Wikipedia talk:Verifiability/First sentence/V FS Archive 47
 | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Jargon in the first sentence. How about some standard English instead?
Over the years Wikipedia, like any culture, has acumulated its share of jargon. Some of it is natural, and some is unnecessary and legalistically obstructive to understanding. Nowhere is this problem demonstrated than the first sentence in the WP:V policy, where two standard English words are used in a way that is nonstandard, and which tends to invite wiki-wars among people who assume unconsciously that the words they have heard over and over, are meant in their normal senses. Here is the first sentence of WP:V:
The threshold for inclusion in Wikipedia is verifiability, not truth; that is, whether readers can check that material in Wikipedia has already been published by a reliable source, not whether editors think it is true.
This statement is quite deceptive, for it is full of jargon. In failing to define its terms completely and immediately, it sets up an odd dichotomy between "verfiability" (as we usually understand the word) and "truth." (as we usually understand the word). Iin the real world, where verifiable simply means provably-true, and provable is a loaded word, such a dichotomy exists between that which is true and provably true. In the real world, standards of provability are not those acceptable to Wikipedia, but that doesn't mean Wikipedia isn't trying. "Verifiability" is defined by Wikipedia in this opening sentence as consisting of two distinct parts: [1] Whether readers can check that material in Wikipedia has already been "published" [at that point the "material" could be a flat-out mistake, lunatic idea, or a deliberate lie], and [2] That the publication occurs in a "reliable source." A "reliable source" is (in turn) defined lower down within the same policy, in the WP:V section on sources WP:SOURCES: "Articles should be based on reliable, third-party, published sources with a reputation for fact-checking and accuracy."
Now, "fact-checking" has to do with checking facts, which in turn are defined (see any dictionary) as statements that are TRUE. Likewise "reputation for accuracy" means "reputation for being true or close to true." Fact and accuracy are synonyms for truth and "truth-hood." (opposite of falsehood). Calling the idea of "truth" by another name (as well as putting synonyms like "fact" and "accuracy" farther down the policy page) unnecessarily obscures a simple meaning, and a simple goal. Wikipedia's goal is indeed truth. However, the very idea of that it is not, causes Wikipedia editors to go around cudgeling each other with the idea that "verifiablity," trumps "truth," as though they were using these words in their normal meanings, instead of their special "wiki-meanings."
So now, we require an exegesis. Let us find out what the first sentence of WP:V means in standard English, by getting rid of all jargon and wiki-meanings. We will unpack and rewrite the first sentence of WP:V in standard English, using the policy's own definitions for the wiki-words, as necessary. Ready? First, we replace "reliable" by "published, with reputation for fact-checking and accuracy," and then, we replace "fact-checking and accuracy" with "truth-checking and truthfullness." Here's what we get:
The threshold for inclusion in Wikipedia is verifiability, not truth: that is, whether readers can check that material in Wikipedia has already been published by reliable third-party sources with a reputation for truth-checking and truthfulness, and NOT whether the editors think the material is true.
That fixes puncutation and gets rid of the wiki-words and wiki-meanings in the last part. We are now in a position to get rid of the wiki-meanings in the first part, so that now ALL of the words are used in their normal standard English definitions:
The threshold for inclusion of material in Wikipedia is published-source truthfulness, not the editor's idea of truth: that is, whether Wikipedia's readers can check that material in Wikipedia has been published by reliable third-party sources who have a reputation for truth-checking and truth-telling, and NOT whether Wikipedia's editors merely personally think that the material is true.
At this point, we have de-obscured the text, and with it, some of the problems. Wikipedia's own policies are actually all about "truth"- TRUTH is what reliability and fact-checking are about. When the policy says it is about verifiability and NOT truth, this is an obscurantist, for neither word is being used here as in standard English. In standard English, "verifiability" can mean many things, including what the reader can prove it to him or herself by checking by his/her own math, or by going to see for him or herself. Verification also has many possible meanings in the sciences and law, and standards differ from one discipline to another. The statement doesn't actually mean that Wikipedia is about "verifiability" as we all understand the word— it's about a particular documentary version of reality used by Wikipedia, which is narrower, not perfectly well-defined, and very much media-driven. For example, I might be a celebrity with a mole on my right hand. If the New York Times states it is on my left hand, then Wikipedia's use of the word "verifiability" departs from the standards of the English language. For here, Wikipedia doesn't mean the word in any way that has contact with either reality, or standard English usage. In such cases, I can't prove a thing to anybody who reads otherwise in their newspaper.
Likewise, when the WP:V policy states that the standard of Wikipedia inclusion is NOT "truth," it also does not use the term as it is usually understood. In this case, Wikipedia IS actually interested in truth as we normally understand the word "truth" when we speak of public truth, demonstrable truth, scientific truth, legal truth, or even journalistic truth. It simply isn't about personal truth or any other type of truth one could NOT use in a legal or science debate. It's not about your truth and my truth, unless some other "reputable" person or source has published it, and one can cite it, as in a debate or courtroom. Worse, the question of "reputation" of a publisher, is in the eyes of the beholder.
What this policy actually means to say, is that the threshold for Wikipedia inclusion is NOT what its writers think is true, but rather what some OTHER published writers, who have a "reputation" for telling truth, think is true, and have gotten somebody else to "publish" (whatever "publish" means these days). That seems to be a more narrow statement, but at least it is closer to what Wikipedia's policy actually means to say.
Alas, the problem of how a Wikipedia writer, independent of his or her own sense of "truthiness," is to identify these published truth-telling authors, and their truthy works and truthy reputations, is the rub. Some published writers are wrong, some ignorant, some are shills, and others are less than sane. This is the source of many an edit war, as an editor's personal idea of what is true, will naturally have heavy effects on their judgement of the truthfulness of available published sources. Must one include the writings of the confused, the ill-informed, the hacks and the crazies? But at least we're a bit closer to the root of the epistemological problem. First, one must lay out the policy in plain standard English.
