Wikipedia talk:Arbitration/Policy/Update and ratification/Archive 2

Archive 1

Archive 2

Archive 3

ArbPol2 draft 3

Firstly, thanks to everyone who contributed to the previous discussion. Your input was greatly appreciated.

The new draft is based on the last substantive version from October 2009. Some duplications/repetitions had crept in during the time it was being discussed here and these have been cleaned up. Wherever possible, where legalese was previously used, it has been ditched in favour of clear standard English. The new version also sets out to highlight, as is the case with all Wikipedia policies, the broadstroke principles, with the detail set out in the supplementary procedures section. As ever, comments and suggestions are very welcome,  Roger Davies ^talk 07:33, 29 July 2010 (UTC)

General comment

This is beautifully written. I particularly appreciate the plain-English approach, in broad-brush-strokes. Perhaps one of the keys to controlling the level of detail is the sequestering of procedural details elsewhere; if this is the case, it is a practical and proper way to go about setting down the basic rules and concepts on this page. May I put in a plea that members of the community suggest additions only after considering that length and complexity are the enemy of a wiki. At last, we are on course for a professionally conceived and written policy. Tony (talk) 08:03, 29 July 2010 (UTC)

Thank you,  Roger Davies ^talk 08:14, 29 July 2010 (UTC)

I second Tony's sentiments; this is in general admirably clear and free of the legalese that tends to creep in to Committee proceedings. Commendations, Skomorokh 10:05, 29 July 2010 (UTC)

"Conduct of Arbitrators"

One suggestion: in "Conduct of abitrators", Point 1: "Act with integrity and good faith at all times and maintain the high standards expected of them by the community;

It just seemed a bit circular—the high standards expected by the community seem to be encapsulated neatly in the first nine words. Tony (talk) 08:07, 29 July 2010 (UTC)

Agreed and tweaked. Succinctness is clearer here.  Roger Davies ^talk 08:14, 29 July 2010 (UTC)

Comments on "Duties and responsibilities"

Final decision maker

The first duty/responsibility listed is as follows:

To act as a final binding decision-maker for serious disputes which the community has not resolved;

While this is an improvement over the 2nd draft entry, which read

To act as a binding decision-maker for disputes concerning the conduct of Wikipedia editors

I wonder why the traditional notion of the Committee only involving itself in disputes the community has been unable to handle on its own has been dispensed with. I note there is already a "get-out-of-jail-free card" from requiring prior dispute resolution in the "Procedures" section, so I am not sure of the intent behind this omission. User:FT2's proposed amendment to the second draft would seem most apt:

To act as a final binding decision-maker for serious disputes which the community has otherwise failed to resolve

Skomorokh 10:39, 29 July 2010 (UTC)

It was just a copy-editing thing: "has not resolved" is more concise than "has otherwise failed to resolve". Does "has been unable to resolve" work?  Roger Davies ^talk 10:54, 30 July 2010 (UTC)

Yes that works; the point is whether the community need have failed in some way in order for the committee to get involved (see discussion with Tony1 on this). Sometimes it seems that ArbCom jumps in with a motion when you feel ordinary dispute resolution would be "unproductive" or that your involvement is inevitable – which I think ought generally be discouraged by the policy. Skomorokh 11:03, 30 July 2010 (UTC)

Granter of advanced permissions

The fifth duty/responsibility:

To grant and remove CheckUser and Oversight permissions.

I think this may be overly specific. A "constitution"-like document like this should generally speak in principle rather than rule, so as to avoid obstructing unforeseen developments. I would change this clause to stress "advanced permissions/userrights" or "access to non-public data". There are other conceivable positions that CheckUser and Oversight that the Committee may need to oversee; existing examples would be election administrators for SecurePoll who have access to CheckUser-like information, and the privilege to see number of page watchers. One cannot anticipate what permissions the community/developers/toolserver might dream up, and it would be less than ideal to have to have an amendment referendum for non-controversial additions. Skomorokh 10:39, 29 July 2010 (UTC)

Good point (though I suppose it could be handled with a redirect if the role title changed).  Roger Davies ^talk 10:54, 30 July 2010 (UTC)

Just a side point, The Arbitration Commitiee has no jurisdiction over the toolserver, its programs or users, as that is a foundation entitiy owned by Wikimedia DE. And uhh, On a second thought, it would also be proper for arbitrators contesting re-election to abstain from any involvement in Securepoll admins.   «l| Promethean ™|l»  (talk) 10:22, 3 August 2010 (UTC)

Thoughts

Yes, thanks for this draft, which is written in clear and succinct language. I still see contradictions between the concepts, though.

