Wikipedia talk:Verifiability

discussion following the revert (threshold again)

[continued from Wikipedia_talk:Verifiability/Archive_44#.22Threshold.22_again]

• North, there's clearly no consensus for the change you want to make, and the current wording has been widely cited for years, so please don't change it again unless the consensus becomes clear. People have to know, when they quote a policy from memory, that the material is going to be there when they come to check it. "Verifiability, not truth" sums up the core editing approach of Wikipedia, and people find it memorable and easy to understand. SlimVirgin ^{talk|contribs} 12:52, 12 January 2011 (UTC)

People, even some admins, also misunderstand "verifiability, not truth" as saying we don't care about truth at all and that we intentionally make our articles lie (e.g. about things concerning Wikipedia, as in Sam Blacketer affair) in case the "reliable sources" clearly get things wrong. I.e., these editors argue that even an exceptionally strong consensus of editors that information (in this case BLP information) is obviously false is not enough to keep it out of an article once it has appeared in an "RS" and not contradicted by others. This is not what happens in practice, but the widespread misconception has sometimes led to unnecessary disruption.

This formulation has existed for so long that it has become almost impossible to change now, and of course a consensus needs to be obtained before changing it. But it must be changed or at least annotated with an explanation that discourages the fundamentalist interpretation. Let's work for a consensus to do so. Hans Adler 13:32, 12 January 2011 (UTC)

Hello Slim and all. The change was the end result of a 3 week group discussion (above) and then in the end set out for an extra 2 days for any one who had any objections. There seems to be a double standards here. North8000 (talk) 19:36, 12 January 2011 (UTC)

It's a long, convoluted discussion, and it's by no means clear that there is any outcome at all. There can be no strong consensus unless interested editors can see at a glance what is at stake and that a consensus, with which they may or may not agree, is about to form. Without that, there will never be enough participation for a sufficiently robust consensus. Hans Adler 19:49, 12 January 2011 (UTC)

Regarding the substance of it, a part of the discussion was that this is a verifiability policy, and it says that verifiability is absolutely required, and that nothing trumps the requirement. This doesn't change policy, it actually gets rid of flaws which clouded the policy. Specifically, this gets rid of the lack of clarity in one portion of the sentence which just said one particular thing ("truth") doesn't trump verifiability, and in a way that causes it to be widely mis-misapplied to disparage the idea of accuracy (in cases where objective accuracy exists). IMHO this is a good change that should be made. Sincerely, North8000 (talk) 19:55, 12 January 2011 (UTC)

Oh, hells belles, let it go. This doesn't help clarify the policy at all, it just adds more wikilawyer fodder. I can already imagine editors trying trying to subvert NPOV by presenting a biased source and saying it's 'absolutely required' that we use verified sources. please read the points I made a couple of threads up. --Ludwigs2 20:36, 12 January 2011 (UTC)

Last thing I want to do is add to Wikilawyering! But I think that the title of this subsection confused matters.....there was no "absolutely" in what the group developed. The new wording was that of Nuujinn's 18:06 9 January post except with the punctuation change by 66.212, Which I recapped in my 17:38 10 January post.

I read your earlier post. I think that may be a very good discussion to have, but a bit off of this topic. The change did not put anything in promoting accuracy. It just reworded something which had often been mis-quoted as disparaging accuracy. Sincerely, North8000 (talk) 22:51, 12 January 2011 (UTC)

A change we agreed to in this discussion has been reverted with the edit comment, "...please gain clear consensus before changing it".

This discussion began on 25 November 2010, and over the course of three sections has had 163 comments from 21 participants.  Thereby, readers of this page are aware of this discussion and its objectives.  Five comments marked the consensus call on January 10.  More than 46 hours elapsed after the final consensus call before the change was promoted.  I think that this is a reasonable procedure to determine consensus in this context, and clearly there was a consensus with this procedure.  Thus I see no basis for the revert.  RB  66.217.118.187 (talk) 05:50, 13 January 2011 (UTC)

Jayjg, This is a continuation from the message started in the next subsection.  I think there is a difficulty here in consensus building if a VIP editor can wait until after a presumably valid consensus has been established and make a revert.  I'd like to know if you agree or disagree that the consensus process that was followed was reasonable in the context.  If not, what should have been done differently?  Thank you, RB  66.217.117.30 (talk) 05:03, 18 January 2011 (UTC)

Clearly there was a consensus with the procedure that was used.  Are there any comments as to whether it was a reasonable procedure to determine consensus in this context?  If not, what should have been done differently?  RB  66.217.117.16 (talk) 17:17, 25 January 2011 (UTC)

Further discussion on proposed wording change in lead

66 laid it out more thoroughly and completely than I did. Plus it's not even a change in policy, it just more clearly states the policy. North8000 (talk) 11:40, 13 January 2011 (UTC)

There was nobody against the final proposal, and it set out there an extra 2 days for an extra opportunity to comment. So I guess the next step to put it out there for further input//discussion, including any objections. Sincerely, North8000 (talk) 16:53, 14 January 2011 (UTC)

The issue here is that "verfiability, not truth", is a pithy and memorable phrase, which is extremely helpful in starkly outlining what Wikipedia tries to do. Yes, it can be misinterpreted, and of course we want material to be true as well, but "truth" is subjective, verifiability is objective. It might be helpful to further clarify what we mean, but removing the phrase is, I think, detrimental. Jayjg ^(talk) 01:30, 17 January 2011 (UTC)

Jayjg, Please see [Opinion_X.2C_Opinion_Y.2C_and_Opinion_Z].  Many times during the course of this conversation editors have expressed concern regarding Group B thinking.  But you will see that none of the critical opinions; Opinion X, Opinion Y, and Opinion Z; support Group B thinking.  A goal here was to remove part of the basis for the idea that as long as material is verifiable, it doesn't matter whether it is true or not true.  See Argument_from_authority.  In my opinion everyone here agrees and has agreed with your basic concern which is that we don't want to take time to consider the truth of material that is not verifiable.  I agree that we should restore the words "verifiability, not truth" in some form.  I have made a further response regarding process in the previous subsection.  RB  66.217.117.163 (talk) 04:23, 18 January 2011 (UTC)

Jaygig, it's not that it can be misterpreted, it's that it is pervasively mis-interpreted. The proposed change also sidesteps /reduces the other issue in that "truth" is an ambiguous and somewhat pejorative and straw dog substitute for the word accuracy. And also a pejorative way of characterizing efforts towards accuracy, for cases where objective accuracy exists. I say all of this because one of the common uses of the "truth" is referring to what is actually opinion and proselytizing, nothing to do with accuracy. North8000 (talk) 04:57, 20 January 2011 (UTC)

Is there any other opposition to this long-discussed potential change? North8000 (talk) 12:39, 26 January 2011 (UTC)

Not opposition, but I wonder if there's not some way of honoring the older phrasing, perhaps by referencing in a subsequent sentence that the policy has been and will likely continue to be expressed in this way. --Nuujinn (talk) 00:18, 28 January 2011 (UTC)

Where is the proposal located? I don't see it on this page. Cold someone restate it please?   Will Beback  talk  00:30, 28 January 2011 (UTC)

It was a proposal made some weeks ago to remove or change "verifiability, not truth," and/or "threshold for inclusion," and there were several objections. SlimVirgin ^{TALK|CONTRIBS} 00:52, 28 January 2011 (UTC)

With respect, it's a bit more complicated than that. But the short version of the last proposal is here.

Somehow the section got cut in half and the proposal and the arguments for it got archived, even with new material in them. It is:

replace the entire first (one sentence) paragraph with:

"The threshold for inclusion in Wikipedia is verifiability; that is, whether readers can check that material in Wikipedia has already been published by a reliable source. No other consideration, such as assertions of truth, is a substitute for verifiability."

North8000 (talk) 23:43, 30 January 2011 (UTC)

In response to Jayjg's well founded concern, I might suggest this alternative:

"The threshold for inclusion in Wikipedia is verifiability; that is, whether readers can check that material in Wikipedia has already been published by a reliable source. This policy has traditionally been expressed with the phrase "verfiability, not truth," which emphasizes that no other consideration, such as assertions of truth, is a substitute for verifiability."

