Wikipedia talk:Manual of Style/Words to watch/Archive 12

This is an archive of past discussions about Wikipedia:Manual of Style. 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.

Archive 5

←

Archive 10

Close reading of LABEL

I think we need to come to an agreement about the meaning of the current words in LABEL before we try to change them. This requires setting aside any notion you have of what LABEL ought to say (save that for later) and read what it does say. So reach back to your last English literature class, because we need to do some Close reading. Here's the text:

Value-laden labels...may express contentious opinion and are best avoided unless widely used by reliable sources to describe the subject, in which case use in-text attribution.

So we start with "Value-laden labels...may express contentious opinion". This provides a little explanation of why we're talking about this: the subject "may express contentious opinion" – notice that it "may", not that it "always does". We learn from this that the value-laden label, although always laden with values, may (or may not) express an opinion, and that if it expresses an opinion, it may (or may not) be contentious. (Conclusions: LABEL is not exclusively about opinions, and LABEL is not exclusively about contention.)

Next we get "and are best avoided". This puts a general don't-do-it on the subject, but avoidance is not the same as a ban.

The tricky bit comes next: "unless widely used by reliable sources to describe the subject, in which case use in-text attribution."

What does that mean? Well, what it says is that you can do use value-laden labels, including value-laden labels that express a contentious opinion, if and only if:

the label is widely used by reliable sources to describe the subject, and
you use in-text attribution.

This means that if you want to write a sentence like "Peter Duesberg was the most prominent AIDS denialist to [advise] then-president Thabo Mbeki of South Africa", then you can only do that if:

the label is widely used by reliable sources to describe the subject, and
you use in-text attribution.

For the first point, I believe editors have agreed that this label is indeed "widely used", and not merely a couple of weak sources that someone scrounged up.

For the second point, this particular sentence does not provide in-text attribution. Should it? There are no reliable sources that say Duesberg is not an AIDS denialist. What exactly would it look like to provide in-text attribution for the label "AIDS denialist" about someone who openly claims that HIV doesn't cause AIDS?

At the top of LABEL, "denialist" is given as an example of a value-laden label. So we can't finesse this by saying that LABEL doesn't apply.
This value-laden label is widely used, so the exception applies. We should include that label.
But the way LABEL is presently written, if we include the label, then we must also provide in-text attribution.

Are we all agreed that this is what LABEL's current wording say is necessary? Does anyone think that LABEL says (NB: not "is typically interpreted as") something different? WhatamIdoing (talk) 17:22, 25 January 2022 (UTC)

LABEL says that in-text attribution is required whenever a "label" is used. The "case law" on this of course is quite different - people justify this in various ways, but widespread, largely undisputed characterizations are typically presented in Wikivoice (mostly in situations where attribution seems at least somewhat absurd). Newimpartial (talk) 17:47, 25 January 2022 (UTC)

This is my experience, too. WhatamIdoing (talk) 19:35, 25 January 2022 (UTC)

This reminds me of your September 2019 thread In-text attribution and the April 2021 thread Widely vs in text attribution. It doesn't seem to me that most participants thought there should be significant change. As for your words "We should include that label.", well, no, I don't agree that WP:LABEL says that. Peter Gulutzan (talk) 17:51, 25 January 2022 (UTC)

Fair enough: I should have written that when the exception applies, we "are permitted" to use that label. It is WP:DUE and WP:FRINGE that says the most prominent AIDS denialist in the world "should" be described as such. WhatamIdoing (talk) 19:37, 25 January 2022 (UTC)

There was basically no participation and it hadn't been announced at a noticeboard. But my main point in that older thread was that this MOS part makes suggestions that are contrary to more established policy that is already more detailed. —Paleo Neonate – 17:19, 13 February 2022 (UTC)

Setting aside any notions of typical usage, I completely agree with your reading here. It seems the text of LABEL establishes a guideline which applies to value-laden labels (regardless of whether they are contentious or whether they express an opinion) and provides a decision pathway as follows:

♢ Is a value-laden descriptor widely used by reliable sources to describe the subject?

↳ If so: use in-text attribution.

↳ If not: usage [is] best avoided.

If anyone disagrees I would be very interested to hear their interpretation. Srey Sros^talk 17:57, 25 January 2022 (UTC)

I agree that this is the correct reading of what MOS:LABEL actually says. WP:INTEXT seems to be a better explanation of current practice and what people think MOS:LABEL is/should be and would be a good starting point for a rewrite. –dlthewave ☎ 18:50, 25 January 2022 (UTC)

I generally agree with WhatamIdoing, SreySros, and Dlthewave's reading of the guideline. I think the critical ambiguity in LABEL’s current wording is its muddled scope. The title is "Contentious labels", but the text refers to "value-laden labels", which is quite different. A label can be value-laden or neutrally descriptive, and independently can be contentious or non-contentious in its application. All combinations of the above are possible:

value-laden, contentious: "Theresa May was a great prime minister."
value-laden, non-contentious: "Genocide is evil."
descriptive, contentious: "[Moderate social democrat] is a far-left politician."
descriptive, non-contentious: "[Member of a communist party] is a far-left politician."

Read as a whole, I think LABEL applies to 1 and 2, but not to 3 or 4. Regardless of whether they are contentious, value-laden labels should only ever be used with in-text attribution, and only if widely used in sources. That is fully consistent with WP:VOICE (i.e., avoid stating opinions as facts). However, applying the guidance in LABEL to non-value-laden descriptors doesn’t work. For those, we will use them in Wikipedia’s voice if they are generally used by reliable sources to describe a subject (i.e., avoid stating facts as opinions).

In short, the title of LABEL is not consistent with the text, and some of the examples are more descriptive than value-laden, but the body of the guideline is reasonably clear and consistent with governing policies.--Trystan (talk) 19:21, 25 January 2022 (UTC)

Contentious and value laden are being used for a mixture of "definition dependent" and "potentially involving BLP and NPOV problems". Duesberg being an AIDS denialist is not definition dependent. Describing him as a crank is value laden and BLP applies. Anita Bryant being anti-gay-rights is not definition dependent (even among those denying those are human or civil rights, i.e., "gay rights" is an accepted term for the position she opposes), calling her "anti-gay" is contentious. Sesquivalent (talk) 02:14, 26 January 2022 (UTC)

I agree with you on that's what it says and mean and also agree with the implication that the last bit regarding in-text attribution inherently conflicts with the "widely used by reliable sources" part. If it's widely used by reliable sources, then other than attaching those references as references, any in-text attribution would be reducing the meaning and presentation of that wide use. Because you would be attributing it to a specific use in a newspaper or book or scholarly article or whatever. Which would not be properly showcasing that wide use in reliable sources. And any attempt to be vague in saying "Most sources" just runs up against WP:WEASEL problems. The in-text attribution part is very clearly what doesn't work here. Silver seren^C 22:06, 25 January 2022 (UTC)

I agree, attribution can be used to cast doubt on factual statements which is something that WP:INTEXT warns against. It would be appropriate for widely-reported opinions such as the third paragraph of Proud Boys#History and organization where we attribute "alt-right fight club", "hate group", alt lite", "overtly Islamophobic and misogynistic", "transphobic and anti-immigration" etc. –dlthewave ☎ 17:26, 26 January 2022 (UTC)

Let me flip it around to present the issue that we have now, too often which is cherry picking of labels from a few sources but not reflecting a broader segment of sources available for a target. While these are used with attribution, they do not reflect the "widely used" part, but editors are often too far into trying to vilify certain people or topics to care about that and pick whatever they can find as long as it comes from at least one RS. That's definitely not what LABEL says. Labels should only be used if they are the type of thing you can't help but stumble across in doing research on a topic, and as long as they are coming from quality RSes. Otherwise, we're going out of our way to try to include said labels which is a POV problem. --Masem (t) 01:15, 26 January 2022 (UTC)

I think we all agree on the "widely used" part.

Do you agree that LABEL, as written, requires us to provide in-text attribution for all value-laden labels, even when those labels are actually, truly widely used? WhatamIdoing (talk) 01:47, 26 January 2022 (UTC)

Yes... but there are ways to summarize the attribution when they are widely used. "X has been called Y by sources, including RS1, RS2, and RS3", or "X has been broadly considered a Y by the media" and using three or four of the best sources for that - even possibly as a groupped referenced leading "For example". What I know we don't want is to have to list out 20+ sources to support that type of statement, but we do want to give the impression that even though we may be only naming 3 or so, this is not the extent of those sources. --Masem (t) 02:20, 26 January 2022 (UTC)

If you write that Duesbergy "has been broadly considered an AIDS denialist by the media", someone's going to demand that you produce reliable sources that uses words like "broadly construed by the media", not merely sources that actually construe him that way. If you say "He's an AIDS denialist", they will not require you to produce sources that say that other sources say that. WhatamIdoing (talk) 02:39, 26 January 2022 (UTC)

However, you are now speaking a value-ladened label in Wikivoice which is absolutely against NPOV. (There is something to say that some decades after the person has fallen out of the spotlight or has died and RECENTISM would not apply that we could be less concerned about that, but I'm assuming we're talking current coverage). There needs to be some type of wording to bring the statement out a factual claim. --Masem (t) 03:43, 26 January 2022 (UTC)

I don't think that's true. WIKIVOICE has five points. Compared to that list, this statement:

is not an opinion;
is not seriously contested (even by Duesberg and his supporters);
doesn't present this fact as if it were an opinion (on the contrary, labeling it as only what the media says could have the effect of making it sound like an opinion);
is "judgemental" only to the extent that is reasonably necessary for clarity; and
is the only viewpoint held by any reliable source (including Duesberg and his supporters, who are proud of him for denying that HIV has any connection to AIDS).

What would really violate NPOV is trying to hide the facts by dressing them up in some wishy-washy euphemism ("an innocuous word or expression used in place of one that...suggests something unpleasant", assuming that we agree that being famous for promoting a particularly deadly form of pseudoscience is considered something "unpleasant" to say about a person).

(Duesberg's been an AIDS denialist for the last 35 years, so you don't need to worry about RECENTISM in this example.) WhatamIdoing (talk) 04:04, 26 January 2022 (UTC)

The issue around points #3 and #4 in general for labels. Labels are, by their nature, contentious, even if there are no reliable sources that describe the opposing view, so we should never treat labels as fact, at least in the short term. Additionally, labels are 100% judgmental more so in the current social climate and ways the media tends to report on things. --Masem (t) 04:21, 26 January 2022 (UTC)

I do not agree that labels are inherently contentious. I think we should treat some labels as facts, even in the short term. I do not agree that labels are 100% judgmental.

Maybe you're using a different definition of label than the guideline (or most dictionaries, AFAICT).

This guideline is addressing specifically "value-laden labels". The fact that there's such a thing as a "value-laden label" means that there is also such a thing as a "non-value-laden label". Labels such as student, worker, politician, musician, scholar, and parent are all useful words for labelling people, but they are not inherently value-laden. WhatamIdoing (talk) 00:41, 30 January 2022 (UTC)

I've been following this thread and the section above, and for what it's worth, I agree this is a huge problem with the current wording of the page. To adapt what someone said above, if the only policy currently allowing us say the Nazi Party was anti-Semitic or that the American Nazi Party are neo-Nazis is WP:IAR, this guideline has to change. I mean, strictly speaking, the WP:NPOV policy to "avoid stating facts as opinions", instead "directly stat[ing them] in Wikipedia's voice", overrules this guideline, but as SMcCandlish has said about other situations, it's better to revise guidelines to not directly contradict policies (than to leave the contradictory text and rely on everyone to continue to de facto ignore it)... -sche (talk) 04:32, 26 January 2022 (UTC)

This is why I think RECENTISM is important. When labels stand the test of time and academic rigor (as in the case of calling the Nazi Party as anti-semitic) then we shouldn't have to worry about attribution/speaking out of wikivoice. But we don't want to be throwing the same type of labels at a person or group that just has become an issue in the news. When the RECENTISM aspect goes away, I don't know, but I think it's definitely on the order of decades. --Masem (t) 04:52, 26 January 2022 (UTC)

Which would then make it impossible to label any of the fringe pseudoscience topics as pseudoscience or similar terminology, regardless of all the sources and major scientific organizations calling it such, because it would be recent. That doesn't seem like the right way to go at all, Masem. Silver seren^C 05:02, 26 January 2022 (UTC)

"Fringe" in its recent use in Wikipedia plays the same role as "national security", a semantically flexible trump card that can be weaponized to override other principles previously regarded as constitutional foundation. The possibility that exploring and applying those other principles ends up inhibiting some of the labelling as fringe, conspiracy, and pseudo is not an argument against them. The current language is using contentious and value laden to mean some slightly different things, and I think the OP is right that this should be clarified, but by no means should that be done by working backward from some all controlling uberprinciple of having to Protect The Children from the Dread Menace. Sesquivalent (talk) 06:23, 26 January 2022 (UTC)

I also said "academic rigor", which is a lot faster to be shown through for topics that have objective means of evaluation, like demonstrating that something is pseudoscience. Ideological labels have less objective means to characterize and thus require time more than anything. --Masem (t) 20:39, 26 January 2022 (UTC)

But (and this is often a sticking point) there is extensive high-quality academic sourcing discussing modern white nationalists and white supremacists, as well as holocaust denialists and climate-change denialists and so on and so forth, covering most of the labels listed on this page. Is there a point where you would concede that sufficient sourcing exists to describe currently-active real-world groups and even, sometimes, individuals using those descriptors in the article voice, without attribution? Because I feel that once that point is conceded (and such usage clearly reflects current policy and practice, the guidelines on this page notwithstanding), the current wording of WP:LABEL falls apart and definitely needs revising - at that point WP:LABEL is a mere appendage to WP:EXCEPTIONAL (and, for living people, WP:BLP) that adds little to either beyond suggesting a few words that might be considered exceptional and which therefore require strong sourcing to use in the article voice. --Aquillion (talk) 07:28, 29 January 2022 (UTC)

At the end of the day, a lot of this needs to be common sense about how frequently and from what type of respectable sources that the label is being used. There are editors out there that want to disparage the names of BLPs and other topics, and have cherry picked labels from RSes to include in articles, and that's simply not appropriate, and that's what LABEL tells us not to do. As editors we should not be going out to say "we need to classify this topic as a (value-laden label)", and only conclude after doing research that it was near impossible to avoid that label when reading sources so it is an appropriate term to include and attribute to some degree. Otherwise, wikieditor bias is going to come into play in how that label will be introduced, and this I've seen all over the place, and both in negative and positive connotations. There's a natural tendency to want to vilify people that we think are bad, or to elevate people we think are good, but the whole point of NPOV is to write dispassionately and put those aside, and LABEL is meant to support that. So the problem with labels usually becomes that people include them without doing a survey of sources fire and then have to back up with showing that the label is widely used, rather than doing the survey of sources first to prove the label is actually justified. --Masem (t) 14:04, 29 January 2022 (UTC)

"Common sense" is not how I'm feeling about the way this section is written. Common sense says that when someone is (really, truly, actually) widely labeled as an AIDS denialist, you say that he's an AIDS denialist, in wikivoice, per WP:WIKIVOICE. You don't try to soften that by saying "Alice and Bob and Chris and some other people, who are probably all meanies, called him this value-laden derogatory label just because he actually, publicly, repeatedly, unashamedly, and self-promotionally denied the connection between HIV and AIDS". WhatamIdoing (talk) 22:21, 29 January 2022 (UTC)

I agree that we need to use common sense; but that is best achieved by making the requirement for attribution a suggestion rather than a requirement, since that allows us to use common sense in individual cases. And while it is certainly true that there is a danger of people with strong feelings about a topic violating NPOV by presenting something that the sources only justify as an opinion as if it were a fact, there is equally a danger of people with strong feelings about a topic introducing their personal biases by framing a fact as a mere opinion. If high-quality sources universally describe someone as an "AIDS denialist" in their article voice, and an editor who is partial to the subject insists that WTW requires that we downgrade that to "several scholars describe them as an AIDS denialist" (or, worse, "the New York Times describes them as an AIDS denialist" when in fact the descriptor is used universally), that is also introducing bias and is something we need to constantly push back against. This is sort of bias (editors overtly trying to downplay facts they disagree with, are uncomfortable with, or which they personally doubt the sources on based on their own gut feelings rather than anything well-cited) is absolutely something I have seen all over the place myself, and it is extremely important that WTW not provide any support for that sort of WP:TEND editing. --Aquillion (talk) 22:35, 29 January 2022 (UTC)

(this is effectively to both WAID and Aquillon above) However, my experience is that when push comes to shove and an evaluation of what sourcing supposed "widely labeled" proves out that that really isn't the case and that the label is far less than widely used than postulated. Again, in the ideal setting, we'd discover what labels are widely used during the course of research and thus come to use them in that manner, but in reality, editors want labels to apply and try to use research to justify them, which is what WP should not be doing. I absolutely agree that if a RS survey shows the label in use a high percentage of the time in non-opinion articles on the topic from high quality sources, that's far far different from the case where one or two sources solely use the label. But in both, we need to have editors prove out that reasoning on the talk page to show they've done the legwork and the sourcing supports their stance, rather than the reverse to assert the label applies and make others point out the problem. --Masem (t) 23:16, 29 January 2022 (UTC)

but in reality, editors want labels to apply and try to use research to justify them, which is what WP should not be doing Again, you are presenting this as one-sided, which is absolutely not the case - there are at least as many editors who want to downplay facts that they disagree with, even when the sourcing is overwhelming, and who will therefore continue to argue against inclusion no matter how overwhelming or widespread the sourcing is. Obviously per WP:QUO anyone who wants to make a contested change to an article must obtain consensus for it (which includes establishing consensus on talk by doing the legwork, whether that is to establish that there is widespread sourcing for a new addition or to overturn an existing implicit consensus that there is sufficient sourcing that presents longstanding text as fact), and for WP:BLPs we default to removal; but WTW must acknowledge that when there is a consensus (status-quo or otherwise) that a particular label is well-cited as a statement of fact, then it must be presented as such in the article voice, and cannot be downplayed with an attribution. Otherwise, POV pushers will constantly point to WTW to remove or downplay factual statements they disagree with, even when the citations are impeccable. Setting a one-sided standard (the way you seem to be suggesting) encourages people who want to downplay or omit facts that they disagree with to WP:STONEWALL discussions rather than do their own legwork for changes they are proposing or answer legitimate arguments for including facts that they disagree with. --Aquillion (talk) 23:31, 29 January 2022 (UTC)

I agree with Aquillion: We should not write a "known wrong" rule just to rein in bad actors.

As for whether it happens: Of course it does. I have certainly seen plenty of efforts to whitewash bad behavior by claiming that "only" mainstream media believes this or that, and that anyone looking "all the relevant sources" (i.e., extremist chat forums) would discover that the label isn't "widely" used at all.

The solution to POV pushing isn't to obscure the subject by refusing to label anyone. The solution to POV pushing is to accept the mainstream viewpoint as being the mainstream viewpoint, and to accept that this will sometimes mean that we have to do some work to figure out whether a label is, in fact, as widely held as some editors claim, just like we accept that we sometimes have to do some extra work to figure out whether a label that is allegedly from a tiny minority actually is a minority viewpoint.

Also: "widely held" is not the same as "universally held", and holding a view does not mean that the exact terminology must be used in every single source. It is possible for Duesberg to be "widely held" to be an AIDS denialist even though some reliable sources don't happen to mention those exact words, or even if some reliable sources don't mention the subject at all. If he were the victim of a car wreck, one might not necessarily expect a local news report to mention his notorious pseudoscientific views in a report about the cause of last night's traffic jam. Not everything is relevant to every source's subject. WhatamIdoing (talk) 00:23, 30 January 2022 (UTC)

I agree with Masem and disagree with Aquillion. Consider the opposing issues here, a laxed LABEL guideline means that we have people labeled as racist, haters, etc undeservedly so per NPOV because an editor adds the label based on their POV and the ability to find at least some supporting sources. Even worse is if, due to cytogenesis a report in the "real world" follows our lead and applies that contentious label in their own report because, well there was "consensus" to do so on the worlds largest encyclopedia. So what's the harm in erring on the side of avoiding labels? We can still present the facts people used to support the label. We can describe the actions rather than the labels. Not long ago we had a discussion similar to this one and it was pointed out that the intro for Adolf Hitler makes it very clear he was racist without ever actually calling him racist. Another reason to err on the side of avoiding labels is it makes our article's more fact driven and less gossipy tabloid. People who want to say bad things about people they don't politically agree with can still do so but would, god for bid, have to do so by presenting the evidence that underlies the labels rather than the often lazily applied labels. In summary, we can get the same information across without labels that may be applied by those (writing here or in RSs) who are not careful and/or neutral in their use. Adding them when they are undeserved does cause harm, not adding them and instead adding the supporting evidence doesn't cause harm. Springee (talk) 02:59, 30 January 2022 (UTC)

A different way to view this: there should be sufficient onus by editors that want to use a label without attribution by showing a source survey and getting consensus on the talk page for that -- once that's established, then 1) you have demonstrated for future editors what the sources are and why there's no need for attribution, and 2) you provide a pointer to those that want to whitewash that information or the like. But there's still should be an onus that to use such value-laden labels without attribution to prove that there's the more than sufficient amount of sourcing to back it up -- otherwise the label must be attributed at the bare minimum. But there's also factors related to RECENTISM, YESPOV and other factors to take into account. --Masem (t) 03:14, 30 January 2022 (UTC)

Masem, it seems like we're all in agreement on the basic principles: Labels that aren't widely used need to be attributed or left out altogether, and ones that have been shown to be widely used among reliable sources may sometimes be used in Wiki voice depending on the circumstances. However the current text of WP:LABEL requires attribution even if the label is widely used. I think the next step will be to discuss ways to rewrite the section to reflect the intended meaning. –dlthewave ☎ 05:00, 30 January 2022 (UTC)

If I'm not mistaken LABEL says this should be done for "contentious" labels. I don't think we need to use in text attribution for a label like "Senator". Springee (talk) 11:43, 30 January 2022 (UTC)

No, "Senator" wouldn't be contentious, but WP:LABEL currently requires attribution for factual descriptions like "cult", "terrorist" and "myth" even when well-supported. –dlthewave ☎ 13:09, 30 January 2022 (UTC)

Those aren't necessarily factual descriptions though, or at least in the same way. Eg: "terrorist" requires that the crime be defined as "terrorism" by the government, not by the media or other agencies covering it. An alternative medicine idea should only be labeled pseudoscience once MEDRS-meeting sources call it out that way. etc. This is the area of time/academic rigor that we should absolutely be waiting to make sure exist and/or has passed before making these terms into fact, and in the interim treating them as labels with attribution. --Masem (t) 14:23, 30 January 2022 (UTC)

That is not true. LABEL requires that a label such as terrorist be widely used by "sources". LABEL does not care whether that label is widely used by the government, in scholarly sources, or in news media.

LABEL says:

Not widely used? Then don't use that label.
Widely used? Then use (probably misleading) in-text attribution (so readers will think that it's not actually a widely used label, but instead is only used by a handful of named sources, or by vague groups like "mainstream media").

This is bad. This goes against the principles of NPOV and the rules at INTEXT.

What we should be saying is closer to:

Not widely used? Then don't use that label.
So widely used that an article fully complying with NPOV and INTEXT would use this label without in-text attribution? Then use it, with no in-text attribution.

