Wikipedia talk:Consensus

From Wikipedia, the free encyclopedia
Frequently asked questions (FAQ)
Q: When was WP:CONEXCEPT, which says that editors at the English Wikipedia do not get to overrule the Wikimedia Foundation on issues like server load, software and legal issues, first added?

A: It was added in January 2007 by User:Circeus, after a brief discussion on the talk page in the context of whether this page should be a policy rather than a guideline. It has been discussed and amended many times since then, e.g., here, here, and here.


Levels of consensus

The levels of consensus section needs to be elaborated. Right now it names essays, WikiProject advice pages, information pages and template documentation pages (which may or may not represent broad community consensus, or one editor's personal opinion) and policies (which should represent the broadest community consensus). There are a few intermediate levels of consensus that are mentioned in the policy but not in the levels-of-consensus section:

The addition is meant to reduce article ownership behavior where an editor refuses to acknowledge consensus because they didn't approve it, or because there's no local consensus. Bright☀ 11:03, 19 September 2017 (UTC)

Changes in italics. Bright☀ 09:36, 20 September 2017 (UTC) Alternate bullet-point version below. Bright☀ 09:04, 21 September 2017 (UTC)

List in its own subsection (version 2) - 23:56, 22 September 2017 (UTC)
Scale

Though consensus is not a vote, the amount of community participation and agreement indicates the level of acceptance of a consensus. Usually, participation follows this rising order:

Added bullet-point version. Bright☀ 09:04, 21 September 2017 (UTC)

