk page to tell other editors who did the translation.

• A machine translation may be used in the text of the article only if the Wikipedian speaks the source language and confirms the accuracy of the translation. If using a machine translation, please confirm its accuracy in the edit summary or on the talk page.

Rubywine . talk 01:02, 20 August 2011 (UTC)

Q: Why do people who can translate the source language need to use a machine translation at all? Why can't they do it all themselves? A. Some skilled translators find machine translations useful in the initial stages of their work. A machine translation can speed up the process of reaching an initial draft by translating 75%-90% of the text with reasonable accuracy, allowing them to focus on the difficult 10%-25% remaining.

For the purposes of judging the success of this proposal, the RfC started at #3. Since the change between #3 and #4 alters only the wording, not the meaning, I hope that noone will object if I leave the remainder of the discussion below this point. Rubywine . talk 01:02, 20 August 2011 (UTC)

• Support: I think the new version (#3) is better, but I don't think there will be much difference in practice between the two (#2 and #3). I suggest that "speaks the source language" be replaced with "is familiar with the source language". People can sometimes read a language but not be able to speak it. For example, Latin and Classical Chinese, are written languages that few can speak outloud. --LK (talk) 10:11, 18 August 2011 (UTC)
  • Good points. We need a third clause to address extinct languages unless we accept your suggestion.
  • Non-speakers who can read will have a limited grasp of pragmatics and idiom so depending upon what they are translating, they might have difficulty. Then again, so will some non-native speakers. Should we trust that non-speakers will confirm a translation only if they are confident about its accuracy? On reflection, I think we should.
  • I'm willing to change "speaks the source language" to "is familiar with the source language". What do other people think? Rubywine . talk 06:03, 20 August 2011 (UTC)

• Weak support If a wikipedian speaks the source language, why not translate it him/herself? That said, I see no problem with allowing the use of machine translation. Perhaps a friendly ammendment, and if it isn't I won't make a big fuzz: still say that a translation by a wikipedian is better? This is to encourage the translators among us to do it and be recognized.

Also, I might misunderstand, but I think what you are saying can be achieved with Template:Ref series templates, and I have seen done - even to borderline copyvio.--Cerejota (talk) 11:05, 18 August 2011 (UTC)

• Three cheers for encouraging and recognising translators! I'd like to put some energy into supporting that, independently of this RfC.
• "Why not translate it him/herself?" In earlier discussion some translators said they use machine translations for convenience and speed before deciding on the best source to use. If they decide they can't improve upon the accuracy of a machine translation, it effectively becomes their own. If we were to say " a translation by a wikipedian is better" that would leave open the possibility of using a machine translation without human intervention, which is precisely what we're trying to rule out here. So I think that doesn't fly. Is there a different rephrasing that might work better?
• Template:Ref is a great template, but this idea would entail adding notes about source translations to footnotes in the article, when they are really metadata. So I think previous discussion on that points rules this out. But thank you for the useful information. Rubywine . talk 06:03, 20 August 2011 (UTC)

• Comment. I support that reliably published translations are preferred. Not sure about requiring Wikipedians to understand the language of a machine translation. Which Wikipedians would we be referring to? SlimVirgin ^{TALK|CONTRIBS} 01:58, 19 August 2011 (UTC)

I think the point is that we cannot rely on a machine translation to say anything remotely like the original, so an editor adding a machine translation assumes responsibility for the translation being accurate, which is not appropriate if he/she cannot understand both the original and the translation. I take "Wikipedian" in this context to mean the editor adding the translated text." --Boson (talk) 11:00, 19 August 2011 (UTC)

I suppose the thing that concerns me is (a) if the Wikipedian understands the translation, then she could do it herself; and (b) it leaves us open to people saying machine translation is acceptable under the policy simply because they personally understand it. In fact, we should allow editors to approach these things on a case-by-case basis, because machine translations are sometimes fine as sources, sometimes not, depending on many factors. SlimVirgin ^{TALK|CONTRIBS} 01:15, 20 August 2011 (UTC)

But machine translation is acceptable under this policy if the editor can understand the source language. Some translators will start off with a machine translation for convenience (perhaps they want to run English language searches on it to determine relevance). If they understand some text they want to quote and they can't improve upon its accuracy, it effectively becomes their own. If the machine translation matches the best human translation, that's fine. What this amendment is designed to exclude is a machine translation that's not understood. Rubywine . talk 06:03, 20 August 2011 (UTC)

• Oppose as written, as I believe it is not sufficiently flexible. Some information is more complex and open to translation errors than others; some machine translations are better than others; there is wide diversity in degree of familiarity with a language and how people self-assess that. I have encountered people who have claimed fluency in languages where they are plainly not fluent, and I've been told that I myself tend to underestimate my own facility. A machine may well help to adequately supplement with straightforward information where it might fall down in philosophical or subtly nuanced content. I would support discouraging contributors from relying on machine translations where content is controversial or errors are more likely, but not altogether. --Moonriddengirl ^(talk) 17:24, 19 August 2011 (UTC)
  • Does LK's suggested amendment address your issues? Rubywine . talk 06:05, 20 August 2011 (UTC)

• Comment Where the translation is not the editor's own work, I think it should be made clear that attribution is obligatory - in such a way that the author of the translation (i.e. the derivative work) is visible to the reader of a book or PDF file created from the article. In other words, an edit summary or a note on the talk page is not sufficient. The reader should also be able to establish whether a Wikipedia license or fair use applies to the translation. --Boson (talk) 21:25, 19 August 2011 (UTC)

I would agree that third-party translations should be attributed, although I think that this should be discussed in a separate RfC. I'm not sure whether short translations necessarily subject to copyright restrictions. At least in Germany, you can only claim copyright for something that is sufficiently elaborated. A translation of a novel or a complete newspaper article would meet that condition, but a single sentence would probably not.)  Cs32en Talk to me  22:25, 19 August 2011 (UTC)

I am not a copyright expert, but my understanding was that German copyright law is based on "creativity", but with the barrier so low that anything much more than an unordered list or a statement that cannot possibly be phrased differently is subject to copyright. Of course, fair use permits large chunks to be quoted, but that still requires attribution. Without attribution, I would say the editor is probably claiming copyright and releasing it under a Wikipedia license, which explicitly permits commercial use and modification (which is not permitted under fair use). This is OK for original-language quotes, because they are usually attributed (that is the whole point of citation), but the translator's work also needs attribution. Often, the cited source will be the translation of a whole work, so attribution is covered; where it is not, the translator needs to be attributed separately. Perhaps the citation templates should be expanded to include trans_quote and translator. One special case that occurs to me is quotation from laws and regulations in the body of an article, which is quite common (eg. Volksverhetzung, Strafgesetzbuch section 86a). Depending on the jurisdiction, the laws may be free of copyright, but I suspect that does not apply to translations (possibly depending on where they are published). --Boson (talk) 09:09, 20 August 2011 (UTC)

• Question. Ruby, could you say briefly what is problematic about the current policy, i.e. what problem does the RfC seek to resolve? The policy says: "Translations published by reliable sources are preferred over translations by Wikipedians, but translations by Wikipedians are preferred over machine translations." SlimVirgin ^{TALK|CONTRIBS} 01:26, 20 August 2011 (UTC)
  • Good point Slim, this was an oversight on my part. I've added a brief explanation to the start of this RfC. You might also find the earlier discussion prior to this RfC useful. Rubywine . talk 06:03, 20 August 2011 (UTC)
    • Okay, I see now, thank you. SlimVirgin ^{TALK|CONTRIBS} 06:08, 20 August 2011 (UTC)
• Comment  I don't see an example of the problem that is being solved.  Unscintillating (talk) 03:05, 20 August 2011 (UTC)
  • Please see my answer to SlimVirgin. No specific examples have been mentioned, but the nature of the problem has been explained. Rubywine . talk 06:03, 20 August 2011 (UTC)
• Oppose  The article Dictum states, "legal practitioners in the U.S. colloquially use dicta to refer to any statement by a court that extends beyond the issue before the court."  Machine translations are a tool, they can be used and they can be misused, and in this case without examples of misuse we assume good faith.  In the event that examples surface, I would not agree that the current factoring of the proposal works, since we already have a bullet point at WP:NONENG focused on "quoting a source".  And as to what extent we need to discuss who did the translation, this is a different topic than machine translation.  Unscintillating (talk) 12:36, 20 August 2011 (UTC)
  • The proposal seeks to replace one sentence in the existing policy with alternative text with exactly the same scope. Please reread it and if you still have an issue, explain what factoring problem you perceive.
  • This proposal has nothing whatsoever to do with bad faith or misuse of tools. This is a policy amendment. In the scenario that a source is available in a language that the editor does not understand, and where no reliably sourced translation is available, the existing policy gives the nod to an unassisted machine translation. This is based on the implicit assumption that a bad translation is better than no translation. Those of us backing this amendment say that this assumption is wrong - no translation is better than a bad translation. Some experienced and multilingual translators have reported having seen highly misleading machine translations. An editor who doesn't know the source language can't recognise one of these and may add it to Wikipedia - in good faith, as the RfC states. This proposal changes the policy in order to remove that possibility.
  • We also argue that knowingly introducing any material of unknown quality into an article contravenes core principles of Wikipedia: to uphold quality and reliability of sources. Again, this criticism is targeted at WP:NONENG, and not at editors.
  • The requirement to identify the translation source is a separate issue, as you say, but it is a related one, and it is no longer contentious. I'm assuming you mentioned that because you didn't wish to argue the point. Rubywine . talk 14:15, 20 August 2011 (UTC)
    • I think that this proposal over time has had a reduction in the scope such that the scope is now within the bullet point mentioned.  But I think counter-proposals are not constructive when this appears to be a problem that is better solved by waiting until we have an example to facilitate discussion.  The following premise remains speculation: that there is an article that needed a foreign-language quote for which there were no published English translations, and the machine-language quote just happened to be one of the ones that was highly misleading, and the editor who thought the quote was essential for inclusion was at the same time unaware or willful about the nature of the English machine translation, and that no other editors intervened when this happened, and that the best solution for this problem is by changes to WP:V—this premise remains speculation.  Unscintillating (talk) 15:55, 20 August 2011 (UTC)
      • Please explain the intent of the proposal with regard to the machine translation in Jaakko Pöyry.  Thanks, Unscintillating (talk) 02:38, 21 August 2011 (UTC)
• Comment I am sympathetic, and agree that there is an issue that needs to be dealt with, but I'm not sure this is the way to do it. Currently we do not require translations of non-English sources, so an editor could comply with this suggestion by simply not providing a translation and we'd have no idea whether anyone had verified the source. For online sources, or sources that are readily available off line, this is not so much an issue, but I do see a problem with more obscure sources. --Nuujinn (talk) 18:05, 20 August 2011 (UTC)
  • Yes I agree that is a potential problem. Maybe a {{Source translation required|language|subject}} tag would be a good idea. Rubywine . talk 20:08, 21 August 2011 (UTC)
• Support Any machine translation must be checked by someone who can read it at least well enough to tell if it's totally off base. ("Is familiar with the source language" is fine with me.) For some examples of catastrophically bad translations, see the Lost in Translation postings on Language Log. The writers of this blog are all highly trained linguists (i.e., language scientists). --Thnidu (talk) 21:39, 20 August 2011 (UTC)
  • Thanks for replying to my request for comments about LK's idea, and for your interesting and helpful remarks. Rubywine . talk 20:08, 21 August 2011 (UTC)
• Oppose I am not happy with encouraging attribution solely in edit summaries or on the talk page. On reflection, however, I think the question of attribution is a distraction at this point and should be discussed separately. I think the main point is not to give the appearance of recommending the use of machine translation where the editor does not know the meaning of the original text. On balance I think it might be best not to mention machine translation at all, and merely state "Translations published by reliable sources are preferred over translations by Wikipedians." --Boson (talk) 10:05, 21 August 2011 (UTC)
  • Thanks for your comment, although needless to say I can't agree. It leaves the same loophole wide open, but offers less guidance than the current policy. I'd just like to point out that attribution was first introduced into this proposal at your own suggestion. I am now left wondering what attribution format you had in mind when you posted the message below, but never mind, it's not worth discussing further. This RfC has clearly failed to gain enough support.

I don't think machine translation can ever be considered a reliable source. Indeed, the meaning can easily be the opposite of what the original said. If the person providing a machine translation is in a position to check that the translation is adequate, it effectively becomes his/her translation, and the fact that it is (based on) a machine translation needs only be mentioned for the purposes of intellectual integrity. If he/she is not able to verify the correctness of the translation, it should not be used. The editor introducing the translation must assume responsibility for its accuracy, preferably explicitly. Perhaps we could change the text to something like

Translations published by reliable sources are preferred over translations by Wikipedians. If the translation is not the editor's, it must be attributed. If the translation provided by the editor is substantially the result of machine translation, this should also be mentioned, but an editor should never provide a machine translation unless he/she is in a position to confirm its accuracy.

