Wiktionary talk:Votes/2021-03/Polarity of voting proposals and application of supermajority rule

Rationale

As further elaborated at Wiktionary:Beer_parlour/2020/December#Operation_of_"supermajority"_voting_rule, the supermajority voting rule is presently problematic in various ways. Its apparent underlying assumption about the polarity of voting proposals is nowhere stated or explained. It may render the outcome of a vote arbitrarily or unintentionally dependent on the "direction" in which a proposal is phrased, including in cases where the phrasing is not to the eye obviously of an unexpected polarity. It may also result in anomalous situations in cases where there is no clear or viable status quo. Mihia (talk) 20:33, 1 March 2021 (UTC)

The proposed provision for no supermajority rule to apply in the case of no clear or viable status quo is intended to address situations such as the following. There are two alternative but incompatible ways of doing something: X and Y. Presently, some people do X, and other people do Y, and it is desired to choose one or the other. Here it may be impossible to say which of "Do X, implying don't do Y" or "Do Y, implying don't do X" is of the "correct" polarity for the supermajority rule to reasonably apply. If there is some better way to handle cases like this, please suggest it. Mihia (talk) 21:01, 1 March 2021 (UTC)

If we were to adopt "supermajority rule doesn't apply" in certain cases, there would also need to be an "official" way to record this, prior to the vote starting. Mihia (talk) 22:19, 1 March 2021 (UTC)

@Mihia, what's the status of this? If you want to run it, you should fix it to start in the future and add it to {{votes}}. If you no longer want to run it, we can delete it. —Μετάknowledge^{discuss/deeds} 23:59, 16 March 2021 (UTC)

@Metaknowledge: I don't understand. When I created it, it was set to start in the future. Now that future is the past, since the review period has expired without comment. Please do whatever procedural thing is necessary to run this in a timely manner after some sort of sanity check. Thanks. Mihia (talk) 01:36, 17 March 2021 (UTC)

@Mihia: Adding it to {{votes}} is not so much procedural as just the bare necessity to ensure that people are aware of the vote existing. Fixing the start date is just something you should do so this vote gets its rightful duration. —Μετάknowledge^{discuss/deeds} 02:55, 17 March 2021 (UTC)

@Metaknowledge: I did flag it at the BP discussion, but this may have had limited visibility. Actually, I think I may have misread your "If you want to run it ..." remark, or been on drugs at the time or something, but there is still a potential issue. Let's look at the text at the top of Wiktionary:Votes. It says "Edit Wiktionary:Votes/Active to add new votes to the “active” list and remove old ones." Is Wiktionary:Votes/Active the same thing as your {{votes}} or something different? The contents presently look the same. It is very easy to read "active" in that instruction as meaning "in progress", i.e. NOT including pending votes. If pending votes need to be manually added to somewhere at the point of creation then this needs to be more clearly stated, which I will do once I am clear on the procedure. Mihia (talk) 18:39, 17 March 2021 (UTC)

@Mihia, it's the same thing, and I just used the shorthand because I thought you already understood that Wiktionary:Votes/Active is what has to be edited, as you've created a vote before. —Μετάknowledge^{discuss/deeds} 19:52, 17 March 2021 (UTC)

@Metaknowledge: Thanks, yes, I think I have just forgotten what to do and/or was misled by that word "active". But, on that point, can you confirm whether votes should be added to Wiktionary:Votes/Active as soon as they are created, even while pending, and hence perhaps not perceived to be "active"? Or should they only be added when they go live and people can vote? I think you earlier implied the former, but perhaps you can just confirm, and then I can try to make the wording clearer. Mihia (talk) 20:19, 17 March 2021 (UTC)

I would say as soon as they're created, so that people will see it and hopefully offer feedback prior to it starting. —Μετάknowledge^{discuss/deeds} 20:22, 17 March 2021 (UTC)

In fact, it does even say "Planned, running and recent votes" at the top of that list. Anyway, I added a clearer instruction to add it straight after creation. Mihia (talk) 18:32, 18 March 2021 (UTC)

Administrator role

Mihia, you, Metaknowledge, and I spoke about this some time ago. I still don't think it's necessary, but since we're here, let's do it. This is what ruins your proposal for me: "The nature of the status quo will often be obvious, but if editors cannot agree on this, then, as a last resort, an uninvolved administrator may determine it. Editors may also agree, or as a last resort an administrator may determine, that there is no clear or viable status quo, in which case the supermajority rule will not apply and a vote will be carried by a simple 50%+ majority. This must be decided and stated before the vote starts."

Why are we codifying even more powers for administrators? Why should we make them better than ordinary users here? This is an egalitarian community. There is, by the way, no such thing as an "uninvolved administrator." Some of them choose to sit on the sidelines but they don't need to. Votes affect us all and administrators, themselves, are people with their own biases. Everything quoted is, in my opinion, superfluous and creates the dangerous possibility that administrators can short-circuit the existing supermajority rule when it suits them. Please remove the problematic sentences and change the last section to the following:

Proposed additional wording

Voting proposals must be worded so that a "support" vote is a vote to change the status quo, while an "oppose" vote is a vote to leave things unchanged. Moreover, the status quo must not be threatened by default. If no one participates in the vote, or if there are only abstention votes, the status quo must remain unaffected.