You have successfully convinced me you are thinking about something worth improving, but it's not entirely clear how you propose to change the existing text. The best way to communicate your ideas might be to just copy the subsection here and edit it the way you like, then post so we can see exactly what you're thinking. Thanks NewsAndEventsGuy (talk) 11:47, 19 September 2017 (UTC)
Good idea, added. Bright☀ 09:36, 20 September 2017 (UTC)
@BrightR: Excellent, I'll take a look in a day or two. Meanwhile, as a housekeeping matter, please consider deleting my comments in this subtread. No need for others to review this now that you updated the original post. NewsAndEventsGuy (talk) 15:00, 20 September 2017 (UTC)
As long as we are discussing this topic... I would love to see some discussion about where policy related RFCs fit in... these can take place on LOCAL pages (article talk pages, project talk pages, etc), but are not necessarily LOCAL in scope (with wide participation, thus reflecting wide community consensus). One option would be to shift from "local" (focused on the place where the consensus was established) to "limited" (focused on the number of participants). Blueboar (talk) 12:46, 19 September 2017 (UTC)
Three things:
  • Another consideration: The stuff that's in the current Levels of consensus section is really only about the difference between content consensus and policy/guideline consensus and the special nature of consensus needed for policy/guidelines. I'd be extremely loath to change that language or dilute it by inserting inside it material having more to do with general levels of consensus than with policies and guidelines. On the other hand, the current section title is too broad for just policies and guidelines. What if BrightR's material, or something like it, were to be inserted before all the existing material and the existing material (along with the three current shortcuts) were moved into a four-equals subsection just below it named something like "Consensus for policies and guidelines"?
  • Having said that, I'm not sure just how useful or necessary this change is. I don't think that it's wrong or mistaken (though it probably needs some discussion and tweaking), I'm just not sure what problem it's really seeking to address that's not already adequately addressed here and elsewhere. I like it in one way because it makes clear something that's only kind of implied in this policy right now, but I'm not at all sure how needed it is.
  • A word of caution: This proposal could by mentioning consensus by silence become very controversial and easily become sidetracked through that detail. Though the concept is, as noted, already here in EDITCONSENSUS what it means (especially mechanically, if anything at all), and the degree to which it should be emphasized has caused some heated discussions in the past.
Best regards, TransporterMan (TALK) 16:51, 19 September 2017 (UTC)
All very good notes.
  1. No problem.
  2. This is specifically meant to address situations where an editor engages in WP:OWN behavior by saying, for example "you don't have consensus here" or "that guideline is just a suggestion, therefore my version should remain until you have consensus," whether or not the current version has any consensus.
  3. I tried to make it clear that consensus by silence is no consensus the moment it's disputed.
Hopefully this will make the proposal as uncontroversial as possible. Bright☀ 09:42, 20 September 2017 (UTC)
I'd been thinking about this topic for a while and though I reserve my right to comment further on this section in the future, I think the change is an improvement. I would like to see policies and guidelines separated to clearly identify that a guideline is not equivalent to a policy. Scribolt (talk) 11:46, 20 September 2017 (UTC)
Updated. It's worth mentioning that the order is not absolute; some RfCs actually have broader support than some guidelines. Parts of some guidelines (and even parts of some policies, couple of examples above) were written by one person and simply stuck, while formal RfCs always have broad exposure. Some talk-page discussions involve a broader subset of the community than some parts of guidelines and policies, too. Bright☀ 12:22, 20 September 2017 (UTC)
It is fuzzy and that was one of the things I've been thinking about, and is why some of the language used causes me some concern; "cannot" (where do IAR and sensible exceptions fit in), "breadth" & "wider" (for exactly the reason you mentioned above, this doesn't necessarily correlate to a 'level' of consensus arrived at by a particular methodology). However, although I've got some ideas, they may never turn into concrete suggestions, so as I said I think what you've suggested is an improvement.Scribolt (talk) 12:35, 20 September 2017 (UTC)
In general all the "cannot" in policies should be replaced with "should not", but maybe that's a semantics discussion for another time. Bright☀ 12:57, 20 September 2017 (UTC)
  • The proposed change seems to be covered in the sections above the LOC....is the intent to link to the other sections that explains more?--Moxy (talk) 14:17, 20 September 2017 (UTC)
Yup. Bright☀ 14:19, 20 September 2017 (UTC)
oK. ....then I suggest we move the LOC section to before the others....odd to see the intro to the other sections after the sections it mentions.--Moxy (talk) 14:24, 20 September 2017 (UTC)
Good idea. What if we were to leave the current section where it is but rename it "Consensus for policies and guidelines" and then put BrightR's material (I really liked the bullet points better, paragraphed text is too dense) in a new section entitled "Levels of consensus" above the existing "Achieving consensus" section? Again, I'm opposed to changing any of the text in the current section. This is a draft of what I'm suggesting, though the wording of the last bullet point probably still needs some cleanup. Best regards, TransporterMan (TALK) 16:31, 20 September 2017 (UTC)
I like it....but all our templates like {{Information page}} and hundreds of administrative pages like WP:HOWTOPAGES link to "Levels of consensus" because of its content about the type of pages. Is there another name we can use because linking to the new section would be very odd for many incoming links. Perhaps call the section an Overview? Moxy (talk) 16:55, 20 September 2017 (UTC)
Good point. How about "Types of consensus"? Best regards, TransporterMan (TALK) 18:32, 20 September 2017 (UTC)
Added a bullet-point version, probably the one I'll RfC since it's easier to read. Bright☀ 09:04, 21 September 2017 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I returned to load this in my brain, only to find two text boxes in the opening post instead of one. I might be able to decipher the full current proposal reading the subsequent discussion, but.... being a policy page, it will help the final consensus if new comers to the discussion don't have to do that. So as a housekeeping idea... to facilitate discussion... please consider creating a demonstration edit by first changing the policy page, then self reverting and posting the DIFF to the proposal. That would lock it in nice and neat for discussion purposes. If this suggestion appeals to you, I further suggest you name the DIFF anticipating that there may later ones as the ideas are vetted by additional eds. Maybe "DEMO-Ver-01" or equivalent. That way if additional good ideas come along, and the conversation gets convoluted, you can update the proposal with a link to "DEMO-Ver-02", and with proper threading and subsection titling it will be super-simple for people to instantly know what is being proposed and the current status of discussion. NewsAndEventsGuy (talk) 11:53, 21 September 2017 (UTC)

The RfC will not have two boxes; when this discussion concludes, I'll phrase an RfC neutrally and succinctly ("Should the following subsection be added to WP:Local consensus? [...]") Bright☀ 12:24, 21 September 2017 (UTC)
Apologies, if I had sorted out the subsequent discussion I would have learned of your plans to repackage it after initial feedback. Carry on! NewsAndEventsGuy (talk) 23:35, 21 September 2017 (UTC)
Where do RFC's fit into CONLEVEL?