--Boson (talk) 19:25, 7 August 2011 (UTC)

Rubywine . talk 17:35, 21 August 2011 (UTC)

Yes, that's why I wrote " On reflection . . .". I regard the issue of attribution as important , but the discussion here convinced me that we would not come to a consensus on the text in a reasonable time frame unless we limited the discussion to the essentials. If we don't mention machine translation at all, we don't put ideas in people's heads, and attribution can be dealt with separately. We don't have to solve all problems at one go. Attribution for sourced translations can be done in the normal way by means of a footnote, and all translation not done by the editor should be sourced. If a translated work is being cited, we should be doing this anyway. --Boson (talk) 19:23, 21 August 2011 (UTC)

But we're not, and it's not in the policy, and this RfC would have solved that, had it succeeded. Everyone has different ideas about priorities. Some people expressed strong opposition to the idea of putting source translation attributions onto the article page, which is why we ended up with the edit summary/talk page idea, and now there's strong opposition to that idea. I suspect this opposition would not have arisen if I hadn't accepted an unnecessary and wordy rephrasing (#3 > #4) which did not change the meaning to try to reach consensus. In hindsight I agree that an incremental approach would have been better from the start, but back then most people seemed in agreement. After two weeks of trying to move this forward and making compromises to try to reach consensus, and now finding that previous supporters have fallen away, I'm tired. If you would like to propose a new amendment to WP:NONENG then go for it, but I can't support your new suggestion. Yes it would remove the implicit nod to unassisted machine translations, and yes it would avoid suggesting machine translations to new editors, but editors who cite non-English sources aren't going to forget about machine translations just because WP:NONENG goes silent about them. It wouldn't be quite so bad if this were a first draft of WP:NONENG but to amend it to be less clear just strikes me as wrong at every level. Rubywine . talk 21:04, 21 August 2011 (UTC)

• weak support. I'm somewhat fuzzy about the attribution part. If we have a translation by a reliable external source it should be directly attributed (footnote) rather than on the discussion page or edit summary. If there's no such attribution, I'd assume a translation by a wikipedian anyhow, so noting that in the edit summary doesn't harm but is redundant.--Kmhkmh (talk) 19:36, 21 August 2011 (UTC)
  • Thanks. My reply to Boson applies to your comment too. Cheers. Rubywine . talk 21:15, 21 August 2011 (UTC)

Closure

• I removed the RfC tag on the discussion of #4. It has clearly failed to reach consensus. I'm not sure what the protocol is. From my own point of view this discussion topic could be archived. Rubywine . talk 05:19, 23 August 2011 (UTC)

Perhaps proposing an additional sentence offering guidance would be enough, Ruby. For example (new sentence in bold):

"Translations published by reliable sources are preferred over translations by Wikipedians, but translations by Wikipedians are preferred over machine translations. When using a machine translation as source material, editors should be reasonably certain that the translation is accurate and the source is appropriate."

Would that go far enough for you? SlimVirgin ^{TALK|CONTRIBS} 07:49, 23 August 2011 (UTC)

Yes thanks, that's a good suggestion and I support it. However I need to take a break from editing, so I'm not taking this further myself. Rubywine . talk 07:54, 23 August 2011 (UTC)

I think that's acceptable, and better than what we have. Machine translations are often imperfect and sometimes disastrous, and it would be appropriate to acknowledge that. I'll add SV's suggested language. WhatamIdoing (talk) 19:39, 24 August 2011 (UTC)

I'm glad our two weeks of discussion have been productive. SlimVirgin, many thanks for your help. Rubywine . talk 15:40, 25 August 2011 (UTC)

Notability section

Notability section, tweaking

[subsection inserted above] Unscintillating (talk) 03:04, 13 August 2011 (UTC)

Does anyone mind if I tweak the writing of the part in bold? Or do we even need it? I had difficulty understanding it when I first read it.

"Wikipedia's General Notability Guideline says that the subject's notability is established by the publication of independent, reliable sources, regardless of whether the names of any such sources have yet been typed into the article.

I assume it means notability depends on whether published sources exist, not whether they've been added to the article. But that's what the first part of the sentence says already. SlimVirgin ^{TALK|CONTRIBS} 17:00, 12 August 2011 (UTC)

I also believe that's what it means. Considering this is a summary of another policy, if it is to be an actual convenience to readers, it should include the critical points from the other policy. It is a critical point that articles should not be nominated for deletion if nearly everyone knows the topic is notable, even though the independent sources are not cited. Jc3s5h (talk) 17:20, 12 August 2011 (UTC)

In the recent edit history of the page, it went from this to this. What would the "tweak" consist of? It might be best to put the proposed wording here, in talk, first. --Tryptofish (talk) 18:58, 12 August 2011 (UTC)

I was going to tweak it, but then realized it doesn't need to be there in the first place. The first part of the sentence says: "the subject's notability is established by the publication of independent, reliable sources." The bolded second clause just repeats that, but in an odd way. No one would assume that, if I created a stub that said only, "Queen Elizabeth II is the queen of England," that the lack of sources on the page at that moment in time would mean Queen Elizabeth wasn't notable. SlimVirgin ^{TALK|CONTRIBS} 19:07, 12 August 2011 (UTC)

I'd like to hear from the editor(s) who made the change about why they wanted to add that wording, before making a decision. I agree that it sounds odd, as written, but I have a feeling (ie, guess) that the idea was to point out that sources can exist even if they haven't been added to a page, so a page shouldn't be deleted without checking for sources first – as though these editors have had experiences where your hypothetical Elizabeth II stub was nominated for deletion (and stranger things do happen). --Tryptofish (talk) 19:16, 12 August 2011 (UTC)

I agree that it's redundant, but AFD has significant and persistent difficulties with people who do not (choose to) understand this relatively simple point. Repeating it seems to get the point across more effectively than stating it once. The point that needs to be made is that {{unref}} and {{nonnotable}} are completely separate issues. If you want more formal language (which I've been finding less effective recently), then something like "Notability requires only that suitable independent, reliable sources have been published; it does not require the citation of any sources at all in the article" would probably suit.

There was discussion earlier about removing the entire section, since there's no particular reason to for WP:V to say anything about all about WP:N. WhatamIdoing (talk) 19:42, 12 August 2011 (UTC)

I would prefer to go back to the previous text, which simply said: "If no reliable secondary source can be found on a topic, Wikipedia should not have an article on it." That's the only part of the issue that concerns V. It was expanded on August 7, [1] and again on August 8. [2]

If we can't have the old version, I'd like to tweak the new sentence to say: "Wikipedia's Notability guideline says that a subject's notability is established by the publication of independent, reliable sources, not by whether such sources have actually been added to the article." SlimVirgin ^{TALK|CONTRIBS} 20:20, 12 August 2011 (UTC)

Two thoughts... 1) Does not the existence of a tertiary source on the topic (such as an article on the topic in some other encyclopedia or almanac) go towards notability? (if so, we should avoid the word "secondary"). 2) There is a difference between "I can't find any sources" and "No sources exists".

I think the point of this section is to tell editors: "Don't create an article if you (the creating editor) can not find any sources that discuss it"... but it is being misunderstood as saying "I should nominate an article for deletion if I can't find any sources about it". The first is correct... the second isn't. Blueboar (talk) 21:03, 12 August 2011 (UTC)

Well, first of all, the original is one diff earlier, and it says not one word about secondary sources. It refers to WP:Third-party sources (aka independent sources), not WP:SECONDARY sources, and as you know, WP:Secondary does not mean independent. A first-party secondary source (and there are many in the hard sciences) does not demonstrate notability.

"Can be found" has unfortunately been interpreted as "can be found by lazy deletionists who refuse to look any further than a section named ==References== in the current version of the article". I suggested "have been WP:Published" last month, and the discussion was de-railed by a long digression by one editor who wanted to ban unref'd articles entirely, and the distraction of the endless not-truth discussion.

Similarly, I'm concerned that "whether such sources have actually been added to the article" may be more susceptible to misunderstanding than "whether someone has typed the names of such sources into the current version of the article". I had a long and discouraging conversation last month with an experienced editor who seemed to have trouble distinguishing between adding content from a published book (e.g., typing "Only 5% of women die from breast cancer" into an article) and actually WP:Citing the book (=typing the author's name, the title, and the date into the article), so my belief in the average editor's ability to understand plain English is at an all-time low. Someone might well think that "adding sources to the article" referred to including the full text of primary sources rather than to typing up the author, title, date, and publisher for any source that you used. WhatamIdoing (talk) 21:17, 12 August 2011 (UTC)

For years it said "third party sources," and that would be my preference. But someone recently changed it to secondary sources, and I don't mind that either. The only concern I have is extending it to say things covered by the guideline. SlimVirgin ^{TALK|CONTRIBS} 21:22, 12 August 2011 (UTC)

The long-standing version was "If no reliable third-party sources can be found on a topic, Wikipedia should not have an article on it" (e.g. here). I also wouldn't mind: "If no reliable secondary or tertiary sources can be found on a topic, Wikipedia should not have an article on it." But I don't see the need to say more than that here. The details are dealt with in the Notability guideline. SlimVirgin ^{TALK|CONTRIBS} 21:25, 12 August 2011 (UTC)

I think these comments by both Blueboar and WhatamIdoing are very helpful to me in understanding the issue. Perhaps the key issue, then, is the wording about whether WP "should" have an article on a subject. That does, indeed, sound like an invitation to lazy AfD. Instead, maybe the wording should be about whether "content" (as opposed to article) should be "added" (as opposed to exist), with the "added" part what is most important. What I mean is to frame it in terms of adding material, as opposed to responding to material previously added. --Tryptofish (talk) 21:25, 12 August 2011 (UTC)

I think we should go back to third party sources, as the issue of the independent coverage is pretty essential. And I concur with Tryptofish's last observation. --Nuujinn (talk) 15:11, 13 August 2011 (UTC)

How about: What matters is whether such sources exist, not whether the article presently cites them.? Septentrionalis PMAnderson 21:45, 12 August 2011 (UTC)

So, delete the bolded part above, and add that sentence in its place? I like that. I think it's a good idea. --Tryptofish (talk) 19:54, 13 August 2011 (UTC)

Notability section, WP:V is content policy

WP:V is content policy, see WP:N#NNC.  If the purpose of the notability section in WP:V is to repeat parts of the WP:N guideline, the section can be deleted.  Blueboar has mentioned that "the language...originally...was meant to relate to how the concept of WP:Verifiability could be applied at the article level (as opposed to a sentence or paragraph level)."  Given the burden of deletion at Wikipedia, I think we need to be moving in the direction of encouraging more sourcing by article creators, which in the absence of new proposals is best done by leaving the current language.  Unscintillating (talk) 03:04, 13 August 2011 (UTC)

Its a guideline not a policy. Lets not mix the two up. I say eliminate that section altogether. Its also misleading because it only covers the GNG, when if you click on that link it says an article is notable if it meets the GNG OR one of the secondary guidelines listed on the right. Dream Focus 03:43, 13 August 2011 (UTC)

Two points: first, the policy now says that the GNG says something that I can't find it saying. Does anyone have a link to where the guideline says this: "that the subject's notability is established by the publication of independent, reliable sources (regardless of whether the names of any such sources have yet been typed into the article)"?

Secondly, a bit of historical context. This policy contained the sentence about "no third-party/secondary/reliable sources = no article" in some form since at least Feb 2006, six months before the Notability guideline was created. So this sentence of ours—"If no reliable third-party sources can be found on a topic, Wikipedia should not have an article on it"—[3] is a verifiability issue, entirely independent of whatever the Notability guideline might say. We therefore shouldn't say anything here that implies this policy follows that guideline. That's why I think we ought to return to that one stand-alone sentence without elaborating. SlimVirgin ^{TALK|CONTRIBS} 18:11, 13 August 2011 (UTC)

I was wondering where that sentence got to - that's a very critical sentence that outlines fundamental verification policy. That needs to be put back in. --MASEM (t) 21:07, 13 August 2011 (UTC)

That a notability issue was described here when no other suitable page existed does not mean that "we should not have an article" is actually a matter of verifiability rather than notability. It could merely be a legacy of what seemed convenient at the time.

I do not believe that this sentence deals with verifiability. Material can be 100% verifiable without coming from a third-party source—exactly as this policy has said for years. WhatamIdoing (talk) 21:14, 13 August 2011 (UTC)

The point of the sentence is the can be found part. It's find to use first-party, dependent sources as long as they are reliable (SPS comes into play here, for example), as long as we know there also exist third-party sources that discuss the topic; the first party source may be the best source to use in that case (such as using the award organization's list of winners on their webpage than a newsreport reporting the same - the key is that the newsreport has shown interest in it. Without any third-party sources in existence, a topic has no relevance to anyone but those directly involved with the topic, and thus there's appropriateness for an article on that topic. That's a key aspect of verifyability, that someone else has at least considered the first party source as factually correct. --MASEM (t) 21:58, 13 August 2011 (UTC)

I am beginning to think that we made a mistake in trying to tie the sentence in question directly to the WP:Notability guideline. I think the original intent of the sentence was to say:

• If no Verifiable information exists on a topic, Wikipedia should not have an article on that topic.