Just a tupence, --Nuujinn (talk) 00:23, 31 January 2011 (UTC)

Well, that could be a compromise. Recapping a couple of points from the quickly archived discussions, the currently one is widely misquoted as disparaging the idea of accuracy. Also, the current one is structurally bad because it just gives one particular example of a thing that isn't a substitute for verifiability. Sort of implies and 'either or" situation, and weakens the verifiability by seemingly leaving out the other things that aren't a substitute. North8000 (talk) 01:48, 31 January 2011 (UTC)

Yes, I agree that the current version emphasizes the one opposition, and there are others. But history is important as well, and the current version has served us well. --Nuujinn (talk) 01:51, 31 January 2011 (UTC)

Coming to this discussion late, but when I see people arguing that a particular wording "has served us well" or some such, I can normally be pretty sure that they've run out of genuine arguments. I would certainly like to drop or amend the misleading "verifiability, not truth" slogan - we presumably do want truth as an ideal, but the only practical way we know of of ensuring that we approximate truth is by insisting on verifiable statements. And that's actually "verifiability" in a very specific sense, too. What the slogan actually means, I guess, is "truth according to reliable sources, not truth according to the declarations of editors". --Kotniski (talk) 07:38, 31 January 2011 (UTC)

The truth (verifiability) the whole truth (NPOV) and nothing but the truth (NOR) -- PBS (talk) 10:04, 31 January 2011 (UTC)

I like that, PBS, be a nice start to an essay. Kotniski, would you like a nice chablis with your snark? --Nuujinn (talk) 11:42, 31 January 2011 (UTC)

As a sidebar, "truth" is sort of a slightly pejorative straw man substitute word for the word "accuracy". I say this because the word "truth" often refers to proselytizing, faith-based statements, opinions or worse. Objective accuracy does exist when there is 99%+ acceptance for the framework of the statement. So it does exist for the statement "tallest mountain in the world", and it doesn't exist for "the greatest football player of all time". And information is a part of the Wikimedia objectives. Objectively wrong "information" is not information. North8000 (talk) 12:06, 31 January 2011 (UTC)

Proposed solution to this interminable debate

This argument has been going on forever, and shows no real likelihood of stopping anytime soon. So, maybe it's time we just boldly rewrote the lead entirely, and to hell with history. We can add a 'legacy' section below if anyone really wants it. Proposed rewrite (getting rid of all the things that people squabble over, hopefully):

The aim of of Wikipedia is to give neutral, accurate descriptions of the subjects it covers. To ensure that the information presented is accurate, every substantive claim made in Wikipedia articles should be capable of being verified in reliable published sources. Verifiability is an exclusionary principle: Not every claim needs to be explicitly verified, and being verifiable does not ensure that a claim will be used on a given article, but claims that are challenged by other editors and cannot be traced back to published sources should generally be removed.

A claim is verified when it can be found used in a source with the same basic sense and context as its intended use in the Wikipedia article. The claim does not need to be a literal repetition of what is said in the source (and care should be taken to avoid plagiarism), but should be a proper summary consistent with the source's intent and purpose. It does not matter whether what the source says is correct or incorrect (true or false) in the greater world, so long as what the source conveys is a significant aspect of a neutral description of the topic.

Once they have passed verification, different claims may require different kinds of attribution (from uncontroversial claims which require no attribution to claims which need to be directly attributed to a specific source or author) in order to avoid over- or under-representing their significance to the topic.

Verifiability is one of Wikipedia's core content policies, along with No original research and Neutral point of view. These policies jointly determine the type and quality of material that is acceptable in articles. They should not be interpreted in isolation from one another, and editors should familiarize themselves with the key points of all three. All core policies may be applied more stringently on material relating to living persons.

Have at it! --Ludwigs2 18:11, 31 January 2011 (UTC)

You fixed several issues in one swoop! Something that huge and fast inevitably has a few imperfections. Not sure I understand the distinction between verificaton and attribution. Also, I'd rather not diss objective accuracy (for those cases where such exists). North8000 (talk) 22:47, 31 January 2011 (UTC)

Definitely way to go. But not exactly this, though - this lead needs to be written in nice, ordinary-person-friendly language - the worst thing it could do is put people off joining the project by making them think Wikipedia is terribly complicated. The word "exclusionary" here loses the audience as early as the third sentence. And no references to "claims"!! That misleads as to what frame of mind we expect people to come to Wikipedia in. --Kotniski (talk) 22:59, 31 January 2011 (UTC)

@ kotinski: yes, sometimes I am ordinary-person-friendly-language-challenged, I freely admit it. I'm ok with a rewording, but those two point do need to be covered somehow

• the 'exclusionary' bit is needed in some form to kibosh the classic POV-advocate "I have a source, so we have to include this" thing. we need to point out that verification is a test for whether something can be used, not whether it should be used.
• the 'claims' thing is really important, though maybe not the best language. the thing to get across here is we as editors make statements about a topic in the encyclopedia, and it's those statements that we make that need to be verifiable in sources. A lot of editors forget that wikipedia itself always has a voice - whatever is written in any article is what Wikipedia says about a topic - and the struggle here is to ensure that wikipedia's voice (the claims that wikipedia makes about a topic) accurately reflects the voices of sources.

@ North8000: the way I see it (based on debates I've had here over the issue):

• verification is the simple act of making sure that something written in wikipedia accurately reflects what's been said in the real world. For instance, if some editor edited the GH Bush article to claim that the sign behind W said "Mission accomplished, suckers", that would not be verifiable, while the same phrase without the 'suckers' part would be.
• attribution is how we give credit for a particular statement, which is more in the NPOV balance direction. generally there are four levels of attribution:
  • no attribution, implying broad acceptance of a claim (the standard "Mars is a planet" thing)
  • group attribution, implying broad acceptance of a claim within a particular POV ("Hindus see Brahman as the highest spiritual entity")
  • specific attribution using a footnote, implying a claim given by a particularistic source who would largely be considered qualified and unbiased ("The Obama administration is embarking on a new policy of..." <with citation to the New York Times>)
  • specific inline attribution, implying a claim that is questionable, localized to a particular (usually minority) POV, polemic, or otherwise not to be taken as a generally accepted statement ("Linus Pauling argued that large doses of Vitamin C could effective cure a broad range of diseases")

Also, I'll confess I have difficulty with the concept of 'objective accuracy'. Too much Philosophy of Science in my brain... Plus, I think Wikipedia needs to make a bright-line rule blocking concerns with ontological truth no matter what language it appears in. Science itself might aim for something like 'objective accuracy' (as I've said elsewhere, one of the premises of modern science is the assertion is that one can make an induction from collective subjective experience to ontological facticity - yeah, I know, everyone just had a brain freeze; sorry), but wikipedia shouldn't be trying to evaluate science on that level. We should keep it on the level of describing topics neutrally, and leave it up to sources to worry about whether the topics are understood correctly. --Ludwigs2 18:23, 1 February 2011 (UTC)

Hello Ludwigs2 I think that the "the more questionable the claim, the stronger sourcing needed (and vica versa) fills an important hole in the policy. But it is confusing that you seem to be switching between implying that verification and attribution are two different things, and then that they are the same.

I wasn't trying to push the idea of objective accuracy (even in those cases where it exists). I was just saying let's not actively diss it. Sincerely, North8000 (talk) 19:29, 1 February 2011 (UTC)

Ah, I see what you mean about accuracy - I misunderstood you before, sorry.

With respect to the other: verification and attribution are two different things, but they are related. verification asks the question "is what we say an accurate reflection of what's said in the real world?" Attribution asks the question "How careful do we have to be making this statement in Wikipedia's voice". Obviously, one can verify many statements from many sources that should not be stated directly in wikipedia's voice (e.g. statements of White Supremacist organizations about minority groups); Obviously, one can make many common sense statements in wikipedia's voice that would be difficult or pointless to verify from specific sources (e.g. statements about who the president of the US was in a given year). First we need to verify a statement (if only to say that it's too generic to need specific verification); once verified, we need to determine the best way to psent it in the encyclopedia so as not to misrepresent it. --Ludwigs2 20:00, 1 February 2011 (UTC)

A lot of brilliant ideas there that would be big fundamental changes. I'm almost thinking that we should still deal with the tiny proposed change (and which is not a policy change) prior to your proposal and then move on to yours? North8000 (talk) 13:03, 3 February 2011 (UTC)

well, North, as a dyed-in-the-wool tinkerer I have a strong preference for patching things up - duct tape is my bestest friend. But even I recognize that there comes a point where you just have to stop nursing something along and go out and get a new one. People have been wrestling over micro-changes to this policy for years now, with absolutely no progress whatsoever. Whether or not your tiny proposed change goes through, this same argument (IMO) is going to start up again in a week, as though nothing had changed. and that's because - honestly - nothing will have changed.

I'm halfway tempted to edit this in right now with the edit summary "BOLD change to long-standing policy because I'm tired of people squabbling endlessly over petty crap", because that's pretty much how I feel about it. --Ludwigs2 16:44, 3 February 2011 (UTC)

I like many aspects of it. Maybe a few suggestions:

• "Exclusionary" is not a common word, but it does have an explanation. But given that it gets explained anyway, do we need the word itself?
• "over- or under-representing their significance to the topic" . Maybe "...as it is presented in reliable sources relevant to the topic" or something like that? My point being that there is perhaps a hole here concerning how what defines the significance.--Andrew Lancaster (talk) 17:10, 3 February 2011 (UTC)

Ludwig2, you do brilliant work. But I think that this contains a fundamental re-definition of the word verifiability/verification. I still am not clear on what that new definition is, and not sure that you have fully thought it through. Your first definition of verification sort of presumes that a agreed definitive source exists (I think an oversimplificaiton in a major area) and that verification means checking against it. Then your attribution definition introduces the concept that I have been promoting (the more questioned/questioned the statement, the stronger the sourcing required. and vice versa. But thiis deos not define attribution. Your only real definition of attribution seemed to related to giving credit. North8000 (talk) 16:42, 6 February 2011 (UTC)

Back to the smaller / longer-discussed proposal

Ludwig2, again you do brilliant work, but I would like to still deal with the much smaller proposal of the discussions which, recapping, is:

replace the entire first (one sentence) paragraph with:

"The threshold for inclusion in Wikipedia is verifiability; that is, whether readers can check that material in Wikipedia has already been published by a reliable source. No other consideration, such as assertions of truth, is a substitute for verifiability."