I mentioned this above, but it really is an unrelated issue from what is being discussed... so I am hiving it off into a separate discussion. CONLEVEL is missing an important item: where do RFCs fit in? The problem is that RFCs can take place on local pages, but they can sometimes reflect very broad community consensus (indeed some RFCs can reflect a broader consensus than was achieved at the guideline level). Then again, other RCFs reflect consensus of only a few editors. It all depends on the level of participation (not the level of location). Thus, I am not sure how to account for RFCs, but I think we do need to account for them in some way. Blueboar (talk) 12:29, 22 September 2017 (UTC)

RCFs reflect consensus of only a few editors. It all depends on the level of participation - exactly. It entirely depends on the number of participants and percentage of agreement. An RfC with a million participants and a 50% split between two options is a clear no-consensus, and so any local consensus is better than it. An RfC with 20 participants and 90% agreement is pretty strong, and local consensuses shouldn't be able to override it; if it's reasonable to believe the RfC might not represent the general consensus any more, then a new RfC can be held. Exact numbers should not be specified. Bright☀ 14:11, 22 September 2017 (UTC)
The problem is that an RFC held at an article level can result in an outcome that disagrees with something said at the guideline level (this is less common with policy). The RFC then gets dismissed as being a "local consensus" (due to location) when in fact the RFC is hardly "local". Blueboar (talk) 14:59, 22 September 2017 (UTC)
The policy page already specifies that it's not local (inviting others to participate), and in any case whether it's local or "broad" depends on the number of participants and the exposure of the notices of the discussion. I'm getting the feeling this is about the semantics of the word "local" in favor of the word "limited"? In which case, I strongly urge you to avoid quibbling over this, like this seemingly endless quibbling over the word "reason". Bright☀ 15:08, 22 September 2017 (UTC)

RFC's are a tool for obtaining consensus (just like 3O, DRN, or MEDCOM) , not a type or level of consensus. Depending on how they're "advertised" and where they are used they can either form local consensus or community-wide consensus. The difference between local consensus and wider consensus is the type of place (along, again I would argue, with advertisement) that the RFC occurs. If the RFC happens at Talk:My Pretty Pony or Wikipedia talk:WikiProject Radio, it's ordinarily local; if it happens at Wikipedia talk:Verifiability, it's ordinarily community-wide. The best place to illustrate the distinction is at wiki-projects. Ordinarily, RFC's (and, for that matter, ordinary consensus discussions) at wiki-project talk pages only form local consensus and essay-level standards for those projects (as currently stated at CONLIMITED), but with proper advertisement and intent they can create policy or guidelines. Thus, the means of obtaining consensus, whether through ordinary discussion or through RFC, isn't the determinant; it's how the RFC is promoted and advertised. To say it a third way: If an RFC at a wiki-project specifically says that it's proposing to create a guideline or policy and that RFC is, per the Policy policy advertised at places like the Village Pump then it can, indeed, create policy or guidelines. In that case the RFC is community-wide. But if you file an RFC at Talk:My Pretty Pony to determine whether chartreuse ponies can be documented through reliable sources, that's local consensus. Best regards, TransporterMan (TALK) 22:15, 22 September 2017 (UTC)

Hmm, good point, I'll revise the bullet-point list above. Bright☀ 23:52, 22 September 2017 (UTC)

Request for Comment - add subsection to WP:Local consensus about consensus levels

Should the "scale" subsection below be added to the WP:Local consensus section of WP:Consensus? Bright☀ 10:14, 7 October 2017 (UTC)

Scale

Though consensus is not a vote, the amount of community participation and agreement indicates the level of acceptance of a consensus. Usually, participation follows this rising order:

See also the preliminary discussion of this topic. Bright☀ 10:14, 7 October 2017 (UTC)