While that idea is similar to the concept of notability, it isn't quite the same. Blueboar (talk) 13:44, 14 August 2011 (UTC)

Yea, I have no idea why the idea was tied to notability. "Third party" source is one, by definition, would have had to do some research to validate the facts since they were not privy to the first-hand details; the more reliable that source is, the more fact-checking they have likely done. Requiring that some third-party sources exist is a basic metric for verification of a topic, and has little to do with notability which is more about how deeply that topic is covered in sources as to make for a good encyclopedic article. --MASEM (t) 14:00, 14 August 2011 (UTC)

That sentence was never meant to be tied directly to the Notability guideline. It spent a lot of its life under Burden of evidence. I moved it not long ago to a subsection called Notability under the "Verifiability and other principles" header, but the intention was not that it derived from the Notability guideline. Perhaps we should simply move it back to Burden of evidence to break that connection. SlimVirgin ^{TALK|CONTRIBS} 18:49, 14 August 2011 (UTC)

Notability section, proposal to put sentence back under "Burden of evidence"

The proposal is to delete the notability section and move the sentence back to the "Burden of evidence" section.  The sentence is:

“ If no reliable third-party sources can be found on a topic, Wikipedia should not have an article on it. ”

• Support  Unscintillating (talk) 22:49, 14 August 2011 (UTC)
• Support retaining that sentence and not elaborating further about notability; and I have no objection if it's moved back to the "Burden of evidence" section, and the Notability subsection is deleted. SlimVirgin ^{TALK|CONTRIBS} 22:54, 14 August 2011 (UTC)
• Aye.—S Marshall T/C 22:56, 14 August 2011 (UTC)
• Moving it works for me. I still strongly prefer that "can be found" be replaced by "have been WP:Published", to alleviate the problems that AFD is reporting with a small number of editors who apparently want to believe that "can be found" means "can be found without me needing to look any further than the ==References== on the current version of the article". WhatamIdoing (talk) 23:07, 14 August 2011 (UTC)

• Oppose. The only interpretation of "topic" for which this statement is correct (never mind which policy it belongs in) is the article as a whole. But it would be too easy for people to interpret as the topic of a specific addition, or a topic that is broader than the article. Also, the statement only belongs in the notability policy, or a summary of notability in this policy. It is not a statement about verifiability. I could give a perfectly verifiable statement about the location of a state highway culvert, from a first-party source (the state highway department). The reason for not having such an article is that nobody cares, not a lack of verifiability. Jc3s5h (talk) 23:16, 14 August 2011 (UTC)
  • It has nothing to do with notability, that's the thing. A lack of third-party sources means we have no assurance on the reliability and verification of the information on the topic, and thus should limit its coverage from "none" to "part of a larger article". --MASEM (t) 23:20, 14 August 2011 (UTC)

• A Supreme Court decision is a first-party source, but is absolute proof of the decision. It is certainly verifiable, although there are many reasons, which fall outside this policy, why reporting just the decision would make for a bad article. Jc3s5h (talk) 00:02, 15 August 2011 (UTC)

• But at the same time, a SC decision will be covered by third-party sources as well, even though we'd likely to use the actual decision for sourcing statements about that decision. The issue is when only first-party or second-party sourcing is available. --MASEM (t) 00:30, 15 August 2011 (UTC)
• A SCOTUS decision is not a first-party source for everything. It is a first-party/affiliated source for the new decision it contains, but it is is a third-party source for (e.g.) any prior court case the decision describes, laws it mentions, facts relevant to the specific case, etc. WhatamIdoing (talk) 20:33, 15 August 2011 (UTC)

• No, SCOTUS is a part of the US government, so is a first party source for its own decision, and any federal court decisions, quotes from federal prosecutors, federal legislation, etc.

• No, that's not enough to make them first-party. SCOTUS is a first-party, primary source for what SCOTUS says; it is a third-party/independent source for what some independent branch of the federal government, i.e., the Executive or Legislative branches, says. And it's certainly not "affiliated" or first-party with respect to the claims put forward by the opposing sides in the case. When they wrote in Brown v Board that "This segregation was alleged to deprive the plaintiffs of the equal protection of the laws under the Fourteenth Amendment", SCOTUS was not a first-party source for that material. (Primary, yes. Affiliated, no.) WhatamIdoing (talk) 02:09, 16 August 2011 (UTC)

• I would say an appellate court is a secondary source when describing the evidence placed before a lower court, and it's a primary source when discussing the arguments made before it, and its own decisions. SlimVirgin ^{TALK|CONTRIBS} 02:14, 16 August 2011 (UTC)

• If it merely repeats what was said by someone else, then it is not a secondary source. It's not merely a matter of counting up chains in the link. If I quote Masem, and you cite me, then that's still primary material, even thought your step is third-hand. WhatamIdoing (talk) 22:47, 16 August 2011 (UTC)

• Strong support - the removal of this line drastically changes how WP:V is to be applied. --MASEM (t) 23:21, 14 August 2011 (UTC)

• That sentence is about notability, not verifiability, so it does not belong in WP:V. — Carl (CBM · talk) 02:35, 16 August 2011 (UTC)
  • The sentence is not about notability, it is about the existence of third-party sources to provide an unbiased and reviewed facts about a topic, needed for WP:V. --MASEM (t) 03:04, 16 August 2011 (UTC)
    • Here is what the first paragraph WP:Notability says about the sentence.

"... if no reliable third-party sources can be found on a topic, then it should not have a separate article. Wikipedia's concept of notability applies this basic standard to avoid indiscriminate inclusion of topics."

Bob K31416 (talk) 15:32, 16 August 2011 (UTC)

• • • Third-party sources are only needed to show notability. An article written using entirely first-party sources (e.g. a biographical article using only a published autobiography by the subject of the article, or an article about a scientific experiment sourced only to a reliable, peer-reviewed paper where the experiment was published) could meet WP:V without meeting WP:N. There are many reliable first-party sources (and unreliable third-party sources). — Carl (CBM · talk) 15:53, 16 August 2011 (UTC)

• Such an article would violate V, because of the lack of at least one reliable third-party source. The point of requiring this is to make sure someone other than the primary sources (the authors) deem the material worthy of comment; the requirement is also in place to avoid OR. The requirement has been in this policy (a core content policy) for years, since before the Notability guideline existed, and was regularly acted on. SlimVirgin ^{TALK|CONTRIBS} 19:17, 16 August 2011 (UTC)
  • No, WP:V has nothing to do with "third-party" sources, that is purely the domain of WP:N. It may be that this policy had language about notability before WP:N existed, but that is years in the past and no longer relevant. Now that WP:N does exist, we have a distinction in policy between verifiability and notability: it is possible to have a verifiable article on a non-notable topic, and an unverifiable article on a notable topic. These independent considerations are covered by separate policies, and language about notability doesn't belong here. — Carl (CBM · talk) 19:46, 16 August 2011 (UTC)

• N is a guideline. V is a core content policy. It has been the case for as long as I remember that articles had to be based on reliable secondary sources. We are meant to offer an overview of the literature, and to do that a literature has to exist, at least in the minimal sense that a third-party source—someone other than the primary source and the Wikipedian who created the article about it—has deemed the issue worthy of mention. That's not just an N issue. It is a V and NOR issue, which is why this policy has included mention of it since at least 2006.

  You can't (by definition) have an unverifiable article on a notable topic, and I can't think what a verifiable article on a non-notable topic would be. SlimVirgin ^{TALK|CONTRIBS} 20:57, 16 August 2011 (UTC)

• Mostly, I agree with SV's analysis, but it is certainly possible to have a verifiable article about a topic which is non-notable for encyclopedic purposes. Millions of newspaper pages of "society events" and such drivel have been published, rendering the parties described verifiable. Some of these were even written by reliable journalists. But the parties would not on their own merit be notable topics for WP, would they? There needs to be a little more to it if we don't want Emma Smith's 1875 cotillion in New York City. LeadSongDog come howl! 22:17, 16 August 2011 (UTC)

WP:N is only a guideline, not a policy, because if a separate article is created on something that is non-notable but verifiable, no great harm is done (provided other policies, such as NPOV, are followed). Another way of saying this is that if a few facts that really belong as a heading in a broader article are instead a separate article, this is merely a guideline non-compliance rather than a policy violation. Jc3s5h (talk) 22:27, 16 August 2011 (UTC)

Apart from society events (like LeadSongDog mentions) another way to get a verifiable article on a non-notable topic is to take a paper in some academic journal, on a topic that has only ever been studied in that one paper, and write an article that directly summarizes the results of the paper. We often delete or merge articles like that for lack of notability, but nobody argues that they are unverifiable if every claim they make is literally present in the paper being used as a source. Regarding "You can't (by definition) have an unverifiable article on a notable topic, ...", that brings out the fundamental difference between WP:V and WP:N. N is about the article topic, independent of its actual content, while V is about the actual content, independent of the worthiness of the topic. These should be treated separately, which is why we have separate pages for them. — Carl (CBM · talk) 00:10, 17 August 2011 (UTC)

There's no point in tying ourselves in knots, because this isn't rocket science. If someone creates an article on "John Smith's fishing technques," it needs a source that isn't My Fishing Techniques by John Smith, or Smith's wife's blog. That is, we need a reliable, published third party—not John Smith and friends, and not the Wikipedian who created the article—to tell us (a) that this issue is worth mentioning, and (b) what kind of thing we should saying about it to avoid OR. That's the only issue that V has ever commented on. All the details belong in N. SlimVirgin ^{TALK|CONTRIBS} 00:27, 17 August 2011 (UTC)

We both agree that article needs another source besides that one. But I am saying that fact belongs in WP:N, not here, because it is not related to whether the material that is included in the article is verifiable. Knowing whether "the issue is worth mentioning" (that is, whether we should have the article) is entirely a WP:N issue. The other thing you mentioned seems like an NPOV issue. If Smith's book is the only source in existence, then (essentially by our wiki definitions) its viewpoint is the NPOV viewpoint on the matter, and descriptive claims taken from it are verifiable and not original research. My response to the RFC is that this policy should stick to verifiability, not repeat things that belong in other policies. — Carl (CBM · talk) 01:35, 17 August 2011 (UTC)

I do not agree, actually. Per WP:SELFPUB, a primary source is a perfectly good source to V itself and the basic facts about itself. I have a dictionary here. I don't need a third party source to tell me that it exists, nor what year it was published. I can cite the primary source itself. That's neither enough to 1) comment on specific controversies, awards, etc. or 2) demonstrate notability. But the "we need secondary sources to meet V!" mantra is just not correct. No matter what the particular wording of V is changed to, it doesn't make sense to say that a published work is not an authoritative source on its own existence and publication date. Jclemens (talk) 01:39, 17 August 2011 (UTC)

I just meant that article needs more sources to meet WP:N. I agree with what you said. — Carl (CBM · talk) 01:48, 17 August 2011 (UTC)

To take the example of "John Smith's Fishing Techniques", if the only sources are the book and the blog, we have way to consider if this source is even close to reliable, working on the assumption that John Smith here-to-date was a unknown person. Ergo, without any third-party addressing the book, we cannot even assure it meets the basic "verifiability, not truth" aspect. Maybe the entire book is a work of fiction published as a non-fiction title, maybe it plagiarizes Joe Jones's Fishing Techniques, we don't know. We'd not be able to verify that what John Smith has written actually took place; we'd only have John Smith's word on it, which is not sufficient. A reliable third-party source that would at least consider the work, determining if the account is legit or not but without otherwise introducing additional analysis (eg a primary third-party source) would at least provide something towards verifiability. Furthermore, a third-party source would still show the work existed even if only ten copies were made and destroyed without the original text surviving; without a third-party, in such a case, even if you can remember the general ideas of the text in your head, there's no way for any other user to validate it, ergo, it would fail. However, even for a published book, there is nearly always an ISBN number, and therefore some record of the book's existence in a third-party catalog (maybe LOC) even if that is just a primary source.

Note that this is nothing about secondary, transformative sources. Third-party != secondary. WP:N asks for secondary sources as a basis to build a good encyclopedic article. WP:V asks for third-party so that we know we're putting in facts that the reader can check themselves. --MASEM (t) 13:11, 17 August 2011 (UTC)

• Masem commented, "WP:V asks for third-party so that we know we're putting in facts that the reader can check themselves."
• The subject sentence is, "If no reliable third-party sources can be found on a topic, Wikipedia should not have an article on it."
• And here is the subject sentence in the first paragraph of WP:Notability, "... if no reliable third-party sources can be found on a topic, then it should not have a separate article. Wikipedia's concept of notability applies this basic standard to avoid indiscriminate inclusion of topics."