(The above was written after a recent encounter with a few other editors of an article who had no idea what WP:V means, but it wasn't their fault. It was the WP:V policy's fault). SBHarris 01:23, 22 March 2011 (UTC)
- For those interested in the mad scribblings of the editors who do not understand WP:V, the discussion in question is on Talk:Max Gerson. - SummerPhD (talk) 01:52, 22 March 2011 (UTC)
- And there, you will find a mighty lame argument, in which one editor (not me) removes material about an alternative clinic's views on a conspiracy to poison its guru, under the guise that they are not "verifiable" (though they come from a film made by the woman who runs the program, the daughter of the guru, and this statement was sourced and cited). And then you will find the editor above, adding back material from literature from the very same quacks, on the idea that quacks indeed are a reliable source about what quacks think about their own quackery. Neither of these editors are me. However, everybody on that TALK page seems to be convinced that they know WP policy in this area, and have quoted WP:V to me several times, now, as though I needed some teaching. Okay. Yes, it prompted me to come here to find out if everybody who thinks they understand this policy, really understands this policy. What did I miss, please? I put my essay up above, regarding the first line of it. Now it's your turn, Wikipedians. Impress me. SBHarris 06:36, 22 March 2011 (UTC)
- For what it's worth, I agree that this policy (and its cousins) are very poorly written in places, and from a logical point of view very poorly constructed. It's as if the people writing them didn't really want them to be understood. But try to improve them in the normal wiki way, and you come up against a kind of religious cult which practically worships the current wording of these pages as scripture, and will ensure that even the most reasoned changes (even those resulting from lengthy talk page discussion) get blindly reverted without rational explanation. Hence these most important parts of Wikipedia's documentation, which ought to be among the most clearly written, in fact end up among the most confusing (since the normal healthy processes by which we improve pages are not allowed to operate here).--Kotniski (talk) 10:07, 22 March 2011 (UTC)
- And there, you will find a mighty lame argument, in which one editor (not me) removes material about an alternative clinic's views on a conspiracy to poison its guru, under the guise that they are not "verifiable" (though they come from a film made by the woman who runs the program, the daughter of the guru, and this statement was sourced and cited). And then you will find the editor above, adding back material from literature from the very same quacks, on the idea that quacks indeed are a reliable source about what quacks think about their own quackery. Neither of these editors are me. However, everybody on that TALK page seems to be convinced that they know WP policy in this area, and have quoted WP:V to me several times, now, as though I needed some teaching. Okay. Yes, it prompted me to come here to find out if everybody who thinks they understand this policy, really understands this policy. What did I miss, please? I put my essay up above, regarding the first line of it. Now it's your turn, Wikipedians. Impress me. SBHarris 06:36, 22 March 2011 (UTC)
- We don't require third party sources for verifiability; we prefer them, and we need some of them for notability, but for verifiability, information may be taken from primary sources as well, with some caveats (not contentious, not unduly self-serving, ...). The rest of your changes don't really make the meaning of the nutshell any clearer, in my opinion, e.g. "published-source truthfulness" is much harder to decipher than "verifiability". We may of course improve the current text, but your proposal is a step in the wrong direction. Something like
The threshold for inclusion of material in Wikipedia is that it has been published by reliable sources, not that it is the truth that is, whether Wikipedia's readers can check that material in Wikipedia has been published by sources where fact-checking and accuracy is generally expected, and not whether Wikipedia's editors personally think that the material is true.
If we're talking about this sentence, then: (a) not "the threshold" (there are various reasons we might or might not decide to include some statement in some article; this is just one of them); (b) not "that it has been published by" but "that it is supported by" or words to that effect; (c) not "not truth" or anything like that - we don't want sourced information that we know to be false. I would start something like this: "It is a core principle of Wikipedia that information added to it must be verifiable from reliable sources. This policy describes what kind of sources are considered reliable for which purposes, and how they can be used to support the inclusion of information in Wikipedia articles." --Kotniski (talk) 10:43, 22 March 2011 (UTC)
| Current | SBH "un-packed" | SBH "no jargon" | Fram | Kotniski |
|---|---|---|---|---|
| The threshold for inclusion in Wikipedia is verifiability, not truth; that is, whether readers can check that material in Wikipedia has already been published by a reliable source, not whether editors think it is true. | The threshold for inclusion in Wikipedia is verifiability, not truth: that is, whether readers can check that material in Wikipedia has already been published by reliable third-party sources with a reputation for truth-checking and truthfulness, and NOT whether the editors think the material is true. | The threshold for inclusion of material in Wikipedia is published-source truthfulness, not the editor's idea of truth: that is, whether Wikipedia's readers can check that material in Wikipedia has been published by reliable third-party sources who have a reputation for truth-checking and truth-telling, and NOT whether Wikipedia's editors merely personally think that the material is true. | The threshold for inclusion of material in Wikipedia is that it has been published by reliable sources, not that it is the truth that is, whether Wikipedia's readers can check that material in Wikipedia has been published by sources where fact-checking and accuracy is generally expected, and not whether Wikipedia's editors personally think that the material is true. | It is a core principle of Wikipedia that information added to it must be verifiable from reliable sources. This policy describes what kind of sources are considered reliable for which purposes, and how they can be used to support the inclusion of information in Wikipedia articles. |
I've placed the three versions under discussion above for easier comparison. My own first comment about SBHarris' proposals is that I believe the aim of using normal English is a good one, and I agree that this policy is not always well understood, but (a) trying to avoid all internal jargon ("un packing") makes sentences longer, which obviously does not mean easier to understand. (b) the attempt at reducing jargon also does not really work for me. Is "published-source truthfulness" normal English? I do not think so. Fact is that Wikipedia has developed some new concepts which are useful for discussion about Wikipedia and hard to discuss without using terms for those concepts. This is the type of situation in which neologisms and jargons are sometimes justified. (And wikilinking does make our jargon a little easier to learn.) That does not mean we should not be constantly careful of jargon, and constantly looking for policy areas where people are misunderstanding things often.--Andrew Lancaster (talk) 10:49, 22 March 2011 (UTC)
- Yes, I don't think SBH's proposals achieve the stated aim (they look even more jargony than the original, and are not fully accurate in all situations). Can you add Fram's proposal - which is better (though still not as good as mine, of course ;) ) - to the table?--Kotniski (talk) 10:57, 22 March 2011 (UTC)
- Kotniski, one thing I do not like about your proposal is that you remove one of the most "normal English" bits about "not whether [individual] editors think it is true". BTW, the word individual should logically be there because a consensus of editors is good enough also. I'll add Fram.--Andrew Lancaster (talk) 11:04, 22 March 2011 (UTC)
- How about simply removing the 'not truth' from the current version but leaving in everything else to produce
- "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, not whether editors think it is true."