Survey

  • Very strong oppose - Totally unnecessary attempt to impose on the community one editor's idiosyncratic and inflexible views of what consensus is. Beyond My Ken (talk) 20:23, 7 October 2017 (UTC)
  • Oppose without a clear statement of the claimed underlying problem. There is only one "consensus", but there are different ways of estimating the consensus position. Experienced editors can often guess what consensus would be from previous cases. An unchallenged edit shows consensus by silence. Agreement among regulars on a talk page indicates local consensus. However if the conclusions from these methods of estimating consensus varies, a wide community discussion is needed for a better measurement. Johnuniq (talk) 22:25, 7 October 2017 (UTC)
This won't change your !vote but everything you said actually supports this addition, which clarifies exactly that. Bright☀ 09:47, 12 October 2017 (UTC)
  • Oppose per the law of unintended consequences. Suggest it is placed in a WP:ESSAY. --Tom (LT) (talk) 22:26, 7 October 2017 (UTC)
  • Oppose, per others above. Flyer22 Reborn (talk) 22:29, 7 October 2017 (UTC)
  • Oppose: Completely erratic attempt to establish a one editor's view on consensus. KGirl (Wanna chat?) 02:04, 8 October 2017 (UTC)
  • Oppose, appears to be a solution looking for a problem. Stifle (talk) 16:14, 9 October 2017 (UTC)
  • Support when implemented as shown below (i.e. in this edit). While I share some of the concerns mentioned by the opposes, I think this correctly summarizes the policy and is useful as a summary, since the proposal links to the relevant parts of the policy. Regards, TransporterMan (TALK) 16:59, 9 October 2017 (UTC)
  • Oppose per the views expressed above. I also feel it doesn't correctly preserve or explain the current language of the page. Gestrid (talk) 18:21, 9 October 2017 (UTC)
  • Oppose - there are some highly volatile issues that many editors refuse to partake in...such as politics and religion...many don't want to be "labeled" for expressing their views so they say nothing. I agree with BMK regarding the one editor's views and inflexibility. I also believe that unless the topic is expressly about changing a policy/guideline, consensus to waive it in a particular instance should not override existing PAGs. Atsme📞📧 00:25, 10 October 2017 (UTC)
  • Oppose - splitting hairs by the shades of gray. Yes, consensus is not a vote, end exactly for that reason consensus is established by the scope of arguments, not by scope of participation. If someone wants hierarchy, we have WP:Dispute resolution escalation process. Staszek Lem (talk) 00:43, 10 October 2017 (UTC)
  • Oppose. Attempts to define, quantitate, or scale consensuses, as a result following a presume question, misconstrue the meaning of consensus. Consensus decision-making involves re-formulating the question to avoid dividing conflicts, and these almost-but-not-quite consensus usually involve a failure to re-examine the question following later input, and thus should be regarded as unconcluded discussions. --SmokeyJoe (talk) 00:45, 10 October 2017 (UTC)
  • Oppose. I appreciate the intention, but I can envision this being misused for wikilawyering. --Tryptofish (talk) 21:45, 10 October 2017 (UTC)
  • I'll be that guy – Support. Unless I'm completely missing something, this is a perfectly accurate description of how we handle conflicting consensus: global trumps local, and the more participation led to a consensus, the "firmer" that consensus is, in the sense of requiring a higher bar to overturn it later. Right? I really don't understand the wave of opposition here, nor do I see good arguments being made – I mostly see personal sleights and non-sequitur slippery slope arguments, which a sensible closer will ignore. As Bright pointed out, the complete 180 in support from the initial discussion to this one is...very curious. —swpbT go beyond 13:27, 13 October 2017 (UTC)
If this is what we already do without this text, is not adding the text WP:CREEP and liable to provide warriors some other verbiage to weaponize? I have not yet stated a WP:!VOTE because I'm still waiting for someone to explain what problem would be solved by the additional text to make the unknown risks worth it? Open mind here. If you move the discussion part of your comment to the discussion section, @Swpb:, please move my reply as well. NewsAndEventsGuy (talk) 16:31, 13 October 2017 (UTC)
CREEP is a particular type of new guidance, concerned with finicky details that should be left to editor discretion; this proposal doesn't resemble that at all. This is making a fundamental standard practice explicit, because leaving it implicit has invited too much re-litigation. I also don't see anything to "weaponize" here, and I think anyone worried about that has the burden to demonstrate what such abuse would look like, because I can't see it: this guidance emphasizes that wider participation carries more weight, which totally undermines the lone warrior getting their way by browbeating a few users in an obscure discussion. —swpbT go beyond 18:50, 13 October 2017 (UTC)
Would you our @BrightR: be able to show examples of the "relitigation" this might have prevented or at least helped to resolve?NewsAndEventsGuy (talk) 19:49, 13 October 2017 (UTC)
I have about four examples with my own personal involvement (with some of the oppose voters here!) of discussions with arguments of "community consensus doesn't matter because there's no local consensus." To avoid the personal angle, I'll round up some randomly-found "local consensus" examples, but please mind that this issue has no searchable keywords ("consensus" or "local consensus" or "community consensus" give endless results...)
From personal experience, I find it's relatively frequent that an article owner would say something in the lines of "your consensus is not binding!" True, consensus is "not binding", but it's still better than non-consensus or not-consensus, and broader consensus is better than local consensus. This is all the proposed addition is saying. Bright☀ 05:50, 18 October 2017 (UTC)
  • Oppose for taking number of contributors as a too exclusive criterion, see my more detailed explanations below. --Francis Schonken (talk) 13:39, 13 October 2017 (UTC)
  • comment (oppose) What would be the main point of classifying RfC's outcomes as "broad", "local", and so on? To know how strongly it is established and how hard would be to change it or to go wp:IAR about it. I think we could do that much better if we make past RfCs (and RfC-like) discussions easily available. I am not sure about the exact way but, as an example, talk pages could have a linked index of past RfC's with one liner description of the subject, outcome, and date. While policy pages should have "sources" pointing to the latest discussion(s) about each topic. Currently, almost only "the initiated" know how consensus was reached and with what arguments. And I presume having sources should also apply to policy pages for the very same need of verification. - Nabla (talk) 02:59, 14 October 2017 (UTC)
  • Support To me, this seems to be a suitably high level description of what everyone who's been here a while has already realized about how consensus actually works. Does anyone really believe that a non challenged or discussed edit has more consensus than an issue that's been discussed on the talk page? Or that something that's been discussed by broadly attended RfC with high attendance doesn't have more consensus than an informal discussion with three involved editors on the talk page? Or that something in a policy has more consensus than an RfC at the article level? There is a concern about gaming, but BRD relies on a consistent understanding of what the first BOLD edit is, which is at least partially related to whether there was a form of consensus for the content to begin with. I believe adding a little more clarity to what is established practice here could serve to reduce the lawyering that goes on in some of these cases. Scribolt (talk) 06:28, 18 October 2017 (UTC)
  • Agree with Scribolt, this matches my experience on how consensus is established beyond discussion over a particular article, and I think that a similar description could help newcomers understand what's going on in larger discussions. Maybe if the section was written in a way that emphazises that it's being descriptive, not normative, people would agree to it better? I.e. not to be used as a reason on itself to support arguments, but merely as a summary of policy that is expressed elsewhere. After all, the proposed text roughly follows the approach at Consensus-building and WP:Dispute resolution.
I like Nabla's suggestion above to routinely provide links to previous discussions that informed a particular policy or guideline. I've seen this approach working very well in {{Old AfD multi}}, the RS noticeboard archives, and Wikipedia:WikiProject Video games/Sources#Checklist. Diego (talk) 13:26, 18 October 2017 (UTC)