Bob K31416 (talk) 14:31, 17 August 2011 (UTC)

An ISBN catalog is not a third party source, nor is a library catalog. Those help Wikipedians locate the primary source, but that's essentially it. I could probably get an ISBN for my cat (if I had one, which I don't), and then get it cat-aloged (ha ha) somewhere, but those entries wouldn't prove the existence of a work: the ability to look up, purchase, check out, or otherwise acquire the primary source for verification is what's really happening, and catalogs and listings are merely aides to that end in this context. — Preceding unsigned comment added by Jclemens (talk • contribs) 02:00, 18 August 2011 (UTC)

• Comment. This seems to have consensus, so I'm assuming we can go ahead and do it. SlimVirgin ^{TALK|CONTRIBS} 07:38, 23 August 2011 (UTC)
  • So long as "this" still refers to the proposal at the top of this sub-section ("to delete the notability section and move the sentence back to the "Burden of evidence" section"), then I think you're right about the consensus. WhatamIdoing (talk) 19:43, 24 August 2011 (UTC)

Time to centralize discussion

STOP... it seems that there is a debate at GNG that is centered on what this policy says about notability.... and a debate here that is pointing to what that guideline says. I suspect that both pages are being improperly edited in order to skew the debate at the other page. Please, do not edit one policy or guideline in order to win a debate at some other policy or guideline page. Instead, we need to create a centralized discussion so editors on both pages can reach a consensus as to what should be said on both pages. Given this confusion... I am going to revert both pages back to their last stable versions while we centralize the discussion and reach such a consensus... I suggest that WT:Notability is the better venue for that discussion. Once we figure out what the notability guideline should say, then we can come back and make any edits to this page that are needed. Blueboar (talk) 20:07, 13 August 2011 (UTC)

Assume good faith! First off, different people are editing these articles. The debates for each page should be on that article's talk page, not mixed together. Dream Focus 20:29, 13 August 2011 (UTC)

OK... I suppose it is possible that it is not a deliberate thing... but the problem remains... both discussions are pointing to similar bits of language that exist in the other policy/guideline page. Neither page can have a meaningful discussion or resolution when both sections are being edit warred over. Blueboar (talk) 20:56, 13 August 2011 (UTC)

Removing "not truth"

• Discussion moved to Wikipedia talk:V/First sentence. Please continue there.

Let's end the "Verifiability, not truth" topic

• Discussion moved to Wikipedia talk:V/First sentence. Please continue there.

Discussion about the new sub-page

[above header inserted by Unscintillating (talk) 00:21, 20 August 2011 (UTC)]

Is there a reason why all the discussions are being moved to a sub-page? Blueboar (talk) 02:52, 19 August 2011 (UTC)

Following suggestions from several people, it's to let this talk page return to its normal function. SlimVirgin ^{TALK|CONTRIBS} 02:56, 19 August 2011 (UTC)

I don't think that I agree with moving existing discussions to a subpage. It is disruptive to expected archiving. I think highly focused, extremely length (bytes and time) discussions are better handled at a similarly highly focused essay, but that is in respect of moving forward, not refactoring. --SmokeyJoe (talk) 15:12, 19 August 2011 (UTC)

(1) Do you have an example of such an essay/discussion?  (2) The essay you mentioned seems to be a biased starting point since it has language in the title that we want moved to a footnote.  (3)  When I proposed that we pick a subpage name for a future discussion to start on September 15, the model I had in mind was like at the mediation cabal or an AfD discussion—they start such discussions on a fresh page.  Unscintillating (talk) 03:40, 20 August 2011 (UTC)

(1) Wikipedia:Verifiability, not truth is the title of the essay, or did I misunderstand the question? (2) The essay is, yes, POV with respect to the issue. It represents the long standing status quo. The challenge to anyone who seeks to change the status quo is to make the essay NPOV. In general, the best essays are NPOV. They state the issues, the facts, the opinions, without reflecting the bias of the author. --SmokeyJoe (talk) 22:58, 20 August 2011 (UTC)

(1) Do you have an example where people have used an essay to build consensus for a policy page?  Thanks, Unscintillating (talk) 00:21, 21 August 2011 (UTC)

WP:N (not policy-proper but pretty close - it is enforced by policy) was built using essays. Uncle G's was early and the most influential. WT:N (especially 2007) was a complicated, noisy place. Regulars maintained positions. Newcomers (not familiar with the regulars' positions) engaged by writing mini-essays all over the WT:N. Essay writing became popular, and it was useful. Once a regular had an essay, it was so much easier to understand their position. When disgruntled newcomers wrote essays, it quickly forced them to adopt a logical framework. There was competition to have your essay cited by the policy page.

There was no single essay that was used to build consensus, that's true. Here, I am suggesting focus on a single essay because I read a dispute about an extremely focused issue, which is the subject of that same issue. Perhaps multile essays could be used. However, what is important is the use of essays. Threaded dialogue needs something, like an essay, or a proposal, to give it direction. Normally, per WP:Consensus, it is the article, or actual edits to the article, that are the subject. This doesn't work with policy because experimental edits to policy are not welcome. WP:BRD therefore doesn't work with policy, or at least, anymore with established policy. --SmokeyJoe (talk) 23:05, 21 August 2011 (UTC)

I'm not pleased at the prospect of having yet another page to keep track of. WhatamIdoing (talk) 16:18, 19 August 2011 (UTC)

Isn't it just two clicks to add the other page to your watchlist? And this way, people get the option of watching only one of the two pages, if they want. Though the downside is that people who might be newly watchlisting this page won't be aware of the pertinent discussion taking place elsewhere. --Kotniski (talk) 16:28, 19 August 2011 (UTC)

Sure, but with more than 2,000 pages on my watchlist at the moment, the mere fact that the page is listed there does not mean that I notice it. WhatamIdoing (talk) 19:33, 24 August 2011 (UTC)

As much as I dislike the endless discussion... Given that the proposals under discussion have the potential to completely reverse current policy ... I am not at all sure that it is proper for discussions of such magnitute to take place "hidden away" on a side channel. But I suppose I can live with it... as long as it is understood that any consensus that might be reached there will have to be reviewed and fully discussed here before any major change is implemented. Blueboar (talk) 17:34, 19 August 2011 (UTC)

I don't think there's any fundamental change being suggested, just an improvement of the wording (though even that seems unlikely to be accepted, given how religiously people seem to cling to their received scriptures). A better thing to discuss would probably be merging this page and its fork, WP:OR, into one, and hopefully eliminating the odd perversions of language from both at the same time. This will no doubt also raise the ire of the devout, but at least the potential gain would be something large rather than something small.--Kotniski (talk) 17:44, 19 August 2011 (UTC)

If the purpose of the subpage is for people to work out a proposal to present here, that seems fine to me. But this discussion has been going nowhere for many months now, and I think continuing to push for removing "not truth" is tendentious. If editors working there reach consensus there about what to bring here, no problem. But any decisions about what should be changed on the policy page should be vetted here before they are implemented. --Nuujinn (talk) 17:53, 19 August 2011 (UTC)

If the discussion has to be here before it can be implemented, then all discussion must take place here. Using the subpage discussion as an excuse to create extra obstacles to change is unacceptable.—S Marshall T/C 18:16, 19 August 2011 (UTC)

It is not setting an obstacle, it is recognition that editors may not be aware of the discussion. I think it would be fine to work out the wording of a proposal, either for the text or for an RFC on a subpage, even beneficial, but we will need to hold an RFC on any suggested change at this point, I think it's fair to say. --Nuujinn (talk) 22:03, 19 August 2011 (UTC)

A sentence that was added without a RFC ought to be removable without a RFC. The phrase "not truth" was added with a unilateral WP:BOLD edit, which was followed by individual discussions with the individual editors who objected, which discussions were conducted on their personal talk pages. You'll notice that some editors are still free to unilaterally edit the page, and take objections to individual editors' talk pages. Other editors are autoreverted and then directed to this page, where their proposed additions can be mired in eternal discussion. I learned this when I was trying to add the phrases about copyright, and I was only able to achieve change by extreme persistence on this page. The experience has taught me that the only route to success is to individually persuade everyone who will listen until we outnumber those who will not.

The stance that the current version of WP:V is carved in stone, and can only be changed after everyone who's ever edited Wikipedia has been personally notified of the discussion and given a chance to object at a long series of RFCs, is understandably popular with editors who like the current wording. But in fact, that kind of process is not necessary to improve a policy page. A local consensus is quite sufficient.

I don't mind the subpage, on condition that there's a reasonably prominent note directing editors to it. That seems like a suitable way forward to me. However, if there's concern that interested editors might fail to find the subpage, then there's no reason why we shouldn't make the dispute notice more prominent to ensure that it gets noticed. Shall I put a red warning triangle on it and a border around?—S Marshall T/C 22:54, 19 August 2011 (UTC)

• Oh, will you look at this behaviour. She refuses to participate in the discussion; she will not allow the discussion to take place on WT:V; she will not allow her wording to be flagged as disputed; and she will not allow a pointer to be added to the place where she requires the discussion to take place. What manipulative, controlling behaviour. I see this as an extremely bad faith and obstructive edit.—S Marshall T/C 23:46, 19 August 2011 (UTC)

If you feel that way, take it to one of the noticeboard such as ANI. And while I believe that there is obviously no consensus to remove not truth, which would be required to remove not truth, even though there may be no consensus to keep it either (which I myself do not believe), I think it's pretty clear that we will need yet another RFC on the issue, should anyone manage to pull together some semblance of a proposal to hold an RFC on. And at this point, I'm pretty certain that we'll have to take this through dispute resolution or a similar board, because this is clearly going nowhere. Some of us have tried to work on a compromise only to have those efforts rebuffed, and continuing to push a position after it is clear that there is no consensus for it is disruptive. I think we pass that point months ago, myself. --Nuujinn (talk) 00:32, 20 August 2011 (UTC)

(edit conflict)At present there are two brief sections on this page that refer editors to the respective sections on the subpage. When new sections are added to the subpage, new brief referring sections can be added here. When sections on the subpage are archived over there, then the brief referring sections on this page can also be archived here. Perhaps this is sufficient? Bob K31416 (talk) 00:40, 20 August 2011 (UTC)

I think that is a good idea, but given that Blueboar and Nuujinn have both objected, it seems that the new page should only be used for background discussions.  Unscintillating (talk) 03:40, 20 August 2011 (UTC)

I think if this sentence is to be protected through the banning of discussion about it on this page and the suppression even of notices regarding the fact that it is being discussed on some other page, then its supporters ought to be able to do more than "like" the present version, they ought to be able to defend it logically. As far as I recall, no-one has yet given any satisfactory explanation as to:

1. Why we want to say "the threshold", when (if threshold is to be interpreted as "necessary condition") this would imply "the" (only) necessary condition, and hence a sufficient condition, which is not the case;
2. Why we use the word "verifiability", which means something else in English (and is therefore being used unnecessarily and misleadingly as jargon for sourceability - much like the twin phrase "original research" has had its ordinary meaning perverted in the same way);
3. Why we say "not truth", when, out of all the things which the threshold is not, truth is actually one which we don't have any particular objection to (again we are taking a word and making it mean something other than it does in English - what "truth" is supposed to refer to here is something like "editors' unsupported opinions").

The rest of the sentence kind of makes up for points 2 and 3, at least, by explaining what it is we actually mean - but why go to the trouble of misleading people first? --Kotniski (talk) 05:46, 20 August 2011 (UTC)

@Kotniski, I think your points are excellent, but there is an expectation after the last round of RfCs that there would be a break in the discussion for a while.  This break has not yet happened.  Would you be willing to hold off on content discussions until September 15?  Note that both S Marshall and North8000 are currently on Wikibreaks.  Thanks, Unscintillating (talk) 13:57, 20 August 2011 (UTC)

I don't particularly mind not discussing this for another month if no-one else wants to. But I hope that when discussion resumes, people will try in good faith to find a wording of these policies that meets all the genuine concerns raised, not just oppose all change on the grounds of "it's always said this".--Kotniski (talk) 14:59, 20 August 2011 (UTC)

Kotniski, with all due respect, I feel like a large number of us have defended our views logically, and have expressed ourselves clearly and some of use have explicitly and opening tried to work out wording that would address some of the concerns of other editors even though we do not see a problem that needs to be fixed. And those discussions have been derailed by those who are insisting that "not truth" be removed. What I want to avoid is a discussion with a limited number of editors in a subpage who declare consensus and then foist the result on others. I strongly endorse the idea of a break until Sept 15th, and I would strongly suggest that we use the subpages to hammer out proposal to bring back here at that time. Then we can list the proposals and hold an RFC on which is the best of the lot. --Nuujinn (talk) 15:54, 20 August 2011 (UTC)

• tl;dr. So, we are having a discussion about discussion? I'm all in favor of moving that stuff... anywhere else. When someone actually has something that's ready for prime time, as opposed to being a waste of everyone else's time, do please post it here for serious review. --Tryptofish (talk) 16:17, 20 August 2011 (UTC)

Proposal for moving forward productively

Just a straw poll, please !vote yea or nay to see how everyone feels. Who supports the notion that:

• We cease all discussion of the wording of the first paragraph of WP:V on this page until 15 September in order to let everyone collect their thought.
• Editors work on proposals for wording of that paragraph anywhere appropriate, in groups or alone
• We find a venue for discussion of the proposals and ask some uninvolved admins to act as moderators of the discussion
• We bring the proposals to a section here starting on 15 September
• We begin an RFC on the submitted proposals starting 17 September

The goal is to wind up with some number of concrete proposals about which we can gauge consensus without falling into endless discussions as we have for the last months. --Nuujinn (talk) 17:10, 20 August 2011 (UTC)

I prefer a new venue for the discussion on September 15, one that a moderator can organize.  Unscintillating (talk) 17:38, 20 August 2011 (UTC)

Seems like a good idea, amended. --Nuujinn (talk) 17:47, 20 August 2011 (UTC)

Please note that bringing back too many proposals here will have the problem of splitting the support for change. Those who want change should reach consensus amongst themselves on the best available proposal. Individuals should keep in mind that even though it may not be their favorite proposal, they should support it when it is brought back here if it is an improvement on the present sentence that is in the policy. Bob K31416 (talk) 18:31, 20 August 2011 (UTC)