We start off by saying ArbCom serves: "1. To act as a final binding decision-maker for serious disputes which the community has not resolved." That sounds great. But is it really true? I always thought ArbCom shied away from actually deciding disputes - rather, most of the time it looks at editors' conduct, hands out various blocks and restrictions on editors whose conduct is felt to be a problem, perhaps lays down certain ground rules for editing in particular areas - it aims to facilitate the resolving of the dispute rather than actually "decide" it. (Of course, the sentence doesn't literally say that the committee will decide the disputes, but that's how anyone would read it.)

Then we say "consider appeals from blocked...users" - don't admins consider such appeals in the first instance? Shouldn't it be "final appeals" or some such?

Then later: "The Committee has jurisdiction over all disputes on the English Wikipedia concerning user conduct or users who hold advanced permissions." So it turns out that it can only consider disputes about conduct and use of permissions, not about content - which is closer to the truth, perhaps, but is inconsistent with point 1, which talks about (all) "serious" disputes (and if content disputes are not serious, I don't know what is).

Then we start talking about "parties", without defining what that means, and we have the line "3."Remedies" (binding rulings about the parties' future conduct);", which implies that the only decisions ArbCom can make concern this undefined but apparently limited group of people called "parties" - which I don't think is true, because sometimes it makes decisions which all editors are expected to observe.

Just my immediate thoughts; sorry I don't have time to write more or better at the moment.--Kotniski (talk) 12:35, 29 July 2010 (UTC)

I suppose the broad unstated principle guiding ArbCom is that policy is the best way to solve disputes. What often impedes this is editor (mis)conduct, which has the effect of driving less entrenched / less obstinate / less POV-pushing / less uncivil editors away. So if you tackle the conduct and bring the excesses to an end, the underlying issues can be often (but not always) be resolved by discussion and consensus.  Roger Davies ^talk 11:13, 30 July 2010 (UTC)

Point raised is pertinent and accurate: rather than decide matters, ArbCom often (i) delays a decision by imposing remedies that squelch problematic behavior, regardless of the underlying issues, for a limited time, with the idea that things will simply die down as tempers cool, or by the withdrawal of some parties (desirable or not) or (ii) imposes a remedy that requires some "uninvolved administrator" to make their own judgment in response to any future behavior and impose a "reasonable" remedy on their own (which then cannot be overturned, by virtue of recent rulings). These actions are not an execution of ArbCom responsibilities but avoidance of them. They are, however, considered actions, which suggests that many of the matters brought before ArbCom are, in fact, unsuitable for this process and so are being ducked. That suggests the matters brought to ArbCom should be limited to those issues they actually feel comfortable with deciding, and other matters should be handled differently. In particular, vaguely formulated problems that require a lot of fact finding and crystal ball gazing (what's going on here?) should be forced to go through a distillation process before they reach ArbCom. Brews ohare (talk) 16:05, 17 August 2010 (UTC)

Intent

I have a question on the intent behind this draft. I have seen comments before to the effect that "policy is what we do", and that policy documents should be changed to match standard practice. So, is this proposal meant to document what ArbCom does, or to lay down what ArbCom should do? It seems to me that different comments might be warranted, depending on the purpose of the draft. Also, is the intent for the content of the policy to follow the views of present ArbCom members or are the views of the community (assuming those can be determined in some way) supposed to have primacy in considerations of policy content? EdChem (talk) 15:54, 29 July 2010 (UTC)

The underlying intent, in some sense, is to cause "what ArbCom does" and "what ArbCom should be doing" to converge.

The arbitration policy has always been an unusual case. It has, for many years, required substantial updates to reflect the the way ArbCom actually functions; but the method for updating the document has been unclear from the start, since it was created outside of normal community channels, and since the Committee, while nominally empowered to change it, has never actually done so. In some ways, also, the arbitration policy is viewed as a pseudo-constitutional document, since people want it to govern the Committee's role vis-à-vis the community and the project (witness the regular debates about whether the Committee is overstepping its remit). The revised policy is thus intended to document the current role and authority of the Committee, as well as to bind the Committee to that particular role until the community explicitly decides to alter it.

As far as community primacy is concerned: yes, the community is meant to have primacy in determining the content of this policy. However, given ArbCom's role, it is obviously important to ensure that changes are made in keeping with the wishes of the community in a broad sense, rather than being primarily a product of a small group of interested individuals (as they tend to be on many other policy pages). Thus, the policy provides for community modification—but through a formal process for amending the policy, rather than by simply allowing anyone to change it as they please. Kirill ^{[talk] [prof]} 04:14, 30 July 2010 (UTC)

So has this document been written with the aim of describing what ArbCom has done up to now, or with the intention of changing anything in what it does or how it does it? See my objections just above, which no-one has responded to yet - in order to know which way to go to eliminate the apparent contradictions in the policy, we need to know if we are just supposed to be descriptive of existing practice (e.g. alter the sentence that implies that ArbCom will decide all serious disputes, to say that it won't generally decide actual disputes over content), or if ArbCom is proposing modifying what it does (e.g. to decide a wider range of disputes).--Kotniski (talk) 10:29, 30 July 2010 (UTC)