Discussion

Can you be a bit more explicit about how it would be added? As I've said before, I can't support anything which doesn't preserve the current language. If this is an addition to the current language, I'll be able to support it; if it replaces the current language, then I'll oppose it; it it modifies the current language then my position will vary based upon exactly how it does so. Best regards, TransporterMan (TALK) 16:47, 7 October 2017 (UTC)

The first two paragraphs below already exist in the policy. The proposal would add the text in the "scale" subsection. This note added by me.NewsAndEventsGuy (talk) 16:36, 13 October 2017 (UTC)

Levels of consensus

Consensus among a limited group of editors, at one place and time, cannot override community consensus on a wider scale. For instance, unless they can convince the broader community that such action is right, participants in a WikiProject cannot decide that some generally accepted policy or guideline does not apply to articles within its scope. WikiProject advice pages, information pages and template documentation pages have not formally been approved by the community through the policy and guideline proposal process, thus have no more status than an essay.

Wikipedia has a higher standard of participation and consensus for changes to policies and guidelines than to other types of pages. This is because they reflect established consensus, and their stability and consistency are important to the community. As a result, editors often propose substantive changes on the talk page first to permit discussion before implementing the change. Changes may be made without prior discussion, but they are subject to a high level of scrutiny. The community is more likely to accept edits to policy if they are made slowly and conservatively, with active efforts to seek out input and agreement from others.

Scale

Though consensus is not a vote, the amount of community participation and agreement indicates the level of acceptance of a consensus. Usually, participation follows this rising order:

Bright☀ 09:25, 9 October 2017 (UTC)
  • One thing is missing... we need some discussion of RFCs would fit into all this. The problem, of course, is that an RFC may get a small turnout or a large turnout... and yet an RFC is considered an "official" reflection of consensus (and is a recommended step in WP:Dispute resolution). Blueboar (talk) 17:11, 9 October 2017 (UTC)

Pinging NewsAndEventsGuy, Scribolt, and Moxy who participated in the preliminary discussion. Strange how the preliminary discussion had almost complete agreement and now the RfC has almost complete opposition, and even people whose survey rationale supports the addition, oppose it. Bright☀ 09:45, 12 October 2017 (UTC)