Sincerely, North8000 (talk) 14:47, 9 February 2011 (UTC)

well, ok, but the problem with that particular line is that it doesn't make it clear what relation people should have to truth. I can see people following this kind of logic: "sure, assertions of truth are not a substitute for verifiability. but assertions of truth are 'right' and so therefore have their own value and place on project." We have to get something in here that dispels the idea that the project is aiming to express truth, but instead uses verifiability to ensure accuracy. do you see what I mean? --Ludwigs2 17:50, 9 February 2011 (UTC)

I'm with you 100%. (actually 1000%). Right now this is only in the higher level Wikimedia statements, and wp:ver is merely a means to that end. WP:ver CAN semi-successfully be and is myopic of the bigger picture. You are proposing to change / fix that 100% which is huge. What I'm promoting only fixes that 1%, that 1% went a long time in discussion here without dissent, and even then got reverted when I put it in. Gotta start somewhere. North8000 (talk) 18:25, 9 February 2011 (UTC)

As a thought: Accuracy on Wikipedia is objective and is determined per sources, while truth is subjective. When using sources to illustrate truth one is actually creating OR because the editor and his opinions are manipulating the sources to an end "statement". I think one of the hardest things for a new editor is to understand that an encyclopedia is not a research paper, but is a compilation of information. Information by definition has to be found somewhere rather than newly created. Weight is an incredibly important aspect of Wikipedia since it helps determine the the importance/significance of the sources. By determining their weight to the whole subject area accuracy of the article is determined. As an aside editors quite often use weight when it suits them and ignore it when it doesn't possibly because its a kind of second tier. Threshold for inclusion doesn't mean everything can be included just that the basic standard has been met then weight comes into play. Probably what is needed is a chart to set this down... a picture is worth a thousand words...yadyadyada..:o)

Actually the Policy as it is worded now does not define verifiability and its connection to what an encyclopedia is. It starts right in with verifiable as a descriptor of something else - sources.

An encyclopedia is a compilation of information that by definition has been published elsewhere. The threshold for inclusion of that information (sources) in Wikipedia is verifiability; that is, whether readers can check that material in Wikipedia has already been published by a reliable source. An assertion of truth, a subjective consideration, is not a substitute for verifiability.

Just brainstorming with this version. Not attached in anyway.(olive (talk) 19:07, 9 February 2011 (UTC))

Overall I think that yours is better. But structurally I think we need to get in there that no other consideration is a substitute. North8000 (talk) 13:09, 10 February 2011 (UTC)

Making charts... you must be an academic - lol. The difficulty with this is that saying "An assertion of truth, a subjective consideration" is likely to confuse vast ranges of people who don't see truth as subjective at all. The whole 'truth' thing is such an ugly red herring... how about replacing the last line with something like "Wikipedia editors should not write articles from the perspective of what they know to be true, no matter how well founded their knowledge is, but should restrict themselves to information that can be found by anyone." --Ludwigs2 03:20, 11 February 2011 (UTC)

I agree. The word truth shouldn't be used because of its diverse relationship to multiple philosophical positions/theories/meanings... very red herring ish. Your wording is an improvement over mine... I'm not sure about "know to be true" another red herring, maybe "think to be true" would be better, and "found by anyone" reads well, but does't take into account sources that are only available to some editors. Maybe that's not a concern.(olive (talk) 17:27, 11 February 2011 (UTC))

I like the first sentence as-is — with the bolding for emphasis, and with the placement to stand alone as a first paragraph. Clarifications are best left for later. Wishy-washiness and weaseling should be left out.

Also, saying that editors should restrict themselves to information that is easily found would be to back off of longstanding policy. Ease of verifiability is a plus, evan a big plus, but is only a plus. I say this as an editor who lives on and is writing this from a small island having no libraries located in a country having few libraries; "verifiable" for me is restricted to what I have on my bookshelf or can access online. Having said that, I'll also say that ease of verifiability is not and should not be a requirement. Wtmitchell (talk) (earlier Boracay Bill) 03:27, 13 February 2011 (UTC)

I agree with your thoughts on "easily found", but I believe that is not an issue because I don't think that it is in any of the proposals. Sincerely, North8000 (talk) 19:25, 13 February 2011 (UTC)

Dataspaces and OR

Recent policy announcements by NIH and NSF require funded research workers to put results in sustainable dataspaces. See [1], [2]. (This is not mentioned in Dataspaces.

Will WP consider information in the DataSpace of a major research university self-publication or publication by the university?

In conjunction with the copyright policy

See also previous section now archived Wikipedia talk:Verifiability/Archive 45#Copyright in the lead

Given the outcome of this Wikipedia:Village pump (policy)/Archive 83#RfC: What is the status of WP:COPYVIO? RfC. I am adding the phrase:

", in conjunction with the copyright policy,"

to the sentence in the lead:

"These policies jointly determine the type and quality of material that is acceptable in articles."

The new sentence will read:

These policies, in conjunction with the copyright policy, jointly determine the type and quality of material that is acceptable in articles.

--PBS (talk) 20:29, 15 February 2011 (UTC)

To be clear on the "outcome" of that discussion, it appears that one editor agreed with your suggestion and another disagreed. Are there other opinions on this? Location (talk) 21:04, 15 February 2011 (UTC)

I make it two agreeing and one disagreeing, and the one saying he disagreed didn't in fact address the specific issue of the addition of these words. (I can add my agreement as well.)--Kotniski (talk) 07:53, 16 February 2011 (UTC)

The place to discuss changes to the policy is here, and I don't think anyone would argue that Wikipedia:Copyright is a core content policy. And it certainly doesn't determine the type and quality of material that's acceptable in articles. SlimVirgin ^{TALK|CONTRIBS} 08:07, 16 February 2011 (UTC)

According to the first few comments on the RfC it was move there from here because some people objected to it being here. It is also advertised in the last comment on the section above that you just copied here. -- PBS (talk) 21:25, 16 February 2011 (UTC)

The wording does not say it is a core content policy (it says in conjunction with the content policies) and copyright does prohibit certain types of content and force us to use to use certain constructions when using some sources. I would have thought you would have have agreed with that as you have been campaigning for in-line attribution of closely paraphrased text from copyright sources. Such usage is a direct result of copyright prohibitions. --PBS (talk) 21:25, 16 February 2011 (UTC)

I would say it certainly is a core (i.e. very important) policy, and it certainly relates to article content, so it doesn't seem unreasonable to argue (as some have done) that it's a "core content" policy. Perhaps it doesn't determine the "type and quality..." exactly, but then that's not a very good description of what V/OR/NPOV do, either (and there are many other policies, guidelines and customs that do do that). Perhaps we could just drop all this peacock language and make a straight statement that these three particular policies overlap quite a bit (and then set to work trying to see if we really need three separate pages and what the scope of each page ought to be).--Kotniski (talk) 08:33, 16 February 2011 (UTC)

You've posted a contradiction in your reasoning. You are in favor of a wording change that states WP:COPY does "determine the type and quality of material that is acceptable in articles", but then you state that "[WP:COPY] doesn't determine the 'type and quality...' exactly". Location (talk) 17:47, 16 February 2011 (UTC)

As one possible alternative (albeit without any substantial change in wording), move the reference to WP:COPY up to the section "Reliable sources and other principles" like this. By being in that section, it would better stress the importance of WP:COPY its relationship to the others. Location (talk) 18:12, 16 February 2011 (UTC)

The problem is the peacock sentence about V, OR and NPOV, rather than whether COPY should be added to it. There are multiple factors (mostly documented in various policies and guidelines, though not specifically these three) that "determine the type and quality of material acceptable in articles". If we replaced that sentence with a simple factual pointer to the reader, saying that issues very similar to the ones discussed in this policy are also discussed in OR and NPOV, then I don't think anyone could find anything to object to in it. --Kotniski (talk) 20:24, 16 February 2011 (UTC)

Location Would you argue that copyright does not in anyway affect content, if so why do we need the section you refer to? The problem with that section is that it by no means covers what it tries to describe (see Wikipedia_talk:Verifiability/Archive_44#paraphrase). If we include mention of copyright in the lead as I have done, then that section becomes redundant and it can be removed as the nuances of it can be covered in the polices and guidelines covering copyright and plagiarism. If copyright is not mentioned in the lead then we need to expand the section you mention to include all the exceptions, so that it does not contradict copyright policy and the plagiarism guideline. By putting in the phrase I have suggested, or a similar one, in the lead we can simplify the wording in this policy. -- PBS (talk) 21:25, 16 February 2011 (UTC)

I am not definitively opposed to mention of copyright someplace within the lede; however, the current proposal is not accurate since copyright affects the expression of content/material, not the type and quality of content/material that is acceptable. Location (talk) 23:08, 16 February 2011 (UTC)

Is that also not true of this policy also? The content of the article is affected by copyright or once could save a lot of time by just copying modern British biographies from the ODNB (which would change the type and in some cases the quality of the content of our articles). -- PBS (talk) 10:05, 17 February 2011 (UTC)

Self-published sources

A recent discussion at the Aviation WikiProject, renewed concerns I had about the policy on Self-published sources. Let me focus on the spaceflight example of Encyclopedia Astronautica, which is a website made by a spaceflight enthusiast Mark Wade. According to this policy, it appears it shouldn't be used as a reliable source because Mark Wade hasn't been published in a reliable third-party publication on the topic of spaceflight (as far as I know).