Yes. I also think that if we are going to work with a moderator or moderators, we should figure out how to hold the RFC with their guidance. Ideally, I think 1-3 proposals would be most appropriate. --Nuujinn (talk) 22:13, 20 August 2011 (UTC)

Why can't the moderator set thing up here? I really dislike the idea of discussions about changing the policy taking place on some sub-page. Drafting of proposals is one thing... but policy changes should be discussed on policy talk pages. Blueboar (talk) 23:00, 20 August 2011 (UTC)

I oppose ritualised confrontation. It is not the wiki way. It is not WP:Consensus. It implies seniority of editors involved in the pre-agreement. It discourages outside editors from engaging. If progress stalls, better to make your point, and leave it to later editors. --SmokeyJoe (talk) 23:16, 20 August 2011 (UTC)

So, what would you suggest we do? This has been going on here literally for months, with no progress whatsoever. --Nuujinn (talk) 00:06, 21 August 2011 (UTC)

I see virtually nothing for you or me to do here. The policy page works. Editors know that content needs to be verifiable, or it gets removed. The people who want to make a change, they need to offer a better explanation. I cannot find the logic of their rationale. I read flawed logic, but it is hard to hgihlight because I find no agreed explanation. --SmokeyJoe (talk) 00:25, 21 August 2011 (UTC)

Not verifiable, but properly sourceable. I think that's the fundamental point - this page has the wrong title (and consequently the wrong first sentence) to start with. The same applies to WP:NOR - the only way these two principles can be made to make sense is to redefine English words to mean something incompatible with what they mean in real life. And it so happens that we've redefined these two terms ("verifiable" and "original research") to refer to exactly the same concept ("OR" is that which is not "V"). Hence we have two policy pages where we should have only one; both those pages are wrongly titled; and the beginnings of both of those policy pages are misleadingly written, as editors try to contrive a link between the words used in the title and the actual substance of the policy. What we should do is combine the two pages into a single policy, compact and logically coherent, and with a title which, while not necessarily serving the soundbite addicts, properly expresses in real English (or as close as we can get) what the subject of the policy is.--Kotniski (talk) 09:12, 21 August 2011 (UTC)

We even have that page sitting on the shelf, ready to go. Compact and logically coherent, and with a title which properly expresses in real English what the subject of the policy is. See Wikipedia:Attribution. --SmokeyJoe (talk) 11:28, 21 August 2011 (UTC)

Hmmm, "attribution"?  Seems like there has been recent objection to overloading that word in the Wikipedia context.  What about a new attempt at combining the two called WP:Sourceability?  Unscintillating (talk) 13:52, 21 August 2011 (UTC)

I prefer Wikipedia:Attribution. It was very well done. It will define the word "attribution" nicely. Unfortunately, there was a little problem with its implementation, though not so much with its substance. --SmokeyJoe (talk) 22:48, 21 August 2011 (UTC)

Predictably, as one who is never satisfied, I don't believe "Attribution" is the right title either. This isn't (just) about attribution; it's about "attributability" (if that's a word), or "sourceability" (which seems more like a word to me). The fundamental principle (and this page is about fundamental principles) is not that everything has to be sourced, but that everything has to be able to be sourced (which is presumably why we currently call this page "Verifiability" rather than "Verification"). "Attribution" (when, why and how editors should attribute things) is an overlapping but non-identical topic. --Kotniski (talk) 11:10, 22 August 2011 (UTC)

Predictably, the discussion has wandered off the topic for this section. --Nuujinn (talk) 11:23, 22 August 2011 (UTC)

Straw poll, should this process discussion remain here or be moved to WT:V/First_sentence?

• Move  Unscintillating (talk) 17:38, 20 August 2011 (UTC)
• Remain ... I don't mind if those who desire change use some other page as a place to draft a proposal... but discussions as to process, and and discussions about any proposals that result from the drafting should take place here on this page, so the maximum number of people are involved and know what is going on. Blueboar (talk) 22:26, 20 August 2011 (UTC)
• No. Any proposal to modify directly WP:V belongs on WT:V. However, the proposal above to alter the few words has failed. To my reading, the case against "not truth" is not clear, is certainly not coherent. A decent explanation may exist, but it is buried. There is a long standing essay supporting the status quo, at Wikipedia:Verifiability, not truth. I cannot find an essay challenging. This is why I suggest debating the idea at the essay. I think that debate is very likely to be mutally educational. I did not suggest moving focused debate to a subpage. I think that Slimvirgin's refactoring should be reverted, at WT:V/First_sentence deleted. --SmokeyJoe (talk) 22:53, 20 August 2011 (UTC)

Hmmm... I'm not completely happy with people changing the current WP:Verifiability, not truth essay... it sums up the majority view on the the sentence in question very well. But I could definitely see how a separate essay, something that explains that not everyone agrees with the viewpoint expressed at the WP:Verifiability, not truth essay and explains the minority viewpoint, would be beneficial (suggested title: "WP:Truth through verifiability"). Blueboar (talk) 23:13, 20 August 2011 (UTC)

IF "not truth" should be removed, then the essay is flawed and should be improved. You are free to link to your preferred versions. I recommend NOT forking to lock in different POVs. --SmokeyJoe (talk) 23:18, 20 August 2011 (UTC)

Expressing different POVs as to policy is what essays are for (or at least one of their purposes). Essays reflect the opinions of various segments of the community... they don't necessarily reflect majority opinion. Blueboar (talk) 23:36, 20 August 2011 (UTC)

I don't much disagree, except that I think it would be better for the minority faction to try to get traction at WP:Verifiability, not truth, than to taalk in the dark at WP:Truth, not verifiability. --SmokeyJoe (talk) 02:48, 21 August 2011 (UTC)

I don't know what you mean by the "minority faction", when according to the previous RfCs about 50% want change, and 50% oppose change.  Unscintillating (talk) 04:31, 21 August 2011 (UTC)

The "minority faction" are the group who know what specific change they want. --SmokeyJoe (talk) 11:24, 21 August 2011 (UTC)

Well then I think that you are confused, because twice there has been a consensus for a change, only to have one editor that had not been participating in the discussion to arrive and become a lone voice of objection.  Also, I skimmed the essay, the material in the essay is not the current problem, but since I don't want to drift into a content discussion, I don't want to further elaborate, there is already plenty about this in the archives.  We might want to use some of the material in the essay for proposals.  Unscintillating (talk) 13:38, 21 August 2011 (UTC)

You only just notice now that I am confused? This place is very confusing. Are we talking about removing "not truth"? In archives, I find text about an undefined "the issue". I see lots of agreements on things that don't seem to be associated with a problem. If "twice there has been a consensus for a change" then please point me to the records. If the material in the essay is not the current problem, then what is the current problem? Was it a past problem, solved or deferred? I'm also confused because I cannot work out your position in the debates. --SmokeyJoe (talk) 22:45, 21 August 2011 (UTC)

• (ec)(ec)Comment  There is already process discussion taking place at WT:V/First_sentence, so this straw poll exists to decide if we want to have one place for process discussions, or proceed with having two places for process discussions.  I see no basic problem with having such discussions in two places.  Process discussions do not include changes to WP:V, but rather how we are going to organize ourselves for the likely discussion on September 15.  Unscintillating (talk) 23:39, 20 August 2011 (UTC)

And I was simply expressing my dislike of the fact that there is discussion going on at a sub-page... and my opinion that WT:V/First sentence should be used for drafting proposals, while all of the discussion part of the process should be done here. Blueboar (talk) 00:48, 21 August 2011 (UTC)

Agree with Blueboar, except that drafting proposals would go better at WP:V/First_sentence. --SmokeyJoe (talk) 02:45, 21 August 2011 (UTC)

If I understand the concept SmokeyJoe is suggesting: the drafts would be located at wP:v/first sentence, discussions about the wording of the drafts would take place at the related talk page (wT:v/first sentence)... and then, when the drafting is complete, and the drafts are ready to presented as more formal proposals, those proposals would be brought back here (WT:Verifiability) for discussion. That seems to make sense to me... organized but without any seeming impropriety. Blueboar (talk) 14:38, 21 August 2011 (UTC)

Yes. (watch capitalisations though). Drafts can go on a subpage. Very specific discussion about the draft (discussion about improving the draft) goes on the draft talk page. Discussion about actually implementing the draft goes on WT:V. --SmokeyJoe (talk) 22:39, 21 August 2011 (UTC)

• Move the discussions to WT:V/First sentence. Per Blueboar, the policy can't be changed without agreement on this talk page. Then again, it can't be changed substantively without discussing it with the wider community anyway, so that's a moot point. But to continue these endless polls and proposals on this talk page has been extremely disruptive, and it's discouraging people from using the page to discuss other issues. SlimVirgin ^{TALK|CONTRIBS} 03:02, 21 August 2011 (UTC)
• I'm planning on starting a a poll to decide whether this poll should be moved to another page, but I'm not sure whether to start it here or on another page. Can we have a poll to decide? (OK, seriously, have all first-sentence discussions on the subpage, but don't keep removing the pointer from the policy that tells people that the sentence is under discussion there.)--Kotniski (talk) 09:17, 21 August 2011 (UTC)

• Move it there for a few weeks or several weeks This could use a few or several weeks to calm it here, get out of the trap of just trading and parrying talking points, for the "change needed" folks to clarify the issues and goals and come up with a proposal and organized presentation. But I certainly am against any move etc. which could get interpreted/quoted as banishing/burying it to an obscure page from the main page which is this one. Sincerely, North8000 (talk) 02:28, 23 August 2011 (UTC)

new page Wikipedia:V/First sentence/Drafts

Please see Wikipedia:V/First sentence/Drafts. Unscintillating (talk) 23:40, 21 August 2011 (UTC)

new archives

• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 44  WT:V/Archive 44 with only First sentence sections
• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 45  WT:V/Archive 45 with only First sentence sections
• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 46  WT:V/Archive 46 with only First sentence sections
• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 47  WT:V/Archive 47 with only First sentence sections
• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 48  WT:V/Archive 48 with only First sentence sections
• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 49  WT:V/Archive 49 with only First sentence sections
• Archive: Wikipedia talk:Verifiability/First sentence/V FS Archive 50  WT:V/Archive 50 with only First sentence sections

Unscintillating (talk) 04:37, 22 August 2011 (UTC)

I'm not sure what the above is. All the 2011 discussions about this are archived on one page, Wikipedia talk:Verifiability/First sentence polls 2011, as well as in the regular chronological archives. SlimVirgin ^{TALK|CONTRIBS} 07:21, 22 August 2011 (UTC)

1. Was the documentation beside the name of the archive helpful if you weren't sure?
2. As far as the "poll" file having discussion in addition to the polls, it hadn't occurred to me to look for such. 
3. The earliest discussion in the "poll" file is from April, that is not all of 2011; for example we reached a talk page consensus for change on January 10.
4. Anyway, they have slightly different purposes and are being maintained slightly differently, so this will give editors a choice, including reading from the original archives.  Unscintillating (talk) 09:01, 22 August 2011 (UTC)

Long-term sourcing/removal policy

I find a contradiction between different aspects of policy, and between policy and the real world:

WP:PRESERVE says not to remove material unless it's more than merely unsourced. WP:V says "Anything that requires but lacks a source may be removed... You may remove any material lacking a reliable source that directly supports it. How quickly this should happen depends on the material and the overall state of the article. Editors might object if you remove material without giving them time to provide references."

Have any of you read this or similar article?

Wikipedia has many, many articles filled with unsourced material, which are almost certain to never be sourced. The community needs to decide what's more important: having a lot of unreliable content, or letting people remove unsourced material which has been appropriately tagged for a (long) period of time.

I'm of the opinion that if material has been tagged as unsourced for a long time (say a year) it should be removed, and that the Encyclopedia is too large to require the editor who removes material to try to source it (per WP:BURDEN). But I was recently informed that such removal is a major issue for some people. Can you help with this, and can we clarify policy on it so I and others like me will know where Wikipedia as a whole stands? B^e——C_{ritical__Talk} 23:59, 22 August 2011 (UTC)

I think we would deal with such issues on a case by case manner. We're all volunteers so dictating actions can be problematic. Can you point us to the problematic discussion? We may be able to make some concrete suggestions. --Nuujinn (talk) 00:03, 23 August 2011 (UTC)

Oh, I'm a bit embarrassed really... see, I had been just editing according to the principle above: if it was tagged for a year or more, I removed it. I did that on a whole bunch of articles, and was reverted maybe once last time I looked. But when someone thought I was an admin I said I wished I was, and later they said they wouldn't vote for me because I'd been deleting that text. See my edit history. B^e——C_{ritical__Talk} 00:12, 23 August 2011 (UTC)

Well, this isn't really the forum for this kind of discussion, but I don't see much wrong with deleting unsourced material, esp. if an article or section has been tagged for a year or more. But I would say that if you balance the deletions with efforts to find sources and improve articles, too, that's better for the project in the long run. I wouldn't worry about becoming an admin, just focus on being a good editor. Just my two cents, worth less every day. --Nuujinn (talk) 00:26, 23 August 2011 (UTC)

Thanks (; You're right it would be better to source things given time. I'd really like to know if policy can be made clear on this, so is there another place I should post this? B^e——C_{ritical__Talk} 00:40, 23 August 2011 (UTC)

Hi, that part of the policy means only that we should ask for sources, ideally help to find them ourselves, and wait a decent period before removing unsourced material (not counting BLP and other pressing issues). But I would say a week or few days, definitely not a year. That's assuming it's something that really needs a source. SlimVirgin ^{TALK|CONTRIBS} 07:43, 23 August 2011 (UTC)

Hi SV, OMG, the opinions are really all over the place on this issue. Any suggestions for how to go about consensus building for policy change? But I'm not sure what you mean by "really needs a source." For example, does a history section on an article about a high school need a source? B^e——C_{ritical__Talk} 13:42, 23 August 2011 (UTC)

A lot depends on what you mean by "need a source". all information in Wikipedia must be verifiable... but not all information needs to be actually verified ... in other words, a source has to exist, but the article does not necessarily need to contain a citation to that source. That depends on the specifics of what is said in the article and whether the material is "challenged or likely to be challenged".