- This expresses it well enough for me whilst upping verifiability and not confusing people by having not immediately preceding truth. Dmcq (talk) 11:09, 22 March 2011 (UTC)
- It's an improvement, but still has the problems mentioned (and some others): this isn't the threshold (there are other factors too); it's not essential that all readers are able in practice to check something (so those words are a bit misleading); the material itself needn't have been published (in the way that many people will understand that phrase); and frankly, if editors don't think something's true, it shouldn't be going in. I would be happier if we stuck with Andrew's suggestion of saying individual editors.--Kotniski (talk) 11:29, 22 March 2011 (UTC)
- And let's remember this isn't the whole policy, just the first sentence - it's not essential that we say everything here, just that we introduce the subject, say what the page is going to be about, and avoid saying anything contentious or misleading. (Which is what I've tried to do with my version - the contentious bit of course is what counts as "reliable", so I say straight away that this issue is going to be dealt with in detail.) --Kotniski (talk) 11:34, 22 March 2011 (UTC)
- Perhaps "published-source truthfulness" isn't very good, and "publication in a reliable source" is better. I just don't like the word "verifiability," as it means something different to Wikipedians than it does to everybody else, and it's especially bad to use a privately-defined term, up front in a policy. Define it specially much later, if you MUST use it. The resistance to using other words is that we'd have to rename this policy page- but so what?
Somebody has noted that not everything in WP really must be proven (by cite) to have been published in a reliable source, so long as the editors all agree with the material, and there is no contention. So it's theoretically provable-by-cite (if you could find a cite for an obvious thing). But trying to cram that whole concept into "wiki-verifiability" is also a bad idea. Better to just qualify the idea of "needed publication" right at the beginning. Jimmy Wales is wrong that obvious facts are easy to find cites for. Example: you can't find me any published source that says Charles Dickens' novel The Pickwick Papers was originally written by Dickens in English. However, we all believe it since it's obvious to any thinking adult who knows anything about Dickens, and you could add it to an article without a {{citation needed}}. Wikipedia takes "judicial notice" of certain non-controversial facts. SBHarris 19:00, 22 March 2011 (UTC)
- But that kind of fact is verifiable (in the normal English sense of the word) just by looking at the original text of the novel and observing that it's written in English. I don't think the WP use of the term "verifiability" really differs all that much from the normal English use of the term - especially if we write "from reliable sources" after it, as I propose. I see several problems with the first sentence of this policy (as enumerated above), and would be happy to hear anyone's response to those, but I don't honestly think that a jargonny use of "verifiability" is one of them.--Kotniski (talk) 07:20, 23 March 2011 (UTC)
- Kotniski, to me it seems that your concern with the logic of the word "threshold" could also be addressed by saying that maybe verifiability should not be equated quite so simply to being able to be verified in the one way specific way of looking up published sources. If you do not make this over-simple equation of V and RS then cases like the one SBHarris mentions are resolved. For example a consensus that something is commonly known is a kind of verification which does not involve citing reliable publications, and then you can logically say that verification, in this broader sense, is the threshold. Maybe my comment helps explain how others are seeing the problem you point to. This is for example the reason I find the wording about it not just being an editor's personal opinion so important.--Andrew Lancaster (talk) 08:57, 23 March 2011 (UTC)
- Yes but then you end up with a sentence that says nothing, except that the standard for inclusion in Wikipedia articles is something that we have decided to call "verifiability".--Kotniski (talk) 09:06, 23 March 2011 (UTC)
- Actually there are loads of sources saying Dickens wrote the Pickwick Papers. The only exclusions are for absolutely obvious common sense things like the sky is blue and straightforward conversions and suchlike as in WP:CALC, and every day some POV pusher tries to drive a horse and cart through even those. I find a lot more difficulty with finding reliable sources for things like how wooden floors should normally be treated, there's loads of how to sites but precious little in the way of reliable sources even though cleaning floors is one of the largest occupations in the world. I still wouldn't want to change those requirements though. Dmcq (talk) 09:45, 23 March 2011 (UTC)
- Dmcq, I think you are misunderstanding the example of SBHarris, which was indeed interesting in my opinion. He was talking about sourcing which language Dickens wrote it in. I am sure many of us have had someone tag something like this.--Andrew Lancaster (talk) 10:57, 23 March 2011 (UTC)
- Exactly. And I picked The Pickwick Papers for a reason, as the origial written manuscript is lost, so you can't "look it up." By induction it's obvious that Dickins wrote it in English the way he did his other works, but at the same time it's SO obvious that nobody really bothers to say it in print. Thus, I cannot find a cite for it (can any of you?).
Another different kind of problem occurs when a popular concept in (say) science is wrong, but so widely believed that even lower division college texts are more likely to have the wrong thing than graduate level texts and journal articles, which have the correct thing. One example I once fought through is the notion that mass can be "converted" to energy, so that the mass dissappears entirely, and leaves just the energy (this is true for matter, but is not true for mass-- they aren't the same). So "verification" in this case requires Wikipedia editors not only to know their physics, but to be able to sort out the graduate level texts and texts by relativity experts, from all the other physics texts- a thankless task.
When it comes to definitions, it's even worse. In the article on weight (for example) editor consensus has forced me to live with an article that defines weight in a way that differs from the ISO definition, just because most college undergrad physics texts define it another way that isn't as good (by the opening definition, orbiting space station astronauts aren't weightless, they are only "apparently weightless"!). And so on. The judgement of sources is not a job that can be fobbed off, but WP attempts to do it all the time. We are interested only in sources mostly likely to be TRUE, but that in turn cannot be determined without knowing yourself how truth is arrived at, in that area. That's a tricky and knowledge-area dependent process. SBHarris 21:04, 23 March 2011 (UTC)
- Exactly. And I picked The Pickwick Papers for a reason, as the origial written manuscript is lost, so you can't "look it up." By induction it's obvious that Dickins wrote it in English the way he did his other works, but at the same time it's SO obvious that nobody really bothers to say it in print. Thus, I cannot find a cite for it (can any of you?).