Suggest you put the first two paragraphs in a grey box and add a title bar to the whole thing that says Existing text in grey box, proposed new text in green box. I'm not convinced everyone realizes the first two paragraphs are already part of the policy.NewsAndEventsGuy (talk) 10:51, 12 October 2017 (UTC) Also, I share others concerns how this new text will possibly be fodder for arguments where one side may try to frame the debate around he levels in this scale. Do you think that might happen? Is this supposed to be a tool to help resolve disputes? As a said in the protracted discussion I hadn't studiend the wall of text, though I'm sure there were lots of good points made by many. I was waiting for the distillation in this RFC. I kinda expected a new-and-improved succinct explanation what problem this is trying to solve, but all I see is the naked proposal. As it stands, I'd have to oppose, but you haven't provided your reasons yet. In 50 words or less if possible, why? NewsAndEventsGuy (talk) 11:59, 12 October 2017 (UTC)
The problem, as stated before very succinctly, is editors who try to enforce local consensus (or local no-consensus) in the face of broader consensus. To take the Daily Mail example, it's like an editor insisting on using the Daily Mail as a source inappropriately despite broad community consensus that it should not be used. What's more, I don't know if anything I can say can sway any of the oppose not-voters because some of them explicitly repeat what the addition says and yet vote against it: "An unchallenged edit shows consensus by silence. Agreement among regulars on a talk page indicates local consensus. However if the conclusions from these methods of estimating consensus varies, a wide community discussion is needed for a better measurement. Oppose." When someone says almost word-for-word what the proposal says, and yet opposes it, I get the feeling this proposal is not being weighed on its merits... Then there are oppose not-votes that say "look at a the Daily Mail consensus, which was among a limited number of participants" - technically I guess "a hundred" (30 oppose, 68 support, plus several comments and several more closing admins) is limited... anyway I find that the not-votes here are very much divorced from reality. Bright☀ 15:39, 12 October 2017 (UTC)
When people go "meta" (by discussing the discussion instead of the merits) I quickly lose interest. NewsAndEventsGuy (talk) 16:59, 12 October 2017 (UTC)
That's because the proposal is so simple and the rationale is so brief that explaining it takes one or two sentences. It's the replies that I find utterly inconsistent and demonstrably wrong. Bright☀ 10:10, 13 October 2017 (UTC)
  • Oppose – "number of contributors" is the wrong entrance to make this discrimination. Compare the RfC that outruled Daily Mail as a general-use RS (limited number of participants, nonetheless a clear decision); Compare high participation discussions where the majority "wins" with a small margin (may be defined as consensus, but not a "broad" one); Compare Gamergate-related decisions (often high numbers of participants ushered in from elsewhere: despite their numbers they didn't weigh on these decisions); etc. --Francis Schonken (talk) 12:36, 12 October 2017 (UTC)
Disagree, keeping with a very strong opposition to the proposal. --Francis Schonken (talk) 10:21, 13 October 2017 (UTC)
The wording specifically and explicitly mentions that consensus is not a vote... The mention of the Daily Mail example is particularly puzzling since it was discussed hundreds of times by countless editors before a final decision was made. And even if you discount all those presumably no-consensus discussions, the final discussion and consensus was made among a very large number of editors. Not that it'll change your not-vote, but you both mischaracterize what the suggested addition explicitly states, and you mischaracterize the very examples you provided. Bright☀ 15:27, 12 October 2017 (UTC)
Re. Daily Mail "discussed hundreds of times by countless editors before a final decision was made" – yet, after the decision was made commentators outside Wikipedia wondered how come that such a small percentage of Wikipedia editors can make a decision that applies to Wikipedia as a whole. To which was replied that these outside commentators didn't understand how Wikipedia's decision procedures work (read: no clue about the WP:CONSENSUS policy). Let's not give such outside commentators ammunition to torch Wikipedia's decision procedures. So, no, can't accept this proposal, also because of what you point out now: it seems internally inconsistent with the current policy ("not a vote" includes, indeed, not counting the overall number of participants). --Francis Schonken (talk) 15:42, 12 October 2017 (UTC)
How can what non-participants think have any bearing on Wikipedia policy? How can you call a consensus among 100+ editors (with about 2/3 supporters, 1/3 opposers) "limited"? "not-vote" doesn't mean that the number of supporters and opposers doesn't matter. It means that a discussion has to take place, and it's perfectly fine to hold a straw poll after a discussion has taken place, see this very RfC or the Daily Mail RfC, and in fact pretty much any community-wide discussion. See WP:WIKINOTVOTE. This addition does nothing to encourage more voting or not-voting. It collects existing wording from WP:CONSENSUS and arranges them from the narrowest consensus to the broadest consensus, in order to emphasize that a local consensus cannot override broader community consensus.