There could be a footnote that says for borderline cases, in which the label is used very frequently but not widely, or in a situation that is relevant to the article (such as "This celebrity feud was sparked by Bob calling Alice a name"), editors can agree to use the label with in-text attribution. WhatamIdoing (talk) 16:40, 30 January 2022 (UTC)

Certain terms like "terrorist" are tied to legal factors that we cannot let non-governmental sources make the determination about. Terrorist attacks are typically tried and sentenced far more aggressively than similar crimes that did not have terrorism motives, so it would be absolutely inappropriate to use non-government claims of terrorism on WP. Similarly, we should not use non-MEDRS sources that assert that certain alternative medicine is pseudoscience. (In both cases, we can use RSes that repeat statements made from appropriate sources about these terms, eg the NYTimes reporting the govt' statement that a crime was terrorism-related). But that's not the case of all labels, but a very specific subset. --Masem (t) 17:15, 30 January 2022 (UTC)

No. Terrorist is not a word that's owned by the government (which government, by the way?). I can't imagine why you would even think that a press release from an elected district attorney (which is a governmental source) is better than a peer-reviewed article in a scholarly journal about terrorism (which is a non-governmental source). You must not have thought this idea through completely. Your story leads to acts of terrorism not being even theoretically possible until a law was created that used that word.

We absolutely should not require MEDRS's ideal to declare that an altmed product is pseudoscientific. It absolutely should be possible to describe notable subjects even if they don't get mentioned in medical school textbooks or multiple journal articles.

Your views on this subject seem to be very different from most editors'. WhatamIdoing (talk) 17:26, 30 January 2022 (UTC)

We absolutely need to rely on the expert or authoritative sources here, otherwise, we get back to cherry picking in regards to labels that could have legal implications on BLP or related articles, or where laypeople are using science they don't understand to try to pick apart someone's research. But this again gets back to short-term reliance on mainstream media to apply labels rather than following RECENTISM and waiting for long-term, more reasoned-out sourcing that is better quality and reflective of what an encyclopedia should be using to report on these things. --Masem (t) 06:04, 31 January 2022 (UTC)

My largest issue that is while I agree there is a point where contentious labels are used so much by reliable, high quality sources that we can use them without attribution (Alex Jones as a conspiracy theorist being a good example), when that point occurs needs to be a really really high bar that should be demonstrated on talk pages first before implemented. If that bar can't be shown, or hasn't been demonstrated, then the label should be attributed, period. --Masem (t) 14:27, 30 January 2022 (UTC)

According to LABEL, when that bar can't be cleared, then the label shouldn't be used at all.

The problem with LABEL as written is that when that bar is cleared, no matter how amply, it says that in-text attribution is necessary. LABEL as written says that Alex Jones' status as the promoter of conspiracy theories is something that requires in-text attribution.

What you say here aligns with what everyone else is saying about LABEL needing to change. WhatamIdoing (talk) 16:42, 30 January 2022 (UTC)

But that isn't a hard problem to overcome. Instead of using Wiki voice, "Mr Smith is a racist", we can use a general attribution, "Mr Smith is widely described as/considered to be/etc a racist". That takes it out of Wiki voice by giving a general attribution which can be supported with select references. Springee (talk) 17:22, 30 January 2022 (UTC)

If you want to write that someone or something is X, you need sources that say the subject is X.

If you want to write that someone or something is widely described as X, you need sources that say the subject is widely described as X.

There are far more sources that comment on a subject's X-ness than sources that comment on what other sources say about the subject's X-ness. WhatamIdoing (talk) 17:30, 30 January 2022 (UTC)

Also, NPOV says to actually use wikivoice for many such statements. NPOV does not allow "Smith has been described as X" when all or nearly the sources that mention Smith's relationship to X describe him as X. Sometimes Smith really is X according to sources, and encyclopedias should have the courage to say it forthrightly when that's the case. WhatamIdoing (talk) 17:33, 30 January 2022 (UTC)

"encyclopedias should have the courage to say it forthrightly when that's the case." This is the absolutely last thing a neutral work should be trying to do, we are meant to describe conflicts, not get involved. And that sometimes includes taking a side when one side happens to be mainstream reporting that is considered reliable in the short term but doesn't have the impact of time and disconnected reviews from other works that comes from the long term. This attitude is the central problem of wanting to vilify topics that are in the news today, when we really should be waiting to see how they are viewed decades from now before getting into that type of consideration. This is why RECENTISM is extremely central to all this discussion around labels. Once RECENTISM no longer applies, its also less likely that attribution with regards to widely-used labels also applies. --Masem (t) 05:59, 31 January 2022 (UTC)

Accurately summarizing the reliable sources is not "getting involved". If all the reliable sources say that 9/11 attacks were terrorist attacks, then we aren't "getting involved" by saying that the 9/11 attacks were terrorist attacks. We should write a neutral, NPOV-compliant statement that "The 9/11 attacks were terrorist attacks" and not pretend otherwise with some weak circumlocution like "Most reliable sources say that the 9/11 attacks were terrorist attacks (but Wikipedia is not taking sides in this legitimate debate, of course!)". WhatamIdoing (talk) 02:29, 2 February 2022 (UTC)

And if describing a person as X really is DUE, then it should be trivial to find sources directly supporting the idea of 'widely described as X'. Crossroads ^-talk- 05:37, 31 January 2022 (UTC)

This guideline isn't solely about describing people. Consider how you would apply labels to subjects like Time Cube. WhatamIdoing (talk) 02:23, 2 February 2022 (UTC)

Also, @Crossroads, I wonder if you could find me a source that actually says, in direct and unmistakable words, that other sources say that chemotherapy is a treatment for cancer. NB that I'm not looking for sources that define chemo; I only want to see a source that talks about other sources defining chemo that way. A medical school textbook that says something like "Chemotherapy is widely described as being a treatment for cancer" would do very nicely. Do you think you could find one? Just one? WhatamIdoing (talk) 02:32, 2 February 2022 (UTC)

My argument doesn't apply to non-opinion labels or claims, so chemotherapy has nothing to do with it. And the Time Cube article contains no unattributed LABELs at all, and is much better for it. Imagine how ridiculous it would look if instead we described it as Time Cube is a fringe[1][2] pseudoscientific[3][4][5] and racist[6][7] conspiracy theory[2][5][8][9][10]... Some of the leads on Wikipedia right now almost look like this, even with LABEL the way it is already. Crossroads ^-talk- 04:52, 3 February 2022 (UTC)

Fringe, pseudoscientific, racist, and conspiracy theory aren't merely "opinions". When someone writes that there is a global conspiracy to prevent anyone from finding out about his ideas, that is objectively a conspiracy theory. When someone writes that racial integration destroys the races, then that is a racist statement, not merely someone's opinion or personal interpretation. There are times when interpreting actions might be a matter of opinion (Did that man avoid looking at a person of a different race because he's overtly racist, or because he hasn't had his coffee yet and doesn't want to look at anyone? Different people could form different opinions about that), but when you actually write, in plain language, that racial integration will destroy the white race, we are beyond the point of "opinions" and firmly into the territory of facts. That is a racist statement.

Also, Time Cube isn't "fringe", because that label suggests that it is on the outskirts of real science, rather than being the paranoid and grandiose thoughts of an old man with schizophrenia.

I think it would be a good idea to remember that the formula of "X is a Y-ist Z" is not the only way to label a subject. WhatamIdoing (talk) 17:11, 3 February 2022 (UTC)

I think the problem with those terms is they have at least some level of subjectivity. When does a pond become a lake, when does a mound become a hill becomes a mountain. At the extremes we have no trouble defining Victoria or Superior as a lake vs a pond. At the same time a bristle when a new subdivision calls their retention pond a lake (lake front home sounds better than pond front home). Of course, a developer has an incentive to call their retention pond a lake just as someone who labels someone else a "racist" might be motivated by creating/pushing a narrative rather than just stating facts. Springee (talk) 17:31, 3 February 2022 (UTC)

There are borderline cases, but there are also clear-cut cases. Presumably in borderline cases, we will see sources disagreeing with each other, or only a small number of sources asserting a viewpoint. The existence of borderline cases should not blind us to the existence of the clear-cut cases. Editors should be using their best judgment to determine which is borderline and which is clear-cut (based on the sources, not their personal interpretations). They should not be saying "Well, I think racist is always a subjective opinionated label with no basis in objective facts, so every single use of this word needs to follow LABEL, even when that would contradict NPOV and INTEXT". WhatamIdoing (talk) 18:35, 3 February 2022 (UTC)

Split

"Labels are, by their nature, contentious..." What do you consider a label? Using the definition of "a word or phrase that describes or identifies someone", applying labels is core to WP's mission of describing and identifying its subjects for the reader. Most labels are not contentious — at least not when widely supported by reliable sources (French, Catholic, gay, politician...). Some labels are value-laden, in that they inherently impart judgment, which we avoid to stay on the right side of the is–ought problem. Others will have negative connotations among large groups of people (convicted murderer, hijacker, Holocaust denier), but are at their core factual claims. All we can do is follow the sources. Editors cherry-picking to advance a certain view will always be a problem, but the solution has to be upholding WP:DUEWEIGHT and crafting an article that reflects what all the sources say collectively. If reliable sources as a whole support describing someone as an AIDS denialist, as a factual description of someone who denies that HIV causes AIDS, it is not compatible with NPOV to substitute a euphemism, impose an arbitrary years-long cooling off period, or require in-text attribution.--Trystan (talk) 14:53, 27 January 2022 (UTC)

This seems to be the least intrusive place to make a point I have made in other, related discussions: that a head-count of sources not using a particular term/label should never be used as evidence against the appropriateness/prevalence of that term/label, except for sources that actually use a term editorially that logically contradicts the term/label under consideration or, even better, actually disputes the term/label either editorially or through the presentation of contrary evidence. I have far too often seen editors present head counts of sources that don't use a particular term/label as though that were evidence that the application of the term/label were controversial or contested in a particular case, which is not a conclusion that can be deduced by simple absence of a term. Newimpartial (talk) 02:52, 30 January 2022 (UTC)

Suppose for a BLP 5 sources refer to them as "homophobic", while 50 talk about them without using any such label in their own voice. Is this a good reason to label a person? Sources that are about someone but do not use a term absolutely matter. This is the principle of WP:DUE. Crossroads ^-talk- 06:54, 30 January 2022 (UTC)

It depends. Are those 50 talking about their social or political views? Or are those 50 talking about subjects unrelated to that? Do any of those 50 claim the opposite? It's one thing to balance "He's homophobic" against "He's got a mixed track record for public support of gay rights". It's another thing to claim that "He's homophobic" is offset by "He directed a film about the Vietnam War."

How does the passage of time affect sources? Nobody was an AIDS denialist in the 1970s, because nobody knew that it existed. Would you look at sources from the 1970s and say "Look, I found 50 sources about Duesberg from before the discovery of HIV that talk about him without mentioning AIDS at all, so he's not really, truly, absolutely, 'widely' described as an AIDS denialist"? Of course not. But you could probably find 50 old sources that mention Duesberg without mentioning AIDS.

You can't decide whether a label is widely used until you look at all the facts and circumstances. This is the sort of thing that editors have to settle article by article, label by label.

As for DUE, the NPOV policy says that when a label is widely used, it should be used in wikivoice, without falsely ascribing it to only a small number of named sources. LABEL says the opposite: that when a value-laden label is used widely, it should be falsely given in-text attribution to a few representatives of that wide use. WhatamIdoing (talk) 16:54, 30 January 2022 (UTC)

Attribution can take the form of "widely described as", and if that is true, then it will be easy to find sources directly saying that. Anyway, let me rephrase my point: Suppose for a BLP 5 sources refer to them as "homophobic", while 50 talk about them in context of their LGBT-related views or actions, without using "homophobic" in their own voice. Contra Newimpartial, these sources choosing not to use the term do bear weight when it comes to whether we use it. Crossroads ^-talk- 05:37, 31 January 2022 (UTC)

I agree that all the sources addressing the subject (e.g., all sources talking about adoption rights for same-sex parents), whether by:

using the word homophobic (or very similar words, e.g., homophobia) to describe opposition to civil rights for gay couples
describing a viewpoint that matches the word homophobic but using only other words (e.g., gay-bashing, discrimination against gay people)
describing the opposite viewpoint that disagrees but using other words (e.g., supporting traditional family structures)
using the word homophobic (or very similar words) to deny that opposition to gay rights has any connection to homophobia

should be taken into account. I do not think that sources unrelated to the question of whether opposition to civil rights for gay couples is homophobic (e.g., if there were a source about how politicians voted on a particular law about adoption) should have any weight in this particular question.

I do not agree that if editors are looking at one source saying "homophobic", two sources saying "gay-bashing", and one source that says "traditional family structure", that we have to go with some sort of misplaced majority rule and use gay-bashing in the article. There are many considerations in deciding how to describe a subject, and those include not merely the raw popularity of a word in sources, but also encyclopedic tone (homophobic is more formal than gay-bashing), avoidance of euphemisms (traditional family structure is a modern euphemism for the heterosexual nuclear family; a multigenerational family is the actually-according-to-actual-historians "traditional" family structure), educational value (is homophobic too jargon-y for this part of the article?), and more. WhatamIdoing (talk) 02:51, 2 February 2022 (UTC)

It is important to point out that WP:INTEXT is policy, while MOS:LABEL is just a guideline, so if the two contradict then INTEXT takes precedence (and we should probably start a discussion to resolve the difference by updating the text of one or the other.) I would strongly favor tweaking WP:LABEL to tone down its insistence on in-text attribution to a suggestion, since I feel that always requiring in-text attribution, simply because a word is listed on WP:LABEL, even in situations where it is uncontentious in that specific context, plainly violates WP:NPOV's prohibition on not portraying uncontested facts as opinions. I think the underlying issue is that MOS:LABEL tries to assert that certain words are always and without exception contentious in all circumstances, which is an unworkable standard - if the sources completely unambiguously treat the use of a term listed on WP:LABEL as uncontentious in a particular context, WP:NPOV requires that we do the same and therefore that we use it in the article voice without attribution. I would suggest updating MOS:LABEL to make in-text attribution a suggestion rather than a requirement and to specifically reference WP:INTEXT as the policy that should be used to determine when it is appropriate, as well as perhaps acknowledge the WP:NPOV requirement that if something is clearly uncontested fact then it cannot be presented as opinion. The current wording of MOS:LABEL has lead to the nonsensical situation where people who want to maintain its current wording and application seem to be making the patently absurd argument that "Nazi" or "Fascist" or "anti-Semitic" are not value-laden labels because conceding otherwise would make the conflict between the current requirements of MOS:LABEL and WP:NPOV irreconcilably obvious; but the basic problem doesn't go away either way. --Aquillion (talk) 07:07, 29 January 2022 (UTC)
WP:PGCONFLICT, which is also a policy, tells us that when advice pages conflict, the correct answer is to make them match. Otherwise, POV pushers will declare that LABEL prohibits common sense and presenting mainstream views as being the mainstream views, and we're relying on everyone to know that NPOV and INTEXT say something different and "outrank" the only guideline that says the opposite. WhatamIdoing (talk) 22:26, 29 January 2022 (UTC)

I think we should probably come up with specific tweaks to the wording, since at least based on this discussion it is likely there could be sufficient consensus to turn WTW's requirement for attribution into a suggestion. I would suggest something like ...may express contentious opinion and are best avoided unless widely used by reliable sources to describe the subject; if sources using the term are WP:BIASED or opinion-based, consider using in-text attribution per WP:INTEXT. --Aquillion (talk) 23:41, 29 January 2022 (UTC)

That's a red carpet for POV pushing. The desired label is only in opinion articles or biased sources? Well, just consider using in-text attribution. Crossroads ^-talk- 06:54, 30 January 2022 (UTC)

Obviously the consideration would have to be done according to WP:INTEXT, which is the actual policy governing when to do it; it does require in-text attribution sometimes (and would continue to do so), but only for biased statements of opinion, and it forbids or discourages them under various other circumstances. So we could import that text and say simply that ...may express contentious opinion and are best avoided unless widely used by reliable sources to describe the subject; for biased statements of opinion, use in-text attribution per WP:INTEXT. But it is unacceptable and unworkable for WTW to set a standard that goes past WP:INTEXT; that invites people to cite WTW to justify POV-pushing in the other direction, since anyone who disagrees with a factual statement that falls under one of the words in WTW could cite it to argue that it must be attributed (and therefore framed as an opinion) with no regard for what the sources say at all. That is to say, if someone wants to push the POV that racism, white supremacy, holocaust denial and so on are not real things and are only some people's opinions and must be expressed as such on every part of Wikipedia with no regard for sourcing, they could try and use WTW to support that crusade. (They would fail because WP:NPOV and WP:INTEXT are policy and trump MOS:WTW in that regard, requiring that we eg. describe Neo-Nazis, denialism or racism as such in the article voice when the source is sufficiently overwhelming to establish it as an uncontested fact -- but currently WTW does not acknowledge the relevant policies, which is a problem, and is part of the reason it has lead to interminable arguments on numerous articles, since some people do feel comfortable citing it to argue for attribution with no regard for sourcing.) At the bare minimum it needs to make it clear that in-text attribution is something that depends on the strength of the sources and whether they fall under WP:RSOPINION / WP:BIASED or not; and it should probably reference INTEXT in some form to make it clear that there are also situations where in-text attribution is inappropriate (although perhaps that is a subset of WP:WEASEL and should be covered there instead, ie. weasel wording can be used to obscure who holds a view but it can also be used to present a view that the sources treat as factual as if it is mere opinion.) But either way, even ...may express contentious opinion and are best avoided unless widely used by reliable sources to describe the subject; if only sources using the term are WP:BIASED or opinion-based, use in-text attribution in accordance with WP:INTEXT would be a clear improvement over the current text.--Aquillion (talk) 07:09, 30 January 2022 (UTC)

I think that something like this might work better: ...may express contentious opinion and are best avoided unless widely used by reliable sources to describe the subject; all such uses must comply with WIKIVOICE and INTEXT".

If you wanted to be more explicit, then we could be more verbose: ...may express contentious opinion and are best avoided unless widely used by reliable sources to describe the subject; all such uses must comply with the WIKIVOICE section of the NPOV policy (which permits the use of value-laden labels without in-text attribution under some circumstances) and INTEXT (which explains when in-text attribution of labels to specific groups or individuals is required)"

However, I'm not sure that the verbosity is necessary. WhatamIdoing (talk) 17:01, 30 January 2022 (UTC)

Regarding WhatamIdoing's question, the solution is not to hamstring WP:LABEL by allowing editors to launder opinions into facts with WP:BIASEDSOURCES contrary to WP:WIKIVOICE, but to remove terms from LABEL's example list that are not vague opinion terms. "Denialist" and "neo-Nazi" are not vague terms that vary widely in meaning based on opinion; they have well-defined meanings that academic sources and people in general would agree on. Many people would consider all sorts of well-defined terms value-laden, such as "socialist" or "conservative", but we obviously won't add them here. Consider the difference between vague opinion terms and well-defined ones:

"John Doe is a racist..." -> okay, as a reader this tells me very little as to what he actually did. Is this guy an opponent of affirmative action, did he use the N word, or is he a white supremacist? It could be anywhere on that spectrum, and at least some sources will call him a racist.
"John Doe is a neo-Nazi..." -> oh, this is immediately understandable. He supports the modern incarnation of the ideology of Nazi Germany.

Let's try another:

"Peter Duesberg is a crank..." -> okay, people have no idea what this is getting at, other than that Wikipedia editors consider his views to be wrong. Unsurprisingly, this isn't what it says.
"Peter Duesberg is an AIDS denialist..." -> easy to understand, he denies the existence of AIDS.

Perhaps here we are focusing too much on "values" - itself sort of vague - has obscured more important aspects, which is that terms that are vague or opinionated should not be in wikivoice. This is in line with the WP:WIKIVOICE policy, Avoid stating opinions as facts, and LABEL's own statement that Rather than describing an individual using the subjective and vague term controversial, instead give readers information about relevant controversies - note that the issue is that the term is "subjective" and "vague". Crossroads ^-talk- 07:21, 30 January 2022 (UTC)

My problem with this is that it allows for the situation I outlined above (where someone's personal feeling that eg. "racism" or "white supremacy" or whatever are meaningless value-laden terms can trump overwhelming sources indicating otherwise in a particular context.) Whether a word is specific and factual has to be decided in individual cases based on the strength of the sources, and even when it comes to general guidelines we have to decide by looking at the sources - obviously you cannot simply say "I personally feel, in my gut, that 'racism' is vague and unhelpful", since we have to WP:STICKTOTHESOURCE. And there are, indeed, overwhelming high-quality academic sources that treat racism as a highly specific and well-defined topic to be discussed as fact and not solely opinion. Along those lines, we also have to be cautious about people trying to launder facts into opinions, especially when it comes to concepts (like racism) that are rejected entirely by many highly-opinionated culture-war groups; regardless of their feelings, WP:NPOV requires that we use terms like "white supremacism" or even, yes, racism in the article voice, without attribution, when the sources are sufficiently overwhelming and one-sided to establish it as fact in a particular context. Obviously that bar is often going to be quite high (because it can be WP:EXCEPTIONAL) and is particularly high for WP:BLPs, but WTW cannot (and does not) place it completely out of reach - it is simply not viable to require that "racism" always be attributed, without exception, every single time it appears in Wikipedia, since requiring that without regard for what the sources say is an unambiguous violation of NPOV and therefore unenforceable. If the overwhelming majority of high-quality sources unambiguously say as fact that eg. that Apartheid was racist - and they do - then we must describe it as such in the article voice (as we do, in numerous points on that article.) You can argue that any individual use of the term needs attribution based on the sourcing and so forth, but we cannot and do not universally require it in the general case; and if WTW implies that it is universally required then that is a simple error in WTW. --Aquillion (talk) 07:41, 30 January 2022 (UTC)

Yes, that might be a problem but it's one that should be easy to overcome by adding the facts that support the label. If those facts can't be added then we shouldn't have added the label in the first place. Erring on the side of excluding the label or attributing the label does less overall harm to the IMPARTIALity of the article vs erring in the other direction. Using terms similar to yours, allowing someone's personal feelings to cherry pick sources to smear an article subject via their Wikipedia page. Springee (talk) 12:02, 30 January 2022 (UTC)

Aquillion, I'm afraid you are misapprehending the situation. Only a tiny, FRINGE minority of editors believe that "racism" or "white supremacy" or whatever are meaningless value-laden terms can trump overwhelming sources indicating otherwise in a particular context. The much more relevant case is anti-transgender activism, which a number of editors (including several participants in this discussion) explicitly believe to be a meaningless and/or emotive term to be excluded regardless of its prevalence in the sourcing. We wouldn't be having this discussion if it were only for the neo-Nazis. Newimpartial (talk) 12:35, 30 January 2022 (UTC)

@Newimpartial, we are having this discussion because LABEL contradicts NPOV and INTEXT. But did you mean to suggest that this might be a central subject area that prompts support for having LABEL incorrectly contradict the other policies and guidelines? WhatamIdoing (talk) 17:18, 30 January 2022 (UTC)

No; I think INTEXT and NPOV get it right and LABEL currently gets it wrong. My observation is an empirical one: that the anomalous language of LABEL is seldom (and almost never successfully) employed by editors to WHITEWASH racist and antisemitic labels, but is frequently deployed as an argument to remove "anti-trans" or require in-text attribution for it, even when the term is well-sourced. Newimpartial (talk) 17:35, 30 January 2022 (UTC)