- Dmcq, I think you are misunderstanding the example of SBHarris, which was indeed interesting in my opinion. He was talking about sourcing which language Dickens wrote it in. I am sure many of us have had someone tag something like this.--Andrew Lancaster (talk) 10:57, 23 March 2011 (UTC)
- Actually there are loads of sources saying Dickens wrote the Pickwick Papers. The only exclusions are for absolutely obvious common sense things like the sky is blue and straightforward conversions and suchlike as in WP:CALC, and every day some POV pusher tries to drive a horse and cart through even those. I find a lot more difficulty with finding reliable sources for things like how wooden floors should normally be treated, there's loads of how to sites but precious little in the way of reliable sources even though cleaning floors is one of the largest occupations in the world. I still wouldn't want to change those requirements though. Dmcq (talk) 09:45, 23 March 2011 (UTC)
- Yes but then you end up with a sentence that says nothing, except that the standard for inclusion in Wikipedia articles is something that we have decided to call "verifiability".--Kotniski (talk) 09:06, 23 March 2011 (UTC)
- Kotniski, to me it seems that your concern with the logic of the word "threshold" could also be addressed by saying that maybe verifiability should not be equated quite so simply to being able to be verified in the one way specific way of looking up published sources. If you do not make this over-simple equation of V and RS then cases like the one SBHarris mentions are resolved. For example a consensus that something is commonly known is a kind of verification which does not involve citing reliable publications, and then you can logically say that verification, in this broader sense, is the threshold. Maybe my comment helps explain how others are seeing the problem you point to. This is for example the reason I find the wording about it not just being an editor's personal opinion so important.--Andrew Lancaster (talk) 08:57, 23 March 2011 (UTC)
- But that kind of fact is verifiable (in the normal English sense of the word) just by looking at the original text of the novel and observing that it's written in English. I don't think the WP use of the term "verifiability" really differs all that much from the normal English use of the term - especially if we write "from reliable sources" after it, as I propose. I see several problems with the first sentence of this policy (as enumerated above), and would be happy to hear anyone's response to those, but I don't honestly think that a jargonny use of "verifiability" is one of them.--Kotniski (talk) 07:20, 23 March 2011 (UTC)
- Perhaps "published-source truthfulness" isn't very good, and "publication in a reliable source" is better. I just don't like the word "verifiability," as it means something different to Wikipedians than it does to everybody else, and it's especially bad to use a privately-defined term, up front in a policy. Define it specially much later, if you MUST use it. The resistance to using other words is that we'd have to rename this policy page- but so what?
So how about we combine my first sentence above with something about it not being about editors' opinion, e.g. "It is a core principle of Wikipedia that information added to it must be verifiable from reliable sources. What matters is not editors' personal opinions as to whether something is true, but whether it can be supported through the proper use of sources. But I still feel a need to say straight away, either in the same paragraph or just after it, that the concepts of "reliable" and "proper use" (or whatever words we end up choosing to express those ideas) are not trivial or (necessarily) intuitive, and indeed the main purpose of the policy we're introducing is to explain what we at Wikipedia consider to be reliable sources and proper use of sources in various situations. (Oh, and I don't find it at all obvious that the sky is blue.)--Kotniski (talk) 10:51, 23 March 2011 (UTC)
- You seem to me to be going along with almost equating WP:V and WP:RS. What is the difference between these two policies?--Andrew Lancaster (talk) 10:57, 23 March 2011 (UTC)
- Well, to me IRS is just one aspect of V in more detail. V says "everything should be supportable by reliable sources", while IRS adds some more detail about what we mean by "reliable", leaving V to address mainly the question of what we mean by "supportable". (Wikilawyers will also point out that IRS is a guideline rather than a policy, whatever they may think the difference signifies.)
Are you proposing that we acknowledge any other kind of verifiability which isn't based on sources? It seems generally accepted that we can't ask people to verify something by, say, doing an experiment or even following the steps in a non-trivial logical deduction. --Kotniski (talk) 11:35, 23 March 2011 (UTC)
- I think the key aspect of WP:V is already slightly different from WP:RS. The most important thing is to make sure information can be confirmed not just to be some individual or group's non-notable opinion. So one way to understand why we are not meant to demand proof of widely known things is to say that we should not ask people to verify what is already known to be verifiable. This would make no sense if WP:V just meant the same as WP:RS?--Andrew Lancaster (talk) 11:43, 23 March 2011 (UTC)
- I'm not entirely sure I understand you - are you saying that there's some other kind of verifiability that we accept, other than the possibility of verifying based on reliable sources? Can you give examples? (I know that not everything has to be verified explicitly from RS; that's why we say verifiability rather than verification.)--Kotniski (talk) 13:18, 23 March 2011 (UTC)
- Right, the -able makes the word different, and means we do not really have to verify everything. So this makes common knowledge things verifiable, but possibly not able to be sourced from a reliable source, as per the example of Dickens having written in English.--Andrew Lancaster (talk) 13:56, 23 March 2011 (UTC)
- I'm not entirely sure I understand you - are you saying that there's some other kind of verifiability that we accept, other than the possibility of verifying based on reliable sources? Can you give examples? (I know that not everything has to be verified explicitly from RS; that's why we say verifiability rather than verification.)--Kotniski (talk) 13:18, 23 March 2011 (UTC)
- I think the key aspect of WP:V is already slightly different from WP:RS. The most important thing is to make sure information can be confirmed not just to be some individual or group's non-notable opinion. So one way to understand why we are not meant to demand proof of widely known things is to say that we should not ask people to verify what is already known to be verifiable. This would make no sense if WP:V just meant the same as WP:RS?--Andrew Lancaster (talk) 11:43, 23 March 2011 (UTC)
- Well, to me IRS is just one aspect of V in more detail. V says "everything should be supportable by reliable sources", while IRS adds some more detail about what we mean by "reliable", leaving V to address mainly the question of what we mean by "supportable". (Wikilawyers will also point out that IRS is a guideline rather than a policy, whatever they may think the difference signifies.)