  • Wording that already exists in WP:CONSENSUS:
    • Consensus among a limited group of editors, at one place and time, cannot override community consensus on a wider scale.
    • Wikipedia has a higher standard of participation and consensus for changes to policies and guidelines than to other types of pages. This is because they reflect established consensus, and their stability and consistency are important to the community.
    • Consensus is a normal and usually implicit and invisible process across Wikipedia. Any edit that is not disputed or reverted by another editor can be assumed to have consensus.
    • If an edit is challenged, or is likely to be challenged, editors should use talk pages to explain why an addition, change, or removal improves the article, and hence the encyclopedia.
All this already exists in the policy and the proposed addition merely collects it all into one place, in order to avoid editors who insist that their local implicit consensus (or lack of consensus) overrides broader community consensus. Imagine, if you will, that I collected all these existing policy snippets and put them in a section called "local consensus" that reads like this. And then I get really weird oppose not-votes like "this is one editor's view" (which apparently is already part of Wikipedia policy, albeit spread across four or five sections), or "you can't use the phrase 'number of contributors' because of what people outside Wikipedia think of the Daily Mail consensus, which was limited" (to about a hundred participants, which is a very large discussion by Wikipedia scale), or an oppose not-vote that repeats the proposal almost word-for-word and yet opposes it...
This is already part of the policy, I'm simply trying to make it more accessible so editors would not insist that their local-consensus/local-no-consensus overrides broader community consensus. Bright☀ 16:03, 12 October 2017 (UTC)
None of what you quote above has the slightest indication that all of a sudden starting to count participants in a discussion ("...number of contributors..." as the proposal has it, as an instrument to compare "levels of consensus") has any merit. Keeping to my, in the mean while very strong, oppose. --Francis Schonken (talk) 16:13, 12 October 2017 (UTC)
It's already in the policy: Many of these discussions will involve polls of one sort or another; but as consensus is determined by the quality of arguments (not by a simple counted majority), polls should be regarded as structured discussions rather than voting which parallels Though consensus is not a vote, the amount of community participation and agreement indicates the level of acceptance of a consensus; and Consensus among a limited group of editors, at one place and time, cannot override community consensus on a wider scale. which parallels Local consensus, among a small number of contributors and Broader community consensus, among a large number of contributors. polls [are not] votingconsensus is not a vote; limited groupsmall number; wider scalelarge number. The proposal is not encouraging voting. It discourages local consensus (among a limited group, which is a small number of editors) used to override broader community consensus (which has a wider scale, which means a large number of editors). It's already in the policy. The proposal is collecting it into one place. Bright☀ 16:28, 12 October 2017 (UTC)
"wider scale" does not (necessarily) mean "large(r) number of contributors". Some "wider scale" consensus procedures (e.g. WP:DRN) do not necessarily imply a larger number of participants. They can be even less numerous to come to a "wider scale" decision. --Francis Schonken (talk) 16:45, 12 October 2017 (UTC)
Can you show me a single example of consensus that was reached on a "wider scale" with less participants? Logically, it's inescapable that the second you involve more people, you have a larger number of people involved... By appealing to DRN you involve more editors... Bright☀ 10:10, 13 October 2017 (UTC)
Nah, again, not necessarily. E.g. an editor takes a content dispute to ANI, with multiple editors commenting, after which it is decided it should go elsewhere as a content dispute: wherever it is taken the new consensus may emerge with less participants in the discussion. So, no, your proposal is principally flawed on this point. --Francis Schonken (talk) 10:19, 13 October 2017 (UTC)
new consensus may emerge with less participants in the discussion - but it's not a broader consensus, it's just a consensus. The proposal doesn't say "new consensus HAS TO HAVE MORE PARTICIPANTS". It says that a broader consensus is usually among a large number of participants. You're somehow interpreting that as "any consensus has to have a larger number of participants than the previous consensus". No. The proposal doesn't say that. Bright☀ 10:24, 13 October 2017 (UTC)