The concerning thing is that there are plenty of sources which do count as reliable (see this Google books search) which cite this website for facts. So a published reliable source makes a statement, for which it cites a source which is not deemed to by "reliable". Is Wikipedia allowed to use this statement?

It's not relevant to my question, but in this case, the website Encyclopedia Astronautica is "reliable" in any normal English usage of the word (so much so that NASA points others to it for facts), but it's not "reliable" in Wikipedia jargon. Mlm42 (talk) 03:57, 17 February 2011 (UTC)

The question keeps coming up, and I don't know the answer. It has happened a number of times at RSN that similar sources have been deemed reliable for limited purposes, although some don't credit those discussions. Some point out that WP:V says "whose work in the relevant field has previously been published", not "who has previously been published" ... that is, the claim is made that V doesn't require that the guy with the website get his name on a book jacket, only that his work gets reproduced in other people's books. - Dank (push to talk) 04:28, 17 February 2011 (UTC)

Oh.. interesting point.. Mlm42 (talk) 04:36, 17 February 2011 (UTC)

I think a wording tweak will eventually be needed. This is shown by the fact that if you ask people whether someone cited as an expert in expert reliable sources is an expert, everyone seems to think this question is almost not worth asking. (They do not ask in which type of media the expert's frequently cited work has been published.) The problem is that even though most such experts have published somewhere where other experts publish, some have not. For example they might have a widely cited website of their own. Using the wording of WP:SPS to exclude such sources which are not excluded by experts in the fields we are trying to describe just seems like obvious wikilawyering and loophole seeking and we should try to close that gap in the logic. --Andrew Lancaster (talk) 10:13, 17 February 2011 (UTC)

The SPS guidelines are pretty clear that there are two separate criteria that must be met before a SPS can be used: (i) the person is recognized as an expert in the field; and (ii) their work in that field has been independently published in reliable sources. The first criteria can be met by examples of the person's work being cited in other sources; citation by third parties does not strike me as satisfying the second requirement of independent publication. To argue otherwise essentially eliminates that requirement. That being said, I can see an argument for a change in the policy such that, if we have multiple independent reliable sources which recognize the person as an expert, that their self-published works can be cited notwithstanding that they have never been previously published by an independent reliable source. I can see arguments pro and con for such a change in policy, but I would view it as precisely that - a change in policy, and not a mere clarification of existing policy. Fladrif (talk) 21:10, 17 February 2011 (UTC)

For better or worse I started a new subject line below on the same subject, as I felt this one was more about a specific matter. It would be good to collect some opinions on this. To remind: my basic point is that WP:SPS does not cover all the cases that WP:RS's basic wording implies it should cover. It does cover most and it must have seemed very reasonable when written.--Andrew Lancaster (talk) 10:13, 18 February 2011 (UTC)

Challenged or likely to be challenged

FYI - Wikipedia:Village_pump_(policy)#Challenged_or_likely_to_be_challenged, JeepdaySock (AKA, Jeepday) 11:59, 17 February 2011 (UTC)

SPS implied definition of an expert

I think the SPS implied definition of an expert needs to be refined. I've just posted on RSN about this, here. This is one of those questions which keeps coming up, and logically that strongly implies that the current text is NOT a good consensus. So please let's not have too many responses which try to bin the concern by saying "this has come up many times and we've never changed it so we should never change it". Here is what I wrote on RSN...--Andrew Lancaster (talk) 16:41, 17 February 2011 (UTC)

I think it is clear that being cited by experts as an expert can be considered as evidence that an author is an expert. If someone's personal webpage is widely cited in expert publications, then I think that also means that website has a reputation for fact checking. I've never seen an argument to the contrary based on either normal word meanings, or the spirit of Wikipedia policies and guidelines and norms. I have OTOH seen the argument now presented here, which is based on taking a literal reading of one bit of wording on the current WP:SPS page to be the last word on this in Wikipedia. But where there is an obvious difference between the intention of our policies overall and one bit of wording, then that one bit of wording looses out (and eventually should be changed). The wording on WP:SPS which defines an expert in terms of publications probably covers 99.9% of cases, but we keep seeing cases where it excludes people who are obviously considered experts outside of Wikipedia, but who for whatever reason do not publish in an orthodox way. And here on Wikipedia we are not supposed to over-rule outside experts, which is what we would be doing if we decide here that some widely experts are more equal than others. Wikipedia can also not consistently define an expert in any way which conflicts with WP:RS which makes it clear that the most important thing is a reputation of fact checking, and if there are a small number of cases where the current wording of WP:SPS is in conflict with that, then now that we've seen it a few times those words need to be fixed. Until that time there is an inconsistency, and WP:IAR can and should be invoked.

Per the discussion at RSN, I disagree (no need to repeat it all here). SPS covers this situation fine with its present wording; there is no need to weaken it to include hobby sites from editors who are not published in the relevant field, when more reliable sources are available and the hobby site contains errors. SandyGeorgia (Talk) 17:12, 17 February 2011 (UTC)

SandyGeorgia, with all due respect I see nothing in your posting on RSN or here which is not entirely being shaped by your position on the specific case there. Let's say that you are right in that case (I have explained there how that might in fact be the case) but let's consider whether SPS covers all cases of experts properly. Do you think it is impossible for an expert in a field to be widely cited as an expert but only published in places which might raise question marks if that person was not an expert? For example, let's say that a particular website is very widely cited by academics, but maintained by volunteers. --Andrew Lancaster (talk) 17:39, 17 February 2011 (UTC)

I posted a counterexample over there. How about if we let the multiple discussion on multiple pages wrap up before we attempt to change a core policy page? We have some "opinions" from various editors that Baugher is an "expert", we have many questions about his expertise that are covered now by SPS, and that discussion seems to have become very personal to his supporters, but the issues here extend beyond one website and one WikiProject. SandyGeorgia (Talk) 17:51, 17 February 2011 (UTC)

Exactly, so why do you keep framing it in terms about Baugher?--Andrew Lancaster (talk) 18:33, 17 February 2011 (UTC)

Have you another example? SandyGeorgia (Talk) 18:36, 17 February 2011 (UTC)

The same basic problem comes up on RSN every now and then, indeed. I am not sure it helps to focus on any specific examples. I think the most typical format of any such case is going to be widely cited website, but not obviously connected to anyone publishing elsewhere. Imagine a volunteer group collecting bird sightings, or counting car accident in a city. Of course if such groups are not taken seriously anywhere, we can't use them either, but what if they are cited quite often in specialist literature?

One case I am familiar with is human population genetics. Population genetics is obviously basically an academic field, but as it happens there are at least two non-academic websites which are cited in most articles about human Y DNA types. The background reason is that Y DNA happens to also be used by genealogists these days and genealogists are of course a very varied group of people, often retired, and often well educated. And they've become a part of the landscape in terms of the information resources the academics use in their articles.

• One of the two widely cited websites is http://www.isogg.org/tree/index.html which is basically keeping record of Y haplogroup discoveries made elsewhere, and is run by an organization of genetic genealogists. It is cited sometimes also on Wikipedia, and I think this is uncontroversial when it is used as intended, the same way academics do.
• Another is http://www.hprg.com/hapest5/ which is a webpage by Whit Athey that has a tool that is extremely frequently cited as the standard tool academics use to predict SNP haplogroups based on STR haplotypes. This website is not so relevant to WP because it is a tool used by researchers, and WP should not contain original research. (I think it is cited as a useful tool on genetic genealogy though.)--Andrew Lancaster (talk) 19:20, 17 February 2011 (UTC)

isn't one solution to this to cite an academic source that cites the web source WP:SAYWHEREYOUGOTIT. That way the citation to your "two non-academic websites which are cited in most articles about human Y DNA types" can be justified because you have cited an academic source and it is the judgement of the expert academic whether the content of the "non-academic website" is usable, rather than the judgement of a Wikipedia editor. -- PBS (talk) 12:02, 18 February 2011 (UTC)

No, that would only give a solution in cases where the exact information you want to cite has been cited elsewhere. The question is which is the more basic rule in WP:RS - reputation for fact checking, or is WP:SPS so unquestionable that it stops us using sources which are considered to be at the highest level of expertise by the highest level of experts?