The WP:V policy page actually has an RS section which is pretty good, in part because it is vague. It says:
Articles should be based on reliable, third-party, published sources with a reputation for fact-checking and accuracy. Sources should directly support the material presented in an article and should be appropriate to the claims made. The appropriateness of any source depends on the context. In general, the best sources have a professional structure in place for checking or analyzing facts, legal issues, evidence, and arguments; as a rule of thumb, the greater the degree of scrutiny given to these issues, the more reliable the source.
That ends up leaving the standards in each field up to the experts in that field. The only problem for Wikipedia is that it takes some expertise to know who the experts are. The idea of "wiki-verifiability" is "you could in theory go look it up and verify that it's true" but the question then is "where do you go to look it up to be most confident?" The rest of the RS section in WP:V attempts to say more on this question, and says some questionable things. What is a "mainsteam newspaper"? Does that mean obituaries in the New York Times are reliable? And how do you tell that you have a "respected publishing house"? I pulled at random from my large shelf of quack health books and found Fit for Life, which was published by Warner Books, owned by Time Warner. It'a hilarious bunch of food combining nonsense, but it sold well (3 million copies already in 1985, which is when my edition was printed in paperback). So is it "reliable"? Reliable for what purpose?
The article WP:RS (now WP:IRS) attempts to expand on this theme, and succeeds in some areas and fails in others. If you want to know how complex this can be, look at WP:RSMED, which looks at the problem just in medical material. This is one of WP's better guidelines (no, I didn't write any of it; I do agree with most of it). However, there is no corresponding RS section for the physical sciences.
When it comes to journalism, the section in WP:IRS mentions churnalism (please read this), and that alone should either make you nod grimly or your hair stand on end, depending on your prior view of newspapers. But knowing that, NOW what do you do? You're stuck. There's really no way you can see what's going on behind the scenes at a newspaper, the way you can a science journal. Peer-reviewed journals or book publishers like Springer require many credentials from editors, and have many cycles of fact-checking that go on between author and outside degreed editors, before they publish. A newspaper journalist by contrast often will not read the entire article over the phone to the person it's about, or who supplied its expert-information, before print. And so on. Verifiability needs reliable sources, and WP:IRS can't walk people through all the problems with finding them, if they have no knowledge themselves. SBHarris 22:24, 23 March 2011 (UTC)
- I don't see any need to change this policy: most people can, and do, figure out what we mean. If people are really concerned about the phrase being widely misunderstood, then they might consider writing an essay to provide a jargon-free explanation of it. WhatamIdoing (talk) 23:36, 23 March 2011 (UTC)
- Actually most people do NOT have any idea what the policy means, and there is endless arguing about it. You could indeed write a jargon-free essay about it. And then replace the policy with the jargon-free essay! Two problems solved in one blow. Of course, you're still stuck with the IRS problem, but at least all you now have is one problem. SBHarris 00:35, 24 March 2011 (UTC)
- Yes, it seems a bit strange to propose that the way to deal with a potentially misunderstood passage is to write an essay explaining it elsewhere; the thing to do is to rephrase it so as not to admit of the misunderstanding in the first place (or at least, reduce the likelihood of it). We can do that; that's what we would do on any other page; why not do it here? Just say as far as possible, in ordinary language free of euphemism and hyperbole, what this policy is about. --Kotniski (talk) 11:23, 24 March 2011 (UTC)
- Actually most people do NOT have any idea what the policy means, and there is endless arguing about it. You could indeed write a jargon-free essay about it. And then replace the policy with the jargon-free essay! Two problems solved in one blow. Of course, you're still stuck with the IRS problem, but at least all you now have is one problem. SBHarris 00:35, 24 March 2011 (UTC)
Copy-edits
I did some rephrasing to the Lead, particularly including a reference to reliable sources and what that means, in the first two paragraphs. Maybe it is a bit clearer? I agree that a sentence explaining the common misunderstanding and what it means could also be useful.
Some other things I'd like to see in the lead:
- Explanation of what V means in practice: This means that individual editors may disagree that a source is accurate, but we still don't dispute the truth of the claim itself, only whether or not the source meets our standards to make such a claim.
- Burden of proof: The burden of evidence is on those who want to keep material not those who want to remove it, although it is okay that Wikipedia if claims are unsourced but neither challenged nor controversial; they can improve over time.
- Substantive reference to NPOV: Through properly summarizing all relevant reliable sources, Wikipedia approaches the best version of the truth that can be trusted. It might turn out to be false, but we wait for reliable sources to cover that change. Ocaasi c 02:08, 24 March 2011 (UTC)
- Actually I think we do sometimes dispute the truth of a claim without necessarily throwing the entire source out as unreliable - even reliable sources make errors, and we don't have to parrot them where we have good reason to suppose they've got it wrong. But generally, yes, it would be more useful if the lead contained more practical information.
Another absurd thing is that when we propose merging WP:V with WP:NOR on the grounds that they're the same thing, people claim "oh no, not the same thing at all" - but now you have, in the second paragraph here, this policy being summarized in terms of original research - effectively defining original research to be exactly the same thing as what this policy forbids! Can you all make up your minds - if original research is a separate subject, then let's not say that this policy is about it - and if it's not a separate subject, then let's simplify and merge the two pages into one. --Kotniski (talk) 11:30, 24 March 2011 (UTC)
- Actually I think we do sometimes dispute the truth of a claim without necessarily throwing the entire source out as unreliable - even reliable sources make errors, and we don't have to parrot them where we have good reason to suppose they've got it wrong. But generally, yes, it would be more useful if the lead contained more practical information.