— Dentonius 11:44, 17 March 2021 (UTC)

Why shouldn't an uninvolved administrator be entrusted to determine a close call as a last resort? Isn't that what is done in most instances where there is no clear outcome? An arbitrator – preferably higher ranking – gets to make a decision. It seems perfectly sensible to me, so I see no reason why Mihia should reword it. --Robbie SWE (talk) 17:16, 17 March 2021 (UTC)

High-ranking? Kind of like, a private compared to a colonel? Let's dispel that notion right away. This is what is done? No, this is not what is done. Community consensus is what is done here. We are all equal here. You, Robbie, as an administrator, have been entrusted with special functions: deleting pages, blocking users, and limited grant permissions. It doesn't mean that you're any different from any other user or that you should be treated any differently. The bureaucrat has the power of one vote. The administrator has the power of one vote. The autoconfirmed user has the power of one vote. What Mihia is proposing will allow administrators to decide which votes can be made easy and which votes can be made harder to succeed. This "last resort" wording means nothing. It's a decision to relegate us -- non-administrators -- to second tier users. I oppose this move vehemently. YES to community consensus. NO to special privileges on voting matters. — Dentonius 17:43, 17 March 2021 (UTC)

Dispel what, the very real and universal notion of a hierarchy?! I've never said that one user is more valuable than another, but your position does not take into account the special roles of admins and bureaucrats, who by the way, got their roles by their fellow peers. We're expected to be level-headed, fair and, most of all, be able to make decisions where there is no clear consensus. Whether that has been the case in every situation, maybe not, but we've always been able to talk it out. I honestly believe that you're making a mountain out of a molehill here – you're once again talking about "us" versus "them" as if you belong to and represent a persecuted and marginalised group of users, bullied by the big bad admins with ulterior motives. Don't go down that rabbit hole again, Dentonius. --Robbie SWE (talk) 19:17, 17 March 2021 (UTC)

You keep referring to things like an “egalitarian community,” “democratic forums,” and everyone’s “right to edit.” But we’re just an Internet dictionary, not a liberal-democratic state. Clearly we’re not truly egalitarian because roles and permissions are stratified; newer users are subject to greater supervision than experienced ones; etc. The same principle goes here: A "higher-ranking" neutral party solves a dispute between opposing opinions. There is little difference between this and the well-established ways that we conduct business at Wiktionary. Imetsia (talk) 18:12, 17 March 2021 (UTC)

• If editors cannot agree on what is the status quo, then what other resolution would there be, other that what I suggested or something similar? A vote to determine what is the status quo? On the other point about "If no one participates in the vote, or if there are only abstention votes, the status quo must remain unaffected", to me, it is blindingly self-evident that nothing changes if there are zero votes, but if this does need stating, I suggest it goes in a separate regulation about minimum levels of participation. For example, if there is only one vote, a support vote, then does the proposal pass? This is reasonable to ask, but, and hopefully for the final time, this is not within the scope of this proposal, which deals with a different matter altogether. Mihia (talk) 20:53, 17 March 2021 (UTC)

Simpler formulation - failed/NC votes have no effect

What about just saying something like "A vote closed as "failed" or "no consensus" has no effect."?

This would still leave open the possibility that someone constructs a vote on what is essentially a no-op proposal. This would be silly but not harmful or exploitable, and would, I hope, be typically caught while the vote is being drafted.

If the status quo is that the community disagrees whether foo is allowed in the case of bar, and a vote on "Do foo in the case of bar" is split 11-10, I don't think that should be taken to mean "Do foo in the case of bar" now has the force of consensus-backed policy. Rather, it should be taken as a signal to try again. Maybe we can at least find consensus for "Do foo in the case of bar, unless baz". Colin M (talk) 21:41, 17 March 2021 (UTC)

"A vote closed as "failed" or "no consensus" has no effect" seems self-evident to me. I suppose it could be stated also, but it does not address any of the issues relevant to this proposal. The point of this proposal is to address unstated assumptions and potential anomalies in the supermajority rule concerning which voting numbers CAUSE a vote to fail or be no consensus. Mihia (talk) 21:56, 17 March 2021 (UTC)

Okay, I guess I'm not understanding. Your original BP thread started with this hypothetical:

Suppose that we presently do X, but this is disputed, and, in fact, 15 people want to stop doing it, while 10 want to continue. Suppose we vote on whether to "Stop doing X". 15 supports, 10 opposes, no consensus, so no change, and we carry on doing it. Correct? On the other hand, suppose we vote on whether to "Continue doing X". 10 supports, 15 opposes, what happens then?