I think it's a parallel issue in both topic areas; I've mostly encountered it on the racism side. A May 2021 RfC resulted in no consensus to include RS-supported attributed labels of racism from Tucker Carlson, with several editors citing WP:LABEL. This isn't the place to rehash that discussion but it does show that LABEL has been used to exclude content. –dlthewave ☎ 21:11, 30 January 2022 (UTC)

Are you suggesting that the changes proposed here would have flipped the outcome of that RfC? Per the closing a number of policies/guidelines were cited and LABEL doesn't appear to be a critical one even though it was mentioned. Springee (talk) 21:40, 30 January 2022 (UTC)

I think that this might just be a matter of you being more familiar with discussions you frequent. WP:LABEL is absolutely cited in an effort to try and insist on attribution for things like white supremacy or racism. The implication that it could be used to insist that something like Apartheid cannot be described as racist is obviously more extreme, but our policies shouldn't be written in ways that have such nonsensical implications either way; the point is to establish that large numbers of places do exist where WP:NPOV and WP:INTEXT require that we terms on the WTW list without attribution, and that it's therefore unreasonable (in addition to being against policy) for WTW to imply that it is forbidden. --Aquillion (talk) 06:41, 31 January 2022 (UTC)

Interestingly, looking at Apartheid, the word "racist" doesn't appear until about 1/3rd into the body. The intro relies on statements of fact and allows the reader to draw their own emotional conclusions. The intro is a good example of using impartial terms to describe a very racist institution. It's an example of what we should be doing instead of trying to apply the labels in question. Springee (talk) 14:14, 31 January 2022 (UTC)

I don't think that racist is an "emotional conclusion". I think it is an objective, indisputable fact that an explicit, intentional, systematic separation of people according to their race is racist. WhatamIdoing (talk) 02:55, 2 February 2022 (UTC)

Is it racist to oppose affirmative action? To oppose defunding the police? I don't think anyone should answer these - my point is that there is actually a significant amount of disagreement across society as to whether some things are racist (even while almost everyone would agree that, say, apartheid is racist). The same issue of heavy disagreement applies to transphobia and Islamophobia and every other accusation of bigotry. We should not be making labeling with these sorts of terms easier, as it will lead to more tendentious arguments for certain labels. Wikipedia can then be used to advance an agenda that something controversial really is racist or phobic or whatever. Crossroads ^-talk- 04:46, 3 February 2022 (UTC)

I'm not seeing how additional tendentious arguments for certain labels would be worse than the current language, that gives policy cover for tendentious arguments against the use of terms that are well-sourved and are not really controversial except for extreme sources and for certain Wikipedia editors. The current text of LABEL is easily weaponized in the service of POV whitewashing, which actually happens, in contrast to the hypothetical argument you put forward here. Well-sourced terms that are not in dispute among reliable sources should be in wikivoice per WP:INTEXT, and the current text of LABEL (which rests on a lower CONLEVEL than INTEXT) gets in the way of encyclopaedic writing. Newimpartial (talk) 15:34, 3 February 2022 (UTC)

Almost everything you claim can be turned around if we make it easier to apply contentious labels. What counts as "well sourced"? If we make it too low a single so-so source is all we need. The other extreme would be if any RS doesn't say it then we say it must be attributed. You argue this leads to POV whitewashing but why shouldn't we be worried about POV blackwashing? I would argue blackwashing is worse since it gives UNDUE weight to typically negative characterizations that can harm a BLP's reputation if someone does a quick web search and sees just the first 3 sentences of their bio (oh, he is a X-ist!). If they deserve the label then the rest of the article will make that clear with evidence. However, if the label is questionable then we are failing if we err on the side of inclusion. I've seen a number of cases where it's clear editors were trying to get a label into the lead ("far-right, alt-right, provocateur etc) even though the body of the article doesn't really support the label. Other times editors are more interested in inserting the label (source calls Ms Doe an X-ist) but aren't really interested in adding the examples where X-ism was displayed. No, this is supposed to be an encyclopedic work, not a news article, not a gossip column and certainly not a source that should be used to persuade vs inform. Springee (talk) 17:03, 3 February 2022 (UTC)

Re: I've seen a number of cases where it's clear editors were trying to get a label into the lead ("far-right, alt-right, provocateur etc) even though the body of the article doesn't really support the label - any instances where this problem actually occurs are WP:NPOV problems, not LABEL problems, and run afoul of WP:INTEXT as well. If the justification of the current text of LABEL is to provide a shortcut to avoid actual discussion arriving at policy-based consensus on POV issues based on DUE representation of quality RS - well, to me that isn't really a justification at all. Newimpartial (talk) 18:50, 3 February 2022 (UTC)

Again, you have said LABEL is a tool of whitewashing but I don't see that you are addressing the reverse, where label is used to prevent blackwashing? Why is it better to allow more blackwashing vs prevent contentious labels and force people to stick with the details/actions that support the label vs the label itself? Springee (talk) 18:55, 3 February 2022 (UTC)

I thought I had been sufficiently clear about this, so I'll try again: any time what you call blackwashing is an actual problem, we have NPOV and INTEXT to deal with it. LABEL isn't needed for that, and doesn't actually help. Newimpartial (talk) 19:36, 3 February 2022 (UTC)

I agree with Newimpartial. If mentioning a value-laden label isn't warranted, then its use is a violation of NPOV. It would be a violation of NPOV even if LABEL had never existed.

The problem this section is meant to address, though, is that LABEL is demanding that some widely used, fully warranted, NPOV-compliant, and (if relevant) BLP-compliant labels always have INTEXT attribution, even when NPOV and INTEXT says not to do it and even when that attribution would be misleading.

The goal here is not to change anything about when/whether you can use a label. The goal in this discussion is only to take out the incorrect advice about always providing INTEXT attribution.

So far, this discussion sounds like this:

Me: LABEL should give advice about INTEXT that matches what INTEXT says.
Two editors: People just shouldn't be using biased and opinionated labels anyway.

This sounds like a non sequitur to me.

Is the actual argument meant to be something like "LABEL should misrepresent INTEXT because that will discourage editors from using labels that I don't want them to use"? WhatamIdoing (talk) 20:13, 3 February 2022 (UTC)

Newimpartial, any time you feel a label is removed via LABEL+whitewashing the facts that support the label can/should be in the article thus making the label redundant. If the content of the article isn't there then the label certainly shouldn't. As for WhatamIdoing's concern, perhaps INTEXT needs to be updated to match LABEL? If an article is well written the use of these labels really shouldn't be needed. Perhaps we can say, in "sky is blue" cases the label doesn't need attribution? Springee (talk) 21:04, 3 February 2022 (UTC)

Since I believe that, broadly speaking, the language of INTEXT currently gets it right while LABEL gets it wrong, I don't agree with your proposal. I also dispute your apparent premise - that if the article demonstrates through sourced examples that (e.g.) a BLP subject is an anti-trans activist, that somehow it would be redundant for the article actually to say that the person is an anti-trans activist. This is precisely what I am calling whitewashing, and perhaps also illuminates why LABEL issues arise frequently in relation to the lead - in the body, it may be possible to achieve a well-sourced and clear depiction by piling on examples without using a label (though I don't accept that it is preferable, or even encyclopaedic, to do so). But this would never be possible in the lead, which may explain why some of the more eggshell-walking ("because LABEL") lead language can be so tortured and false-balance-ey. Newimpartial (talk) 02:16, 5 February 2022 (UTC)

But you jump to the conclusion that such sources apply the level of scrutiny we should expect before using a contentious label in what is supposed to be an encyclopedic article. Your position doesn't address the very real concern regarding editors searching for sources based on a few keywords, effectively having a conclusion and then trying to find the sources that support it. Again, you are concerned about whitewashing but ignore the reverse. If we have a situation that is perfectly knife edged, that is, half the sources say a comment shows racism, half the sources say it was not racist. For argument sake, I'm going to assume we have a binary choice, call it racist or apply no label. You are, in effect, saying we should err on the side of saying it was racist because we don't want to let a case where 60% say it was go without being called racist. Effectively you would prefer to accept that we might err on the side of calling a BLP subject racist out of fear that if the rules go the other way, we might not call a racist a racist (or a transphobe a transphobe). I think that is wrong and at least in consistent with the BLP ARBCOM view of avoiding harm. If the racist, transphobic, etc actions can't be adequately described then they shouldn't be labeled. We don't want whitewashing but it doesn't cause harm, blackwashing does so we should always make sure to err on the side of not applying contentious labels. If we can't reasonably describe the issue in the lead then we shouldn't apply a label in the lead. Springee (talk) 04:51, 5 February 2022 (UTC)

In a case where half the sources say a comment is racist and half actually say it is not, and both sets of sources are of comparable quality, then of course DUE requires a BALANCEd statement and anything we say should be given in-text attribution. Even if the sources for are clearly of higher quality than those against, some kind of attribution is required in instances where there really are two sides DUE for inclusion. And of course, BLP considerations (which are barely unaddressed by LABEL) mitigate for caution where they apply.

But what I have seen much more often than this (essentially hypothetical) case is the real (and fairly common) one where we have multiple, quality sources offering a specific characterization, no independent sources against that characterization (though perhaps a WP:MANDY denial), and a few sources that don't weigh in one way or another. I am not saying that we should go without attribution for cases that are actually perfectly knife edged - and in fact, one of the problems with the current version of LABEL is that it doesn't allow a graduated approach where uncontested labels are given without attribution and contested ones are attribured. So your statement of my position, Effectively you would prefer to accept that we might err on the side of calling a BLP subject racist out of fear that if the rules go the other way, we might not call a racist a racist, is entirely a straw man.

What I am saying is that a guideline that tells people that nobody should be characterized as a "racist" in wikivoice, regardless of the sourcing, because the word itself is too emotive (or whatever) is just wrong, unencyclopaedic and actually works against BLP, NPOV and INTEXT policies. The current text of LABEL doesn't err on the side of not applying contentious labels - if followed consistently, it would prevent the appropriate use of labels where NPOV requires that attribution not be used.

In the cases I am talking about, it is possible (even easy) to reasonably describe the issue in the lead using well-sourced characterizations in terms of racism, or antisemitism, or anti-trans activism or what have you. Actual RfCs on these issues in regard to specific articles actually examine the sourcing and come to a judgement of what NPOV requires. But LABEL is a thorn in the side of those RfCs, by giving a set of editors a pretext to opt-out of source-based discussion by arguing that term X can't be used "because LABEL". Sometimes that argument succeeds, sometimes it fails, but it always uses up valuable time and it never IMO draws any attention to the policy-based, encyclopaedic criteria on which such decisions ought to be based (primarily WP:V and NPOV, and sometimes BLP). Newimpartial (talk) 00:21, 6 February 2022 (UTC)

We are circling. You feel that subjects that rightly should be called transphobic are not and this is because LABEL offers too much protection. Of course others don't agree else their wouldn't have been a RfC. However, you don't seem to be concerned that perhaps a subject that shouldn't be called transphobic (or some other contentious label) will be if we relax the rules. Why is that not an equal problem? As I've already said, Wikipedia should always err on the side of do no harm. We see that in policies like BLPCRIME. If the label is applied the facts may not matter as the reader may take the label as gospel. However, if we err on the side of avoiding the label the reader still has the option to read more and decide for themselves. That seems more aligned with Wikipedia principles to me and, more importantly, less likely to be used to push a POV via the words we use in Wiki articles. Springee (talk) 00:42, 6 February 2022 (UTC)

Circling, indeed. My belief is that LABEL in its current form is being used to push a POV by whitewashing articles, and I do not believe that the possibility for unjustified labels to be included would be an equal problem without LABEL, because this is already prevented by the effective use of NPOV, WP:V and BLP policies, which cover these issues both more effectively, and with more nuance, than does the current LABEL guideline. Newimpartial (talk) 02:53, 6 February 2022 (UTC)

Crossroads, this section isn't about vagueness or opinions. It's about value-laden labels. The fundamental point behind labelling (value-laden or otherwise) is to divide people into a group. When someone gets labelled as a climate denialist, that means that they have been judged as belonging to a particular group of people. It could be a small group ("freedom fighters") or a large group ("men"), but the person is being judged as belonging to that group.

The point behind this section is to say that when the labels in question are value-laden, then we should avoid using them unless they are "widely" (not universally) used. This is the difference between "Paul Politician was a climate denialist who spent his career supporting the petroleum industry" (which we could use if sources widely use that label) and "Paul Politician spent his career supporting the petroleum industry" (which we would use if the label isn't used widely in sources). WhatamIdoing (talk) 17:16, 30 January 2022 (UTC)

Even if a label is widely used, it is far more informative to summarize what the person actually said. For a pair of hypothetical examples:

"Paul Politician, who spent his career supporting the petroleum industry, asserted in 2020 that climate change does not exist."

This establishes beyond question that he has denied climate change. It is clear to any reader. Adding the label "climate denialist" would serve no useful purpose.

"Paul Politician, who spent his career supporting the petroleum industry, asserted in 2020 that climate change is largely caused by non-anthropogenic sources."

This establishes the person's position on the subject, which some sources (and many readers) might call denialism, and others might not. It is more information than the label would provide to the reader.

This gives the reader the significant benefit of replacing a value-laden label with a fact (e.g., a sourced quote from the subject) to which readers can apply their own values. Lwarrenwiki (talk) 21:39, 30 January 2022 (UTC)

This is exactly how we should be handling criticism geared towards BLPs or other topics when labels can be avoided (in the situation where they are not widely used); explain what exactly the issue is, and assume the reader is smart enough to make the connection, as well as, when it can be sourced, more accurate to the person's views (which is an issue with the "anti-trans" "transphobia" section above). --Masem (t) 06:08, 31 January 2022 (UTC)

The problem with this is that you are essentially asking that editors perform WP:SYNTH using primary sources to imply that the article's subject believes something controversial. That is actually a more serious WP:BLP violation than just citing reliable sources saying what they believe. Now, for BLP specifically the standard to avoid in-text attribution is sky-high, but I would always, without exception, go with "these sources described X as a white supremacist" as opposed to trying to make the argument that X is a white supremacist implicitly ourselves in the article voice. And in cases where the sourcing is high-quality, near-universally in agreement, and where it is a major aspect of their notability, it should, yes, be used in the article voice (eg. David Duke.) Your interpretation invites editors to say basically "ignore all those sources, if you look at his source he didn't literally directly say this", which is clearly a violation of WP:NPOV - and even worse, it invites them to say "here's a pile of everything bad this person ever said, which proves they are a bad person", which is WP:SYNTH. We describe BLPs according to what the sources say (including, as WP:BLP specifically spells out, unflattering things or things that they may not want said about them). The standard for negative BLP material is simply higher-quality sourcing and more weight on it being WP:DUE, not "ignore the sources, WP:SYNTH up stuff using quotes." If someone is described as a neo-Nazi in a significant subset of sources, but not all, and that description is WP:DUE, we should describe and attribute it; if they are described as a neo-Nazi near-universally, we should state it in the article voice. But we should never be saying "here's a bunch of quotes, with no secondary interpretation or analysis, that imply that this person is / is not a neo-Nazi", that is absurd. In fact, I would actively discourage quotes in that situation (vs. paraphrasing and citing secondary sources) because the danger of WP:SYNTH or pulling quotes out of context to make someone look good or bad is so severe. Our articles, especially on such sensitive topics, need to be based on the conclusions of secondary sources and not quotefarms thrown together by editors to guide readers to an uncited conclusion. --Aquillion (talk) 06:41, 31 January 2022 (UTC)

I don't disagree that there's a point that when more than a sufficient portion of sources uses a label that we can use that label and without attribution, but that should be an onus on those wanting to use the label that way to show that the near universality of the label exists in sources about a topic - including past, present, and likely future ones too as to avoid a burst of coverage that changes tone later down the line. And no, avoiding labels to describe a point of view does not fall back to quote farming; we are able to summarize and paraphrase without introducing interpretation as well, though we should be using statements made in reliable sources as the origins for these statements. Most good RSes, even when they apply labels, explain why the label applies, and that "why" is what we should be using, not the label itself, otherwise we're far too much focused on the mudslinging that labels bring. --Masem (t) 13:31, 31 January 2022 (UTC)

If you'd want any changes to specify that the default position is to attribute, and that the onus is on those who want to leave it out, maybe that's a possible path to come to agreement. In that case, one approach might be to use the wording ...with in-text attribution if in doubt. That would address the violation of core policy in the current wording, while indicating that any uncertainty would default towards attribution. It's also the same language as is already used one paragraph below in the same context. Sunrise (talk) 14:12, 31 January 2022 (UTC)

@Sunrise, your suggestion of "...with in-text attribution if in doubt" seems reasonable to me.

Alternatively, we could just refer people to the actual NPOV policy and the actual INTEXT guideline, and tell them to follow those. (Presumably they say something about what to do for borderline or uncertain cases?) WhatamIdoing (talk) 02:56, 2 February 2022 (UTC)

No, absolutely not. At least on my end I would never consider a default presumption to attribute to be acceptable - again, I feel that attribution has at least as much potential to be WP:POV; making it the default would encourage NPOV violations by people who could stonewall facts in the sources. Default outcomes need to be handled extremely cautiously because they complicate consensus-building by giving some people in a dispute less of an incentive to come to the table and negotiate or to engage with the sources; the default should always be "we go with what the sources say", outside of very specific situations where there is significant risk of real-world harm. The situations where that would be the case here are already covered by WP:BLP; situations outside of that shouldn't (and realistically cannot) have a default that would potentially allow an editor to frame longstanding text and well-sourcex as mere opinion and eg. insist on extended discussions or an RFC to turn it back. --Aquillion (talk) 01:10, 5 February 2022 (UTC)

@Masem, should I interpret your statement that "there's a point that when more than a sufficient portion of sources uses a label that we can use that label and without attribution" as you agreeing that LABEL's current requirement that in-text attribution always be used, even when we are obviously well past that point, is a problem? WhatamIdoing (talk) 02:58, 2 February 2022 (UTC)

I don't think the current wording is a problem. I consider the scenario of "no need to attribute when near-universal agreements of a label" as the exceptional case and if anything were to be written into this page, it should explain that should be when certain conditions are met. It may seem somewhat annoying to present it this way, but for novice editors we want to stress and maintain that labels should never (save for that exception) be used without attribution. I know that some are presenting that the argument should be "labels that have near-universal agreement can be presented without attribution but otherwise labels should have attribution" (that is, in conflict with LABEL) and while that means the same thing but I am afraid of how novice editors will take that. It is better to maintain the cautionary language of LABEL but establish the needed carveout. --Masem (t) 04:12, 2 February 2022 (UTC)

LABEL doesn't admit of exceptional cases, and editors shouldn't have to invoke IAR over a guideline. WhatamIdoing (talk) 04:49, 2 February 2022 (UTC)

I'm saying that I think the current wording of LABEL should stay, but add to explain the exceptional case, rather than saying that LABEL needs to be completely rewritten. --Masem (t) 05:04, 2 February 2022 (UTC)

How about we just take that out of LABEL entirely, and direct people to the actual policy and the INTEXT guideline? We'd never have to worry about keeping them synchronized again. WhatamIdoing (talk) 05:16, 2 February 2022 (UTC)

I'm not seeing the synchronization problem, though LABLE should include language that says "should not be used without in-line attribution as outlined at INLINE". I think the question is more "when does a label (defined per LABEL) lose its subjective/contentious nature and become a reaonably objective descriptor to be stated in Wikivoice?" which is absolutely a guideline-type of consideration (there's no hard-fast rule) and should be outlined at LABEL where it makes the most sense or perhaps another guideline page "when is a label not a label". --Masem (t) 13:36, 2 February 2022 (UTC)

INLINE is Wikipedia:WikiProject Inline Templates; I think you meant INTEXT, which is part of Wikipedia:Citing sources.

The problem with "should not be used without in-text attribution as outlined at INTEXT" is that INTEXT says "it should always be used for biased statements of opinion" (with a link to WP:ATTRIBUTEPOV). INTEXT also warns about the risk of "an inadvertent neutrality violation" by implying parity between majority and minority views, suggesting that a widely held view can be attributed to a single source, and cluttering articles "with information best left to the references".

To compare them, LABEL, as currently written, says:

Peter Duesberg has been called an AIDS denialist by Alice, Bob, Chris, Dave, and more.^[1]^[2]
Time Cube was a website that Alice, Bob, Chris, Dave, and others said espoused a pseudoscientific conspiracy theory.^[1]^[2]
JK Rowling has been called as transphobic by several trans advocates, including Alice and Bob.^[1]^[2]

but INTEXT and NPOV, as currently written, say:

Peter Duesberg is an AIDS denialist.^[1]^[2]
Time Cube was a website that espoused a pseudoscientific conspiracy theory.^[1]^[2]
JK Rowling has been called transphobic by several trans advocates, including Alice and Bob.^[1]^[2]

LABEL should give the same advice, leading to the same results, as the other pages. WhatamIdoing (talk) 17:07, 2 February 2022 (UTC)

You've repeatedly brought up Peter Duesberg so I did some close reading, of all the sentences in the article that say denier/denialist and have any cite at all ... Chigwedere et al: never says denier/denialist. Nattrass: says "group of AIDS denialists" clearly including Mr Duesberg. Cohen: never says denier/denialist, says skeptic. Kalichnan: says "denying AIDS" clearly including Mr Duesberg. Goertzel: never says denier/denialist, says dissenter and mentions conspiracy theories. Lenzer: never says denier/denialist herself, says he and others have been labeled "denialists". Sithole: never says denier/denialist. Schoofs: dead link. McGreal: says denier/denialist about Thabo Mbeki and about a group that included Peter Duesberg, says dissident. Clifton: says denial of the link between HIV/AIDS. Karon: never says denier/denialist, says skeptic, mentions flat-earthers. JournAIDS (Author Unknown): contains the phrase "AIDS denialists" and says Mr Duesberg is a prominent denialist. Durban Declaration: dunno, paywalled. Corbyn: never says denier/denialist, doesn't mention Mr Duesberg. Goldacre: never says denier/denialist, mentions dissidents. Enserink: says Mr Duesberg is a so-called "AIDS denialist". Miller: dead link. Cartwright: says Mr Duesberg denies that HIV causes AIDS. Enserink: never says denier/denialist. Summary: a small minority of the sources say denier/denialist, an even smaller minority use the phrase AIDS denier/denialist, the Peter Duesberg example isn't simply showing a failure re WP:LABEL and WP:ATTRIBUTEPOV, it's a failure re WP:BLP because in most sentences the contentious label is poorly sourced or unsourced. Peter Gulutzan (talk) 15:15, 3 February 2022 (UTC)

I'm sure the article could use some updates. Here's a few sources that might be relevant if you wanted to do that:

Mbeki was influenced by the AIDS denialist Peter Duesberg
AIDS denialists like Peter Duesberg
Peter Duesberg, a brilliant virologist, became an AIDS denialist
rogue scientists, headed by Peter Duesberg, held fast to failed theories. Although they consider themselves dissident scientists, they are actually denialists
a Presidential advisory panel, which included Duesberg and other denialists
Mbeki...assembled a "Presidential Aids Advisory Panel" of 30 people, half of whom were denialists, including Duesberg
"Duesberg, the denialists’ founding nutjob"
joined by another HIV/AIDS denialist, the granddaddy of them all actually, Peter Duesberg
At the top of a list titled "HIV/Denialist Scientists": "Peter H. Duesberg is the single most important figure in HIV/AIDS denialism...In every respect, HIV/AIDS denialism starts and ends with Peter Duesberg" (ISBN 9780387794761, p. 174)

This is not a minority viewpoint, and regardless of what's currently in the article, it can be sourced to scientific journals, popular magazines, newspapers, and books. WhatamIdoing (talk) 17:52, 3 February 2022 (UTC)

One issue I see is that when one says that LABEL is not consistent with NPOV or INLINE because (for example) they would require attribution of calling someone an x-denialist when that is widely used in sources, is that this presumes that x-denialist is not a label. Even in that case that type of term is still a type of subjective statement, and a label. Broadly WP should never have a subjective statement in wikivoice, and always should be with attribution. BUT there are times where there is near universal agreement on a subjective term that we don't need the inline attribution, and this not only applies to labels but to other things eg "Mozart is considered one of the greatest pianists of all time." Meaning what is at issue is something more at INLINE to clarify when mass-agreement on a subjective matter is had, in how that can be said without attribution. There would still need to be some updates to align LABEL to that but thats the larger issue here. --Masem (t) 21:47, 3 February 2022 (UTC)

Broadly WP should never have a subjective statement in wikivoice, and always should be with attribution. I agree. But we determine what is subjective solely, and exclusively, by looking at the sources. Nothing else matters one iota. If the sources treat something as an objective statement of fact, we must do so, and it would be a WP:NPOV violation to do otherwise. We absolutely must never allow an editor to say "well, yes, the sources state this as fact, but I personally feel this kind of statement is subjective because I don't think racism / neo-Nazism / whatever is objectively definable" - that is an outright WP:TEND / WP:POV argument in which an editor is directly requesting we put their unsourced views in the article text. We must report things the way our best-available sources do, fullstop; my problem with WP:LABEL is that people have repeatedly added words to it based on their personal sentiments and without regard for any sources at all, then turned around and argued that this allows us to downplay or ignore sources that they personally disagree with - again, often without reference to any contradictory sources at all, simply based on their personal gut feeling that "these words are not objective." That is not acceptable - we follow the sources, fullstop. If you feel that something is subjective, and you want it treated as such in the article voice despite widespread sources treating it as objective, you must at a bare minimum find sources of comparable weight that treat it as subjective. Otherwise you're arguing for a violation of WP:NPOV. --Aquillion (talk) 01:15, 5 February 2022 (UTC)

Subjectivity is a function of language, not sources. Just because lots of sources use a subjective term for X does not make that an objective term, it still a subjective term, just one that is widely used, and I think we do need a mechanism that we don't have to mess too much with inline attribution when it has been shown beyond a doubt the term is widely used. But in this, we still need to keep in mind that that term is a subjective term and there are aspects about its use to be careful about.