In other words... we allow removal of unsourced information, but we don't require it. Whether to remove or not depends on the specifics of the article, and the nature of the information in question. Blueboar (talk) 15:00, 23 August 2011 (UTC)

If you have a bunch of facts stated, and those facts are not common knowledge that any reader will have, doesn't it require a citation? I know you don't have to cite that the sky is blue. I see a problem with the following text: "To show that it is not original research, all material added to articles must be attributable to a reliable, published source," because without a citation, how do you know it's "attributable?" I have been interpreting that to mean "To show that it is not original research, all material added to articles which is not common knowledge must be attributed to a reliable, published source or it may eventually be removed." B^e——C_{ritical__Talk} 17:01, 23 August 2011 (UTC)

I think it's largely a matter of personal judgement, to be applied in good faith and a good dose of common sense. Vast quantities of very useful (and perfectly sourceable) information has been sitting around uncited in Wikipedia for a very long time - by removing that type of information, you'd be making the encyclopedia worse, not better. But if it's not cited because it's wrong (or even if it is cited and it's wrong, which is also very possible) then by removing it you're doing a service. The more you know of the matter, and the more you've made the effort to look for sources yourself, the more capable you're likely to be of judging whether it's case A or case B (though often anyone with common sense will have a pretty good idea).--Kotniski (talk) 17:14, 23 August 2011 (UTC)

Okay, that's good common sense. I went and looked at one of the articles whose unsourced text I removed. What I see there is that the information is outdated, might have been original research or copyvio to begin with (see the very first edit), and includes so many facts that full sourcing would be a pain. On the other hand, it's fairly detailed information. It might be of use to someone... if it's accurate. I also notice that if it wasn't original research to begin with, it would have been extremely easy to cite (as in one cite per section). What do you guys think of it?

Kotniski, what about the good-faith argument that non-subject-matter-experts need to be able to determine the reliability of Wikipedia material, and that the lack of such citation is a basic problem that needs to be dealt with, so that WP can become a reliable encyclopedia, and not just another site? B^e——C_{ritical__Talk} 18:47, 23 August 2011 (UTC)

Just an aside regarding your example article Salmon High School: The source of the verifiability problems was lack of notability and I added a template. Bob K31416 (talk) 19:51, 23 August 2011 (UTC)

OMG, well in that case there are just hundreds and hundreds [4]. I've always had a problem with NOTABILITY, in that it sets the standard way too low: so low that a subject that meets the criteria doesn't have enough sources for a well-rounded article. B^e——C_{ritical__Talk} 20:29, 23 August 2011 (UTC)

Yes, OMG. I wonder how many of the high schools in that list are not notable. Regards, Bob K31416 (talk) 21:00, 23 August 2011 (UTC)

Sorry Kotniski, but "even if it is cited and it's wrong" is exactly what wp:NOTTRUTH is about. Unless we have reliable sources to show that the cited one is wrong, simple removal is purest wp:OR. LeadSongDog come howl! 19:06, 23 August 2011 (UTC)

I didn't think he was questioning NOTTRUTH, just saying that text has to be in accord with citations. B^e——C_{ritical__Talk} 19:19, 23 August 2011 (UTC)

Yes, I had in mind the case where a citation was given but didn't support the text; but there are many other reasons for removing text besides that one: not relevant; undue weight; source not sufficiently reliable - and yes, source got it wrong (though in the last case it might be necessary to contrive some other reason, if confronted with some wikilawyering goon who thinks we have to include information in Wikipedia even if we know it's wrong).--Kotniski (talk) 10:23, 24 August 2011 (UTC)

And as a further response to Kotniski, those of us who hang around and edit a lot of different articles are the ones responsible for the overall health of the information on WP. We can't be responsible for citing information that might or might not be reliable. That's the responsibility of the person who added the text. Why is it that we wouldn't remove information after at time simply because it hasn't passed the basic criteria for inclusion, which I would say is proof that it is not original research? In other words, it seems to me that it's an impossible task to try and source the contributions of others, because of the volume of text to be sourced and the fact that there are fewer and fewer editors. Isn't that why we have WP:BURDEN? Perhaps we need to slightly strengthen BURDEN:

"You may remove any material lacking a reliable source that directly supports it. How quickly this should happen depends on the material and the overall state of the article. Editors might object if you remove material without giving them time to provide references. It has always been good practice to try to find and cite supporting sources yourself, but it is better to remove text which is uncited than to let it remain indefinitely." B^e——C_{ritical__Talk} 21:09, 23 August 2011 (UTC)

I don't think it's reasonable to expect an editor to go searching for sources for someone else's text - I commend anyone who does. The other options mentioned above are good: 1) tag unsourced material (if questionable); 2) wait a reasonable amount of time (1 year?) for the original editor to provide a source; 3) revise (if possible) or remove the tagged text. Sometimes, because the text was interesting even though unsourced, I've moved it to the talk page with an explanatory note. I strongly disapprove of any removal of properly sourced text, even if it's "wrong". WCCasey (talk) 21:38, 23 August 2011 (UTC)

One thing to remember is that WIkipedia in the past was not as strict about sourcing as it is now, and in-line citations were even unusual. Many good articles written in those years do not meet today's sourcing standards. Gutting such articles on that basis is not to the general benefit of Wikipedia. The correct approach is to fix them, bring them to the attention of some project, tag them, or leave them alone for someone else to do one of those things. Zero^talk 09:45, 24 August 2011 (UTC)

Yes, some of the above comments about "responsibility" rather miss the point. This is a cooperative project where we're all trying to make the "encyclopedia" better. You can in practice do pretty much anything you like around here until someone objects, but I think humanity would be grateful if people's edits were directed towards improving that encyclopedia, not enforcing some half-baked rules. The fact that no citation has been given for something for a very long time is not in itself conclusive evidence that it is not good (i.e. potentially sourceable) information. It's quite destructive to the encyclopedia just to remove information at random due to the lack of citations - you ought also to have some reason to expect, based on your own knowledge, research, common sense or something, that it really is wrong or unsourceable. Particularly since once information's gone, it's gone - whereas if it's left (say with a citation-needed tag) it remains visible to other editors who might know what ought to be done with it.--Kotniski (talk) 10:23, 24 August 2011 (UTC)

I essetially agree with Zero0000 and Kotniski. There is no point in deleting unsourced content just because it is unsourced alone, as this is not is spirit of an collaborative project and it leads to a waste of good content and work of others. If you come across content that it is unsourced but looks ok otherwise, you should simply tag it. But if there are is an additional reason such as the content looks fishy, content contradicts your context knowledge, you have reason to distrust the author, the content is controversial, etc. then should delete it, but only then.--Kmhkmh (talk) 11:46, 24 August 2011 (UTC)

I think we are all saying the same thing... Sometimes it is best to remove unsourced information, and sometimes it isn't. We each draw the line between when we remove and when we retain in slightly different places. That's OK too... The line between Removal vs Retention really is a matter of editorial judgment, based on the specific situation in a specific article. I don't think we can (or should) try to draw that line as a matter of policy. Blueboar (talk) 12:12, 24 August 2011 (UTC)

That's fine with me and I agree that doesn't need to be codified in the policy. However the policy should make clear that formalistic (mindless) removal of unsourced content is not wanted nor any crusades in that manner.--Kmhkmh (talk) 12:32, 24 August 2011 (UTC)

I think we make it clear enough. And remember, going through a bunch of articles and assessing whether to remove material that has sat unsourced for a long time is not necessarily "mindless"... or a "crusade". I go into "clean up" mode from time to time... going to random articles and seeing if there are problems. As long as each individual removal/retention decision is made on a case by case basis (remaining open to the idea that sometimes it is best to leave the material in the article with the tag, and asking whether it would be better to remove or retain this specific material in this specific article), doing such "clean up" sweeps isn't wrong. Blueboar (talk) 12:51, 24 August 2011 (UTC)

I was talking about going through articles and systematically removing any unsourced content on the sole reason that it is unsourced and not about any systematic (fesired) cleanup. The discussion here is relatively clear, but it also shows that we have editors misreading the policy in the sense that is described above.--Kmhkmh (talk) 13:54, 24 August 2011 (UTC)

Yes, as I have observed before, this policy (and many others) seem to be written so as to be almost deliberately unclear. (And SOFIXIT doesn't apply, since there's a group of editors here who have no intention of allowing anyone to FIXIT.)--Kotniski (talk) 14:23, 24 August 2011 (UTC)

And yet how is the policy as you-all describe it to be reconciled with the oft-spoken-of goal of being reliable? What I'm getting from the above could be summarized as "remove only what you have reason to doubt." In such a system, it's guaranteed that vast amounts of original research of dubious reliability will remain (and could be vandalized by anyone smart enough to insert false info in a convincing way, thus raising BLP issues). It seems to me like WP is basically conflicted between the need to retain useful information and the need for reliability. But I'm not really convinced that removing only overtly suspicious material is an adequate compromise (and sourcing it one's self isn't practical).

I also think that there is a possible technical solution: have a  Removed template or marker of some sort. Then one of two things happens: 1) When an editor tries to edit the page, they are presented with a message saying that large amounts of text were removed for [reason], and they might want to review that before editing, or 2) The template simply says that text has been removed for [reason] and gives a link to the page prior to when the template first appeared.

When I get time I think I'll canvass around a little bit. This is a long-term kind of thing, and brought to a head, for me, because of statistics which say our editorship is at best probably not going to expand. B^e——C_{ritical__Talk} 14:55, 24 August 2011 (UTC)

If people ask themselves whether the material is likely to face a reasonable challenge, it's usually clear whether something needs a source, and how long to wait for it. The problem with trying to generalize is that everything will depend on the particular case. SlimVirgin ^{TALK|CONTRIBS} 15:11, 24 August 2011 (UTC)

Exactly, but that way of doing things overall means that WP has vast amounts of unreliable information. So it's a choice: do you prune the branches which have fruit on them to make the tree strong, or leave the fruit and have a tree that is less sturdy overall. WP is set up on the principle that information is guilty until proven innocent, but doesn't seem to follow through on that. It places the burden on the person who wants to remove text, and what I'm saying is: that might have been a bearable burden in the past, but it's not any more. To improve, WP needs to prune as well as refine and expand. B^e——C_{ritical__Talk} 16:36, 24 August 2011 (UTC)

Most of the time, as a reader, I'd rather see something than nothing, so long as it's not harmful or very silly. As editors we might cringe, but as readers we might be grateful for it. SlimVirgin ^{TALK|CONTRIBS} 16:38, 24 August 2011 (UTC)

Well as far as the pruning analogy goes, "mindless" pruning is not yielding a healthier tree or a better harvest, it just kills the tree.--Kmhkmh (talk) 16:49, 24 August 2011 (UTC)

I don't think you need to set the bar at "information we doubt"... I may not have any reason to doubt the unsourced information, and still remove it... if I think the information constitutes Original Research. Of course that isn't "mindless" either. Blueboar (talk) 20:11, 24 August 2011 (UTC)

well you can formulate it in a positive manner, one should not delete unsourced content one assumes to be correct.--Kmhkmh (talk) 21:37, 24 August 2011 (UTC)

Isn't "assuming information to be correct" exactly what we're not supposed to be doing? And how can anyone know what is original research minus a source? You can't tell from the level of detail. Why do we have to have a positive opinion that something is OR, why not just insufficient reason to think it isn't? And given that each editor is informed of our sourcing policy, can't we pretty much assume that anything which is blatantly lacking in sources is OR? As far as SlimVirgin saying she likes something rather than nothing, of course that's true, but is it encyclopedic? Isn't all this "benefit of the doubt" stuff actually, when it comes right down to it, basing Wikipedia content on what we like or agree with? B^e——C_{ritical__Talk} 23:56, 24 August 2011 (UTC)

You write as if policy requires everything to be sourced, but that is not and has never been the policy. It only requires sources for content "challenged or likely to be challenged" (emphasis in the original). So, no, you don't have policy support for assuming anything unsourced is OR. Zero^talk 00:22, 25 August 2011 (UTC)

Common knowledge doesn't need to be sourced, and I'm only talking about removal of material which has been challenged by way of a request for sources. Material which is not common knowledge does need to be sourced: "a source must exist even for material that is never challenged." Posted a question/request on Jimbo's talk page [5]. It seems like people here are putting the burden on the person who wants to remove text, whereas the overall gist of policy is that it's the person wanting to include text who has the burden to prove that it should remain. B^e——C_{ritical__Talk} 00:33, 25 August 2011 (UTC)