I've tried with this sequence of edits to remove some of the most obvious and hopefully uncontroversial problems with the lead (as explained in the edit summaries and in the above posts) - I hope (perhaps too optimistically) they won't be reverted blindly, but (if at all) then for good reasons that will be presented to us.--Kotniski (talk) 11:49, 24 March 2011 (UTC)
- Maybe Kotniski's work at on end of the boldness spectrum will help pull the evolution situation here more towards the middle ground from the other extreme which it is at now which is Wp:ownership through application of a double standard......an incredibly high bar set for any changes that are not made by an owner. North8000 (talk) 12:12, 24 March 2011 (UTC)
- Kotniski, I haven't had time to look closely at your changes, but I do think that a description of what OR and NPOV in the lead is important as long as the policies are separate. There's way too much interpretation and application of them in isolation. People forget that V means no Syn and V doesn't mean automatically to include unless Weight and Fringe are met, and NPOV doesn't mean you can include OR opinions which haven't been published and OR doesn't meant individuals are or are not RS for claims, and V always depends on the claim being made, in the context it is being made, which is also about NPOV presentation.... these things all intertwine. Not to even mention Copyright and BLP. Maybe the policies should be physically separate or maybe not. But they must be explained in a way that lets readers see how they are not isolated. Maybe each core policy should have an explicit section on: what this means in practice; what this means for OR; what this means for NPOV (respectively).
- You also mentioned that there are times we do challenge the accuracy of a claim (such as in an obvious journalistic error?). If so, I don't see anything in policy that gives literal license to do such a thing, aside from IAR and editorial discretion, and maybe Weight (if it's clearly disputed by other sources and not worth including by itself). Do you think this should be couched more directly? Ocaasi c 14:56, 24 March 2011 (UTC)
- Perhaps not, at least, not as a priority in the lead; but I think it's a good reason not to place so much emphasis on the "we don't care about truth" meme, as I think we do care. And I wasn't objecting to mentioning OR and NPOV in the lead - what I objected to was telling people "in order to avoid original research" and then effectively summarizing this policy - that makes no sense at all (unless we agree that this policy is just a fork of NOR, in which case they should certainly be merged). If we can achieve a sensible logical division between the three policies (which we certainly do not have at the moment - the whole thing's a jumbled mess) then certainly we should be summarizing the relevant points of the other two in this one - but surely after we've explained the key points of this one?
I also notice that the old nutshell has been restored - can anyone explained why? I personally find nutshells a redundant and fairly silly gimmick - the policy should be summarized in the lead, not in a special box - but if we're going to have one, it should surely summarize the policy in as much generality as possible, not just pick one sentence from it.--Kotniski (talk) 15:24, 24 March 2011 (UTC)
- Actually I can perhaps see a glimmer of sense in this nutshell, since I suppose this is the most important practical everyday consequence of the policy; but even so, it seems misleading to present it as the summary, and it just causes unnecessary clutter at the top when exactly the same thing is repeated two paragraphs down. Also we seem to have reverted to saying "readers can check whether material has been published..." - I find this misleading in two ways (as said before): firstly because it's not absolutely essential that the sources we use be practically available for all readers to check (I suppose they should be in theory, but this makes it sound like everything should be a click away on the Web); and secondly because the material we create has not been published before - it only needs to be based on published sources (we know what we mean, but the uninitiated reader will be left guessing).--Kotniski (talk) 16:31, 24 March 2011 (UTC)
- Perhaps not, at least, not as a priority in the lead; but I think it's a good reason not to place so much emphasis on the "we don't care about truth" meme, as I think we do care. And I wasn't objecting to mentioning OR and NPOV in the lead - what I objected to was telling people "in order to avoid original research" and then effectively summarizing this policy - that makes no sense at all (unless we agree that this policy is just a fork of NOR, in which case they should certainly be merged). If we can achieve a sensible logical division between the three policies (which we certainly do not have at the moment - the whole thing's a jumbled mess) then certainly we should be summarizing the relevant points of the other two in this one - but surely after we've explained the key points of this one?
- You also mentioned that there are times we do challenge the accuracy of a claim (such as in an obvious journalistic error?). If so, I don't see anything in policy that gives literal license to do such a thing, aside from IAR and editorial discretion, and maybe Weight (if it's clearly disputed by other sources and not worth including by itself). Do you think this should be couched more directly? Ocaasi c 14:56, 24 March 2011 (UTC)
The threshold
Threshold. The word "threshold" implies movement and the beginning of a process. The valuation term "most important" suggests neither movement nor process. I have added the previous wording as a note because the word "threshold" is essential in the editing disputes which I have encountered -- see here; and such disputes are likely to continue to arise in articles about something to do with East Asia. --Tenmei (talk) 21:56, 24 March 2011 (UTC)
- I agree that "most important condition for inclusion" was not a good change. The writing isn't as good; it's three words instead of one; and it doesn't conjure up the imagery that "threshold" does. SlimVirgin TALK|CONTRIBS 22:17, 24 March 2011 (UTC)
- Please allow me to emphasize the arguable utility of this word, the prospective usefulness of a "threshold" concept which was restored here. --Tenmei (talk) 23:22, 24 March 2011 (UTC)
- My problem is not so much with the word "threshold", it's with the word "The". This could imply that the "verifiability" condition is a sufficient one (rather than a necessary one) for inclusion, which is certainly not the case. I don't like "most important condition" either, but I was trying not to change the wording too much. I would prefer, :as I said before, to phrase it simply and unambiguously: "All information must be verifiable from reliable sources..." If you insist on "threshold", then it should be qualified in some way, like "the first threshold", to show that there are other factors besides this. --Kotniski (talk) 11:12, 25 March 2011 (UTC)
- Not sure if this helps, but anyway. Obviously this point of logic is the same discussion that has been going on for some time, and I have recently started trying to understand your concern, and why others don't agree. I think that others are saying that in a sense they agree with your logic, and that this is only one threshold, the threshold concerning "material". That your logic is not obviously wrong suggests to me that this point should not be considered a closed case. But perhaps more importantly in a practical way they are scared to try to define that it is only for "material" because it would seem to weaken the policy somehow? So to some extent the resolution would seem to require a satisfactory wording being that says something like "Concerning X (e.g. "material"), the threshold for...." ??--Andrew Lancaster (talk) 11:44, 25 March 2011 (UTC)
- I don't think we are on the same wavelength... "material" is a very vague word that I would sooner replace with "information" or "statements" or something, but that's another topic; what I'm saying here is that material (or information or whatever) doesn't automatically have the right to be included just because it meets the condition of verifiability. That's the first condition it must meet (at least, once we've managed to word the condition correctly), but then there are other conditions - it must be relevant, notable, not undue weight, etc. There's no reason to assume readers will interpret "threshold" as meaning necessary condition; they're just as likely to read it as meaning sufficient, or necessary and sufficient.--Kotniski (talk) 11:59, 25 March 2011 (UTC)