That brings me to another point is that a focus of this discussion has been on what I've brought up before, "characterization labels" like "conspiracy theorists", "x-denier", and so on, those that directly characterize their words and actions rather than their attitudes, which are the types of labels that I think can reach the level of wide-spread usage to avoid the need for inline attribution. But unless we're talking a person long day or well beyond their retirement/out of the public spotlight, these must be taken as subjective characterizations, no matter how many RSes use the term. In a situation like this I think we can find a way to use "X is a conspiracy theorist.(2-3 high quality refs)" (note the lack of inline citations) but this 1) first needs to have consensus agreement on talk pages that the sourcing fully supports this and 2) should immediately be followed by context to support this, which stands in place of in-line attribution by providing why they are considered that way. Exactly how to do this, that's beyond the scope here, but I am in agreement there's a time and place where we should be able to do this and details need to be worked out for how to handle this, but key is that this case must be a high bar to evoke, otherwise we need to have editors default to in-line attribution.

These characterizations need to differ from more banal labels, like "evil", "racist", etc. which are far more subjective, can never be taken as fact (at least while BLP/RECENTISM is involved), and which should always be accompanied with inline attribution. But in the cases where there is widespread agreement on these labels from sources, we should be able to adapt language that again helps to make it clear this is a wide-spread opinion from the media. eg "X is broadly considsered to be racist. (3 or 4 high quality sources)" followed by context for why. There's ways to reflect capturing near-universal agreement here in as close to a matter-of-factly voice but without making a subjective term factual in wikivoice. --Masem (t) 02:20, 5 February 2022 (UTC)

Subjectivity is a function of language, not sources. No, this wrong. Per WP:YESPOV, whether a particular statement is fact or opinion, and therefore whether we should present it as objective or subjective, is based entirely on what the sources say. Even when it comes to BLPs: These characterizations need to differ from more banal labels, like "evil", "racist", etc. which are far more subjective, can never be taken as fact (at least while BLP/RECENTISM is involved), and which should always be accompanied with inline attribution - no, that is likewise flatly wrong. If high-quality sources are completely unanimous in calling something racist in the article voice, and treating it like a fact, then we must do the same, even if it would fall under BLP. This is non-negotiable, since it is a central part of NPOV. Neither does it contradict BLP, which is ultimately about following highest-quality sources and erring on the side of caution when there is doubt. Your position here seems to be "sources should never treat 'racism' as something that can be a clearly-defined fact, so if they do, we can ignore that." But that directly goes against WP:NPOV - no matter how strongly you personally feel that racism is far more subjective, it is completely inappropriate to try and impose your feelings in that regard on article content in contexts where the sources unambiguously disagree. I am fine with setting a high bar (especially for BLPs, where it is already high), but ultimately the only thing that matters when determining whether we frame something as a fact or opinion is how it is treated by the sources, which means that absolute statements like you're making here cannot be imposed. And I literally mean cannot be imposed, ie. NPOV directly instructs us on this and is not subject to consensus. Even if you write a policy that directly and unambiguously says "never call any individual's actions or statements racist in the article voice", NPOV will override that in the (rare, but far from nonexistent) situations where the sources are sufficiently high-quality, unanimous and unambiguous. Per WP:YESPOV we are strictly forbidden from treating something as subjective or as opinion when it is unambiguously treated as objective and as fact in the sources, and that cannot be changed or overruled. --Aquillion (talk) 22:16, 5 February 2022 (UTC)

There is nothing in YESPOV that says that even with a near agreement of reliable sources assert a subjective statement about a topic in a factual tone, that we must accept that agreement as fact. That would be making a subjective statement into an statement of fact which we absolutely cannot do. In fact, YESPOV includes "Prefer nonjudgmental language" which says exactly not to take that factual tone in such cases. NPOV does say that we should treat that as the majority position and that unless there are reliable sources that present counterarguments to that point, shouldn't play games around false balance or equivalence. But presenting such a situation where there is near-agreement about how a topic is characterized or how a subjective term is used is far different than something that has gained objective, factual stance. In a case where there is near-universal agreement that a person is an "x-denier" as "Y is widely considered an x-denier" or "Y is an X-denier", as long as that is appropriately followed by context to explain that, falls 100% in line with NPOV, neither weakening the position of the majority position while maintaining the fact it is still something that is, implicitly, a subjective term that never should be stated factually in wiki-voice without context. (and of course, after consensus has agreed that there is sufficient universal agreement that the term can be written that way). This works for positive/praise language as well as non-controversial subjective terms too (such as "philanthropist", "savant", etc.), so its not just an issue for how we write for those treated negatively by media, but how we write period to stay neutral as an encyclopedia. --Masem (t) 22:48, 5 February 2022 (UTC)

I don't see where WP:YESPOV says that the distinction between fact and opinion is based on "what the sources say". It states, Avoid stating opinions as facts. This even goes so far as For example, an article should not state that "genocide is an evil action" but may state that "genocide has been described by John So-and-so as the epitome of human evil." And the rule to avoid stating opinions as facts is distinct from the following separate point to Avoid stating seriously contested assertions as facts. So no, an opinion does not transform into a fact simply by being stated in enough sources. Something can be not "seriously contested", yet still be opinion. I know of no source or policy that says the distinction between facts and opinions is just a matter of how widely they are held.

Just saying 'if enough sources say it, it's a fact' is also a problem because sources can be WP:BIASEDSOURCES. They may mix fact and opinion, have political or ideological biases, and so on. Crossroads ^-talk- 05:09, 6 February 2022 (UTC)

If we can't base distinctions between facts and opinions on sources, what could we base them on? Underlying your reading of LABEL seems to be the idea that the terms mentioned are always to be understood as reflecting the opinion of sources, rather than facts, but this reading is not based on WP policy. Your apparent view goes well beyond what LABEL actually says, namely, Value-laden labels - such as calling an organization a cult, (or) an individual a racist ... may express contentious opinion. The guideline does not state that value-laden terms are necessarily contentious or they they are necessarily matters of opinion, as opposed to fact. What quality sources write about "cults", for example, are typically not A.J. Ayer-style statements of personal opinion but rather (in many cases) claims based on rigorously defined concepts and documented empirical reality. The naïve epistemology that would separate "value-laden" concepts from "factual" statements is, thankfully, not assumed by WP policy, and should not be inserted as a way to overrule quality sources and "win" Talk page discussions, either. Newimpartial (talk) 18:26, 7 February 2022 (UTC)

Value laden labels are implicitly contentious by definition so there's no reason LABEL needs to spell that out. And again there is a point well beyond where BLP or RECENTISM would apply, so that we would be fine calling pre 20th century cults as such in wikivoice, for example, but not some group founded in 2021. --Masem (t) 18:37, 7 February 2022 (UTC)

I'm okay with calling a recently founded group a cult, assuming that there are high-quality sources that say so. There's no reason to obscure this for 100+ years. Wikipedia:Call a spade a spade may not be a functional behavior when dealing with a dispute, but it's perfectly good practice for encyclopedia articles. Charles Manson and Jim Jones and David Koresh founded cults, and it is not too recent for us to say so. WhatamIdoing (talk) 05:39, 10 February 2022 (UTC)

RECENTISM tells us not to. And the cases you name fall outside bound that RECENTISM would cover. (20-some years). --Masem (t) 01:47, 11 February 2022 (UTC)

RECENTISM does not tell editors not to label these men as having founded cults. NPOV and INTEXT does not tell editors to provide in-text attribution for that label in those cases. Why should LABEL tell editors that they must do so? WhatamIdoing (talk) 00:58, 12 February 2022 (UTC)

(I mention those three because you said "we would be fine calling pre 20th century cults as such in wikivoice". Pre-20th-century means 1899 or earlier. We are obviously perfectly fine with calling late-20th-century cults as such in wikivoice.) WhatamIdoing (talk) 00:59, 12 February 2022 (UTC)

I'm not sure this appeal to RECENTISM is valid. We are perfectly fine with calling QAnon a cult in Wikivoice, despite it not existing prior to the first posts by Q in 2017, a mere five years ago. I'm also not seeing anything in WP:RECENTISM about a hard date for when a subject stops being recent. The closest to advice it provides for that is the WP:10YEARTEST, and that is primiarily as a corollary/supplement to WP:GNG. Sideswipe9th (talk) 04:14, 12 February 2022 (UTC)

RECENTISM has to be kept in mind within context of the amount of coverage and other factors, and yes, there's no hard cutoff date where RECENTISM no longer applies. My point from above is that if a group was founded tomorrow, and the day after we had papers calling it a cult, that's probably not our place to be calling it a cult in Wiki voice, that's definitely where RECENTISM would apply. But the QAnon is a good example where there's vast amount of coverage that connect it to a cult (though a surface examination of the refs suggest a bit better preciseness is needed, as it a group compared to a cult but there's not universal agrteement that is a cult - eg a source survey should be done to verify that that is the case or not). --Masem (t) 05:11, 12 February 2022 (UTC)

Calling Qanon a cult in Wikivoice is a great example of RECENTISM failure, NPOV failure, NOR failure, and of course political bias to accept the statement on incredibly weak sourcing. I searched the web for "Qanon", looked at the first 10 RS, all were major media sites plus one think tank paper, and not a single one states as a fact that it's a cult. Most of the articles do not say "cult" at all, a few had comments on how it is cultlike in various ways. The sources cited for "cult" in QAnon, and for Qanon in cult, do not state that as a fact either. Sesquivalent (talk) 07:55, 12 February 2022 (UTC)

Moving forward

In the discussion above, there seems to be broad agreement that:

Labels that are not supported by reliable sources should not be used at all.
Labels that are widely supported as statements of fact by the majority of reliable sources should be used in wiki voice.
If reliable sources present a label as an opinion, or if there is significant disagreement among sources, then in-text attribution must be used.
The current version of WP:LABEL violates WP:INTEXT and WP:NEUTRALITY, which in some cases require labels to be presented in wiki voice.

The main points of contention seem to be where to draw the line between attribution and wiki voice, and how to determine whether a label is widely supported in a way that's not prone to editors abusing/gaming the system.

Obviously a change to MOS will require an RfC, but I'd like to discuss possible ways to word it before going that route. Here's my suggestion as a starting point:

Value-laden labels – such as calling an organization a cult, an individual a racist or sexist, terrorist, or freedom fighter, or a sexual practice a perversion – may express contentious opinions and should be used with caution. Labels may only be used in Wiki voice if they are widely used as statements of fact by reliable sources. If sources present the label as an opinion, or if there is disagreement among sources, then in-text attribution must be used. However, per WP:INTEXT, unnecessary attribution can lead to a neutrality violation. Avoid myth in its informal sense, and establish the scholarly context for any formal use of the term.

Again, this is just a starting point, so feel free to suggest changes. –dlthewave ☎ 19:05, 6 February 2022 (UTC)

I don't see that there is "broad agreement" about those things, so I feel free to suggest that WP:LABEL should be left alone. If there's an RfC, I suppose participants in prior related discussions should be pinged. Peter Gulutzan (talk) 20:29, 6 February 2022 (UTC)

I agree with the first one but I think #2 is a questionable summary, #3 isn't correct because in some cases like that the answer is to not use the label. #4 is not something that has consensus above, especially the part about we are "required" to use any label much less in wiki voice. No, we aren't. Springee (talk) 02:41, 7 February 2022 (UTC)

Sadly, dlthewave, it's going to be a situation that those who disagree with the consensus will be against your summary and those who agree with the consensus will be fine with it. Thus litigating the discussion all over again. C'est la vie. Silver seren^C 03:03, 7 February 2022 (UTC)

I do not at all see "broad agreement" with those bullet points except for the first one. The rest are too in favor of affixing labels. The whole idea of making it easier to affix controversial labels, let alone encouraging editors to do so, is a bad route to go down and will lower our quality and muddy our objective tone. Crossroads ^-talk- 06:04, 7 February 2022 (UTC)

Chiming in with not seeing agreement, except for #1. However, this does not need any change here, as WP:Verifiability policy already requires that *content* that is not supported by reliable sources should not be used at all. I don't see why labels have to be called out as a special subset of content; this is already covered by policy. This just seems like WP:CREEP to me. Mathglot (talk) 07:22, 7 February 2022 (UTC)

I agree with previous comments that the above "agreement" is flawed, hinging on #2 as being what is really the problematic conclusion. I fully understand the need that when a large majority of sources use a label to describe a topic that it is not efficient to use in-line attribution to ascribe that label to 10-20+ different sources, but whatever way we want to end up presenting it, it still needs to be presented as a subjective label or characterization that is the majority opinion of sources, as long as BLP and/or RECENTISM applies. --Masem (t) 13:18, 7 February 2022 (UTC)

There may be no way to implement this, but if 90% of sources agree with the label, it's informative and fine for Wikipedia. If only 70% agree, the it fits the common meaning of "value laden" and is something that is just a statement of somebody's opinion or a concept or neologism that they are promoting and needs in text attribution if even used. So, in "the arctic has a cold climate", "cold" is not value laden (even if someone who lives in Antarctica would disagree) . In "Chicago has a cold climate", "cold" is value laden. North8000 (talk) 14:42, 7 February 2022 (UTC)

North8000, that's basically what I was trying to get across. There are times when a label can be used in Wiki voice and times when attribution is needed, and our guideline should reflect this. –dlthewave ☎ 15:27, 9 February 2022 (UTC)

@Dlthewave, I think there is some level of broad agreement, but I don't think you've quite captured it. For example, "a majority of reliable sources" (i.e., a bare majority) is probably wrong. I think it'd be safer to say that LABELs, if used at all, should be used in compliance with NPOV and INTEXT.

Also, RFCs aren't required to change the MOS, but one might be required to convince editors that this change is needed. WhatamIdoing (talk) 05:44, 10 February 2022 (UTC)

I think that given some of the fundamental disagreements above an RFC is definitely going to be necessary here (especially the objections to point two, which I think is a simple restatement of WP:WIKIVOICE and therefore non-negotiable), but it's worth spending the time to make sure we have an ideal proposed rewording. --Aquillion (talk) 07:29, 10 February 2022 (UTC)

WhatamIdoing, I like it! NPOV and INTEXT are the overriding policy/guideline, and simply referencing them would be an elegant solution. –dlthewave ☎ 18:46, 12 February 2022 (UTC)

A simpler change that might get broader support immediately is to switch the word order in LABEL from "value-laden labels such as .... may express contentious opinion" to "(use of) labels such as ... may be value-laden and thus express (potentially) contentious opinion". This would clarify that the mere appearance of a word on the list is not the thing that is contentious but rather its application in particular cases; that the contentiousness comes from the value-ladenness (whatever the latter means); and would solve or weaken pretty much all the particular examples suggested above where use of a term on the list is not particularly value dependent. Sesquivalent (talk) 12:02, 10 February 2022 (UTC)

I think that could be a useful change. (I consider it separate from the other points, and unless someone objects to it in the next couple of days, I'd encourage you to make that change soon.) WhatamIdoing (talk) 01:01, 12 February 2022 (UTC)

Thus the clear statement that some terms e.g. the listed ones are value-laden becomes the unclear statement that they might be value-laden? I object. Peter Gulutzan (talk) 15:38, 12 February 2022 (UTC)

@Peter Gulutzan, some of those words aren't always value-laden. Consider the word cult: it might be value-laden (e.g., Personality cult), but it might not be (e.g., Cult (religious practice), Cult following). WhatamIdoing (talk) 21:32, 12 February 2022 (UTC)

I think the Lauren Southern article is a good example of why LABEL's strong stance on attribution is needed. The opening sentence of that BLP says "Lauren Cherie Southern (born 16 June[4] 1995) is a Canadian alt-right[a] political activist, and white nationalist[b] YouTuber." Both "alt-right" and "white nationalist" are stated in Wiki-voice. In defense of those labels, a RfC in late 2020[1] found consensus for inclusion of a sentence saying "she has been described as..." (ie the labels are attributed). This is the sort of area where LABEL is important because it makes it clear Wikipedia should not put such claims in wiki-voice. Springee (talk) 17:56, 6 March 2022 (UTC)

So are you advancing the idea that LABEL in its current form is necessary so that RfCs like the late 2020 one on Southern are unnecessary? If not, I am really not clear what you are arguing. Newimpartial (talk) 20:45, 6 March 2022 (UTC)

I disagree with your interpretation of the close. The sourcing for those labels are substantial, and I'm sure that there are more sources that could be added if it was needed. While those two terms are contentious, and are often denied by those who receive them, they do seem factual based on what reliable sources have said about that subject. As per the current wording of LABEL, the only reason why this is allowed to be stated in wikivoice, which I believe should continue both per my opinion and also per the closure of that RfC, is ignore all rules. That article lead would not be improved by stating <insert author here> writing in <insert work here> (times eight plus seven) has stated that Southern is <insert labels here> Sideswipe9th (talk) 21:35, 6 March 2022 (UTC)

That's an interesting one because although the proposed text included attribution, the closer wrote "As words to that effect are already in the lead, there is (in my view) no need to make edits solely to insert the exact wording below." The lead at the time used the same wikivoice text as it does now, so I'm not sure if the closer didn't notice the difference (since the question was more inclusion/exclusion than attribution/wikivoice) or if they thought attribution was unnecessary based on comments to that effect. In any case the stable version implies consensus to violate MOS:LABEL in that particular example. –dlthewave ☎ 22:30, 6 March 2022 (UTC)

You are correct, I misread the close since the question asked included attribution. That said, per LABEL this should not be in wiki-voice. Without reviewing all the arguments made in the RfC I will not claim the close was correct or not. I will say the current sources for white nationalist in the lead certainly would be consistent with the claim but they do not, in my view, rise to the level of wiki-voice, especially since this is a BLP article and as such we must err on the side of caution and attribution. I think only a few specifically state she is a white nationalist (vs makes claims that are consistent with white nationalist talking points). This is exactly the sort of case where we need LABEL (or similar) to make it clear that we always err on the side of attribution when making contentious claims about living people. This is also just the sort of example Masem has warned about. Yes, some RSs specifically call her a white nationalist. Others say she is presenting ideas consistent with (or similar pharsing) but they don't specifically call her a white nationalist. Does this really rise to the standards where we should be able to call her a white nationalist in Wiki voice? Springee (talk) 22:42, 6 March 2022 (UTC)

How is "white nationalist" a contentious claim? –dlthewave ☎ 23:47, 6 March 2022 (UTC)

Is this a serious question? Per Label, "Value-laden labels – such as calling an organization a cult, an individual a racist or sexist, terrorist," It's contentious because it's value laden. It does not have a clear definition as there isn't an obvious line in the sand definition. It seems we are largely in agreement that if the description is universal and stands the test of time we should use it in Wiki voice. I think (hope) we can agree that "white nationalist" is a value laden label. Has it been demonstrated that the label is absolutely universal? I did a news search for her name and found plenty called her far/alt-right but white nationalist? None said she was though a Guardian OpEd said she made white nationalist claims. In fact this Guardian source [2] said, "Southern, a Canadian film-maker and YouTube personality frequently described as a “rightwing provocateur”, has previously rejected claims of being a “white nationalist”, while speaking out against the “cult” of multiculturalism and criticising Muslim immigration into western countries." So they acknowledge it has been said of her and she has rejected the label. They didn't say it in their own voice. This is exactly the sort of cherry picking use of sources to put in value laden labels that editors have been concerned about. Regardless of where the change is made, we shouldn't allow editors to take non-sky is blue evidence and turn it into Wiki voice, especially when we are dealing with BLP subjects. Springee (talk) 02:18, 7 March 2022 (UTC)

The idea that value-laden terms are inherently contentious is debatable; MOS:LABEL even says "may express contentious opinion" (emphasis mine). Consensus at the Southern RfC was that we can still use "white nationalist" even if the subject rejects the label because of course she would. –dlthewave ☎ 03:59, 7 March 2022 (UTC)

You can take issue with my use of the word "contentious" but I'm not sure how you would reasonably argue that "white nationalist" isn't a value-laden label given that "racist, sexist, terrorist" are on the list. The RfC correctly concluded that the lead could include white nationalist. It is less clear if the non-admin close was correct to say attribution was not required. Non-admins closures are not supposed to ignore policies or guidelines. Regardless, this is a perfect example of a value-laden label being applied in wiki-voice based on sources that often don't apply it directly or in their own voice (see my Guardian example). Per the BLP Arbcom case, [3], "do no harm" needs to apply. Attribution vs wiki-voice is part of how we "do no harm". Springee (talk) 04:18, 7 March 2022 (UTC)

I never said that "white nationalist" isn't value laden. –dlthewave ☎ 04:38, 7 March 2022 (UTC)

I'm glad you agree. Do you feel that, as applied in the Southern article, the use of the value laden label, "white nationalist", is in compliance with WP:LABEL? Springee (talk) 04:43, 7 March 2022 (UTC)

I'm not Dlthewave, but the lead of Lauren Southern respects WP:NPOV, the actual governing policy in this area, and that should decide the question per level of consensus, no matter what tortured argument you want to make about non-contentious value-laden labels. Newimpartial (talk) 00:25, 8 March 2022 (UTC)

That the lede includes mention of "white nationalist" which is a term/label used in several RSes is reasonable, but the problem is as Springree is trying to point out, and to what my point is, is that it is not a wide-spread term used near universally among sources, and because we are still in a period that RECENTISM would apply, it would be inappropriate to take a smaller subset of sources stating the term as reason to use the term in wikivoice without attribution. The threshold of when we can make that distinction between when attribution is needed and when it is not is what is very much absent from NPOV even though the implication is there that a near-universally agreed on view can be said in such a manner -- having the establishment of this threshold in NPOV, not only to deal with LABELS but for other viewpoints, would help tremendously to prevent the problem with situations like at the Lauren Southern article. --Masem (t) 01:46, 8 March 2022 (UTC)

As I eventually get around to saying in most of these discussions, it is factually incorrect to construe sources not using a particular label as creating "contention" or dispute about that label, unless they use a contradictory or otherwise opposed label. I don't see any (non-MANDY) dispute about "white nationalist" as a label for Southern, certainly not among reliable sources. Newimpartial (talk) 02:42, 8 March 2022 (UTC)

You are correct, we can not assume that sources that don't use the label would disagree with its use. However, you would be incorrect if you claim that sources that don't use the label and don't dispute its use thus agree with the label. That is an informational blank which we just don't have. Furthermore, it does raise the question, is this label widely used? As an experiment I did a news article search for Southern and then looked at the first 10 hits. The Atlantic more or less supported calling her a white nationalist and so did an OpEd in the Guardian. A non-OpEd in the Guardian instead said that others had called her a white nationalist and that she denied it. Note that isn't in Guardian voice so it would never support putting such a label in wiki-voice. Note that most of the sources I found called her either far or alt right in their own voice. This is the problem Masem mentions, yes, when you do a key word search you can find examples that support "white nationalist" does that reach the level of near universal use? Not in what I can see. As you said, we have a gap where sources don't refute but they also don't directly support. This is a BLP (which is a policy equal to NPOV in importance) and right at the top of that policy it says, "the possibility of harm to living subjects must always be considered when exercising editorial judgment. ". That is why we err on the side of attribution or not mentioning a label even if we have sources that say the person is a white nationalist (or transphobe). MANDY doesn't really apply here (not that it should ever apply). First, we have RSs that say Southern disputes the label thus per NPOV her disputation should be noted. Second, MANDY doesn't say anything about putting a value-laden label in wiki vs attributed voice.