It can surely be a broader consensus: in the example above, the ANI discussion may have yielded a 4 against 2 consensus on the content matter, while in the subsequent discussion on a more appropriate content noticeboard, with 5, in the end they all agree on the prior minority viewpoint: by all means a "broader" consensus, although there was one participant less. --Francis Schonken (talk) 10:37, 13 October 2017 (UTC)
  • Lets consider the following scenario.... a dispute arises at an article over whether X should be considered an exception to guideline Y. An RFC is filed to resolve this dispute. 50 editors respond to the RFC, with 44 saying "Yes, make an exception" and 6 saying "No, don't". I think most would agree that there is a clear consensus for making the exception.
Now, let's assume that the dispute is subsequently taken to the relevant guideline talk page or noticeboard... where only 5 editors discuss, but unanimously say "No... don't make an exception". A clear counter-consensus for Not making an exception.
So here's the question: Which discussion reflects the consensus of the "broader community"? The far larger one formed on the "local" article page... or the far smaller one at the "non-local" guideline page? Blueboar (talk) 11:52, 13 October 2017 (UTC)
"Where" the consensus is reached is not an exclusive criterion either. The RfC OP may have been tendentiously worded, The RfC may have been inappropriately canvassed, etc.; In the second discussion the fact that the discussion was going on may have been insufficiently announced, its section title may have been misleading about its true objective, the editors of the article where the preceding RfC was held may have been unaware that the second discussion was targeting that article specifically (then they can still claim a local exception was possible and agreed upon), it may have been left unmentioned that a preceding RfC already determined consensus on the matter (so that the new discussion is in fact a WP:CCC determination), it may have been closed too soon, etc...
No, all these attempts at defining consensus levels by a too limited set of "countables" are going nowhere. The wider a discussion is announced with a clear presentation of the discussion at hand, without being rigged in any way (no socking, inappropriate canvassing, etc, etc..), the more it is presumed to be representative of the editor community as a whole. The larger the part of the editing community that *had an opportunity to participate* the broader a consensus generally is (although also there, there are exceptions: does an ArbCom decision establish a "broader" consensus than an ANI discussion with more participants? Yet, ANI decisions can be taken to ArbCom if a participant doesn't agree with the outcome at ANI...).
It happens that I arrive at a discussion about an issue about which I have no strong views. The option which I would probably favour most has about 60% support. In such event I might decide not to participate in the discussion (really, when that is the situation I usually don't): does that mean that the outcome of the discussion would have a less "broad" consensus? I would support the outcome, whatever that outcome is (I had the opportunity to participate), wouldn't I? That means that a broader part of the community supports the consensus, which makes the consensus broader without my participation in the actual !vote. --Francis Schonken (talk) 13:22, 13 October 2017 (UTC)
by all means a "broader" consensus ArbCom is explicitly not part of the Wikipedia consensus process, but even if it were, what you are describing is a new consensus, not a broader consensus. If 4 people agree on a topic, and then a marginally larger or smaller group (2, 3, 4, 5, 6, no need to exactly quantify, but on a similar scale) later come to a different agreement on the topic, it's merely a new consensus, not a broader consensus. But regardless this is the third time you bring up an irrelevant "counterpoint" since ArbCom is not part of the Wikipedia consensus process—WP:CONEXCEPT. Bright☀ 02:08, 18 October 2017 (UTC)
does that mean that the outcome of the discussion would have a less "broad" consensus? the RfC is not remotely worded in the way you suggest ("a consensus with n+1 people is broader than a consensus with n people"). While a broader consensus will have more participants, a consensus with more participants is not necessarily broader. This "n vs n+1" attitude does not exist in the text, you are forcing it where it doesn't exist. Take this existing wording: The goal of a consensus-building discussion is to resolve disputes in a way that reflects Wikipedia's goals and policies while angering as few contributors as possible. Now imagine someone raising, like you, an objection to this on the grounds that "if consensus A angers 60% of people, and consensus B angers 60%+1 people, then this policy implies A the true consensus." Of course not. Consensus is not a vote. And the suggested addition to policy literally starts with a link to WP:WIKINOTVOTE. Bright☀ 02:30, 18 October 2017 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Consensus&oldid=805917780"
This content was retrieved from Wikipedia : http://en.wikipedia.org/wiki/Wikipedia_talk:Consensus
This page is based on the copyrighted Wikipedia article "Wikipedia talk:Consensus"; it is used under the Creative Commons Attribution-ShareAlike 3.0 Unported License (CC-BY-SA). You may redistribute it, verbatim or modified, providing that you comply with the terms of the CC-BY-SA