I'm saying the answer to "what happens then?" is "nothing". We remain in the original state (we do X, but it's disputed). Why do you see that outcome as problematic? Colin M (talk) 05:13, 18 March 2021 (UTC)

Brilliant, Colin. I concur. I would have called such a vote an invalid vote. This supposed "clarification of the supermajority rule" (which it is not) is frivolous and is a solution in search of a problem. — Dentonius 06:29, 18 March 2021 (UTC)

The whole point of this proposal is to explicitly make such votes invalid. Maybe you think they already are, but we had one, and nobody noticed. Mihia (talk) 18:59, 18 March 2021 (UTC)

Nobody noticed? I called you out on it. — Dentonius 19:01, 18 March 2021 (UTC)

You did make comments on certain related issues, that's true, though your original complaint was that the status quo would be threatened "in the absence of any participation at all". No (or insufficient) participation is one thing; the direction of the vote is another, and I hope we have now established beyond doubt that the present proposal addresses only the latter. However, after certain wording changes had been made, you agreed "I'm happy now. It looks good to go", even though the vote still contained a proposal that you now say should be invalid. I do tend agree with you one about one thing. This is arguably not so much (or not purely) a "clarification of the supermajority rule", as presently titled, as a general clarification of voting procedures. I may change the title accordingly. Mihia (talk) 20:01, 18 March 2021 (UTC)

(Reply to Colin M) Sorry, I went off slightly on the wrong track with my original response to your comment, and let me also retract what I said earlier about "A vote closed as 'failed' or 'no consensus" has no effect" seeming self-evident. In fact, in my view, in the absence of any written rules about this (that I am aware of), it is far from indisputable that majority opposition to continuing doing something would or should result in no action. I will need to think some more about whether your proposal to codify this would in fact address all the issues that I see as problematic. My feeling is that there probably is an underlying unstated assumption behind the voting system that voting proposals should be proposals to change the status quo, which it would be as well to make explicit. Just one other thing that comes to mind is the issue of mutually exclusive alternatives X and Y, voting on "Do X (implying don't do Y)" vs "Do Y (implying don't do X)". How would your proposal handle this scenario? Mihia (talk) 20:35, 18 March 2021 (UTC)

In fact, let's look at another scenario. Presently we do X, but actually 20 voters want to stop, while 10 want to continue. Your proposal would still allow the vote to be framed with either polarity, essentially "Do X" or "Don't do X". In the former case, 20 oppose, 10 support, vote fails, no action, and we carry on doing X. In the latter case, 20 support, 10 oppose, vote passes and we stop doing it. So, the outcome just depends on how someone phrases the vote, quite possibly even unintentionally in the absence of (as far as I am aware) any existing stated stipulations. With my proposal, the vote would be forced to be of the "Don't do X" polarity. Mihia (talk) 21:00, 18 March 2021 (UTC)

I don't see this example as problematic. If the proposal "do X" fails 10-20, then anyone is welcome to start another vote on "don't do X". It's not guaranteed that this second proposal will even pass. It could be that some of the 20 people who voted against the 1st proposal believe that "X" should be neither universally condoned nor forbidden. Maybe they think we should do X but only in certain narrow circumstances. Maybe they think the decision of whether to do X should be handled on a case-by-case basis. This is why I think it's important not to automatically interpret consensus against some proposal as consensus for some inverted form of that proposal.

These wouldn't necessarily need to happen in serial. Someone could open a single vote page with separate votes on the options "do X" and "don't do X". Colin M (talk) 18:03, 22 March 2021 (UTC)

I Support Colin's proposal as simple and well-thought-out, Oppose Mihia's as instruction creep. MuDavid 栘𩿠 (talk) 03:10, 26 March 2021 (UTC)

Incidentally, has this vote started yet? It's 3 days after it was supposed to. --Droigheann (talk) 13:22, 28 March 2021 (UTC)

Multi-option votes

The problem Mihia is talking about came up in the context of a multi-option vote. When there's a single thing to vote on, it's clear that "fail"/"no consensus" means nothing changes. However, in the "pronunciation spelling" vote, as it was originally written, the conditions were as follows: There were three options. The original premise was that there could only be three possible meanings of what the pronunciation spelling label is and we had to choose which one(s) we wanted to keep. The original wording suggested that if all options were opposed, it would mean that we no longer wanted to use the pronunciation spelling label and we would delete them. I objected to this because I thought it was unfair that a vote which failed would lead to the removal of the labels. In other words, right from the start, the vote had a pre-determined outcome which would change the status quo. I pointed out the wording problem and we came to a compromise on the wording. What I had really wanted, however, was for people to vote on which senses of the pronunciation label should be removed since it seemed that was the intention of the vote: to restrict how people could use it. I don't think there's any vote we can come up with to fix this kind of problem. It just requires vigilance on the part of the community. — Dentonius 11:07, 27 March 2021 (UTC)

To save others some digging, here is the mentioned vote, and here is the initial version with the wording that Dentonius objected to.

My feeling is that an n-option vote should be basically treated as n independent votes. Hence, if all of them fail then (per the section above) nothing should happen. If options X and Y pass, then the effects of both those proposals go into force. (Though it's possible X might be constructed to be strictly stronger than Y, hence making X be the controlling outcome. e.g. Option 1: Allow entries for the names of US presidents. Option 2: Allow entries for the names of world leaders. 3: Allow entries for the names of people with Wikipedia articles. If 1 and 2 both get >=2/3 support, then we do 1 and 2, which in effect just means we do 2.) Colin M (talk) 16:43, 27 March 2021 (UTC)