So to summarize, BLP is policy and says err on the side of caution. That sources don't refute the use of a label can't be taken as support for use in wiki-voice. MANDY is a crap essay. Springee (talk) 03:09, 8 March 2022 (UTC)

Re: we have RSs that say Southern disputes the label thus per NPOV her disputation should be noted - what passage of NPOV leads you to that conclusion? It seems rather hasty (as well as flatly contradicting MANDY :P). Newimpartial (talk) 03:21, 8 March 2022 (UTC)

Common sense contradicts MANDY. Fortunately it's an essay and often one used as a last hope when good rational doesn't exist (if it did why would anyone suggest using it?) Using my previously mentioned quick survey of 10 sources I found I think two included a statement that Southern disputed the WN label. I found one that said she was in their own voice (an OpEd thus not a wiki-RS), one that attributed the claim and I think two which talked about her associations with/promoting of tropes that are frequented by white nationalists. So based on that quick survey I would say the sources that support the label are just as strong as the sources that feel the need to point out that she denies it. Certainly we can't honestly say it is NPOV to say "RS say it" but ignore that RS also point out that she denies it.

MANDY is a last hope essay when the sound arguments have left the room. It presumes we are well enough informed to know that a subject is guilty and thus we don't have to given them a fair hearing in the court of Wikipedia. For example if female swimmers who are concerned about a trans women swimmer are denounced as transphobes we should take their denial as further proof they are transphobes. Trotsky would be proud. Let's take this to it's slippery conclusion. If I am accused of walking on the moon and deny it that must be further proof that I have walked on the moon. Value laden labels are subjective even if we can agree on the extremes (how may grains of sand in a hill? Where is the racist line? Where is the x-phobe line?) This isn't a case of using ABOUTSELF to report her reply. This is RSs saying she denies it so yes, per NPOV and BLP we should include it. Again, BLP is policy. Springee (talk) 04:57, 8 March 2022 (UTC)

I'm afraid that wasn't my question, though. I understand your source analysis and balance argument (which doesn't seem to relate to LABEL but proceeds quite well without it). But you also said, we have RSs that say Southern disputes the label thus per NPOV her disputation should be noted, and I am not seeing the logic there. We have sources saying "Southern is X" and we have another (smaller) set of sources saying "Southern denies she is X". What is it in your reading of NPOV that would require the inclusion of the latter piece of information? I understand that to be what you want, but I'm not seeing the policy rationale. If we had RS saying "Southern is not X", or course NPOV would require that we present the matter differently, but that is exactly what we don't have in this, and similar, cases. Newimpartial (talk) 05:05, 8 March 2022 (UTC)

WEIGHT would cover it. Hers is a clear minority view but since sources decided to mention it as such, we should as well. That also complies with BLP since it allows the reader to know this is a claim she disputes. Hopefully we can then provide the evidence sited by sources. Of course it's better if we put more emphasis on the actual things that caused the label to be applied rather than, as is too often the case, trying to make sure the label is front and center. Springee (talk) 05:23, 8 March 2022 (UTC)

So this is the ~~time on Sprockets when we dance~~ part of the discussion where editors propose that Hopefully we can then provide the evidence sited by sources, in order to second-guess the RS, presumably. Now in the case of Lauren Southern, the sources have very good evidence - she made one of the most prominent white nationalist "documentary" films of the last decade or so - but even in this case it just isn't our job to relitigate attributions that are uncontroversially made in well-established, reliable sources. Newimpartial (talk) 14:06, 8 March 2022 (UTC)

Since I haven't watched the film I will have to take your word for it. Still, NPOV, LABEL, BLP etc don't say we use a value laden label because we know it to be correct. That is the gap here. We are applying a label in wiki voice that most sources do not apply in their own voice but that is OK. Our rules don't need to apply here because she's a bad person. Springee (talk) 14:45, 8 March 2022 (UTC)

As far as I am aware, nobody in this discussion is proposing that Our rules don't need to apply here because she's a bad person. That is at least a straw man, and more likely an AGF fail - if not a CIVIL violation. Does anyone besides possibly Southern herself question that she is a white nationalist? I am not aware of any RS going so, but Springee, if you were to point to some, that would move the discussion forward. Sources that simply do not use that term do not, however, call its use by other sources into question.

And while I recognize your belief that MANDY is a last hope essay when the sound arguments have left the room, I don't think the community of WP editors generally agrees with this characterization. The (quite rational) kernel of MANDY is that of course people deny things about themselves that are factually accurate but that might cause loss of reputation to admit, so the denial does not necessarily create a "both sides" situation about what reality is. You can disagree with reality as much as you want, Springee, but as an epistemological realist I'm here to tell you that it is what it is. Newimpartial (talk) 17:27, 8 March 2022 (UTC)

Does anyone besides possibly Southern herself question that she is a white nationalist? It is not whether anyone questions that, it is an issue of how frequency or common in the sources is she referred to that way, particularly those purported to be factual reporting pieces. If nearly every source about Southern calls her a white nationalist then we're probably right to say it without attribution. But if its only a small fraction,, we better use attribution, or even consider how DUE that might be (which probably still is due in her case). Instead, we have cherry picking of a couple of sources to make the leap to use labels without attribution which is a major problem. See my comment immediately below for where this needs to be resolved first. --Masem (t) 17:47, 8 March 2022 (UTC)

The thing is, I am not taking an extreme position here: I agree, for example, that more well-attested terms should be presented in wikivoice and terms that are undisputed but not widely-used should receive in-text attribution. I'm not sure what side "white nationalist" falls on in the Lauren Southern case: to me, a "documentary" that is widely recognized by RS as "white nationalist" provides additional support (re: WEIGHT) for the use of "white nationalist" for its major creator, even beyond the sources that literally say "Lauren Southern is a white nationalist filmmaker" (we do, of course, need the latter kind of sources to avoid SYNTH, but they aren't the only relevant sources).

What I don't ever agree with is comments like Springee's over at Talk:Lauren Southern, where he recently proposed that not only the less frequently used "white nationalist" but also the near-universal label "alt-right" should be given in-text attribution because of an apparent MANDY denial. Newimpartial (talk) 18:05, 8 March 2022 (UTC)

You have confused unrelated paragraphs. Please note the difference in indentation and the two signatures. My comment on MANDY was related to the comment immediately above mine. It only related to excluding Southern's reply to her critics. Note the next paragraph is less indented indicating it was in reply to a different edit/editor. It says nothing about MANDY but does cover the question of attribution. If you looking a diff/see changes sometimes the indentation is easy to miss. Springee (talk) 18:30, 8 March 2022 (UTC)

So then, what is your rationale for deeming alt-right and white nationalist "contentious", if it isn't the MANDY denial? Facts about which all RS agree can't be "contentious" no matter what one may otherwise believe about epistemology. (My interpretation that you were using the MANDY denial as a rationale for "contentious" was one of my frequently off-base attempts to AGF; clearly I should rely on Occam's razor more often.) Newimpartial (talk) 04:43, 10 March 2022 (UTC)

Facts about which all RS agree can't be "contentious" no matter what one may otherwise believe about epistemology. That's not really true, particularly given that we have systematic severed much of the right-leaning sources from being RSes. For example, in the current debate about Florida's "Don't Say Gay" law, nearly all RSes are critical of the bill, but the right-leaning side of America praise it, which is documented in these sources. Any views about the bill from what we have deemed as RSes will be contentious because doesn't consider all possible views. Now, what is important to consider is RECENTISM as in the short term we should not be trying to decide which is the most prevailing view in RSes, but instead documenting such controversies and avoid going too far into opinion and analysis (which usually also means avoiding labels as well). With the passage of time, and looking at the ten-year view, then we can start looking at how the situation around a controversy has gelled in RSes published since then, at which point it would become reasonable to focus on one side of the issue if that's the only major view given in that long-term view. It should onlybe in the long-term view that we can decide it is appropriate to state views (including the use of labels) in wikivoice w/o attribution. --Masem (t) 05:11, 10 March 2022 (UTC)

Newimpartial, I'm sorry you confused how talk page indentations work. Again the paragraph critical of MANDY had nothing to do with the paragraph that said "alt-right" and "white nationalist" are value laden labels. Perhaps if I did those as separate edits it would have been more clear? Alt-right and white-nationalist are clearly labels to which LABEL applies. Sadly we are simply spinning and spinning when replying to one another. Perhaps you can re-read Crossroads well thought out reply to your question. They said it better than I was. Springee (talk) 05:24, 10 March 2022 (UTC)

Springee, I do understand how indentations work (and have since before Usenet days); I simply do not understand your logic. LABEL says that valie-laden labels ... may express contentious opinion - not that they necessarily do so - but you wrote (in the diff I linked above): This is a BLP and such contentious claims are absolutely unacceptable in Wiki voice. But you haven't shown any claim to be contentious ... if you are under the missapprehension that repeated assertion is a form of proof, it ain't. Newimpartial (talk) 06:01, 10 March 2022 (UTC)

Again this is getting very repetitive. The labels in question clearly qualify as value laden. Often we will say they are contentious claims for this reason. I probably could have been more clear if I said value laden at that time. They are also contentious for the reasons Masem laid out just a few edits back. I can grant that "Atl-right" might have sufficient sourcing to use in Wiki voice (but LABEL suggest otherwise) but "white nationalist" does not. There is no misapprehension here, at least not on my part. I guess assuming MANDY was related to the value laden labels thing was a misapprehension but following the indentations should have addressed that. Again, this is getting repetitive. Springee (talk) 06:50, 10 March 2022 (UTC)

The labels in question clearly qualify as value laden. Often we will say they are contentious claims for this reason So, if it is value-laden, it is automatically contentious? We cannot say Hitler was a Nazi, we cannot say the Unabomber killed people, we cannot say Putin started a war, we cannot say anything that makes people like or dislike someone except with attribution?

If that is not what you are saying, then you did a very bad job in saying whatever you are saying. --Hob Gadling (talk) 11:49, 10 March 2022 (UTC)

HG, what brings you to this discussion? I'm sorry that in your opinioin I've done a bad job. The Hitler Nazi example has actually been previously discussed and your argument is actually rather poor since we can show things like Hitler was a member of the actual NAZI party and the Unabomber actually sent bombs that killed people. Your understanding here is quite confused. Springee (talk) 12:53, 10 March 2022 (UTC)

I have been in this discussion since March 2. Please WP:FOCUS.

Hitler was a member of the actual NAZI party and the Unabomber actually sent bombs that killed people It does not matter what is true, only what is verifiable. See WP:TRUTH. But this discussion is about whether value-laden-ness is even stronger than verifiability. If you are right, those things are automatically contentious and must be attributed. In other words, if it matters whether they are true, then you are wrong and they are not automatically contentious. --Hob Gadling (talk) 14:32, 10 March 2022 (UTC)

Looking at the comments from editors who wish to discourage the use of labels and who insist that editors must always "use in-text attribution", I think that there is a general theme of equating "value-laden labels" and "words indicating a rejection of political positions promoted by the US progressives/lefties".

The examples given are about calling people white nationalist/transphobic/neo-Nazi. There are no examples given to object to labeling people as holding the opposite positions (e.g., anti-racist, straight allies, anti-fascist). The underlying assumption seems to be that it's disreputable to hold right-wing positions. I do not think we should make right-wing views some sort of dirty secret that can't be mentioned, or that can only be mentioned if the sources are overwhelming, and even then, you need to pretend that it's merely the personal opinion of a small group. That's not my idea of being neutral. WhatamIdoing (talk) 17:47, 10 March 2022 (UTC)

This is an apples to oranges analogy. The latter group of labels are usually proudly claimed by those who hold them, while the former are a mixed bag, but are often or even almost always used as snarl words or by ideological opponents. Crossroads ^-talk- 04:48, 11 March 2022 (UTC)

I'm sure that's true in your culture, but in other parts of the world, publicly supporting gay rights can get you shunned or worse, and people who oppose fascism are considered unpatriotic and anti-government. Positions endorsed by US progressives are, in many cases, the opposite of the positions endorsed in all but the richest of countries. Look at LGBT rights in Nigeria: almost everyone in the country opposes gay rights. In a sizeable part of the country, gay men can be stoned to death; in the rest, they are at risk of spending 14 years in prison. I would not be surprised if the tiny number of people who support gay rights want to hide their views. Sixty million Nigerians speak English and are therefore likely readers of this wiki. The words you see as proudly claimed virtues would look like snarl words to them. WhatamIdoing (talk) 04:07, 12 March 2022 (UTC)

I think that someone who took a negative view of others supporting gay rights would be unlikely to refer to those supporters as "straight allies". (I have occasionally seen "homosexualist" used as a derogatory label in this context.) Cheers, gnu 57 15:08, 12 March 2022 (UTC)

I can't find it quickly, but does anyone else remember the editor who turned up at one of the dramaboards some years back to demand that editors allow him to put "I am homophobic" on his user page, because he really was homophobic, and he thought it was important for other editors to know that? We don't permit such messages in userspace. It wasn't Sla20va (whose userpage I'll blank in a moment), but I can't find the discussion I'm thinking of. Anyway, the relevant point is: It's a big world. A label you think is derogatory is one that others will proudly claim, and vice versa. WhatamIdoing (talk) 18:34, 12 March 2022 (UTC)

we have systematic severed much of the right-leaning sources from being RSes So, suddenly, unreliable sources get a vote because almost all American right-wing sources have become unreliable? Here, too, if that is not what you are saying, then you did a bad job in saying whatever you are saying. --Hob Gadling (talk) 11:49, 10 March 2022 (UTC)

Sources that mix facts with left-wing opinion are much more often judged reliable at RSN, while sources that mix facts with right-wing opinion are almost always judged unreliable if not deprecated. Even if one were to believe that editor bias at RSN doesn't affect these judgments, and that errors and leaving out "relevant context" are forgiven equally on both sides, this still creates a problem for any analysis of biased or opinionated labels. This is not to say that RW and LW sources are equally reliable, but rather that I think that LW sources at RSP need more cautions about identifying and attributing biased opinion, and some should have their scope of claimed reliability reduced. WP:ROLLINGSTONEPOLITICS was a good start. Crossroads ^-talk- 22:37, 14 March 2022 (UTC)

Sources that mix facts with left-wing opinion are much more often judged reliable at RSN, while sources that mix facts with right-wing opinion are almost always judged unreliable if not deprecated. If that is the case, then you should try to remedy that situation by solid reasoning in individual cases of sources at RSN, instead of rewriting the rules to reintroduce all unreliable sources through the back door here. --Hob Gadling (talk) 06:38, 15 March 2022 (UTC)

(edit conflict)No AGF or civil violations here. It is perfectly reasonable to expect that many/most editors when presented with an "on the fence" sort of decision will lean in the direction they subjectively feel is correct. As for MANDY, the problem is you are suggesting we are sufficiently informed to make subjective decisions. When does someone become "tall"? Is anyone above average tall? What about 75th percentile? Do we need to wait for the 90th percentile? Does the answer change if we ask people who are in the 25th percentile vs 80th percentile? Here MANDY is being called when the subject isn't binary. We have a value laden binary label (is or isn't labeled "white nationalist" but is based on a gray scale. You say that shade of gray is basically black. Someone else says it's not black. Per MANDY we can ignore a claim that dark gray isn't black because well, we think it is close enough to black. You are arguing that we can do that even though RSs say "she refutes the label". If this were a case where the line is clear (people who have been convicted of a felony) it would be easier to accept a MANDY like justification. When we are dealing with shades of gray and RSs say she disputes the label, well we should to. That falls nicely in line with BLP's "do no harm" philosophy. Do you think BLP is wrong to say we should use caution? Springee (talk) 18:11, 8 March 2022 (UTC)

I certainly do think we should use caution, and I don't see anything in this section (or in MANDY) agitating against caution, either. But I don't accept shades of gray as a reason to change our WP:V or other core policies. Whether a label is clear or a matter of dubious certainty is a matter for the RS like everything else, and I don't see "caution" in your attempt to promote a rigid version of LABEL while denying the principle of MANDY. If Southern says she is charcoal, not black (to use your awkward gray scale metaphor) that wouldn't be any more germane than if she said she were green: reliable sources (like those about her being part of the "alt right") still get to decide, not the BLP subject.

As far as things that are or aren't binary, let's stick to your height metaphor. If the sources describe, say, a female soccer/football keeper as "tall" at 5'9" (or 1.75m or whatever) and the subject says in an interview, "compared to male keepers I'm not tall", that would not affect either her height itself or the way WP should describe her height, as far as I can tell. "Tall/not tall" employs the same logic as your shades of gray metaphor, so I hope this example clarifies things for you. Newimpartial (talk) 18:36, 8 March 2022 (UTC)

Height works as an example but I don't think you really made a point with your example. Yes, "tall" can be context dependent. A "tall" 8 year old is unlikely to be "tall" compared to an adult but more to the point, many of these labels, even ones like "white nationalist" lack a clear, binary definition. Is 1.8m tall, what about 1.85 or 1.9? I think we would all agree that 2.3m is tall and 1.45 is not. That is our gray scale but we don't have a line in the sand when something becomes "tall". The same is true for something like "white nationalist". We can certainly agree when things are going in that direction but the location of the actual line is not clearly defined. Many will say 1.85m is tall but someone who it 1.85m may not agree and say they aren't tall. Is that because they aren't or just because tall isn't exactly defined. Note that tall is typically not seen as a negative but if we were talking about short, well perhaps someone would dispute a claim that they are short. If Tom Cruise says he isn't short should we claim MANDY? Springee (talk) 19:01, 8 March 2022 (UTC)

I'm not sure you grasped the point of my example. Wikipedia describes people as tall/alt-right because reliable sources describe them that way, not based on any absolute scale. We don't conjure up an on-wiki contextual filter to decide what is or isn't tall/alt-right; all we do is evaluate the WEIGHT of the sourcing. And re: your final example, if the RS were to describe Cruise as short, then BLP policy would not entitle him to a denial, and attribution for both sides, in article text. In that sense yes, MANDY would apply. Newimpartial (talk) 20:15, 8 March 2022 (UTC)

We have two topics, MANDY and attributed labels. Yes, when sources describe a subject as tall/short we should as well. However, should we do it in Wiki-voice or with attribution. We have some sources that say, "she is tall" others say "she has been called tall" other say "she". So do we put "tall" in wiki-voice or not? I would say no because sources don't uniformly call her tall in their own voice (1/3 of sources in my example). 2/3rds either make the claim or say others have made the claim. 1/3 don't mention it at all. In my book that means we have enough to make an attributed claim. When a BLP says, "I'm not short" and several sources note they said they aren't short, do we ignore that or say RSs make the statement of fact that the BLP subject has made the semi-subjective claim that they aren't short? Yes, they felt it was important to report her denial of height, we should to. Springee (talk) 20:28, 8 March 2022 (UTC)

Well, your book isn't Wikipedia, and you haven't offered any policy support for your personal thresholds re: attribution. In my book, if one third of sources call a person "short" in editorial voice, another 1/3 gives attribution and the final 1/3 says nothing about height - then the statement that the person is short is, in fact, uncontroversial so we should say "short" in wikivoice. This doesn't change if the BLP subject offers a MANDY denial of shortness - attribution would only be required if the number of sources saying "short" or "said to be short" were much smaller in relation to the total, or if some reliable sources stated that the subject "is actually not short at all" lol. Newimpartial (talk) 20:35, 8 March 2022 (UTC)

Fortunately LABEL is part of the wiki book and it says attribute when dealing with contentious claims. That a person denies being short can count as a disputed claim but we would discount the weight we give it since the person is involved (that doesn't mean automatically give it zero weight). Anyway, given the hypothetical height sources I can see you have ignored BLP because you err on the side of putting things in Wiki voice when only a minority of sources said tall/short in their own voice. Consider another case. Let's suppose 2/3rds of sources say "Tom C is often descried as short". 1/3 say nothing about height. Would you suggest that is sufficient to write Tom C is short in wiki voice? It would be logically consistent with your previous 1/3 1/3 1/3 based argument. Springee (talk) 21:18, 8 March 2022 (UTC)

No, I'm afraid I have more angels on the head of my pin than that: if no sources say "short" in their own editorial voice, then neither should we, but we should say "often described as short" where that follows the clear WEIGHT of the sources - which would also be a LABEL violation if short were considered "value laden", since it doesn't actually provide proper in-text attribution. Which, for me, is yet another place where LABEL breaks but BLP, NPOV and INTEXT have it right. And I don't see any justification in WP:BLP itself for sources not mentioning an undisputed fact being used as a pretext to couch a fact as an attributed opinion. But you seem to have an, err, "unique" reading of what BLP requires.

Lest anyone think my support for "often described as" is some kind of pretext, I have previously argued for precisely this style of language for real cases similar to the hypothetical one at issue here. Newimpartial (talk) 00:40, 9 March 2022 (UTC)

Let's go back to that 'one third of each' thought experiment. At minimum, if there are just as many sources putting a LABEL in attribution than there are using it in their own voice, then per BLP's requirement to err on the side of caution it is absolutely not proper to then seize on that 'own voice' half and demand that the article side with them against the others. It is against WP:WEIGHT. And how to account for that last third depends on how they discuss the person. If they have nothing to do with the aspects of the person's activity from which the LABEL arises, they can be disregarded; but if they do discuss it but do not feel the need to bring up the LABEL at all, then yes, they must be accounted for.