- Yes, I understand your point that way. I think your logic is correct. It is a necessary but not sufficient condition, and a threshold only with respect to one type of issue. I am just trying to work out how to explain it to others.--Andrew Lancaster (talk) 12:23, 25 March 2011 (UTC)
- I don't think we are on the same wavelength... "material" is a very vague word that I would sooner replace with "information" or "statements" or something, but that's another topic; what I'm saying here is that material (or information or whatever) doesn't automatically have the right to be included just because it meets the condition of verifiability. That's the first condition it must meet (at least, once we've managed to word the condition correctly), but then there are other conditions - it must be relevant, notable, not undue weight, etc. There's no reason to assume readers will interpret "threshold" as meaning necessary condition; they're just as likely to read it as meaning sufficient, or necessary and sufficient.--Kotniski (talk) 11:59, 25 March 2011 (UTC)
- Not sure if this helps, but anyway. Obviously this point of logic is the same discussion that has been going on for some time, and I have recently started trying to understand your concern, and why others don't agree. I think that others are saying that in a sense they agree with your logic, and that this is only one threshold, the threshold concerning "material". That your logic is not obviously wrong suggests to me that this point should not be considered a closed case. But perhaps more importantly in a practical way they are scared to try to define that it is only for "material" because it would seem to weaken the policy somehow? So to some extent the resolution would seem to require a satisfactory wording being that says something like "Concerning X (e.g. "material"), the threshold for...." ??--Andrew Lancaster (talk) 11:44, 25 March 2011 (UTC)
- My problem is not so much with the word "threshold", it's with the word "The". This could imply that the "verifiability" condition is a sufficient one (rather than a necessary one) for inclusion, which is certainly not the case. I don't like "most important condition" either, but I was trying not to change the wording too much. I would prefer, :as I said before, to phrase it simply and unambiguously: "All information must be verifiable from reliable sources..." If you insist on "threshold", then it should be qualified in some way, like "the first threshold", to show that there are other factors besides this. --Kotniski (talk) 11:12, 25 March 2011 (UTC)
- Please allow me to emphasize the arguable utility of this word, the prospective usefulness of a "threshold" concept which was restored here. --Tenmei (talk) 23:22, 24 March 2011 (UTC)
Here is an idea. If I look at the opposed opinions, others do not see the word "threshold" as a logical breakpoint in a simple way, but the beginning of a process. Or maybe another way to word it is that they see it as the necessary and sufficient condition for words to be good enough for potential inclusion. Obviously people love the word threshold and they also want to avoid a big logical construction like "necessary but not sufficient condition for inclusion" or "necessary and sufficient conclusion for potential use given other conditions are met" so I am wondering if this observation about how people seem to be reading it can lead to a small tweak that covers all valid concerns. For example:
- A threshold for potential 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.
Does this make any sense to others?--Andrew Lancaster (talk) 11:28, 26 March 2011 (UTC)
Readers can check...
...readers can check that material in Wikipedia has already been published by a reliable source...: I think this should be changed to something like "material in Wikipedia can be supported by reliable, published sources", for reasons I've already given - readers' actual ability to check is not as fundamental a requirement as this makes it seem to be; and the material in Wikipedia need not (indeed usually should not) itself have been published before - what we mean is that it ought to be based on, supported by, published sources. Does anyone disagree, or was this revert collateral damage?--Kotniski (talk) 11:28, 25 March 2011 (UTC)
- It sounds correct to me.--Andrew Lancaster (talk) 13:10, 29 March 2011 (UTC)
- Sorry - which version sounds correct?--Kotniski (talk) 13:15, 29 March 2011 (UTC)
- If you switch from "readers can" to "wikipedia material can", you alter the meaning completely. The latter essentially says sourcing is optional which is a no-go.--Kmhkmh (talk) 13:18, 29 March 2011 (UTC)
- I'm not sure what you mean - the context of these words is that of a requirement, so it's being required that material be able to be sourced. Sourcing is "optional" in the sense that not everything has to be explicitly sourced; but the proposed wording still makes it clear that sourceability is not optional.--Kotniski (talk) 13:24, 29 March 2011 (UTC)
- If you switch from "readers can" to "wikipedia material can", you alter the meaning completely. The latter essentially says sourcing is optional which is a no-go.--Kmhkmh (talk) 13:18, 29 March 2011 (UTC)
- The only issue I have with "material in Wikipedia can be supported by reliable, published sources" does not emphasize that the source must already be published; articles that have been submitted but not printed yet, movies that are not yet in theaters, etc. are not acceptable. Perhaps ""material in Wikipedia can be supported by reliable, previously published sources" (new word underlined). Jc3s5h (talk) 13:18, 29 March 2011 (UTC)
- Sorry - which version sounds correct?--Kotniski (talk) 13:15, 29 March 2011 (UTC)
- It sounds correct to me.--Andrew Lancaster (talk) 13:10, 29 March 2011 (UTC)
Nutshell
Do we in fact want a nutshell? Do we want the current one? As I said above, though I can now sort of see why this one was chosen, it still seems to be misleading (by implying that this sentence sums up the whole policy, when what it actually does is state the most important practical consequence of it), and to provide clutter by simply repeating what's stated in the second paragraph of the lead.--Kotniski (talk) 11:32, 25 March 2011 (UTC)
- The main aim of policy writing must surely be informative, and that means informative to people who are clicking around on the internet and do not all have perfect attention spans. I think that nutshells help a lot in this task?--Andrew Lancaster (talk) 11:38, 25 March 2011 (UTC)
- Can't our attentionally challenged readers just as easily read the first paragraph or two of the policy (which should sum up the content) rather than having it duplicated in a nutshell? We don't put nutshells on our articles, after all. (What it would look like: Italy is a country with lots of art where they eat pasta; Adolf Hitler was a German who did some pretty bad things; etc.) But more to the point, does this particular nutshell serve any useful purpose?--Kotniski (talk) 11:46, 25 March 2011 (UTC)
- I am sure some of them can. :) The nutshell is even shorter.--Andrew Lancaster (talk) 12:21, 25 March 2011 (UTC)
- Can't our attentionally challenged readers just as easily read the first paragraph or two of the policy (which should sum up the content) rather than having it duplicated in a nutshell? We don't put nutshells on our articles, after all. (What it would look like: Italy is a country with lots of art where they eat pasta; Adolf Hitler was a German who did some pretty bad things; etc.) But more to the point, does this particular nutshell serve any useful purpose?--Kotniski (talk) 11:46, 25 March 2011 (UTC)
"Threshold" in a nutshell? This is a practical issue, not an abstract one. IMO, this thread wanders too far afield from lessons learned the hard way and the school of hard knocks. Yes, we can all acknowledge and agree that parsing distinctions between "necessary" and "sufficient" are valid concerns, but the "rubber meets the road" at a far more gritty level:
- The distinction between any sourced statement and a dispute informed by zero cited sources is a recurring problem across a range of articles.