(BTW according to me your proposed method of basically pretending your real source is not your real source, which is a typical method used for challenged sources on Wikipedia, is plagiarism and arguably a copyvio, so it is a problem with current WP habits. Just because someone's SPS is not considered an RS on WP does not mean it has no right to be named as the real source? Should SPS's at least be named sometimes as places where better sources are cited? But this is another subject.)--Andrew Lancaster (talk) 12:35, 18 February 2011 (UTC)

I do think examples are helpful-- in this case, I would be concerned about using a lay organization to report medical info, when MEDRS requires very strong sourcing. I may be missing your point? SandyGeorgia (Talk) 18:37, 18 February 2011 (UTC)

I am not quite sure I follow. Y haplogroup phylogenies, if you are specifically homing in on the examples you asked for, have no medical significance. But even if they did, or if we construct examples where there are medical implications, the same question arises as to why Wikipedia would have its own legalistic definition of an expert which censored mention of a source that all the highest experts use, all due to a historical choice of words and a growing traditionalism. If you saying that any wording change to WP:SPS would have a special effect on medical articles what is it? And can it be fixed by giving medical articles a special mention? But then do we need to put special mentions on every section of WP:V? --Andrew Lancaster (talk) 21:07, 18 February 2011 (UTC)

It affects politics as well. Suppose that an academic at Queens University Belfast issue a paper about the troubles in Belfast and in it she cites the opinion of a member of a previously proscribed organisation using a document published on their website. Does that give a cart blanch to use that web site as a reliable source? I think not. To cite the fact that it was mentioned in a paper put out by Queens is very different from citing it directly, let alone all the other material published on the website. Without a doubt the contributes to such a the website will be experts on the troubles (having caused a substantial percentage of the havoc), but that does not mean we should be using such websites as reliable source just because it has been cited a few times in sources we consider relaible. -- PBS (talk) 08:23, 19 February 2011 (UTC)

I think this example is different from the ones I gave, where the sources in question are not just treated as primary sources, but actually treated as expert sources, by experts.--Andrew Lancaster (talk) 09:11, 19 February 2011 (UTC)

How is WP:SAYWHEREYOUGOTIT "basically pretending your real source is not your real source, which is a typical method used for challenged sources on Wikipedia, is plagiarism and arguably a copyvio, so it is a problem with current WP habits"? In the example I gave no one is pretending. How is WP:SAYWHEREYOUGOTIT plagiarism? How is it a copyvio? -- PBS (talk)

That is not what I intended. What I wrote was, I would say, in defense of WP:SAYWHEREYOUGOTIT, which I fear is being consistently breached on Wikipedia. And remarkably it is a case where the advice one sees from experienced editors is to break it. If you watch RSN you will advice given routinely to pretend you got information from a different source than you really got it from. But this is a side subject. Please start a new topic if you wish to discuss this?--Andrew Lancaster (talk) 09:11, 19 February 2011 (UTC)

@Andrew Lancaster, well, this got confusing because of the SAYWHEREYOUGOTIT issue, but back on your question about medical sources, no, we wouldn't need separate wording for SPS in that area, because we have WP:MEDRS already as a guideline for medical information. As I read it, I thought your example wouldn't be affected by any change to SPS because under MEDRS, we'd typically reject self-published sources anyway in favor of secondary journal reviews. If you're saying that site isn't used for medical info, my post was just an intrusion, but I read it as a medical issue related to genetics. SandyGeorgia (Talk) 08:53, 19 February 2011 (UTC)

OK. I think I understand.--Andrew Lancaster (talk) 09:11, 19 February 2011 (UTC)

The current implied definition in WP:SPS is:-

CURRENT TEXT. Self-published expert sources may be considered reliable when produced by an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications.

This is followed by two sentences, one in a separate paragraph emphasizing caution, which no one is proposing to water down or remove. But compare to WP:RS, which I understand to be the guiding principle behind this sentence (which has a mirror there also):-

FROM WP:RS. Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both. [...] Articles should be based on reliable, third-party, published sources with a reputation for fact-checking and accuracy. [...] Questionable sources are those with a poor reputation for checking the facts, or with no editorial oversight.

I believe that WP:SPS should be tweaked so as to allow for other ways of having a reputation for expertize, than just having a publishing history apart from one's self-published works. Such cases might be rare but there seems no logical reason to block Wikipedians from using them. It would not be consistent with our basic principles in WP:RS and other policies. My proposal also would make clear that the implied definition of an expert is meant to be one. (The wording is actually not very clear now.):-

PROPOSAL. Self-published sources may be considered reliable when produced by an established expert on the topic being cited. An author or publication is an established expert source if it can be shown to be treated as one by clearly identified experts in a field. Concerning authors, common indications of expertise are citations as expert sources in reliable third-party publications, and having had work published by reliable third-party publications.

Comments please?--Andrew Lancaster (talk) 09:30, 19 February 2011 (UTC)

Case in point - Edward Tufte, whose work is almost entirely self-published because he's very particular about the technical side of the publication process. As the policy stands he barely scrapes into the acceptable SPS definition because of occasional third-party publications e.g. [3]. But I tend to agree that even without those third-party publications, he ought to qualify as a RS because of his reputation in the field.

However, I'm not entirely comfortable with the "it's an expert source if other experts treat it as one" definition - I think that opens us up to "well, A is published so he's an expert, and he vouches for B, and B vouches for C..." arguments. --GenericBob (talk) 10:29, 19 February 2011 (UTC)

• I find agreement with you, I think, that calling someone an expert for "indirect" reason should require strong indirect reasons. In one of the examples above I cited a website which is cited in the methods sections of nearly every academic article on a subject in the years since it was set-up. The author of the website has published, but just not in the highest academic journals, except a letter or two. So it is a case of saying that citations are a type of evidence which can be called upon to make a case that someone is an expert, but not maybe the best. If you look at where these sentences would fit into WP:SPS though, they are surrounded by cautions, don't you think?
• Coming to your second point, I wonder if Wikipedia can ever go beyond saying that things are expertly written because experts say they are. Isn't this just another version of our normal approach to WP:TRUTH? This approach is often considered controversial and counter-intuitive, but it seems to work?--Andrew Lancaster (talk) 11:03, 19 February 2011 (UTC)

There's always going to be some level of circularity, yeah. I'd just like to see the self-reference made a bit less explicit (to deter abuse) and maybe the wording tightened a little so it requires a consensus among experts in the field (rather than just "well, I found one guy who'll vouch for him"). If I understand correctly, that's pretty much what you're already aiming for, just that I don't think the wording quite matches that intent. --GenericBob (talk) 00:22, 20 February 2011 (UTC)

I do think we are agreeing on the general principle that "I found one guy who'll vouch for him" is not enough, but I don't think "consensus" is a practical definition of what's required for a few reasons:

• As you know, stating that something is consensus in a field on WP normally requires us to actually find a source that says there is a consensus. In reality you often see soft wordings implying consensus when the sources are "clear enough". But I don't think the community is to keen on the approach where you sometimes see editors try to list up a bunch of sources themselves and demonstrate a consensus, because that is effectively original research.
• Actually, I do not think being called an expert by lots of experts makes one more or less of an expert. If Albert Einstein's work is full of references to someone's website, then I do not think it matters a lot if Stephen Hawking does not ever talk about that website. I think that, consistent with the normal WP approach on things like this, what we are trying to define is the minimum requirement that someone can be considered a possible expert source by involved editors, but as with any source that satisfies the basic criteria of WP:RS and WP:V, whether that source is an appropriately notable source for a particular subject is something that always requires review in the context of what is being sourced. I don't think we can write a rule to cover all cases.
• Remember that even the personal website of Albert Einstein himself is not something we are encouraging to use here. The general recommendation is to avoid self-published sources if we can.

To get a feeling for how tentative the general proposal still is, below is the current text, now expanded to show the context where the implied definition of an expert appears, and then a proposal for a new version. I've tweaked a bit. In my opinion this is hardly a license to go out and use self published sources by anyone who has been mentioned by anyone, so to speak?--Andrew Lancaster (talk) 10:08, 20 February 2011 (UTC)

Self-published expert sources may be considered reliable when produced by an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications. However, one should take care when using such sources: if the information in question is really worth reporting, someone else is likely to have done so. Self-published sources should never be used as third-party sources about living persons, even if the author is an expert, well-known professional researcher, or writer.

— CURRENT WORDING showing more context

Self-published sources may be considered potentially reliable enough for use when they are produced by an established expert on the topic being cited, and represent their considered opinion. An author or publication is an established expert source if they are clearly treated this way in an expert field. Concerning authors, common indications of expertise are citations as expert sources in reliable third-party publications, and having had other work published by reliable third-party publications. However, one should take care when using such sources: if the information in question is really worth reporting, someone else is likely to have done so. Self-published sources should never be used as third-party sources about living persons, even if the author is an expert, well-known professional researcher, or writer.