Sources choosing not to endorse or even mention a LABEL absolutely is a choice and must be accounted for. Disregarding such sources completely rigs the analysis in favor of LABELing and is simply not neutrality.

And WP:MANDY is indeed a crappy essay. Some editors cite it like it's policy, but it is the furthest thing from it. It was only created in November 2019, well after the eruption of the culture wars in the mid-10s, and has been used as a cudgel in service to them. It contradicts actual policy, found at Wikipedia:Biographies of Living Persons#Denial. It notes that "Newspapers typically give the subject the last word" but advocates we ignore what these RS say and not include it. This isn't to say that a person's denial must have equal weight in an article - just that 'we can exclude their denial' is not tenable. Crossroads ^-talk- 03:07, 9 March 2022 (UTC)

Crossroads, your premise appears to be that if an RS attributes a label - even without noting any contrary view or disavowal by the subject - they are somehow contesting the application of the LABEL. This seems to me to be ~~nonsense~~ highly speculative. I don't think there is anything uncautious about adopting in wikivoice any labels that are not contested within RS discourse. If a reliable source, to use the earlier example, provides the precise height of Tom Cruise without referring to him as "short", that cannot reasonably be interpreted as evidence against that descriptor - it is simply non-evidence, the same as a source that doesn't mention his height at all.

As far as WP:MANDY is concerned, I know that the essay is controversial among editors with certain perspectives. However, I have seen much more support for it than opposition on-wiki. Indeed, it would be fun to test the hypothesis - which I find quite plausible - that more editors support the MANDY principle than support LABEL in its current form. Newimpartial (talk) 21:20, 9 March 2022 (UTC)

Yes, if half or more of sources mention a label but don't use it - a use/mention distinction - then it is not WP:DUE to use the label in wikivoice. It's cherry-picking. And such sources almost always discuss those who disagree with the label anyway.

Suppose person X is called a transphobe by PinkNews and LGBTQ Nation, but X's comments on transgender topics are discussed in the New York Times, Guardian, Washington Post, and CNN. These four mention that X has been accused of transphobia by such and such LGBTQ advocates, but report all sides in the controversy and never call X that themselves. Should Wikipedia contain the text "X is a transphobe"? If not, why not? Crossroads ^-talk- 05:05, 11 March 2022 (UTC)

In your example, I don't think Wikipedia should include "X is a transphobe" because that would under all circumstances be poor encyclopaedic writing, just as we would not say "X is a homophobe". In this instance, I would use "X has widely been accused of transphobia" because this is a factual statement backed up by all sources.

This really does depend on the case, though. If the pattern of (editorial voice or attributed) sourcing was the same but the term in question was "anti-gay activist", I would go with "anti-gay activist" in the absence of sourcing that actually casts doubt on the applicability of the label. To me, personally, the term "homophobic" is more difficult to use objectively (outside the context of psychological analysis) than "anti-X activist" regardless of what X is. Newimpartial (talk) 12:35, 12 March 2022 (UTC) Newimpartial (talk) 12:23, 12 March 2022 (UTC)

Based on the hypthetical (two biased sources making the claim, several neutral sources nothing the claim) I think "widely" would be overstating. "X has been accused of..." is more appropriate. Springee (talk) 14:35, 12 March 2022 (UTC)

For the record, I don't accept the characterization of LGBT sources as biased and mass-market outlets as neutral. I doubt that many editors would accept this characterization. Newimpartial (talk) 14:42, 12 March 2022 (UTC)

Ok, let's accept that. You still have the majority of sources saying described as, not "widely described as" thus we can't imply this is a widely held view. Springee (talk) 15:20, 12 March 2022 (UTC)

While this isn't the first time I have heard the argument that "even if all available sources either apply term Y or note that others apply term Y, we cannot say, 'widely described as' term Y", I have never been able to understand the rationale for this. If all of the RS said either that the earth is flat or that the earth is described as flat, I don't see why we couldn't note that the earth was widely described as flat. Newimpartial (talk) 15:32, 12 March 2022 (UTC)

Perhaps a simplified example will help. We have 10 people (P1 to P10). P1 says it's a hot day. P2 through P10 say, today was described as hot because P1 said it was. Would you say the day widely described as hot? No because only 1 person described the day as hot. The others simply referenced the first. We don't multiple sources describing what a single or few sources/people said as "widely". The reason is, as an encyclopedia we should never overstate a claim. We err on the side of caution when making claims, even ones that aren't value-laden like Wednesday was a hot day. Springee (talk) 17:32, 12 March 2022 (UTC)

That's a bad example. How do you differentiate between people who agreed with P1 because P1 simply said it, versus people who agreed with P1 because it is hot and P1 simply said it first? Sideswipe9th (talk) 17:35, 12 March 2022 (UTC)

That is what makes this a good example. We don't know why P2-P10 said the day was described as hot. They might have done it because they agree. They might be neutral on it and are just passing along the information. They might be doing it to show disagreement. We don't actually know unless they say as much. It's OR on our part to assume and then edit based on that assumption. If the source isn't willing to put the claim in their voice we cannot upgrade it from an attributed claim to one in Wikivoice. In effect we are putting the fact that it was described rather than the fact itself in Wikivoice. Springee (talk) 21:25, 12 March 2022 (UTC)

(e-c) This is a great example, except that in your example the remaining sources only offer the description because P1 said it was. This isn't true of my flat earth example, though, nor is it true of any of the "anti-X activist" or "Z-phobic" examples we have been discussing. In the set of cases I have been taking about, we have roughly as many sources making the attribution in their own editorial voice as we have attributing it to (specified or unspecified) others. Newimpartial (talk) 17:40, 12 March 2022 (UTC)

If the source says they endorse "it was hot" then it's no longer P2 reporting on what P1 said, rather it's P2 saying it but that takes more than just saying, "sources described it as hot". With our 10 people what percentage do we need before we describe it as hot in wiki voice? I'm not sure, certainly more than 1 (assuming "hot" were a value laden term vs just a description of the weather) but less than all 10. Springee (talk) 21:25, 12 March 2022 (UTC)

This is an unrealistic, artifically constructed example. If the day was not hot or even the hotness is dubious, then there will be sources among the ten which say that P1 is wrong or dubious. There will also be Wikipedia editors who disagree with P1 and find other sources which do. As a result, P1's opinion will definitely not end up being stated in Wikipedia voice, no matter how this discussion ends. On the other hand, if it was hot, then there is nothing wrong with stating it in Wikipedia voice. --Hob Gadling (talk) 07:40, 13 March 2022 (UTC)

Simplified examples are often an effective way to to illustrate a point. Pick up a copy of Cooter and Ulen's Law and Economics text book and you will find many such examples. You are correct, if "hot" is disputed we should expect a source to specifically dispute it. However, what if hot isn't seen as overly important? What if there is a source that disputes it but we haven't found it? What if "hot" was defined by someone who was used to Norwegian vs Saharan weather? We can speculate why sources chose to attribute vs state in their own voice but we need to be careful we don't tread into OR when doing so. I will grant that "hot vs warm vs cool vs cold" is unlikely to be considered value laden or contentious but, as I said, simplified examples are often helpful to illustrate/discuss a point. Springee (talk) 15:43, 13 March 2022 (UTC)

There will always be theoretical cases where the rules will lead to a false result. For those cases, we need flexible rules. The current strict wording of LABEL is not. It uses the multiply false equation "value-laden" = "label" = "contentious" = "needs to be attributed", which forces us to say "Alice and Bob say that David Irving is a holocaust denier" instead of "David Irving is a holocaust denier". This is a real-world example where the current wording has bad consequences. A more flexible wording can handle this one right, and it can handle your imaginary, unlikely example right if someone makes a case for it. It is better if unlikely cases require discussion than if existing examples require discussion. --Hob Gadling (talk) 14:53, 14 March 2022 (UTC)

How bad are these consequences? Your concern would be valid if the core information couldn't be included in the article unless we accept in in wiki-voice. That isn't the case here. Attribution doesn't keep the label out and if we are going to include the label we can (and should) include the substance that supported the label. If we put the facts before the reader and that sources have called X a denier what loss is suffered because we don't say it in Wiki voice? I'm not familiar with the Irving article so I make these as generalized comments. However, consider BLP and it's idea that we err on the side of doing no harm. Consider that Wikipedia doesn't pick sides (IMPARTIAL). All this supports the idea of erring on the side of keeping things out of Wiki-voice in cases like these. Springee (talk) 15:24, 14 March 2022 (UTC)

Providing attribution in such cases violates WP:INTEXT and WP:NPOV. It is fundamentally not true that "some people say David Irving is a Holocaust denier". Irving actually is a Holocaust denier, and the English Wikipedia should not try to pretend that it is only some people's opinions (including the subject's own). WhatamIdoing (talk) 15:47, 14 March 2022 (UTC)

As I said, my comments were generalized. That said, looking at INTEXT says, "should always be used for biased statements of opinion." That section also points to part of NPOV (WP:ATTRIBUTEPOV), "Biased statements of opinion can be presented only with in-text attribution. " So are these labels statements of fact or opinions? That is where the subjective part comes in. The example used in ATTRUBITEPOV is a positive one, "John Doe is the best baseball player". I think we could agree that even, "John Doe is a great baseball player" is subjective. Replace "great baseball player" with "racist person". Is that now a statement of fact or opinion? The idea that LABEL isn't supported by NPOV is wrong in so far as NPOV says we should attribute biased statements of opinion. So the question is when do these labels move from opinion to fact? Springee (talk) 16:54, 14 March 2022 (UTC)

The main flaw of ~~your~~ some of you people's reasoning is the concatenation of syllogisms. Because someone put "denial" on the list, and i is a list labels, "denial" is a "label". Because it is a "label", and all lables are subjective, it is "subjective". Because it is "subjective", and subjective words must be attributed, it must always be attributed. This is a polysyllogism, logic rookie mistake. You can do stuff like that with mathematical terms where every set-subset relation is exactly true. But with real-life words outside of mathematics, it rarely works because a statement like "labels are subjective" is only true in most cases, but not in all. (By using several non-quite-syllogisms in a row, the inaccuracies pile up and the number of exceptions grows.) And it does not work in this case, as can be seen from the fact that putting in true statements leads to the false statement that "David Irving is a holocaust denier" is just an opinion. This one example disproves your whole logic.

You ask, "what is the harm?" The harm is that we give credence to clearly wrong worldviews. It decreases Wikipedia's reliability. It introduces a false balance. You can still "err on the side of caution" in the individual cases, but if you put the error as a dogma into the very rules that have to be applied in every case, you will do real damage because you force everybody to err.

my comments were generalized Exactly. Read Overgeneralization. --Hob Gadling (talk) 07:03, 15 March 2022 (UTC)

I appreciate the use of big words. However, you didn't use straw man even though it was a key part of your argument. I never claimed all labels are subjective and more critically that they are value-laden or biased statements of opinion. That was your own creation hence a straw man. Your second paragraph sounds like the justification of an inquisitor presuing~~perusing~~ the heretics. You are certain the label is correct thus it is a sin to not use it in Wiki-voice. Is that policy some where? So you are saying we can never be wrong or it isn't possible to have a fence sitting case or a case where reasonable people disagree. Thanks for the link to overgeneralization. I hope it was useful to you. Springee (talk) 11:00, 15 March 2022 (UTC)

The strawman's name is Masem. He uses that reasoning. Of course, defending the current LABEL (all labels must be attributed) implies that sort of logic, but nevermind. I struck out my faulty accusation, in line with the essay User:Hob Gadling/Admit mistakes I am writing, and replaced it by a better one.

An inquisitor perusing [sic!] the heretics does not reason with people who disagree with him, he imprisons, tortures and kills them. I probably do not need to point out that I did nothing of the sort. I know someone who would save that difflink of yours for later use as an example of a personal attack, then put a number of them on the Talk page of a friendly admin (not a public admin board, it could backfire!) in order to get you banned, but I am not that sort of person. So, I'll just tell you to WP:FOCUS.

Back to the subject. I am not saying we can never be wrong or it isn't possible to have a fence sitting case or a case where reasonable people disagree, I am saying it is possible to have a non-fence-sitting case or a case where reasonable people agree. LABEL currently turns everything into a fence-sitting case and demands we treat everything as a case where reasonable people disagree. So, guess what, you made a strawman! I don't expect you to retract it, like I retracted my faulty statement. --Hob Gadling (talk) 11:54, 15 March 2022 (UTC)

My dearest, humble apologies for my typo. I'm sure that totally negated my points. I don't agree that LABEL turns everything into a fence-sitting case. Springee (talk) 12:07, 15 March 2022 (UTC)

I cannot image how one would get the idea that was about the typo. It was about this strawman:

I am not saying we can never be wrong or it isn't possible to have a fence sitting case or a case where reasonable people disagree, I am saying it is possible to have a non-fence-sitting case or a case where reasonable people agree.

But I repeat: I don't expect you to retract it. --Hob Gadling (talk) 12:27, 15 March 2022 (UTC)

You can repeat it all you want. I'm glad you agree that we can have fence sitting cases etc. So for that reason we should err on the side of caution. If that means we are less aggressive than you might feel is correct when dealing with labels that is an acceptable tradeoff. You may not have used the words "we can never be wrong" but your actions and arguments suggest otherwise. This discussion isn't productive. Have a nice day. Springee (talk) 12:53, 15 March 2022 (UTC)

So Springee: are you arguing that, because fence-sitting cases exist, we should therefore have a guideline that treats all cases as fence-sitting cases? Because when all the snark is chipped away from your last several comments in this subthread, that's what remains. Newimpartial (talk) 13:06, 15 March 2022 (UTC)

Yeah, a lot of snark was bouncing around in that last back and forth. Still, if that is all you found your chisel went too deep. I was replying to HG's grandiose statement that we have to protect from false information. Such a statement didn't allow that, perhaps, editors might not have the perfect dip stick of truth when deciding on such matters. Springee (talk) 13:48, 15 March 2022 (UTC)

We may or may not have a dipstick, but we have the core policies of WP:V and WP:NPOV, which work a damn sight better than MOS:LABEL, IMO. We also have WP:NOTTRUTH, which you as well as HG should both be keeping in mind. (And I find it hilarious that my phone kept trying to correct "damn" to "darn" - let me swear, dammit!) Newimpartial (talk) 14:18, 15 March 2022 (UTC)

Darn it, what's wrong with your damn phone? :D Springee (talk) 14:53, 15 March 2022 (UTC)

But WP:BIASEDSOURCES do exist. So in this matter, which sources are BIASED, and how do we tell? I agree that sources can be BIASED and reliable - but that does affect how we relay their claims. Crossroads ^-talk- 05:42, 13 March 2022 (UTC)

And this is precisely why we should not change LABEL - it enables this sort of article slant in favor of contentious claims. "X has widely been accused of transphobia" because this is a factual statement backed up by all sources - except none do, as I explicitly noted that 2/3rds of sources say who specifically is describing X as such - LGBTQ activists - and that those same sources "report all sides in the controversy" - i.e. explaining why some disagree with applying the label. The minority of sources using the label in their own voice clearly have an editorial position in line with LGBTQ activists, consistent with the description in the other sources. Widely is therefore 100% unsourced and hence an NPOV and BLP violation, and you supported this instead of an actually accurate attribution to LGBTQ activists. And all this at a guideline RfC where one might an expect an editor to be more careful, let alone in the field. I have zero interest in enabling such edits or disputes and this illustrates why I oppose this effort to water down LABEL. Crossroads ^-talk- 05:42, 13 March 2022 (UTC)

This is actually a strength of your example, because it points out an unexamined assumption of your argument. You assume that sources that report all sides in the controversy and never call X that themselves (your earlier description) therefore explain why some disagree with applying the label (your new specification). In other words, you assume that disagreement over the label is discussed explicitly among these sources. In line with the actual examples underlying this discussion, I did not make that assumption; rather, I specified the absence of sourcing that actually casts doubt on the applicability of the label. I don't think anyone is proposing that the principle underlying LABEL be set aside in cases where there is any debate whatsoever, among quality, reliable sources, of the veracity of the applicability of the label in question. If reliable sources offer a rationale why a label doesn't in fact apply to a subject, then Wikipedia should never make that statement in Wikivoice - the problem with the current text of LABEL is that it requires intext attribution even in cases where all sources agree. Your defense of the current text is looking more and more like a slippery slope argument. Newimpartial (talk) 10:42, 13 March 2022 (UTC)

"Casts doubt" is ambiguous here because that could mean you require sources casting doubt in their own voices, or merely reporting that some doubt it.

Let me try to remove all ambiguity: Suppose person X is outright called anti-trans by PinkNews and LGBTQ Nation, but X's comments on transgender topics are discussed in the New York Times, Guardian, Washington Post, and CNN. Each of those 4 says something like "X has been called anti-trans by LGBT advocates, but this has been rejected by X as well as by radical feminists and conservatives." Details for each side are given equally in surrounding material in those sources. How would you describe X on Wikipedia? Crossroads ^-talk- 22:23, 14 March 2022 (UTC)

In the case of that particular, complicated example, I would support a complicated, attributed summary statement. Newimpartial (talk) 22:55, 14 March 2022 (UTC)

per BLP's requirement to err on the side of caution it is absolutely not proper to then seize on that 'own voice' half and demand that the article side with them against the others: Not exactly. First of all, you're assuming that the label is at least "contentious", which may not be true at all. It might not even be derogatory. For example: Elizabeth II is kind of short. No sources contradict this. Also, no sources say this is derogatory or a problem in any way.

Should we include it? Hmm, that's a matter of due weight, of course, but I wouldn't usually include height in an article about a head of state (the existence of List of US Presidents by height and a paragraph about Napoleon notwithstanding). It's quite the opposite for an athlete: in those articles, you would usually default to inclusion.

What LABEL currently says, though, is that if you decide that it's due weight to say "She's short", then you have to include in-text attribution for it. You can't just label her with "short"; you have to say "She's short, according to <list of people>". This would be silly. WhatamIdoing (talk) 03:17, 10 March 2022 (UTC)

I think this analogy needs to be put out of its misery. Why would Wikipedia start using vague terms like "short" to describe people anyway? Crossroads ^-talk- 05:11, 11 March 2022 (UTC)

We use that language in hundreds of articles: Mireille Mathieu, Portrait of Isabella d'Este (Titian), Malcolm Marshall, Malcolm Marshall, Abdullah ibn Masud, and more. Why wouldn't we use that language? WhatamIdoing (talk) 03:47, 12 March 2022 (UTC)

In re: many of these labels, even ones like "white nationalist" lack a clear, binary definition

So what? Nobody ever said that labels can only be used if they have a clear, binary definition. There's no clear, binary definition for lots of labels, such as artist or politician. If the sources use them, then we can use them. If the sources use them widely (e.g., as a primary characterization), then we should present that information as a fact. We might present that fact in different ways (e.g., "Jim Jones was a cult leader" vs "Jim Jones started a cult" vs "Jim Jones founded an organization that is now considered one of the clearest examples of abusive religious cults in modern history"), but there is no rule anywhere that says we can only present facts when all of the words in the sentence have clear, binary definitions.

If we had such a rule, then I'd wish you good luck writing anything about economics. Or music genres. Or at least eight of the 10 articles listed at Wikipedia:Vital articles/Level/1. If you think there's a clear, binary definition that separates mathematics from science, or that separates mathematics from philosophy, then you are wrong. WhatamIdoing (talk) 02:59, 10 March 2022 (UTC)

The lack of a clear definition matters when we are trusting that sources were using the same definition/same standards when applying a prejudicial/ negative value laden label. Few subjects would consider it highly problematic to be labeled an "artist" even if the justification was less than ideal. I do some drawing and have produced art. If an art magazine called me an artist I wouldn't protest even though I think most people would find the description innacurate. Would it harm me? Likely no. However, if we stick "white nationalist" on someone's name there is a real concern that, thanks to cytogenesis, news articles may call them a white nationalist because we said so. Such a label most certainly could hurt their prospects of interacting with many mainstream companies. Anyway, like tall/short, white nationalist is not a clearly defined label and some sources may err on the side of using the label because they want readers to dislike the person. They want to scare readers away from the ideas the person is discussing or promoting. As an encyclopedia we should report the facts, not so much the subjective labels in part because those who applied the label may have done so for reasons other than it was a 100% obvious fit. Springee (talk) 04:24, 10 March 2022 (UTC)

I don't think we should second-guess our sources that way. There are times when we should carefully consider whether a specific source is actually reliable for a specific claim (e.g., if a source labels a religion as pseudoscience in passing, it's probably not reliable for that claim; on the other hand, if it spends some time explaining that this religion claims to be both scientific and factual but it is actually incompatible with the scientific method, then that source could be reliable for the claim that the religion is pseudoscientific).

But: when a label is used in multiple sources that are actually reliable for that specific claim, we should not look at all of these sources and say that my personal opinion is that this is a negative value-laden label, and it's my personal opinion that if this already-widely-used label gets used even more, then it might hurt the person's career prospects [even if the person is already dead? even if their career prospects are primarily among people who consider that label to be a badge of honor?], and it's my personal opinion that this definition is unclear, and it's my personal opinion that the applicability of this label hasn't been proven to me beyond any reasonable doubt, so editors shouldn't use this label. That's substituting your personal opinions instead of following the sources, and I know you will agree that this would be unacceptable.

We do have to trust the reliable sources. Trusting reliable sources is part of the deal here at the English Wikipedia. If that label is DUE in an article, then you should include it in the article. If INTEXT says that you shouldn't include in-text attribution because that could make a widely used label appear to be just one person's opinion, then you shouldn't add in-text attribution (even though LABEL currently and IMO wrongly says you should). WhatamIdoing (talk) 04:58, 10 March 2022 (UTC)

Even taking that position into account, we still need to be considering the consistency/frequency of use of a label. Take a case of an article where we have 100 sources, all from high quality RSes, which are reasonable in-depth about the person, and that's the exhaustive source search for the person because we're well outside of RECENTISM. If only one source used a label, would we even include it with attribution (per UNDUE?) Probably not. What if 5 do? Then we'd probably mention it, but definitely include attribution. On the other hand, if half those sources mention the label, then we are probably in the range where we can use the label withouh attribution. So at what threshold of sourcing does it take for us to reasonable drop the need to include attribution? We have zero guidance on this at all, and right now, too many editors side on just having to have a few sources to justify this. (Keeping RECENTISM in mind helps to aalso limit disproportionate focus on recent sources). That's what is the issue at Lauren Southern.

And I will maintain that as WP editors we should not be blind to all possible sources, including those we know aren't reliable. We absolutely should know the big picture and general views on all sides, and while we are only going to use sources that are reliable and not create false balance, we need to document those views from the dispassition position and not treat them as "absolutely correct", which is what this stance takes. In time, those views stratify and at which point we have a far better reason to write from the 10-year stance, but we should be tip-toeing around any viewpoint s and not treating any as absolutely in recent events, meaning avoiding such statements in wikivoice and using attribution when necessary. --Masem (t) 04:14, 8 March 2022 (UTC)

It's not just a matter of counting the number of sources. Several short news articles that talk about "X" do not outweigh an entire book about "Y". When talking about labels, we have to do the difficult work of evaluating what's DUE in each case – not just count up sources that say X and sources that don't, and definitely not just write zero-exception rules that say every single use of the word neo-Nazi requires in-text attribution no matter what. WhatamIdoing (talk) 18:40, 12 March 2022 (UTC)

This is also where RECENTISM comes into play. A book is likely going to be published some years after an event so it going likely be a more dispassionate coverage, and so if it uses a label, that may carry more weight than news sources written at the time of the event. But if the book is the only source that uses that label compared to all sources, that's still a minor view that we should not include per DUE.