- There is only evidence that these kinds of disputes are likely to increase in the future.
When "fact" (supported by WP:V + WP:RS) is defined as indistinguishable from "factoid" (supported by nothing), what next? This becomes an irreducible question, a shared "threshold knowledge" inquiry.
The words "threshold" and "verifiability, not truth" are married; and these words offer perhaps the only arguably constructive step forward. Have you not seen this for yourselves? --Tenmei (talk) 00:02, 26 March 2011 (UTC)
- Can you re-state your point please? I do not see how it contrasts with any other position. Or is it just an observation?--Andrew Lancaster (talk) 11:30, 26 March 2011 (UTC)
- Yes, sorry Tenmei, I don't see what point (if any) you're actually making about the wording of the policy.--Kotniski (talk) 12:07, 26 March 2011 (UTC)
- Yes, I have a point; and yes, my point is amplified by a counter-intuitive observation. I oppose any re-wording of which does not include these words:
- (a) "threshold" and
- (b) "verifiability, not truth".
This is a practical issue which affects day-to-day editing "where the rubber meets the road". This is a recurring problem and it is likely to get worse in the near future. The pointed clarity of words on this page will only become more important as our project continues to grow.
In the alternative, I support any arguable effort to enhance the effectiveness of WP:V as long as these few words are unaffected by editing changes.
At Wikipedia:Nutshell#Verifiability, our policy can be summarized succinctly: The threshold for inclusion in Wikipedia is verifiability, not truth. Maybe the simple act of repeating the same thing over and over again is always necessary and never sufficient? --Tenmei (talk) 21:31, 28 March 2011 (UTC)
- So you're saying the nutshell (i.e. the bit in the box that says "this page in a nutshell") should actually be the same as the first sentence of the policy? That seems to me slightly more logical (though just as redundant) - however as pointed out above, the actual wording of the statement is not quite right - firstly it's not "the threshold" (it's not the one and only condition), and secondly that in actual fact we do care about the truth of statements and don't mindlessly copy apparent errors from sources (as "not truth" implies).--Kotniski (talk) 09:48, 29 March 2011 (UTC)
- Kotniski -- Redundant? Yes, absolutely. Your editing experience is different than mine; but that in no way diminishes the fact that redundant repetition for pedagogical emphasis is needed in contexts I encounter. You seem unwilling to take in this point, but some things are significant regardless of belief.
As for your second point, no. Do not try to "spin" my words to contrive a straw man. Again, no matter how many different ways I re-write using different words, you seem unwilling to take in this point, but some things are both valid and significant despite your belief or disbelief. The WP:V threshold focuses irreducible attention on the pivotal distinction between fact (which supported by WP:V + WP:RS) and "factoid" (which is associated with zero cited support). Your "re-framing" addresses follow-up issues.
Perhaps this can be explained by our edit histories. Although my 50,000+ edits are more than yours, I see that the range of unique pages you have edited is a little more than twice mine. I do not want to impede anyone's attempt to enhance the effectiveness of WP:V; but four words only must remain unaffected by changes: "threshold" + "verifiability, not truth". --Tenmei (talk) 13:49, 29 March 2011 (UTC)
- I'm not entirely sure what you mean with this "spin" and "straw man" stuff - but you presumably agree that "verifiability" (as defined in this policy) is not the only condition for inclusion in WP? And that we don't want to reproduce statements from reliable sources if we know those statements to be untrue?--Kotniski (talk) 13:57, 29 March 2011 (UTC)
- Kotniski -- Redundant? Yes, absolutely. Your editing experience is different than mine; but that in no way diminishes the fact that redundant repetition for pedagogical emphasis is needed in contexts I encounter. You seem unwilling to take in this point, but some things are significant regardless of belief.
- So you're saying the nutshell (i.e. the bit in the box that says "this page in a nutshell") should actually be the same as the first sentence of the policy? That seems to me slightly more logical (though just as redundant) - however as pointed out above, the actual wording of the statement is not quite right - firstly it's not "the threshold" (it's not the one and only condition), and secondly that in actual fact we do care about the truth of statements and don't mindlessly copy apparent errors from sources (as "not truth" implies).--Kotniski (talk) 09:48, 29 March 2011 (UTC)
Copy-edit
SV, re this, look at the paragraph directly above where it says "published in a reliable source". Using "reliable published source" immediately afterward is redundant and missing an opportunity to start introducing readers to what our jargon actually means. Why don't we wikilink the first reference and replace the second with 'reputation for fact checking and accuracy' (or reputation for fact checking and accuracy that is appropriate for the claim being made...)? Ocaasi c 19:14, 7 April 2011 (UTC)
- That makes sense to me (we can still include the word "published", if desired, along with the bit about the reputation).--Kotniski (talk) 10:54, 8 April 2011 (UTC)
- Sorry, Ocaasi, I've just noticed this. I'm fine with your version, but I'd prefer to link to WP:SOURCES (part of V), [1] rather than to IRS, the guideline, so the policy doesn't look as though it's deferring to a guideline for a definition of "reliable source". SlimVirgin TALK|CONTRIBS 00:48, 9 April 2011 (UTC)