It says "must exist", not "must be given". I believe that is deliberate, otherwise it obsoletes the first part of the sentence. My main concern is that the hard work of editors who wrote articles years ago when there was no culture of sourcing everything should be accorded some respect. The culture has changed now, but we should update those articles in preference to gutting them. Only material unlikely to survive a sourcing attempt should be automatically deleted. Zero^talk 00:42, 25 August 2011 (UTC)

Yes that's a valid concern, however, I'm not aware that the culture has changed for at least 5 years, and I'm not talking about older material- my example was started in 2008 [6]. This is from an essay, but I think its statement is worth quoting here: "Responses must be forthcoming: Editors who wish to respond to the challenge [such as a tag] should do so in a timely manner. If no response is forthcoming, the challenger may tag or remove the statement in question... the challenger should await a timely response prior to removing material.[7]" And WP:OR says "The only way you can show your edit is not original research is to cite a reliable published source that contains the same material." [8] Again, I'm only talking about material which has been tagged or otherwise challenged for a good long time. B^e——C_{ritical__Talk} 01:04, 25 August 2011 (UTC)

I have noticed that you are now stating that you are "only talking about material which has been tagged or otherwise challenged for a good long time"; however, your proposal is still that "You may remove any material lacking a reliable source that directly supports it." Which view do you really take, because there is a big difference. Ryan Vesey Review me! 03:07, 25 August 2011 (UTC)

There is also something to be said regarding the reliability as a motivation for "direct sourcing". There is big danger here confuse one tool (direct sourcing/footnotes) with the goal (reliability). Having a lot of of footnotes doesn't make an article more reliable per se. For that we would need reliable editors to check/confirm the sources themselves, i.e. actually reading the the sources rather than just checking whether a source is given. And top of that we need a revision control (flagged versions) to manage which article versions have been proof read.--Kmhkmh (talk) 18:32, 25 August 2011 (UTC)

Interjection by another editor

This thread has become so hard to follow and there are many issues that need to be addressed so I am just going to do that in this section. First, there are some notability issues about high schools. While slightly disputed, it has been accepted that while all schools (elementary and middle schools) are not notable, all high schools are. To address some questions on removing unsourced information. I think this is a ridiculous idea. Information should not be removed for the sole reason that it is unsourced. Information should only be removed if there is a valid reason to challenge the information. If a citation needed tag has been added it has been challenged. The key assumption to take when viewing unsourced information is WP:AGF. Assume that all information that was added was added in good faith. Deleting this information hints at an assumption of bad faith. If there is reasonable belief that the information is false, or if it is an unverified statistic it can be deleted. It is also important to remember that the citation needed tag helps encourage editors to cite unverified information. Before removing the information, it is often useful to add the citation needed tag instead. Ryan Vesey Review me! 02:02, 25 August 2011 (UTC)

I have commented on the notability of high schools at Talk:Salmon High School. Ryan Vesey Review me! 03:04, 25 August 2011 (UTC)

I was fairly sure this was all covered by Wikipedia:Verifiability#Other_issues. Removing material without tagging first, leaving a reason=, or a talk page note should only be done under limited circumstances: obvious errors, such as "the Sun is a green gas giant"; vandalism; blp claims that may be libellous or otherwise defamatory; flammable material that might cause a volatile page to descend into chaos, "this country attacked first and committed genocides, the other country was simply defending itself when it invaded them back"; etc.

It seems that fewer and fewer people are reading MoS before starting on campaigns to "clean up Wiki" - I found one instance where an article was tagged with a citation needed and within two days the editor had removed both paragraphs, even though most of the material was fairly innocuous and easy to ref. A reasonable period of time is really related to the amount of editing the page receives - if the last edit was over a month ago and the last talk page post was three months ago, a reasonable period might well be six months. Page views need to be considered also, if such an article is getting 10 views a month it is less necessary to change it than if it is getting 1000 views a day. I also think that editors should try and source before even tagging with a cn, and least put a reason= into the template. Drive-by-tagging is becoming an issue and will only increase as Wiki approaches the point where less and less articles can be created and maintenance becomes the only way for new editors to measure their worth. In the old days it was possible to create a hundred articles in a month and do that for several months, nowadays new topics that are not already covered are much less frequent. Chaosdruid (talk) 04:55, 25 August 2011 (UTC)

it has been accepted that while all schools (elementary and middle schools) are not notable, all high schools are'. Perhaps it has been accepted, but that doesn't make it policy or a guideline--is there a policy or guideline supporting that notion. I think it is safe to say one can generally assume that most all high schools meet GNG, but if notability is challenged, sources would need to be produced to show that the school in question has some significant coverage in reliable sources. Failing that fails GNG, I think. --Nuujinn (talk) 10:47, 25 August 2011 (UTC)

That's true, but if someone challenges the info and no one is around to respond to it (for a myriad of reasons), that doesn't mean that that person can then redirect or try to delete the article without bothering to check for sources themselves. I personally feel that WP:BURDEN goes both ways. People should only be challenging something if they have proof that it is wrong, via a source, or can't find any sources for it after checking. If they're challenging something with no backing, then they should be thoroughly trouted. Silverseren^C 10:55, 25 August 2011 (UTC)

I agree with Silverseren. Unsourced material is fair game for anyone to challenge - but it should still be up to someone to challenge it. They should have some iota of suspicion that the information is actually wrong, and not just take it out solely due to the lack of a recognizable inline citation.

This is also important because when people are free to take out unsourced information with no further examination, it's very likely that they're going to start taking out information with a citation in the next or previous paragraph... or the next or previous sentence... those old articles with a bibliography at the end are right out. The only way to ensure your information will stay in (at least for this one reason) will be to cite the end of every sentence, even if they're all from the same source. Wnt (talk) 14:05, 25 August 2011 (UTC)

Notability is about article topics... Verifiability is about article content. While both are established by citing reliable sources, they are quite different concepts. Blueboar (talk) 13:18, 25 August 2011 (UTC)

Wnt, that's a very good objection. How do you differentiate between sourced and unsourced, when sources exist for some of the text (what I did was to remove entirely unsourced content)? I think that if there are sources given at the end of a section, it's usually incumbent on the person removing text to know that the source offered doesn't have the info. And I think that's a reasonably easy principle to put into policy. But, I don't think that one should have to have reason to believe information wrong before taking it out: that requires, for example, that before I took out any unsourced information on a local high school, I should basically live near or be a student there. Requiring that level of expertise isn't practical. B^e——C_{ritical__Talk} 20:40, 25 August 2011 (UTC)

I didn't say you need to have proof something is wrong; merely an "iota of suspicion". Maybe an unsourced section offers information you can't really picture having proof of, maybe it sounds like someone might have had a partisan perspective; maybe you found something else in the article that was wrong and now you're ticked; maybe the account that added it had a record of Wikitroubles; maybe you think it's wrong. But the point is, whatever reason you choose to challenge a fact, it is a reason. Siccing a bot on the task, or acting like a bot, goes beyond that, and that's a bad thing. Wnt (talk) 02:31, 26 August 2011 (UTC)

Restoration of "accuracy is an objective" from a previous version of the policy to the lead

I think that the recent attempt to restore (from a previous version of the policy) that accuracy is both an objective and also the reason for wp:ver to the lead was a good (albeit bold) idea. North8000 (talk) 11:16, 25 August 2011 (UTC)

I have not been following discussion about this but, IMO, it is a bad idea for WP to make a judgement of accuracy re assertions made by cited sources, vs. what I would say is a good idea — making a judgement that cited sources actually did make those assertions.

This has nothing to do with judging accuracy, quite the opposite. Essentially it says that accuracy is a goal of wp:ver, and that wp:ver rules. North8000 (talk) 12:03, 25 August 2011 (UTC)

Yes, it's important to remind editors of this fact. Occasionally we get editors who insist on making articles on Wikipedia-related matters contradict our server logs because a "reliable" source got it wrong, or who insist that we have to be agnostic about the existence of Santa Claus because there are so many apparently serious reliable sources that appear to take him seriously. (They just happen to appear mostly around Christmas... .) These exceptions are when we need a reminder of the purpose of the policy. It's not OK to take our policies literally for the LULZ. Hans Adler 12:08, 25 August 2011 (UTC)

I agree. Maybe the edit made was not the best way to do it, but I agree that saying accuracy is still an aim could very likely help achieve the aims of those of us who are concerned about this issue (people saying WP does not even aim at accuracy) without needing to break the now established tradition concerning the word "truth".--Andrew Lancaster (talk) 12:35, 25 August 2011 (UTC)

What do you mean by "accuracy"? The problem here is that "accuracy" can be equated with "truth". If by "accuracy" you mean "we should present information accurately", then I agree... if you mean "we should only present information that is widely held to be accurate", then I can not agree. NPOV often requires that we present theories, claims and information that the mainstream believes are inaccurate... the key is to present such information accurately (or "with accuracy"), which includes phrasing opinions as being opinions, attributing controversial ideas to those who hold them, noting that minority opinions are minority opinions, etc. Blueboar (talk) 12:42, 25 August 2011 (UTC)

We are not supposed to present fringe theories, which is where you almost heading, so we are only supposed to present things understood to be accurate, and yes we are supposed to do it accurately also. The word is quite simple and does not seem to me to require any digression about how to define fringe, notable and various other jargon about how we know when something is really understood to be accurate. That is clearly for other policies and not something where any normal reader is going to be seeing assumptions written between the lines?--Andrew Lancaster (talk) 12:50, 25 August 2011 (UTC)

Where do you get the idea that we are not supposed to present fringe theories?... We have lots of articles about fringe theories. We also have a guideline (WP:FRINGE) that is all about when and how to present fringe theories. Blueboar (talk) 13:32, 25 August 2011 (UTC)

Describing a fringe theory in context IS accurate. E.G. "The flat earth society asserts that the earth is flat" is an accurate statement. Ditto for where there is an objective answer but we don't know it (How did the Edmund Fitzgerald Sink?....report on the theories as being theories). Accuracy is irrelevant on matters of opinion, where no objective answer exists (e.g. Is Obama a good or bad President?).North8000 (talk) 14:14, 25 August 2011 (UTC)

OK, yes, and we could have a whole discussion about it. But it appears to be a different subject, and that is my point. Accuracy seems a pretty clear and neutral word for what WP is trying to achieve. The word does not seem to imply anything about where WP sits concerning notability, fringe, mainstream etc.--Andrew Lancaster (talk) 14:08, 25 August 2011 (UTC)

The two of you seem to be using "present" in two slightly different ways. The problem is: Given that the majority of reliable sources speak about Santa Claus as a real, superhuman person, how can we justify not making absurd claims such as that he is one, or that people disagree about him? The simple answer is that Wikipedia isn't a mindless game but an encyclopedia striving for accuracy, and in this case it's obvious what is accurate and what isn't. Hans Adler 14:17, 25 August 2011 (UTC)

Rather than Santa Claus, I am more concerned about the verifiability of allegations presented as fact in biographical articles like Lot (biblical person). Should the section on what counts as a reliable source explicitly mention that scriptures are not reliable sources for statements of fact? --Boson (talk) 15:16, 25 August 2011 (UTC)

As a literary character, Lot is a quite different issue compared to Santa Claus, a modern myth. But in both cases it's not necessary to use excessively guarded language once the context has been set up correctly. The problem isn't that someone with a fringe theory (Santa Claus exists) or dubious theory (the story of Lot is literally true exactly as presented in Bible and/or Quran) might not be disabused while reading the article. The problem occurs when such ideas are actively pushed by the article with explicit statements. Hans Adler 16:25, 25 August 2011 (UTC)

The answer to "What do you mean by "accuracy"? " would be the definition from any dictionary. There are numerous complexities in applying it, but we should not use those to derail the basic concept and statement of objective. North8000 (talk) 14:06, 25 August 2011 (UTC)

OK... then let's clarify the basic concept and objective... Is the objective to present information accurately ... or is it to limit the kind of information we present? Blueboar (talk) 14:22, 25 August 2011 (UTC)

I'd say simply that the objective is accuracy. Several things further that goal, especially wp:ver. Often or sometimes both of the things you mention further that goal, although both are ambiguous because the term "information" is ambiguous. Sincerely, North8000 (talk) 15:02, 25 August 2011 (UTC)

I think it is clear WP aims at accuracy, but I would not equate that aim to the policy wp:ver. This is why the edit which was reverted was not appropriate. It put this aim into this policy page in the first words, before the core of the policy itself. I think accuracy as an aim is context of the policy, not the policy itself. Is this also how others in this discussion would see it?--Andrew Lancaster (talk) 15:07, 25 August 2011 (UTC)

(added later) Agree. North8000 (talk) 20:56, 25 August 2011 (UTC)

In general, it's an important statement (implicit in the common meaning of "information" which is informative and accurate) that is homeless at the moment. Putting it back into wp:ver somewhere would also be good and appropriate for a couple of reasons. First, wp:ver is the policy that most furthers that goal. Second, it would help resolve the common mis-statements of wp:ver which have been behind much of the ongoing controversy here. But it doesn't necessarily have to be the first sentence. North8000 (talk) 15:41, 25 August 2011 (UTC)

If verifiability wasn't a decent approximation to accuracy in most cases, then it wouldn't be policy in the first place. By not making clear that accuracy is the reason why we are so obsessed about verifiability, we make the policy vulnerable to wikilawyering that tries to pervert it by making it a tool to force blatant lies into the encyclopedia. Unfortunately this is not a theoretical issue at all. Hans Adler 16:29, 25 August 2011 (UTC)

Indeed--Kmhkmh (talk) 18:22, 25 August 2011 (UTC)

An important point well said. North8000 (talk) 20:56, 25 August 2011 (UTC)

The problem is, the edit included more than just the objective of accuracy. The problem I had was this information. "We can't be sure of our accuracy if we include information which cannot be verified." I believe the information should read

The goal of Wikipedia is to become a complete and accurate encyclopedia. Therefore, 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.