The problem with have right now with this proposed change to label is that there's far too much focus on articles written where RECENTISM would apply and to use the weight of those articles to try to state labels as fact. This is where we should not be trying not to be jumping to these conclusions as fact no matter how much RSes claim to be the case, in the short term. This is the caution that is lacking in NPOV and INTEXT when it comes to expressing views as fact. It obviously makes sense for something that happened 20+ years ago but not something from last week. --Masem (t) 18:51, 12 March 2022 (UTC)

LABEL doesn't say anything about RECENTISM.

I don't know why you say "there's far too much focus on articles written where RECENTISM would apply and to use the weight of those articles to try to state labels as fact". My focus is on the problem of our advice pages giving contradictory advice. LABEL says that National Socialist Movement (United States) cannot be labeled as a neo-Nazi group unless the editors use in-text attribution for that claim. INTEXT and NPOV say that editors should not include in-text attribution for that claim. Editors should end up with the same result no matter which page they read. Which result do you think is the correct one for that sentence in the lead of that article? WhatamIdoing (talk) 22:45, 12 March 2022 (UTC)

It's not that LABEL doesn't talk about RECENTISM, it is that NPOV doesn't talk about RECENTISM, when that is where it should be a key point. Wikipedia should only adapt to accepting these widespread views as attribute-less facts (including if labels should be applied as facts) many many years after events have settled and not in the midst of it. But the steps in how to do that should enshrined in NPOV in the same manner as DUE/UNDUE here. As well as of course making sure some view is wide-spread and not just a cherry-picked stance, label-related or not. Once that is enshrined in NPOV (as it should be) then LABEL automatically follows, fixing the whole issue of this debate. But right now, there is no guidance of when we can make something into a attribute-less fact in wikivoice in NPOV, and that leads to countless problems right now, which fixing LABEL as suggested will not resolve. We need to set the NPOV thresholds first, and then LABEL will naturally follow. --Masem (t) 13:44, 13 March 2022 (UTC)

Masem, you've mentioned several times that INTEXT and NPOV need to be updated. Have you considered starting a proposal for those changes? –dlthewave ☎ 15:33, 13 March 2022 (UTC)

Perhaps it would have been better to not start this RfC until we had a better idea what a universal solution might be. Springee (talk) 15:44, 13 March 2022 (UTC)

Experience has told me it would be very much inappropriate to start suich a discussion while this one is ongoing as to sploit cthe matter. I'd wait until this one is closed. --Masem (t) 19:05, 13 March 2022 (UTC)

@Masem: To be absolutely clear, your position on this matter seems to be the following:

You agree that there is a conflict between the texts of LABEL and NPOV
You think that LABEL gets it right and NPOV gets it wrong
Based on (1) and (2), you think that we shouldn't change LABEL but that we should change NPOV

Am I understanding you correctly? Srey Sros^talk 19:22, 13 March 2022 (UTC)

That's not correct. My stance: there are times when we can state, in wikivoice without the use of attribution, the near mass-agreement of a subjective viewpoint (including when a label applies), but when this threshold is met, both in terms of the proportion and siognificant of sources, and the recentism aspect of those sources, is not at all documented at NPOV. Because labels can be on-wiki troublesome in their use, it is better to adust NpOV to include this guidance when we can speak in wikivoice w/o attribution first and foremost, which then by necessity must filter to LABEL. The suggested change to LABEL are not addressing this core problem. --Masem (t) 19:33, 13 March 2022 (UTC)

If you believe that there are times when we can state, in wikivoice without the use of attribution, the near mass-agreement of a subjective viewpoint (including when a label applies), then LABEL (as currently written) is wrong and needs to be changed.

How would you change LABEL so that it stops saying "use in-text attribution" and starts saying that there are times when we can state, in wikivoice without the use of attribution, the near mass-agreement of a subjective viewpoint (including when a label applies).

(Also, you're back on that idea that all LABELs are subjective, which is still not true.) WhatamIdoing (talk) 04:19, 14 March 2022 (UTC)

You are missing my entire point. That we do not provide any guidance at NPOV when, in general terms, we can state a widely-shared viewpoint in wikivoice without attribution. We say it can be done but do not caution as to the when and where, and that needs to be laid out in stronger terms. INTEXT would have to be updated to follow from that, and the current advice i LABEL would automatically follow from its current wording (though I could see a case to make sure we address for such cases). Trying to change LABEL without addresses this root problem is not the right solution to get to a working result that I know you're trying to get to. Labels are by their very definition subjective, period, because they have vague and non-specific definitions. But there is a reasonable place to say that when more than enough time has past (RECENTISM) and a sufficient majority of sources agree that a label applies, then we can state that in wikivoice without belying the subjective nature of the label itself. --Masem (t) 12:27, 14 March 2022 (UTC)

Are you saying that all the other guidelines and policies need to changed so we have to write "Alice and Bob say that David Irving is a holocaust denier" instead of "David Irving is a holocaust denier"? "Denier" is in the list of labels and, according to you, by their very definition subjective, period. --Hob Gadling (talk) 15:07, 14 March 2022 (UTC)

WP:CLUEFUL editors will continue to follow NPOV (core policy) and ignore the fuck-up that is the MOS. Same as it's been as long as I can remember. Alexbrn (talk) 15:14, 14 March 2022 (UTC)

David Irving is an example of where RECENTISM is well passed, and there are a majority of sources including a court decision to state that he is a denier in wikivoice w/o attribution. This is a case passing this threshold to allow for this. But we have situations like calling Qanon as a cult where RECENTISM still applies and there is not majority agreement in sources to allow us to make that claim in wikivoice without attribution. Describing when that threshold can be met is a function that belongs in NPOV, not in LABEL. --Masem (t) 15:39, 14 March 2022 (UTC)

I feel like you are missing everyone else's point. According to you, LABEL currently tells editors to do the wrong thing for these subjects. Therefore, LABEL needs to change.

Sure, you can say that you'd like to postpone the correction as long as possible and fix every other omission or lack of clarity first, but the fact remains that LABEL currently tells editors to do the wrong thing for these subjects and LABEL needs to be changed.

So imagine that magical world when you've been able to re-write all of the other policies and guidelines to match your own views. When and if you finally reach that point, how exactly would you change LABEL to tell editors to do the right thing? Because, to be very clear, even if those other policies and guidelines were completely re-written by you personally, so they said exactly what you want, LABEL would still be wrong and LABEL would still need to be changed. So pretend you've convinced everyone else of your righteous cause to keep them from calling QAnon a cult and re-written NPOV and INTEXT and everything else to say that. What happens then? The only answer I won't respect is the one in which you say that once everything else is fixed up to your satisfaction, then it's perfectly fine to keep known-wrong advice in LABEL. WhatamIdoing (talk) 15:52, 14 March 2022 (UTC)

INTEXT points to WP:ATTRIBUTEPOV which says biased statements of opinion need to be attributed. So to the extent that these value laden labels are "biased statements of opinion" LABEL is exactly following NPOV. I guess the question is when is "John is a racist" a statement of fact vs a biased statement of opinion? Springee (talk) 16:58, 14 March 2022 (UTC)

Speaking only for myself, the key term in biased statements of opinion appears to be "biased". Not all "value-laden" labels are "biased" in the sense intended by policy. For example, "cult" in the sense of Jonestown can be (and has been) used by scholars in a precise sense, although (because of the context) it cannot be divorced entirely from values: this usage can be "unbiased" in the sense of policy. The same is true of NAZI as an historical term: it seems ridiculous to suggest that it is not in any way "value-laden", but it can nevertheless be used in an "unbiased" sense (what mature epistemologies, though not naïve scientism, would call "objectivity"). The point should not be to have a pin-dancing discussion about "fact vs. opinion", but to recognize that it is actual, not imaginary bias that steers sources towards making statements that are really, not rhetorically, controversial and which therefore demand attribution for encyclopaedic purposes. Newimpartial (talk) 17:09, 14 March 2022 (UTC)

Assuming that NPOV and INTEXT are changed to describe the threshold when a majority held view can be stated without attribution, then LABEL should be modified to say something like "...using in-line attribution when required by NPOV/INTEXT." As to defer to those cases of mass source agreement. perhaps a line to note that labels that apply well beyond the issue of RECENTISM may be present in wikivoice when sources show majority agreement (reiterating what should be in NPOV). --Masem (t) 17:17, 14 March 2022 (UTC)

Here's how I might put it - changing LABEL by just pointing to NPOV without fixing NPOV first makes the situation as a whole much worse, enabling bad labeling, even if NPOV is fixed down the road and the gap is then closed. To avoid this gap, we need to fix that first. Not sure if that's exactly what you had in mind, but it makes sense to me. Crossroads ^-talk- 22:52, 14 March 2022 (UTC)

(For all that's holy, please respect the outdent) - I don't see any consensus that NPOV needs to be fixed, and surely this wouldn't be the place for that discussion if it did. Newimpartial (talk) 00:20, 15 March 2022 (UTC)

Closer reading of "value laden"

I think the flowchart, as intended and as used in practice, is different than the reading given by the OP.

"Value laden" here refers to whether the applicability and correctness of the label are dependent on value judgements, assumptions or definitions that vary considerably among sources or (importantly) readers of the article.

"may ... contentious" here means likelihood of contention arising from that value dependence.

The running example of "AIDS denialist" applied to Duesberg is not value laden provided that it is accepted as a generally understood term of art. The phrase is not 100 percent clear as to whether it means denying AIDS entirely or the role of HIV in causing it, which the associated Wikipedia article resolves by using "HIV/AIDS denialism" as its title. If you accept that bluelinking to the article is enough to resolve the ambiguity then there is no further value (or assumption, definition, culture, language etc) dependence and the MOS:LABEL algorithm as I read it does not stop use of the phrase in Wikivoice.

On the other hand if one tried to label Duesberg with something meaning "responsible for numerous deaths of South African AIDS patients" (whether the exact phrase or a proxy like lethal counselor or angel of death) then even if all sources had such language, it would be contentious by reason of value dependence. Since Duesberg never went out and personally killed people, connecting mass death to him depends on a chain of assumptions and inferences about responsibility, agency, influence, cause and effect; no matter how much the sources may treat the full chain as obvious, a reasonable person or hypothetical reader might not, and the flowchart correctly says to use in text attribution. A wikivoice statement that "numerous sources describe him as..." might also be in the spirit of the current MOS:LABEL wording.

The discussion that ensued here strikes me as discovering an unexploited security hole and then insisting on repairs that create much larger problems. Rewriting MOS:LABEL to allow carte blanche for a consensus of sources to dictate reality where analogous chains of inferences are involved opens a can of worms. It's one thing if sources all agree Duesberg was born in a particular year and quite another if they all agree his actions led to mass death, or that Trump is a serial liar, or other such inferences. The problem is not that these descriptions are negative or necessarily wrong but that they are conclusions which import a lot of logical and conceptual baggage, which I think is what is current summarized as "value laden". Formerly it said "biased" with essentially the same intent. Sesquivalent (talk) 08:10, 9 February 2022 (UTC)

What you describe is reasonable enough, but it's not what LABEL is about. The label denialist is called out in LABEL as an example of a value-laden label. The "value" it is laden with is that the view has been judged incorrect. Duesberg is not "the holder of a minority viewpoint", which carries little or no value. He is not a "dissident", which assigns some positive value to him. He is a "denialist", which indicates that his views are entirely wrong, and that he has been judged incorrect to the point of his views being irrational and suggestive of a psychological defense mechanism.

My goal is to get LABEL out of the business of saying when INTEXT attribution is appropriate. We've got a perfectly good guideline over at WP:INTEXT that says when and how to use INTEXT. It happens that what LABEL says about INTEXT does not match what INTEXT says about INTEXT. The simplest solution is for LABEL to just stop saying anything about when and how to use INTEXT, and merely to point to INTEXT as the canonical guideline on when and how to use INTEXT. WhatamIdoing (talk) 05:54, 10 February 2022 (UTC)

LABEL, including the part about INTEXT, is a gloss on some of the instructions from NPOV, mainly the one about "avoid stating opinions as facts" that has an in text attribution directive, so the latter is not logically detachable. The problems you are claiming with "denialist", and what Aquillion is calling a utilized exploit, are based on interpreting "value laden" as (A) an intrinsic property of the words on the LABEL list, and (B) often having something to do with the term being derogatory or "judgemental" as NPOV calls it, so that people can scream about that. My analysis of what LABEL means to say indicates that (B) can be separated from INTEXT by giving it as an additional and distinct reason to avoid words on the list (basically just a pointer to the judgemental part of NPOV); and that (A) should be resolved by indicating that the concept of value laden (or whatever better term might be used in its place) here applies or fails to apply only to particular instances where such a word is used, and is not a property of the word itself. "Denialist", for example, is only used in contexts where it is said that X is a Y denialist/denier, and whether that is value laden depends on X and how well defined is the meaning of "Y denialist". This still leaves a directive to use INTEXT for what NPOV calls statements of opinion, which LABEL calls value laden, and might more accurately be described as "dependent on unstated baggage". Sesquivalent (talk) 09:31, 10 February 2022 (UTC)

Rewriting MOS:LABEL to allow carte blanche for a consensus of sources to dictate reality where analogous chains of inferences are involved opens a can of worm - to be clear, we are required (not merely allowed) to follow the consensus of sources; any interpretation of MOS:WTW that would allow a clear consensus among high-quality neutral sources to be disregarded or downplayed via attribution is incorrect, and any wording that would support such an interpretation is in error, since WP:WIKIVOICE, as part of NPOV, is core policy and ultimately requires that we follow the sources. Even if an editor personally feels that what the sources say is value-laden and therefore does not reflect their personal subjective view of reality, that is not a valid policy argument that would allow us to ignore an overwhelming consensus among sources, and therefore has no weight whatsoever in discussions. The only way NPOV allows us to determine reality is by looking at the sources. And I certainly do not agree that this issue is "unexploited" - many people, in the past, have leaned on the error in this policy to falsely argue that we can attribute things that they disagree with or personally feel are value-laden without presenting any sources to support that reading and without articulating specific issues with the sources that present it as fact. That has to be clearly stopped - arguing that the sources are WP:BIASED is fine; arguing they're not strong enough is fine; showing that there are other sources of equal weight that treat something as subjective is fine (those are the sorts of arguments people should be encouraged to make!) But an argument along the lines of "this is a word on WTW and therefore I'm going to argue that it is inherently subjective and cannot be stated in the article voice regardless of what the sources say" (something that has become unfortunately common recently) directly violates WP:NPOV and is, beyond that, an unproductive way to engage in discussions, both because it has no grounding in policy and because it distracts from examining the sources that we must ultimately defer to to decide article content. --Aquillion (talk) 07:21, 10 February 2022 (UTC)

NPOV distinguishes opinions from facts and requires they be treated differently, irrespective of the proportion of sources holding the opinion. Whether something is fact or opinion is, for most purposes, a question of fact. Whether something is objective or subjective ("an editor's ... personal subjective view of reality" as you put it) is, for most purposes, an objective question. Questions which are, in most cases, easy to answer. The supposed conflict of in text attribution with NPOV is with the instruction to state in Wikivoice things that are "uncontested and uncontroversial factual assertions". But the point of value laden as explicated above is that by definition a value laden assertion in that sense cannot be taken to be uncontested fact (e.g., it depends on various inferences or assumptions, as in Trump being a serial liar or Duesberg a killer by proxy), no matter how little controversy it generates in available sources. There is no NPOV problem in using attribution that describes the extent of agreement among sources. Sesquivalent (talk) 10:25, 10 February 2022 (UTC)

Thankfully, WP:NPOV distinguishes between "fact" and "opinion" and not between "objective" and "subjective", which would have been a false dichotomy in this context. I am disappointed to see the assertion here that facts that relate to underlying values (like, for example, historical statements that certain policy initiatives were antisemitic or white supremacist) are supposed by Sesquivalent not to be facts even when (as with much of NAZI policy, or the Jim Crow regime in the US South) their status is essentially uncontested. The idea that such uncontroversial descriptions cannot be made in wikivoice is not supported by actual NPOV policy and strikes me as simply a hangover of unacknowledged Vienna Circle metaphysics on the part of certain editors.

One key aspect that NPOV does identify, which would benefit editors if given more attention in actual discussions about article text IMO, is the injunction to prefer nonjudgemental language. I have the impression that editors frequently appeal to "opinion" or "value"-based objections to terms when their actual objection is that they find a term judgmental, regardless of how well it is sourced. This is presumably how "cult" finds itself on the list at LABEL, for example, since it does certainly not always imply an "opinion" rather than a factual claim.

If we were to be more explicit about the role of judgmental language, I hope we could have discussions that distinguish among multiple terms applied to roughly the same phenomenon and try to evaluate what is more or less judgmental in tone. For example, there might be many instances where homophobic or transphobic appear judgmental to editors in a sense in which "anti-gay" or "anti-trans" do not, in spite of denoting very similar phenomena. I'm not convinced that there are always alternative terms that can be compared in this way, and editors can reasonably differ about what terms are more or less neutral, but if we could shift the terrain to this part of NPOV it might at least allow policy-based discussion. The idea that all value-laden terms are to be avoided or attributed (and the debates I have seen about the inclusion of false or falsely in articles represent an extreme version of this, but correctly reflect that "false" is in fact a value-laden term) is a quagmire into which LABEL would lead us. And the solution cannot be only to carve out areas like pseudoscience where we require that value-laden terms be used when they are, in fact, the correct terms - MOS guidelines should reflect the reality that factual and value-laden statements routinely overlap, according to reliable sources, and get on with making an encyclopaedia. Newimpartial (talk) 13:28, 10 February 2022 (UTC)

I generally agree with this. Words to "watch" are not words to avoid in all circumstances. It's often good to take care when using a term, but sometimes, the right way to represent the facts is to go ahead and use that term. XOR'easter (talk) 01:19, 11 February 2022 (UTC)

That's really a poor poor excuse and shortcut from good encyclopedia writing as to explain why a label may apply to a person or group. That's a major issue with labels in general in that they are terms that may overly simplify complex positions, and using a label without any other context (regardless of attribution or not) is not helping the reader to understand that position. We really should be explaining that instead of someone being, for example, a climate change denier, that they have stated that humans have not contributed to climate change and that current planetary warming is part of a natural cycle. (Assuming we can source this to an RS) That gives far more information than just "climate change denier" does, and is more appropriate for an encyclopedia. I'm not saying that we then can't also use "climate change denier" as long as that label is sufficiently used, but using that label alone without further context overly simplifies the situation. And I will say that if we're talking a case that a label like "climate change denier" is used so frequently in RSes, it is pretty much assured at least one or two of those sources are going to explain the person's position, we don't have to go looking for SPS or the like to include that. Its the cases where editors like to cherry pick labels where sources usually don't justify why the label applies, but as LABEL suggests, we shouldn't be using the label in these cases anyway.

Or the tl;dr - labels should never be used without context for explaining the reason the label applies; we want editors to focus on that rather than just outright inclusion of the label and move on from that. --Masem (t) 01:44, 11 February 2022 (UTC)

Perhaps we are talking at cross-purposes here. I don't see how anything either Newimpartial or I said contradicts the idea of presenting the context for why a term like "climate-change denier" (or "Holocaust denier", etc.) is applied to a given individual. XOR'easter (talk) 02:05, 11 February 2022 (UTC)

An issue I've seen far too much in articles is the focus on "well, all theses RSes use this term so we should just present it as fact", but do not address anything about context, which creates the problem with NPOV issues if that context isn't addressed. The whole issue of "a label used by a near majority of sources shouldn't need in-line attribution" could be better answered by focusing more on how we present the context where a label would be used by describing the elements that lead to a label. If we focused more on the label use in the larger context of describing why the label applies, rather than just term-droppping it, it would make more sense what conditions we could present that label without the need for direct in-line attribution. --Masem (t) 02:19, 11 February 2022 (UTC)

- to be clear, we are required (not merely allowed) to follow the consensus of sources There is nothing that says we are lemmings with respect to mainstream media. We are required to make sure the majority view point is well established per DUE/WEIGHT, but that does absolutely does not mean we take that as fact. Especially in this day and age where the bulk of our sourcing leans left, which is a known systematic bias that we are required to try to work against. This is not saying that the bulk of RS are not reliable, simply that when you put that together with 24/7 news reporting and the rise of accountability journalism that mixes editorial with factual reporting, we absolutely should not be taking every work published by mainstream as fact, we should be able to analyze these sources from a 60,000 ft view of the overall situation to understand where statements are likely opinion rather than fact, and particularly in the case of labels, handle those with care. I'm still of the opinion there's ways we can handle a near-universally shared label among sources to avoid direct in-line attribution listing out 10-20+ works, but the wording has to be clear its not a fact but a consensus derived from multiple sources. --Masem (t) 14:03, 10 February 2022 (UTC)
  How would you present "a near-universally shared label among sources to avoid direct in-line attribution listing out 10-20+ works"? Cult is in the list. How would you describe Jim Jones in a way that indicates that (a) every reliable source written in the last several decades agrees that he really was a cult leader but (b) you don't want to say that it's a fact that he actually was a cult leader?
  
  Maybe I'm skipping a step. Do you even agree that it's possible for someone to be a cult leader in actual fact, and not merely as an opinion? WhatamIdoing (talk) 01:07, 12 February 2022 (UTC)
  - Let's start with the fact that the word "cult" in Jones' article outside of In Popular Culture, See Also , and references, appears only 3 times: once in the lead, and twice in the Reactions and Legacy section before the Pop Culture section. We actually do not give any context for why Peoples Temple was considered a cult, or why Jones should be called a cult leader (And Peoples Temple only uses "cult" once.) This is the context issue here. Now we know at the end of the day that yes, Jones and People Temple are templates of how later 20th cults formed and behaved in the US, and RECENTISM doesn't apply at all here, so at some point is likely right that we should likely use "cult" without attribution but that context for its use has to be in the article and its just not there. --Masem (t) 01:20, 12 February 2022 (UTC)
    You say: We should likely use "cult" without attribution. (I agree with you; especially if those articles were developed to FA-level, this is an obvious use for such labels.)
    
    LABEL, however, says: "use in-text attribution."
    
    It is this specific conflict that I want to see resolved. LABEL should not tell people to do something that you wouldn't recommend them to do.
    