— PROPOSED WORDING showing more context, and including a few tweaks compared to last proposal

Verifiability and inline citations

I believe that inline citations greatly aid in verifiability, and thus that all sentences (barring obvious case) require an inline citation. Some editors disagree. A discussion is ongoing at Wikipedia_talk:Citing_sources#On_citing_every_sentence; comments from verifiability experts would be appreciated. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 18:15, 18 February 2011 (UTC)

Piotrus boldly added an assertion to CITE saying "referencing only at the end of a paragraph composed of multiple sentences is insufficient". There are accusations that he did so to game a dispute at DYK. WhatamIdoing (talk) 20:49, 18 February 2011 (UTC)

Verifiability from another article

In Talk:Nude swimming#Trivia and reliable sources I'm asking for citations in the article itself for various trivia and another editor is saying they are not needed because they can be verified by looking at the appropriate wikipedia article. Could someoe come there please and make comments one way or the other please? Thanks. Dmcq (talk) 23:27, 18 February 2011 (UTC)

This talk section is the best place to keep the discussion as the results should be relevant to all articles and not one in particular. The question at hand is this: if a particular fact is sourced in a main article - say a biography - do we have to repeat that source in every other article that mentions that fact and links to the main article? For example, if a biography states that someone was born on August 18th and sources it do we really need to provide that same source in the August 18 article next to that person's name under the "births" section, or is it enough that users have an article link to click on containing the appropriate source? I'm opposed to imposing extra work in the name of "verifiability" when a fact is already verifiable and only one obvious click away. Rklawton (talk) 01:17, 19 February 2011 (UTC)

The same, by the way, would apply to trivia sections or other lists containing article links as noted above. If a fact is verifiably (and reliably) sourced in the linked article, then it would be redundant to reproduce that reference again in the list. For example, the Silver Star article contains a lengthy list notable recipients. Absolutely none of these are sourced in the Silver Star article, but every one of them is sourced in the recipient's article. Dmcq's proposal would have us copying all these references over and adding them to the Silver Star article, and that's a lot of work with zero benefit. Rklawton (talk) 01:25, 19 February 2011 (UTC)

Its an old argument, and one that has been rejected. Information should be cited where ever it appears... For two reasons: 1) Readers should not have to click through multiple articles to find verification. 2) The article where the information is cited may change, and the relevant citations may be edited out.

That said... it does not take long to copy and paste a citation... have you considered "Fixing the problem" by going to the other article, copying the citation and slapping it into the Nude swimming article yourself? Blueboar (talk) 01:29, 19 February 2011 (UTC)

If this is an old argument, please provide a link so we can review it. The likelihood that a birth date will change is minimal, and yes, when it does change, someone eventually catches up to the relevant date article and makes the appropriate correction. And besides, if a source for a fact changes, we'd still have to update all the places where that source has been copied, so your argument that someone would have to update the fact is countered by the argument that someone would have to update the source - it's a wash. Now, do you proposing adding redundant citations to the hundreds of thousands of articles that use the convention I described - starting with all the date articles, city articles, and many more.?

Birthdates are nice and clear-cut, and we're used to having pages that link by birthdate, so the chances are indeed good that somebody will see a change on a bio and think "hey, maybe we should change the 'born on X date'" list". But with some other types of claim, it becomes problematic; if I discover a factual error in an article about John Doe ("wait, he wasn't left-handed!"), I'm unlikely to notice that he happens to be in List of left-handed people with made-up names, so I probably won't fix that. Combine with the fact that there are a lot more John Doe experts following John Doe than there are following List of left-handed people with made-up names, and the list article is a lot more vulnerable to error, so we do need to be careful about sourcing in list-type articles.

This is one reason why I recommend using categories in preference to 'list' articles for this sort of stuff. When we add, say, Ron Kovic to Category: Recipients of the Bronze Star Medal, that edit and the accompanying cite happen in one place where people knowledgeable about Ron Kovic are very likely to be reading. If somebody finds a problem with that claim, it can be thrashed out in one place, and when it's fixed on Ron Kovic the category page will automatically update to match it.

I'm not an absolutist on this. WP:V doesn't state that all claims must be sourced wherever they appear, only that they should be sourceable and that if challenged or likely to be challenged, then a cite should be given. A birthdate is not usually contentious (there are exceptions) so I wouldn't think it was automatically necessary to include birthdate sources in, say, List of people born on July 4. But if a claim in a list is likely to be challenged, then that claim needs to be cited there, no matter how well-sourced it might be on some other article.

As for the "the source might change" point - yes, this does happen, but external reliable sources are generally a lot less prone to change than WP articles, so citing the same external source twice seems to me like a lesser evil than internal citation. --GenericBob (talk) 05:59, 19 February 2011 (UTC)

One of the problems with 'cited in the linked article' is that the linked article may not be included in offline copies of Wikipedia. I believe that Wikipedia:Version 0.7 contains only 1% of our articles. WhatamIdoing (talk) 06:40, 19 February 2011 (UTC)

I did have a look at the first article in the list and it had a citation which wasn't of that fact but might corroborate it. however I have no access to it even with google and therefore do not feel entitled to copy over the citation. The other references there did not even mention the business. Dmcq (talk) 13:32, 19 February 2011 (UTC)

So what are we to do with all our date articles that include events/births/deaths - and all our city articles that include "Notable citizens" sections (and much, much more)? Rklawton (talk) 19:00, 20 February 2011 (UTC)

TV and radio

It seems to be unofficial policy that TV and radio broadcasts meet WP:V, and it would require quite a lot of re-editing of WP if they didn't. However, the wording of WP:SOURCEACCESS would appear to suggest not: "anyone should be able to check that material in a Wikipedia article has already been published by a reliable source". You can't necessarily do this with a TV or radio broadcast.

I had thought that the policy was that the material has to be archived somewhere, but if the archive is not accessible to the general public then too bad.

Opinions anyone? Is there a guideline somewhere that clarifies this? Should the policy covering this be tweaked? Should TV and radio broadcasts not be considered to be verifiable sources?

--FormerIP (talk) 21:58, 19 February 2011 (UTC)

Remember that "archived" does not mean "on line"... Most TV broadcasts (not sure about radio) are archived in places that are accessible to the general public... for example, in the US there is the Paley Center for Media (which has branches in both NY and LA). It has extensive archives. Blueboar (talk) 22:13, 19 February 2011 (UTC)

Sure, if there is a public archive that contains the broadcast then obviously that's fine. But what if there isn't? A lot of content from the BBC that is used on WP would appear not to be accessible at all: [4]. The Paley centre appears to have copies of 120,000 shows - but wouldn't it be necessary to establish that a show is in their archive before using it as a source? --FormerIP (talk) 22:16, 19 February 2011 (UTC)

Note that some material that's not accessible in its original form might still be available as transcript. But if there really isn't any way for a WP reader to go back and confirm that the source says what we're claiming it says, then no, by definition it's not a verifiable source and we shouldn't use it. This is inconvenient sometimes but uncheckable sources do pose a serious risk - no matter how reliable the source might be in itself, editors are quite capable of misinterpreting or deliberately misrepresenting such things. (I've encountered at least one editor who made a habit of using uncheckable sources as a way to insert bogus material into a BLP.) Whether an editor is obliged to check an outlet's archival practices is another matter - I think in some cases (not all) it would be reasonable to assume archival until demonstrated otherwise.

I wouldn't be surprised if there are a lot of TV/radio sources in articles that don't satisfy WP:V for this reason, and if anybody challenges them they're likely to be removed. But the reality is that this will only happen when and where some editor cares enough about the issue to do something about it. Uncontroversial material that's not doing any harm is likely to stay there until editors run out of other things to worry about, which might take a while... --GenericBob (talk) 00:11, 20 February 2011 (UTC)

That all sounds fair. I would say though, that it is not reasonable to assume there will be a public archive in the case of TV and radio material. It seems very clear that in most cases such material is not archived.

I think there may be a significant problem that many editors, including experienced editors, think that TV and radio are generally okay as sources (in fact, I think this is pretty much a stock attitude at RSN). Either that or those editors are right and policy is wrong. --FormerIP (talk) 02:35, 20 February 2011 (UTC)

Well, editors are supposed to have seen a source before they cite it... and the rest of us are supposed to assume good faith about that. Blueboar (talk) 04:05, 20 February 2011 (UTC)

I think there are some real problems here. Any source should, automatically, be specific enough, so I saw it on TV just now or even it was just on CNN shouldn't qualify. Also, it's not appropriate for exact quotes, when the distinction between fact from opinion is not clear, when it's not obvious whether there is original reporting or reporting on someone else's reporting, or any situation where the source of the source is unclear, the context of the report, or the role of the reporter/commentator. I think it should be a minimum requirement that an editor looking to use TV or radio sources a) knows which show, which host, on which day; b) knows who was speaking at the time and what they said; c) can distinguish between a quotation, summary, commentary, etc; and d) is aware of the context of the report or discussion. Assuming good faith does not take the place of another set of eyes being able to verify details or confirm for an interpretation. So, nothing against TV/Radio inherently, but I think we should not encourage their use unless its from an easily accessible online clip. And just as a practical matter, I doubt there are many reliable TV or radio shows cover material that really couldn't be sourced to a major newspaper article? Ocaasi (talk) 04:45, 20 February 2011 (UTC)