Any new thoughts? Ryan Vesey Review me! 21:06, 25 August 2011 (UTC)

I like it overall. Might need a few tweaks. North8000 (talk) 21:51, 25 August 2011 (UTC)

It's not the purpose of individual policies to state the "goal of Wikipedia".

"Complete" is at odds with WP:NOT#DIRECTORY. Writing this in the opening sentence of a core policy invites people to argue that all sorts of things should be included on the basis of completeness.

"Accurate" is problematic vis a vis "verifiable". What is "accuracy" beyond "verifiability"? It invites editorial conjecture (aka WP:OR), and a place for belief.

"Therefore"? This is an attempt at a logical structure A implies B, but the logic is loose, or at least I don't see it. How does the first sentence imply the second? --SmokeyJoe (talk) 23:18, 25 August 2011 (UTC)

WP:ver clearly precludes insertion of unverifiable material based on "belief", "truth" etc. So let's stop shaking that boogeyman once and for all. North8000 (talk) 01:27, 26 August 2011 (UTC)

North8000, I understand that you're trying to work under the constraint of not changing anything in the current first sentence of WP:V. Are you OK with "accurate" in the first sentence here and "therefore ... not truth" in the second sentence here? This might appeal to some people who already understand the policy, but I think it would be confusing to those who are trying to learn about it. Regards, Bob K31416 (talk) 01:54, 26 August 2011 (UTC)

The easiest way to misread something is to start at the top and to stop as soon as you have want you want. Sticking "complete and accurate" in a stand alone sentence (it will be quoted, not even "paraphrased") ahead "verifiability, not truth" is to invite confusion, including willful confusion. Completeness and accuracy must not be encouraged at the expense of verifiability.

Some mention of complete and accurate may be useful, but not before stating the purpose of this policy.

Disagree that WP:ver clearly precludes insertion of unverifiable material based on "belief", "truth" etc. "Belief" has no accurance on the page. "truth" and "think" have exactly one occurance, in the first sentence. The clearness of the distinction between verifiability in reliable sources and the true believer's interpretation of truth is very easily muddied in the mind of that true believer (an important audience of this page). --SmokeyJoe (talk) 02:13, 26 August 2011 (UTC)

Wp:ver precludes the insertion of unverifiable material, period. If you think that it doesn't clearly say that, then we have work to do. A good place to start would be to stop confusing the issue by not attempting to list the things that don't bypass that rule. (e.g. truth)  :-) North8000 (talk) 10:02, 26 August 2011 (UTC)

Answering Bob K, personally I'm open to anything (compromise etc) that at least partially solves the problems. And a mention anywhere in this policy that accuracy is a goal, or that wp:ver is intended to help achieve accuracy would be such a step. Sincerely, North8000 (talk) 10:07, 26 August 2011 (UTC)

Wikipedia:Verifiability, not truth, an essay

Bob K31416 (talk · contribs) has twice now reverted the addition of Wikipedia:Verifiability, not truth, an essay, to WP:V.

This is not a controversial slogan, it is a respectable, informative, well written essay. It is directly relevant to explaining and understanding the opening sentence. The opening sentence, while subject to long standing challenge, is in no danger of being substantially altered. A recent proposal to alter it was overwhelmingly rejected. Proponents for change have agreed to not attempt a new proposal for change until a future date, and this is in effect demonstrates a very strong consensus for the current wording for the current time.

Given that that the opening sentence is a source of controversy and ongoing debate, it is nothing but good that an explanatory essay be included. To do otherwise is to deny the readers easy access to the explanation. If, as alleged, the sentence is misleading, then the hiding of explanation only worsens the situation.

I propose that the essay be listed, alphabetically, in the "See also" section, because it is relevant and needed advice for readers, and because it provides information on a surprising, and possibly-for-some confusing statement. --SmokeyJoe (talk) 07:51, 26 August 2011 (UTC)

I agree; while I am firmly on the side that believes "verifiability not truth" is a stupid slogan and should be abandoned in favour of something that says what we really mean, as long as we are forced to have this stupid slogan in bold in the lead, it's right that we should link people to an essay that does a pretty good job of explaining what that slogan is intended to mean.--Kotniski (talk) 08:09, 26 August 2011 (UTC)

Yes, there's no reason I can see to not link to it. --Nuujinn (talk) 08:47, 26 August 2011 (UTC)

Me too - no reason at all not to link to an essay which explains it better, and every reason to link to it. Pesky (talk …stalk!) 09:49, 26 August 2011 (UTC)

http://getsports-tv.blogspot.com/2011/08/australia-vs-new-zealand-live-stream.html

Australia vs New Zealand Live Stream Rugby Tri Nations Hot Match Live HD Video On Your Pc

RUGBY lovers and fans Welcome to watch Australia vs New Zealand live on PC. Australia vs New Zealand live streaming online with High Quality Definition (HQD) Videos in your PC without any additional hardware so don’t miss this match join here watch and enjoy live streaming online on PC

Again inform all RUGBY lovers about this game. You can easily find live match here don't waste your time Watch and enjoy all live  RUGBY Tri Nations match. Here is live streaming link this game HD online  RUGBY live broadcast. Your subscription will grant you immediate access to the most comprehensive listing of live RUGBY, National RUGBY feeds from anywhere in the world where you have access to an internet connection.

Welcome! Welcome! to Watch Australia vs New Zealand live match. Don’t miss this hot match . Australia vs New Zealand watch live streaming online on your pc. Today start Australia vs New Zealand live match it’s really very interesting and hard competition . So why you miss this interesting match. All the best watching Australia vs New Zealand. 100% trusted>>>>>

Match Details Date: Saturday, 27th August, 2011 Competition : Australia vs New Zealand Venue/Loication: Suncorp Stadium, Brisbane, AUS

Play Australia vs New Zealand Live HD Video

If you’re a Rugby fan, there’s nothing more frustrating than not being able to watch the big game, especially when that game was promised to you. With Stream Direct Sports you can forget about empty promises forever

Watching sports on the go has never been easier. Watch all the big games, match highlights and player interviews live on any computer! Access all sports all the time! Waiting in the airport or on a long train ride? Watch your favorite sport on your laptop! Download now and enjoy all the pleasures of the sporting world from the comfort of your own living room. Enjoy your sports on your time - Watch what you want, when you want! Get Instant Access to all sports channels over the Internet.

Australia vs New Zealand live online, Australia vs New Zealand live streaming online on pc, Australia vs New Zealand live with p2p all support, Australia vs New Zealand watch live, Australia vs New Zealand live telecast, Australia vs New Zealand live in justintv, watch live stream Australia vs New Zealand all online tv, Australia vs New Zealand live online, Australia vs New Zealand live sports tv, Australia vs New Zealand live free game RUGBY on online — Preceding unsigned comment added by 117.18.231.13 (talk) 10:08, 26 August 2011 (UTC)

http://getsports-tv.blogspot.com/2011/08/green-bay-packers-vs-indianapolis-colts.html

Green Bay Packers vs Indianapolis Colts Live Stream NFL Regular season HD video Online Tv On Pc 2011 VISIT HERE TO LIVE WATCH>>>>http://getsports-tv.blogspot.com

NFL lovers and fans Welcome to watch Green Bay Packers vs Indianapolis Colts live on PC. Green Bay Packers vs Indianapolis Colts live streaming online with  High Quality Definition (HQD) Videos in your PC  without any additional hardware so don’t miss this match join here watch and enjoy live streaming  online on PC
Green Bay Packers vs Indianapolis Colts live HD Video

Again inform all NFL lovers about this game. You can easily find live match here don't waste your time Watch and enjoy all live NFL match. Here is live streaming link this game HD online NFL live broadcast. Your subscription will grant you immediate access to the most comprehensive listing of live NFL, National NFL feeds from anywhere in the world where you have access to an internet connection. Welcome! Welcome! to Watch Green Bay Packers vs Indianapolis Colts live match. Don’t miss this hot match. . Green Bay Packers vs Indianapolis Colts watch live streaming online on your pc. Today start Green Bay Packers vs Indianapolis Colts live match it’s really very interesting and hard competition . So why you miss this interesting match. All the best watching Green Bay Packers vs Indianapolis Colts.100% trusted>>>>>

Match Details Date: Friday, 26th August, 2011 Competition : Green Bay Packers vs Indianapolis Colts Time: 08:00 PM

Green Bay Packers vs Indianapolis Colts live online, Green Bay Packers vs Indianapolis Colts live streaming online on pc, Green Bay Packers vs Indianapolis Colts live with p2p all support, Green Bay Packers vs Indianapolis Colts watch live, Green Bay Packers vs Indianapolis Colts live telecast, Green Bay Packers vs Indianapolis Colts live in justintv, watch live stream Green Bay Packers vs Indianapolis Colts all online tv, Green Bay Packers vs Indianapolis Colts live online, Green Bay Packers vs Indianapolis Colts live sports tv, Green Bay Packers vs Indianapolis Colts live free game NFL on online,

If you’re a sports fan, there’s nothing more frustrating than not being able to watch the big game, especially when that game was promised to you. With Stream Direct Sports you can forget about empty promises forever! SDS carries every game of the year, regardless of the sport. SDS covers it all, period. Truthfully, Stream Direct Sports is the next best thing to being at the game! Watching sports on the go has never been easier. Watch all the big games, match highlights and player interviews live on any computer! Access all sports all the time! Waiting in the airport or on a long train ride? Watch your favorite sport on your laptop! Download now and enjoy all the pleasures of the sporting world from the comfort of your own living room. Enjoy your sports on your time - Watch what you want, when you want! Get Instant Access to all sports channels over the Internet. — Preceding unsigned comment added by 117.18.231.13 (talk) 10:11, 26 August 2011 (UTC)

St. Louis Rams vs Kansas City Chiefs Live Stream NFL Regular season HD video Online Tv On Pc 2011

St. Louis Rams vs Kansas City Chiefs Live Stream NFL Regular season HD video Online Tv On Pc 2011 Click here >>> http://getsports-tv.blogspot.com

NFL lovers and fans Welcome to watch St. Louis Rams vs Kansas City Chiefs live on PC. St. Louis Rams vs Kansas City Chiefs live streaming online with  High Quality Definition (HQD) Videos in your PC  without any additional hardware so don’t miss this match join here watch and enjoy live streaming  online on PC
http://getsports-tv.blogspot.com/2011/08/st-louis-rams-vs-kansas-city-chiefs.html

Again inform all NFL lovers about this game. You can easily find live match here don't waste your time Watch and enjoy all live NFL match. Here is live streaming link this game HD online NFL live broadcast. Your subscription will grant you immediate access to the most comprehensive listing of live NFL, National NFL feeds from anywhere in the world where you have access to an internet connection.http://getsports-tv.blogspot.com/2011/08/st-louis-rams-vs-kansas-city-chiefs.html Welcome! Welcome! to Watch St. Louis Rams vs Kansas City Chiefs live match. Don’t miss this hot match. . St. Louis Rams vs Kansas City Chiefs watch live streaming online on your pc. Today start St. Louis Rams vs Kansas City Chiefs live match it’s really very interesting and hard competition . So why you miss this interesting match. All the best watching St. Louis Rams vs Kansas City Chiefs.100% trusted>>>>>

Match Details Date: Friday, 26th August, 2011 Competition : St. Louis Rams vs Kansas City Chiefs Time: 08:00 PM ET

St. Louis Rams vs Kansas City Chiefs live online, St. Louis Rams vs Kansas City Chiefs live streaming online on pc, St. Louis Rams vs Kansas City Chiefs live with p2p all support, St. Louis Rams vs Kansas City Chiefs watch live, St. Louis Rams vs Kansas City Chiefs live telecast, St. Louis Rams vs Kansas City Chiefs live in justintv, watch live stream St. Louis Rams vs Kansas City Chiefs all online tv, St. Louis Rams vs Kansas City Chiefs live online, St. Louis Rams vs Kansas City Chiefs live sports tv, St. Louis Rams vs Kansas City Chiefs live free game NFL on online,

If you’re a sports fan, there’s nothing more frustrating than not being able to watch the big game, especially when that game was promised to you. With Stream Direct Sports you can forget about empty promises forever! SDS carries every game of the year, regardless of the sport. SDS covers it all, period. Truthfully, Stream Direct Sports is the next best thing to being at the game! Watching sports on the go has never been easier. Watch all the big games, match highlights and player interviews live on any computer! Access all sports all the time! Waiting in the airport or on a long train ride? Watch your favorite sport on your laptop! Download now and enjoy all the pleasures of the sporting world from the comfort of your own living room. Enjoy your sports on your time - Watch what you want, when you want! Get Instant Access to all sports channels over the Internet.