    I think one way to resolve it is to have LABEL demand compliance with INTEXT (and NPOV, too), rather than having LABEL assert "use in-text attribution" in every single situation. WhatamIdoing (talk) 21:28, 12 February 2022 (UTC)
    Again, to read NPOV and INTEXT, those do not support converting "mass agreement on the use of subjective language of any type" in "factual statements in wikivoice" either, whether talking about use of labels or other types of quantifications. (WP:SPADE is an essay) Its not an issue with LABEL being wrong, but that we don't have language in NPOV or INTEXT to cover this type of case - but I agree we should. When a significant number of reliable sources share a common opinion about a topic (whether its using a label or other aspect), and this not just from flash-in-the-pan coverage but has been from enduring coverage of that topic (the RECENTISM line has been passed), then there's a good chance we can likely present that in a manner that eliminates the need for inline citation. Whether that can be said directly "X is Y" or alluded to as in "X is commonly said by media to be Y" or some other form is a separate question to be resolved and I am certain there are many possible answers here for this when certain approaches can be used or not. LABEL fits into that but it is not the only part of NPOV or INTEXT that would be effected, and hence why it is a problem at the NPOV/INTEXT level, not a problem at the LABEL level. --Masem (t) 00:33, 13 February 2022 (UTC)
    Perhaps it is a nature of the discussions I have been involved in, but while I have seen both NPOV and LABEL used as Newimpartial succinctly put above to whitewash certain subjects, I have never seen INTEXT cited as a justification. I have to agree with WhatamIdoing, the problem is LABEL, not INTEXT or NPOV. Sideswipe9th (talk) 00:48, 13 February 2022 (UTC)
    To take an example that does not involve LABEL, if we have 100-some sources that said "Mozart is one of the greatest pianists of all time", both NPOV and INTEXT state that we cannot simply repeat that in text without some type of in-text attribution because that still is a subjective claim. This is why I say this is a problem beyond just LABEL, because it would obviously and unnecessarily clumsy to have to name drop 100 sources in-line. We should have ways to handle a case like that, as welll as in cases for LABELs where there has been shown to be wide-spread agreement and endurance of the applicability of the label. --Masem (t) 01:37, 13 February 2022 (UTC)
    
    Sidewipe9th, the claim of whitewashing is always subjective. Take a case where there is debate if a subject should be called "transphobic" or even be labeled at transphobic with attribution. Newimpartial might view the use of LABEL to as whitewashing. However, other editors, acting in equal good faith, see it as a questionable label. Essentially they would argue it's black washing. Why is one better than the other? How many times has LABEL been used to prevent an article from unreasonably having a contentious label applied in Wiki-voice? Isn't that a good thing? To use a legal analogy, which is better, a legal system that catches more baddies but also captures more innocent people or the other direction, one that almost never falsely accuses or prosecutes but leaves more baddies on the streets? I guess that depends on your own views on morality. Ideally we want a set of rules that "always gets it right" but, often unlike criminal law, this is a big case of gray. If one activist says X is racist, is that DUE? What about when 5 or 10 make the claim? What if 5 or 10 news articles report that the subject was called racist by activist? What if 1 news source says it in their own voice? What if one academic publication says it in the subject is racist? With so much gray which way should we err? I feel in the direction of do no harm since descriptions and justifications can always be substituted for subjective labels assuming the sources actually support it. If the sources don't then we shouldn't use the label. Springee (talk) 14:34, 13 February 2022 (UTC)
    @Springee: I can think of about 5 or 6 discussions that I have been involved in, where there is sufficient sourcing to state in wikivoice that the subject of the article is transphobic, and LABEL has been used to block those changes. In two of those discussions, it resulted in a truly astounding amount of words being committed on the talk page before participants either left because they were exhausted, or an RfC was opened and consensus was found for that addition.
    It is not possible to "both sides" this, as that is a false equivalence. If there is insufficient sourcing to establish whether or not the subject of the article is a negative value laden label, then through NPOV and INTEXT it is impossible to blackwash. However because of the way that LABEL is currently written, it is very easy to meet the criteria of NPOV and INTEXT, and still use LABEL to whitewash. That is not a good thing. Sideswipe9th (talk) 17:19, 13 February 2022 (UTC)
    It would be helpful to see those discussions where LABEL is being used to whitewash, both as to judge if "sufficient sourcing" has been shown (more than cherry picking) and how the arguments are presented that say LABEL says not to include. If sufficient sourcing is met (with potentially other caveats like RECENTISM), then the label can be included but attribution should be included. There's also the potential alloways for if there's clearly show near universal use of a label by sources, a potential way to stat that without the need for in-line attribution, but that's the issue we need to discuss with any subjective aspect at INTEXT, not just LABELS. --Masem (t) 17:47, 13 February 2022 (UTC)
    
    "Transphobic" is used largely by the gay or trans press with unstated definitions that vary between sources and are much broader than what someone not in the original audience of those articles would normally assume. If past usage of "homophobic" is the average Wikipedia reader's mental yardstick then most would interpret transphobic to mean something beyond vocabulary disputes, JK Rowling, exclusion from women's sports, TERF, or a lack of interest in dating trans. If there are cases clearly in the intersection of how all sources and most readers could reasonably interpret the term, such as a violent criminal who exclusively attacks transgender people, then there is no dependence on definitions or value judgements (the dreaded Value Laden), but for things to reach that point there would probably be better options like citing a judicial verdict of hate as a motive.
    
    Also, because the term is not used in the strict sense of phobia, and its meanings and common usage are still evolving, RECENTISM is relevant not only to particular applications of the label but to the word itself. Sesquivalent (talk) 20:14, 13 February 2022 (UTC)
    Sesquivalent, you have repeatedly made reference to sections of academia, medicine and now even the press that you seem to believe are influenced by some sort of LGBTQ movement or agenda and thus are unreliable.
    
    You have previously argued that we should discount official statements from the world's leading transgender health organizations because you believe they're trans-affiliated and therefore untrustworthy. In that same diff you call for an RS author's transgender identity to be treated as a COI. Here, on an article that provides biomedical information, you argue that we should give equal weight to the opinion of the Stanford chief fellow in child and adolescent psychiatry (an MD and published expert in the field of transgender health) as we do to the opinion of a freelance conservative journalist with no public health experience or training, because you think the Stanford MD has vested interests. And now in this comment above you argue that we should discount RS usage of a term because you believe those RS belong to some overall grouping of gay or trans press.
    
    I bring up these past comments of yours because I think they reflect an overall misunderstanding of how we evaluate sources on Wikipedia. Evaluations of source reliability are based on community consensus as to editorial practices of fact-checking, journalistic integrity and other factors as defined by WP:RS. Accusing mainstream media, academia, or medicine of a systemic bias and arguing that we should attempt to push back against this reeks of RightGreatWrongs thinking.
    
    To be clear: the arguments "RS agree that X is Y but I don't think those sources substantiate their arguments enough", "RS agree that X is Y but I don't think they clearly define Y enough in their articles", "RS agree that X is Y but I don't trust the gay/trans press" are not valid arguments against inclusion of a characterization in article space. We follow the sources, regardless of editors' personal opinions.
    
    As a side note, you may wish to consult a style guide (this is a very commonly used one) for the way you refer to transgender people. In your comment above (and at least once before) you seem to refer to transgender people as trans, as a noun (the plural of "tran", I guess). You've also referred to the state of being transgender as transgenderism. Neither of these are common or accepted terminology, and in the last decade or so the latter has been almost exclusively used by anti-trans activists (see e.g. here). Srey Sros^talk 22:02, 13 February 2022 (UTC)
    This wall of text does not respond to the comment whatsoever, is totally off topic, and is POVRAILROAD. The summaries of my past comments are extremely tendentious, the attributed positions false, but again, all of that is far off the topic here. Take it to my talk page if it matters so much; it certainly has nothing to do with editing MOS:WTW.
    
    Personal levels of trust in "gay and trans press" are irrelevant to my comment, and I made no statement about their trustworthiness. The actual claim is that the language used much more frequently in a subculture does not (in this case) share meaning for the same word in the larger population, and Wikipedia writes for the latter. Which is a question of linguistics and usage, not bias or politics. Sesquivalent (talk) 22:20, 13 February 2022 (UTC)
    
    In re "mass agreement on the use of subjective language of any type":
    
    Masem, why do you keep going back to this subjective–objective thing? NPOV doesn't. WP:SUBJECTIVE is about "aesthetic opinions", which isn't what LABEL covers. That's the sole use of the word subjective in the whole of NPOV, and it is irrelevant. Whether Jim Jones founded a cult is not an aesthetic opinion. Whether Jim Jones founded a cult is not a subjective question. It is a question of plain hard facts: There are accepted scholarly definitions for cults, there is a factual description of what he did, and either his factual actions line up with the accepted definitions – in which case, he founded a cult – or not – in which case he didn't.
    
    Every time it seems like we have an agreement that value-laden labels for strictly factual situations should be handled per NPOV, which is unfortunately rather different from what LABEL says, you seem to jump to something irrelevant like "But Mozart might not be the greatest ever" (which is not a value-laden LABEL) and you pound on subjectivity (which is irrelevant). What are you afraid of? Are you afraid that if we say that Jim Jones doesn't need to provide a list of a hundred sources that call him a cult leader, then suddenly WP:PUFFERY will no longer apply to Wolfgang Amadeus Mozart? That if LABEL says to follow NPOV and INTEXT, then editors might be less afraid of adding well-sourced labels when the policy says it's appropriate? (Not every editor wants that; some editors are think it is rather impolite to call a spade a spade in the mainspace.) What's your real problem? Can you give some examples of articles that you think would be changed for the worse if LABEL said that all articles using value-laden labels should scrupulously follow the directions at NPOV and INTEXT? WhatamIdoing (talk) 01:37, 17 February 2022 (UTC)
    A significant problem across the board is the trend of editors wanting to rush to characterize a topic rather than write on the objective, neutral aspects of a topic. Characterization of a topic can appropriate but that should be based on sufficient coverage in the sources (eg when DUE is met) rather than the current trend of cherry picking to get the characterization one desires. But even when this basic level is met, characterization should always be presented in a way to take it out of Wikivoice. Now, there can be a point where with the passage of time and near-universal agreement of sources that characterization can become as close to fact as possible (this is the same as a scientific theory being taken as fact despite still being a theory). But there should still be the demonstration of sourcing through a source survey to support that, rather than cherry picking from one or two sources. This is the issue on the Jones pages - the sourcing to call what he led as a cult is nowhere close to impecible to be appropriate for Wikivoice. I don't doubt more sources exist for that, but that needs to be far better documented on our pages rather than being taken as matter-of-fact-ly by giving context to why his group was considered a cult. And no, we don't need a full source list on the page itself, but that source survey should be established on the talk page so that if an editor in the future has a problem, you can point to that. The reason that you should be doing that for Jones is that when we turn to Qanon, as noted above, the sourcing that describes the group as a cult (rather than having cult-like behavior) is not as strong as one may think. A proper source survey would help determine if 1) Qanon should be described as a cult and 2) if so but not by a wide range of sources, which attribution to use.
    
    Editors think they know what sources say particularly when it comes to labels (including claims 'everyone uses this!', but I've seen analysis of sources prove out differently. Labels are terrible shortcuts from actually explaining to the reader actual events and the like (that can be explained objectively) that led to why a person or graoup is criticized, and we should not be seeking out the use of labels to shortcut that process. --Masem (t) 02:50, 17 February 2022 (UTC)
    Writing that a bona fide cult leader like Jim Jones is actually a cult leader is exactly the same thing as writing "the objective, neutral aspects" of the cult leader. Right?
    
    You said above that describing Jim Jones as a cult leader should not be taken out of Wikivoice (even if you'd like someone to upgrade the sources on the page, which wouldn't be difficult but which probably is immaterial, because no such upgrade would change the wording here). Right?
    
    Are we agreed on this much? WhatamIdoing (talk) 17:12, 17 February 2022 (UTC)
    What is the def of a bone fida cult or cult leader? There is none because it is a value laden term with imprecise, subjective bounds. But in a case like Jones, it should be possible to demonstrate via sources over time that there is near universal agreement that what he ran was a cult and that he was a cult leader, to the point we take that as accepted knowledge. (This is barring the sourcing problems on the articles). But we should make sure that when we state that he was a cult leader it is in immediate context with sources that support these factors, which indirectly serves as attribution for this statement without actually giving that attribution. If it were the case that there wasn't universal agreement about him being a cult leader, but enough to be DUE, then we'd need inline attribution. --Masem (t) 20:33, 17 February 2022 (UTC)
    The definitions of cult depend on the academic field you're writing for. A sociologist will write in sociological terms (e.g., "threaten the basic values and cultural norms of society at large"), and a theologian will write in theological terms (e.g., "denies one or more of the central doctrines"). They do not necessarily have imprecise or subjective bounds. Many of them have relative bounds, but relative ≠ subjective; you can be taller than your neighbor (a relative statement) without height becoming a subjective factor. But all of the scholarly definitions agree that Jim Jones' organization was a cult.
    
    You agree that this particular cult leader deserves that particular label, presented in wikivoice as a fact.
    
    Do you understand that a literal reading of LABEL, as written, says that this particular label cannot be presented in wikivoice as a fact in any article, under any circumstances, no matter what the sources say? WhatamIdoing (talk) 02:42, 18 February 2022 (UTC)
    What I'm saying is that sufficient time and sufficient dispassionate analysis has taken place since the time of Jones' leadership that while what a "cult" is is still relative or subjective -- and most importantly not objective -- that there is widespread agreement that what Jones ran fell into what nearly all fields considered to be a cult. So yes, this would be a case where we should be able to state that in Wikivoice without having to direct attribute inline. (in contrast, we would have to be careful with Qanon).
    
    But my point has been is that this is not an issue limited to LABEL, but something that should be addressed at INTEXT to handle a wider set of cases that encompass not only LABEL issues but other cases, such as the "Mozart one of the greatest pianists" cases. Statements that involve language that is not objective but which represented the majority of sources well beyond the period of RECENTISM, where it would be otherwise painful and awkward to have to use inline attribution to support that statement. Limiting it to just LABEL is too narrow a solution. --Masem (t) 05:02, 18 February 2022 (UTC)

I think as editors we can be expected to distinguish between clear statements of fact and statements of value in reliable sources. "Mozart was great" and "Hitler was evil" are clearly, on their face, aesthetic and ethical value judgments, respectively. But when it comes to statements that are ostensibly factual claims, but where the language choice may imply a certain value judgment, we have to follow the language use in the sources. It is not compatible with WP:NOR or WP:NPOV to ask editors to second-guess factual claims in sources. If sources truly agree that something is a cult, whether it is Jonestown or Qanon or the Cult of Artemis at Brauron, so must we. If there is significant disagreement in sources as to how to describe something, that should be reflected in the article. But trying to correct for perceived errors in the body of reliable sources as a whole is well into WP:RGW territory.--Trystan (talk) 14:20, 18 February 2022 (UTC)

I would disagree that editors can distinguish between statements of fact and value nowadays; we are at a stage and partially here because of editors often cherry picking the use of a label from one or two sources and expressing that must be a fact. Coupled with accountability jounralism that places objective writing on the backburner, we need to be more aware when sources, particularly those close to an event, are speaking in opinion rather than fact. That all said, I am not saying that we shouldn't allow Wikivoice to state as fact a point that nearly all RSes agree on if it came to identifying a topic as a label, but we first of all should make sure that that is what sources in majority say as part of enduring coverage of a topic (the RECENTISM issue) as to make sure we're not capturing heated ideological battles from spur of the moment coverage, and that it should be very much demonstrated via talk pages and reflected in what sourcing is used in the article(s) that this is reflecting a majority of sources through a source survey, instead of editors cherry picking a handful this type of result to claim as fact. --Masem (t) 14:44, 18 February 2022 (UTC)

The antonym of "fact" is "opinion", not "value".^[1] If editors really cannot distinguish between fact and opinion even after a discussion, then we should just give up, delete Wikipedia:Consensus, protect all the pages, and go home.

You suggest above changing INTEXT. I think that having this information split across on multiple pages leads to drift between the rules. This division is one of the reasons that LABEL and INTEXT give opposite advice for Jim Jones. What I'd like to see at LABEL is that it tells people to follow INTEXT and NPOV, without telling people what INTEXT and NPOV say. Right now, LABEL and INTEXT say different things. If LABEL says to follow INTEXT (no matter what INTEXT says now, or will say in the future), then there will be no discrepancy between the two. I'm here to resolve the WP:PGCONFLICT, not to say that X or Y is the correct approach. The correct approach to INTEXT is something that should be discussed on the talk page for INTEXT. My goal here is only to make LABEL stop contradicting INTEXT. WhatamIdoing (talk) 17:41, 18 February 2022 (UTC)

But again, to stress: INTEXT does not support the idea that if majority of sources claim a subjective statement is true, that WP should report it as true in Wikivoice (regardless if this is a label or not) (only UNDUE covers this type of situation). There is no current conflict between INTEXT and LABEL otherwise. Hence why it would be better to add in cases at INTEXT to handle broadly how we should handle such cases, including what steps should be done before we place such a claim in Wikivoice. Not just to handle the LABEL cases but other situations (such as describing someone that may be speaking of misinformation by a majority of sources, a situation not covered by LABEL but would still fall under INTEXT). I'm talking a broader solution that still gets to a solution you want but also covers a larger similar set of cases beyond LABEL. --Masem (t) 18:21, 18 February 2022 (UTC)

If you want to start discussions at INTEXT and NPOV about how to improve them, then that's okay with me. My goal here is just to get LABEL out of the business of misrepresenting INTEXT and NPOV. The simplest way to do that is to have LABEL just point to them. WhatamIdoing (talk) 18:32, 18 February 2022 (UTC)

Let's use the Jonestown as the example here. I think we are in agreement that (while our articles don't readily include them) there is sufficient enduring sourcing that call it an example of a cult. This is 100% a case that I would agree that PAG *somewhere* should allow us to express this as "Jonestown was a cult" in wikivoice. But INTEXT in context of WP:WIKIVOICE as it currently stands does not allow that since asserting "Jonestown was a cult" still is an opinion - just one shared by a majority of sources as to taken as given. INTEXT should be modified to talk about how to do deal with such "mass agreements vin sources" to avoid having huge in-line attributions. While I understand you think it starts at LABEL, LABEL already points to INTEXT, so adjusting INTEXT to allow this would automatically fix the LABEL issue as well as other cases. --Masem (t) 18:43, 18 February 2022 (UTC)

I do not agree that classifying Jim Jones as a cult leader is an opinion. If that's an opinion, then it's also "just an opinion" that that the letter R isn't a vowel in English. WhatamIdoing (talk) 19:54, 20 February 2022 (UTC)

It's pretty easy - the definition of a "cult" is not precise nor objective, though there is certainly some core concepts of what a cult can be are reasonably shared. As such, we should wait for both time and critical mass of sources to assure that we're using enduring and near universal agreement that a group is classified as a cult before we take that in wikivoice, instead of just taking a few sources at their word. We can do that for Jonestown but we can't do that for Qanon at this time, for example. Its the same process that scientific theories are taken as fact over time, after mass agreement over time and many many sources. --Masem (t) 20:07, 20 February 2022 (UTC)

The existence of precise and objective definitions is not what makes something a fact or an opinion. WhatamIdoing (talk) 21:53, 20 February 2022 (UTC)

To repeat what was noted above, LABEL is mainly a gloss on a couple of principles enumerated at NPOV. One of those, to avoid stating opinions as facts, says to use INTEXT attribution, so that can hardly be removed here. Both the list of example words and a MOS:LABEL that spells things out beyond merely suggesting to abide by NPOV and INTEXT are necessary. "Cult" is the leading example of a word that should or even must be on a WTW list. Your examples of Jonestown and the like, to the extent they work at all, work because there is a clear and non valueladen (in the sense I explicated above) dividing line, i.e., the leader is able to command the followers to perform murder or suicide, with a nearly 100 percent rate of success. It seems that clarifying what valueladen and contentious are supposed to mean here, maybe by replacing both words with more specific explanations, would solve the problems you are rushing to address by removing material that belongs in LABEL. Sesquivalent (talk) 02:57, 21 February 2022 (UTC)

I agree that cult should be in this list, but I do not agree that the statement that "Jim Jones was a cult leader" is merely an opinion. Therefore – since it is a fact – the bit about avoiding stating opinions as facts (though excellent advice) is irrelevant in this instance.

I do agree that NPOV and INTEXT should be followed scrupulously. I think the better way to approach that is for LABEL to direct say to follow NPOV and INTEXT, and not to say that in-text attribution must be given for widely accepted facts. WhatamIdoing (talk) 05:09, 21 February 2022 (UTC)

It should not be treated as a "fact", but I would agree as a widely accepted conclusion by most researchers distanced from the event at this point. There is no way to prove that what Jones led was a cult because there are no object tests of defintiions for what a cult is, but we can agree that there likely sufficient sources now that agree it is very much an example of a cult. This is why I keep comparing this to how scientific theory works: we accept the theory of gravity as a "fact" but that's because there's centuries of experiments to show it the most likely explanation and validate it, and there is still a slim chance that there is an alternate theory that may be the true "fact".

It is important that we make this distinction between just saying that we can state a label in wikivoice w/o attribution because its a fact, and we can state a label in wikivoice w/op attribution because it is readily agreed upon by multiple sources is that this then creates the required onus to show that many sources exist to demonstrate this agreement to allow us to use a contentious term in wikivoice. That onus has to be done before this use can be inserted into the article, which currently is not standard practice and instead it is the cherry picking of one or two sources that use a label and thus claim "this must be a fact" that happens way too much. --Masem (t) 06:02, 21 February 2022 (UTC)

Okay: We have agreed that "we can state a label in wikivoice w/op attribution because it is readily agreed upon by multiple sources" (assuming that these multiple sources represent "fairly, proportionately, and, as far as possible, without editorial bias, all the significant views that have been published by reliable sources", which in the case of whether Jim Jones was a cult leader, they do).

Now: What does LABEL actually, currently, exactly-as-written say about providing in-text attribution for a sentence for which you and I agree can state a label in wikivoice without attribution, like "Jim Jones was a cult leader"? WhatamIdoing (talk) 21:30, 22 February 2022 (UTC)

LABEL doesn't because it points back to INTEXT to tell us that all views should be given in-line attribution. That is what is limiting our hands here, and why INTEXT should be updated to consider cases when there is this near universal agreement for a specific view. --Masem (t) 04:35, 23 February 2022 (UTC)

That's exactly the problem. LABEL doesn't "point back to INTEXT". LABEL instead "tells us that all views should be given inline attribution", including views that INTEXT says should not be given inline attribution.

INTEXT, as currently written, does not require that we say "Basically every scholar who's written about cults since the 1970s says Jim Jones was a cult leader". LABEL says that. Not INTEXT. WhatamIdoing (talk) 17:11, 24 February 2022 (UTC)

INTEXT still would require inline attribution, because that is still a viewpoint. Mass agreement of a view in sources does not equate to making it fact. (The only place where mass agreement on a view is called out is UNDUE). INTEXT still would require we spell out who made such claims. Which we agree we shouldn't have to, and why INTEXT should explain how to handle cases of mass agreement of a given view without the need for direct inline attribution, whether that is from a LABEL or from other nonlabel facet. --Masem (t) 18:21, 24 February 2022 (UTC)

What sentence in WP:INTEXT makes you believe that INTEXT always requires attribution for anything that can be called "a viewpoint"? WhatamIdoing (talk) 23:55, 24 February 2022 (UTC)

Are y'all working towards wording an RfC/wording change or is this just editors trying to communicate their understandings of policy with each other (not that I don't think that's worth doing)? --A. C. Santacruz ⁂ Please ping me! 17:59, 10 March 2022 (UTC)

References

^ It is possible for facts to be value-laden. Within medical articles, we see this as the difference between "50% survive cancer" vs "50% die from cancer". They are both objective facts. They are, in fact, both the same objective fact. They both carry values (namely, whether the deaths are more important than the survivals).

[1] It is possible for facts to be value-laden. Within medical articles, we see this as the difference between "50% survive cancer" vs "50% die from cancer". They are both objective facts. They are, in fact, both the same objective fact. They both carry values (namely, whether the deaths are more important than the survivals).

[1]