Let me clarify... When I said to assume good faith, I was not talking about someone saying "but I saw it on TV"... When it comes to citing TV or Radio broadcasts (or films for that matter), we should not cite the original broadcast (although we should note when it aired). Instead, we should cite to a reliable copy of the original broadcast (or an official transcript)... and we need to indicate where we located that copy, so that if someone who wants to verify our citation, they can obtain it as well. Blueboar (talk) 15:39, 20 February 2011 (UTC)

[Edit conflict with Blueboar -- who has now answered my concern :)] If the material is not permanently archived -- and even if it is archived but not accessible to the general public -- isn't there a logical problem here? OK, we assume one another's good faith; or we may set up a system for us to verify one another's citations while still verifiable. But that doesn't help our readers. We don't ask them to assume our good faith: we tell them that Wikipedia is not a reliable source. So they should not rely on what one or two Wikipedia editors claim a TV programme said, if that claim is no longer verifiable. So the claim and the dead citation are no use to our readers. So why include them? Andrew Dalby 15:52, 20 February 2011 (UTC)

OK... I have added a caution about this to the policy (see: WP:V#Broadcast media)... I am sure it could use tweaking and revising... feel free. Blueboar (talk) 17:07, 20 February 2011 (UTC)

(ec)Yes, Andrew, I think I agree. To put it another way, the spirit of WP:V is that readers should not be asked to assume the good faith of editors. If there are no means by which to verify a source, then it fails. This means including in the citation a mention of the location, online or offline where the material is available to the public.

Slight issue is that we have a template Template:Cite episode which doesn't inlude any field for an archive.

Subsequent question: is broadcast material verifiable if it is illegally on YouTube? --FormerIP (talk) 17:25, 20 February 2011 (UTC)

The YouTube issue has been answered many times before... we require that things be verifiable to a reliable source... videos on YouTube are not considered reliable - unless hosted on an official sub-channel (such as the BBC sub-channel). Blueboar (talk) 18:05, 20 February 2011 (UTC)

Thanks, Blueboard (and FormerIP), for the added section in the policy: it works for me. Andrew Dalby 18:11, 20 February 2011 (UTC)

We're not supposed to link to a YouTube copyvio. But it wouldn't seem entirely unreasonable for someone to cite a TV broadcast and, if challenged, say that it's findable on YouTube. I tend to think the general principle you are talking about ought to apply, but I don't think it's clear-cut.

Perhaps more importantly though, I think a change to the citation template to include space for archive info should be proposed. Anyone agree with that? --FormerIP (talk) 18:14, 20 February 2011 (UTC)

New section

I don't think this works as written:

Radio and television broadcasts present unique problems with verifiability. A TV or radio broadcast is ephemeral in nature, occurring in the moment and then over. For this reason, its contents will not be verifiable at a later date. However, a transcript or recording of the broadcast can be verified. Editors taking information from a radio or television broadcast should cite a reliably produced transcript or publicly accessible recording rather than the original broadcast (although the original air date should be included in the citation).

We don't require that people send off for transcripts before citing a broadcast. It's the person who wants to verify it who has to do that, as with any other source. We can insist that either a copy or a transcript be available in principle, i.e. that the broadcaster is a reliable, professional one. But we can't insist that people have already seen the transcript and know how to cite it.

Also it's unclear what "Editors ... should cite a ... publicly accessible recording rather than the original broadcast" means. SlimVirgin ^{TALK|CONTRIBS} 18:22, 20 February 2011 (UTC)

I don't think I'm reading it the same way as you (fair enough, that may be an indication that something needs rephrasing). I think the general idea is that an editor needs to provide a means of verification, whatever that happens to be, and that "trust me, I saw it" is not enough. I don't think it says that editors need to "send of for transcripts" (I'm not sure many TV stations would respond positively to such a request in any case). The onus can't be on the person doing the verifying, because that would require them to prove a negative. --FormerIP (talk) 18:32, 20 February 2011 (UTC)

If you have to cite the transcript, rather than "Soldier Arrested," ABC News Nightline, February 20, 2011, it means you need to have seen the transcript, though it will likely say exactly the same thing. So the section really doesn't work as written. (And yes, you can always buy transcripts.) Also, the last sentence doesn't mean anything. How could I cite a second recording, and not a first broadcast (or whatever is being said)? SlimVirgin ^{TALK|CONTRIBS} 18:40, 20 February 2011 (UTC)

I do understand what you are saying Slim. ... but I am not sure you are correct that you can always buy a transcript. This will probably be true of most major network broadcasts... but not of every TV or radio broadcast everywhere (local station?). At minimum, I think we need to have a reasonable expectation that a particular broadcast was reliably recorded or transcribed (ie that it is possible to verify it) before we cite it. Blueboar (talk) 18:54, 20 February 2011 (UTC)

It feeds into the issue of reliability, because I think any halfway-major broadcaster would make transcripts available. Because you have to be able to prove what you said, prove what you didn't say, be part of the public discourse, as with any other reliable source. Any broadcaster not doing that would be an amateur, self-published as it were (or unpublished in this case). SlimVirgin ^{TALK|CONTRIBS} 20:22, 20 February 2011 (UTC)

I rewrote it. Is this okay?

Radio and television broadcasts are acceptable as sources, provided the broadcaster is a reliable one. Editors should cite the title of the segment if it has one, the name of the show, the name of the reporter as appropriate, the broadcaster, and the air date; the time of the particular segment or statement can be included if available. Transcripts are often available online, and can usually be purchased or otherwise obtained from broadcasters. Editors should note that anything contentious that relies on a broadcast or transcript that other editors find difficult to view or obtain risks being deemed unverifiable for that reason.

SlimVirgin ^{TALK|CONTRIBS} 18:50, 20 February 2011 (UTC)

When addressing this subject, I think we also need to differentiate between broadcast media as primary or secondary sources (just as we would with print media). A segment on a news broadcast is a secondary source. A movie or TV show (not a documentary) is a primary source and should not be used per our policy on primary sources. Rklawton (talk) 18:53, 20 February 2011 (UTC)

No, we can not lump all news broadcasts as Secondary sources... a live broadcast from the scene of an ongoing event is a primary source. Blueboar (talk) 18:57, 20 February 2011 (UTC)

Slim Virgin: I think the statement "Transcripts ... can usually be purchased or otherwise obtained from broadcasters" is just plain wishful thinking. It would be nice were it so. Are you able to provide, maybe, the webpage of a major broadcaster where they give the contact details for ordering transcripts? --FormerIP (talk) 19:09, 20 February 2011 (UTC)

Here is the info on CNN's webpage.

I think Slim is correct where major news networks are concerned (and even the better local news stations). But how do we verify a statement that someone said something on a local cable access show, or a local radio program? I am less certain that such outlets would bother to create transcripts or even record their shows. Blueboar (talk) 19:19, 20 February 2011 (UTC)

I can't imagine a professional broadcaster not recording and retaining its own programs. SlimVirgin ^{TALK|CONTRIBS} 19:39, 20 February 2011 (UTC)

Blueboar: That's just the webpage where CNN hosts transcripts. If what I want to cite appears there, then there is clearly no problem. I can cite the transcript. However, if it isn't there, which in most cases it won't be - it looks like they do one or two segments from each episode of a major show - then it is not clear that CNN will get one of their staff to produce a transcript just because I ask for one. --FormerIP (talk) 19:35, 20 February 2011 (UTC)

Here's an example. [5] People send off for transcripts of shows a lot for various reasons. You have to pay for them, but they already exist; they don't create one for you, they just send you a copy. SlimVirgin ^{TALK|CONTRIBS} 19:37, 20 February 2011 (UTC)

Again, that page offers a search facility for 6,500 transcripts. If my show is there, then no problem. But if not, who do I contact at ABC? --FormerIP (talk) 19:46, 20 February 2011 (UTC)

I don't know, but you can definitely order transcripts. They're needed for legal reasons, journalistic reasons (i.e. for sourcing, like us), for personal reasons, and people use them in journalism school etc. Here's where you order them from CNN. SlimVirgin ^{TALK|CONTRIBS} 20:02, 20 February 2011 (UTC)

No it isn't, it's where I can view those transcripts that are available (as discussed just above.

I don't see any reason to suppose that the number of transcripts available to the public is greater than the number of transcripts available online. --FormerIP (talk) 20:06, 20 February 2011 (UTC)

I'm not sure what that means, about the number available. This is a bit like wondering whether publishers keep more than one copy of their books. :) All broadcasters have transcripts of their shows. They study them, they check them, others study and check them. Members of the public can send off for copies. SlimVirgin ^{TALK|CONTRIBS} 20:09, 20 February 2011 (UTC)

So you say. I don't think what you are saying is correct, however.

Your overall suggestion seems to be that we can accept broadcast material because it is almost always transcribed and the transcripts are almost always available to the public. I'm asking for some evidence that this is the case. Clearly, some broadcasters transcribe some shows. But we have no information about how much of the overall field this represents. --FormerIP (talk) 20:18, 20 February 2011 (UTC)

If you look at the closing credits of major network news shows, you will find, amidst all the other legalese and small print flashing by so quickly, a note informing the viewer of the address to write to if they want to order transcripts. It is standard practice (at least in the US). Blueboar (talk) 20:15, 20 February 2011 (UTC)

I think you may be missing the point. What proportion of shows carry this note? Is it high enough for us to say that all broadcast material is in principle verifiable, or is it less high than that? --FormerIP (talk) 20:18, 20 February 2011 (UTC)

Maybe you should look into that, see if you can find a broadcaster who doesn't do it. SlimVirgin ^{TALK|CONTRIBS} 20:24, 20 February 2011 (UTC)

Or maybe I haven't proposed any new text and so I don't need to prove anything. --FormerIP (talk) 20:40, 20 February 2011 (UTC)

You're making edits based on your personal opinion, which seems to be based on you not being familiar with broadcasting. So it would make sense to familiarize yourself. Sorry if that sounds too harsh, but the way this discussion has gone has not been sensible. SlimVirgin ^{TALK|CONTRIBS} 20:53, 20 February 2011 (UTC)

It doesn't sound harsh, although it does sound a little straw-grasping.

I haven't made any edits based on my personal opinion, only based on my unwillingness to accept yours as fact.

I think the best thing would be to bring your text to the talkpage, which is what I'll do. --FormerIP (talk) 21:23, 20 February 2011 (UTC)

(outdent)... Back in college I volunteered as a Classical Music DJ for WRTC in Hartford, CT. (the college radio station). They certainly didn't transcribe or record my broadcasts. The station also had news reports (mostly someone reading the AP wire, but there was some original journalism, especially if it concerned the college), people did live interviews with local officials and other noteworthy people (for example interviews with rock bands that were in town and had a new album to promote)... None of this was transcribed or recorded for posterity. Could we cite these broadcasts? If so, how can someone verify what we say? Blueboar (talk) 21:20, 20 February 2011 (UTC)

But that's the kind of source we wouldn't want to cite, at least not for anything people might worry about. Maybe we ought to leave that section out entirely. I think it's causing more problems than it's worth. SlimVirgin ^{TALK|CONTRIBS} 21:57, 20 February 2011 (UTC)

Maybe you wouldn't want to cite it, but I could certainly see someone else wanting to cite it. Suppose someone from WRTC interviews <insert name of favorite (notable) band here>, and the lead singer says something about what inspired him to write a particular song. I could definitely see the fan crufters wanting to mention this in the band's article. Suppose the College President is being interviewed, and announces that a new dormitory is being planned. I could definitely see someone wanting to add this to the article on the college.

Now, I could agree with saying that, because no recording or transcript was made of the broadcast (and thus there is no way to verify its contents), these interviews are unreliable. But in which case, are we not defining a "reliable broadcast" as one that has been transcribed or recorded? Blueboar (talk) 22:11, 20 February 2011 (UTC)

Yes, indeed, just as books have to be written down, and not only read out once from the mind of the author during a fan club meeting. But that's a very minimal requirement. It's not the same as an editor being expected to prove that a transcript exists. At some point common sense has to kick in. We see here that even ABC News and CNN are being questioned. SlimVirgin ^{TALK|CONTRIBS} 22:27, 20 February 2011 (UTC)

Proposed text moved from page

Broadcast media

Radio and television broadcasts are acceptable as sources, provided the broadcaster is a reliable one. Editors should cite the title of the segment if it has one, the name of the show, the name of the reporter as appropriate, the broadcaster, and the air date; the time of the particular segment or statement can be included if available. Transcripts are often available online, and can usually be purchased or otherwise obtained from broadcasters. There is no requirement under this policy that an editor citing a broadcast be in possession of a transcript.

What's wrong with this is that it does not appear to say anything about verifiability, which is the topic of the policy page. The first sentence is about reliability. The second is about how to cite. The third (factually contested) tells us something about transcripts but nothing about their relationship to the policy. The last sentence is obviously true, but it doesn't really provide anything that helps in applying the policy.

What really ought to be set out on the page is information about the circumstances under which broadcast material can be considered to be verifiable. --FormerIP (talk) 21:34, 20 February 2011 (UTC)

What we need to acknowledge is that some broadcasts are verifiable, but others are not. Blanket statements in either direction are not going to be accepted. So... what kinds of broadcasts are verifiable? and what kinds are not? Blueboar (talk) 21:55, 20 February 2011 (UTC)

I agree with Blueboar on that. It'd be better if it started out something like, Radio and television broadcasts are acceptable as sources, provided the broadcaster is a reliable one and that the broadcast has been archived or transcribed. If it hasn't been archived or transcribed then it's not verifiable except through the use of a time machine.   Will Beback  talk  22:18, 20 February 2011 (UTC)

I think they're verifiable if a record of the broadcast exists and is accessible to the public. That could be a recording in a public archive or online, or it could be a transcript. But we need to be reasonably sure that it exists and is accessible, rather than just proceeding on the basis that this is possibly the case. In other words, I'd add to what Will says that the archive or transcript also has to be available for the purposes of verification by a WP editor or user. --FormerIP (talk) 22:21, 20 February 2011 (UTC)

(ec) Any broadcaster who spouts off into the ether without thinking they might in future need to know what they said, isn't going to be a reliable source for us by definition. The difficulty with saying it must be archived or transcribed is that you'll get editors who know nothing about broadcasting who'll insist the editor prove that a transcript exists. SlimVirgin ^{TALK|CONTRIBS} 22:23, 20 February 2011 (UTC)

That's the whole point. The editor does need to prove - or at least show a good indication - that the transcript, or some other record, exists, otherwise the material is not verifiable. --FormerIP (talk) 22:26, 20 February 2011 (UTC)

See? SlimVirgin ^{TALK|CONTRIBS} 22:27, 20 February 2011 (UTC)

I do know some stuff about broadcasting. I learn it from the BBC website: [6]. --FormerIP (talk) 22:36, 20 February 2011 (UTC)

I think the issue is more: when does the editor need to "prove" that an archive exists... at the point of citation, or when challenged on it? I lean towards "when challenged"... We don't require that someone link to amazon or their library's card cataloge when citing a book (to "prove" that the book exists)... but, if someone suspects that the book does not exist, they can challenge the citation and request such proof. The same should be the standard when citing broadcasts.

The more I think about it, the more I think Will's phrasing above is really all that is needed... 'Radio and television broadcasts are acceptable as sources, provided the broadcaster is a reliable one and that the broadcast has been archived or transcribed. This sets out the condition of reliability succinctly and we trust editors to abide by it. If there is a reason to think that this condition has not been complied with, however, then a challenge can be issued and the problem discussed. Exactly the same as with questionable citations to print sources. Blueboar (talk) 23:03, 20 February 2011 (UTC)

I disagree, because I think the arguments here offer a taste of things to come, with editors being told they're not allowed to cite a BBC program unless they can somehow prove that a copy exists. We're better saying nothing than exposing people to that kind of problem. SlimVirgin ^{TALK|CONTRIBS} 23:07, 20 February 2011 (UTC)

Well, what happens if you think a print source does not exist? You issue a challenge and make the editors "prove" that the source does in fact exist. They don't need to do this in the citation... they can provide a link to amazon or google books on the talk page, but the challenge does need to be responded to or (per WP:BURDEN) the source can be removed. Blueboar (talk) 23:14, 20 February 2011 (UTC)

But we expect people to use common sense. We need a reason to suppose a source might not exist. FormerIP is saying editors need no reason to suppose; that the default position is to assume copies don't exist. What he thinks happens if there's a lawsuit, I don't know. But professional broadcasters do, of course, keep copies of their work, and to ask editors to prove that is silly. SlimVirgin ^{TALK|CONTRIBS} 23:19, 20 February 2011 (UTC)

Are the archives of professional broadcasters open to the public?   Will Beback  talk  23:25, 20 February 2011 (UTC)

I would guess that's not usual. I know that the BBC archives are not. If the archives of any particular broadcaster are open to the public then fine - that's that broadcaster covered. --FormerIP (talk) 23:27, 20 February 2011 (UTC)

This may be a national thing... because in the US they definitely are. Are you saying that if I write to the BBC and request either a copy or a transcript of something they aired, they will not send it to me? Blueboar (talk) 23:36, 20 February 2011 (UTC)

SV, applying WP policy shouldn't be seen as a "problem".

Blueboar: I agree that any requirement to identify means of verifiability should only come into play if the material is challenged - that's normally the case when applying WP:V in any event.

I think it is important, though, to make clear that the archive etc should be publicly accessible, otherwise there will just be disputes about whether the private archive of CBS or whatever counts. As you can see from the OP way above, I had been under the impression that it would. But this goes against WP:SOURCEACCESS. --FormerIP (talk) 23:16, 20 February 2011 (UTC)

Well, according to WP:IRS, all sources need to be publicly accessible (A manuscript that only I possess is not a reliable source) ... so I don't see the need to repeat this specifically for broadcast sources. Blueboar (talk) 23:32, 20 February 2011 (UTC)

That's true, but it's not many bytes just to add the word public. I'd be concerned because that was the misapprehension I was under, and I think other editors will be under it too. --FormerIP (talk) 23:36, 20 February 2011 (UTC)