ArbCom will resolve the conduct issues arising out of content disputes, will occasionally mandate ways of resolving a content issue (RFC, for example); but will not impose a content solution itself. After all, a content dispute is not usually a problem until it becomes a behavioural issue. Anyway, to clarify, I've added "behavourial" into serious disputes.  Roger Davies ^talk 11:28, 30 July 2010 (UTC)

All right, but I hope you know what you're saying and you really mean it - now ArbCom's scope is explicitly limited to deciding disputes about people's behaviour, nothing more (well, apart from the matters listed in 2-5).--Kotniski (talk) 12:42, 30 July 2010 (UTC)

Private evidence

In regard to the part related to private evidence and the provision to opt out of circulating it:

"Evidence may be submitted privately, but the Committee will normally expect evidence to be posted publicly unless there are compelling reasons not to do so. The Committee will make every effort to circulate privately submitted evidence to all parties, except when doing so would result in undue risk of harassment or retaliation against the editors providing that evidence."

Isn't the risk of "harassment or retaliation against the editors providing that evidence" some what mitigated by the implicit threat of being instantly perma-banned should it be shown that such harassment or retaliation had taken place? Now of course that evidence may include personal private information but I really can't see any reason why at the very least a redacted version of the privately submitted evidence cannot be circulated to all parties. The wording implies that the Committee is not even obligated to inform the involved parties of the existence of such evidence should it decide not to circulate it , i.e. it remains secret.

It is a sad and ancient fact of life that people do bear false witness, and without the opportunity to challenge and refute it would really be a travesty of justice. --Martin (talk) 18:29, 30 July 2010 (UTC)

In fact, exactly this point is covered in the Private Hearings section above: "In exceptional circumstances, typically where significant privacy, harassment or legal issues are involved, the Committee may resolve the matter in private. The parties will be notified of the private hearing and be given a reasonable opportunity to respond to what is said about them before a decision is made."

More generally, ArbCom very rarely relies on private evidence for public hearings (I cannot recall a single instance during my stint as an arbitrator) and on the few occasions where I have seen it in older cases the existence of the evidence has been alluded to in the applicable Finding of Fact.  Roger Davies ^talk 19:29, 30 July 2010 (UTC)

I've pulled this around a bit here , to tighten this up along the lines you suggest.  Roger Davies ^talk 07:11, 31 July 2010 (UTC)

And tweaked  for flow and clarity.  Roger Davies ^talk 07:30, 31 July 2010 (UTC)

Yes, it reads more clearly now that you have linked it to the private hearing process. --Martin (talk) 08:27, 31 July 2010 (UTC)

Arbcom and content

I guess I would have liked to have seen (hey, two perfect tenses in the one sentence :)) something along the lines of:

Although the arbitration committee does not rule on content, it can propose a pathway by which means of resolving a content dispute can be facilitated.

Given the arbitration committee is also a problem solver, I saw this as a positive when the committee has recommended courses of action in the Ireland and West Bank/J&S cases. I'd link but someone is nagging me to get off the computer...Casliber (talk · contribs) 03:55, 1 August 2010 (UTC)

PS: Agree it is nice, succinct and easy to follow. Casliber (talk · contribs) 03:56, 1 August 2010 (UTC)

A diff link

It would be nice if there was a way to a get a diff style link to show the difference between the current active policy and the current draft one. Not necessarily showing wiki code differences, but text differences.--Rockfang (talk) 05:03, 1 August 2010 (UTC)

It would be good, wouldn't it? Unfortunately, the software doesn't handle extensive restructuring very well. What I have just done though is to dump both WP:ARBPOL and the latest draft into my sandbox which gives you a diff of sorts. I hope this helps,  Roger Davies ^talk 06:54, 1 August 2010 (UTC)

It doesn't, but thank you for the attempt.--Rockfang (talk) 07:48, 1 August 2010 (UTC)

Do NOT give ArbCom executive power

I very strongly oppose this proposal. It would allow ArbCom "to resolve matters unsuitable for public discussion because of privacy, legal, or similar concerns". This sounds like a direct reference to the dispute regarding WP:Pedophilia (now renamed WP:Child protection), a policy enacted without community support. During discussions over the past two months, several problems with that policy were fixed, but others remain (such as a threat to ban editors who point out that someone has self-identified as a pedophile on ANI or some other forum). I think that the community can write a better policy by the usual process of consensus, and can enforce it more effectively and fairly and with less drama by discussing individual cases in public, transparent forums. After all, this is what they've been doing on Wikimedia Commons, with far better results.

More generally, allowing ArbCom to "resolve matters" would create a toxic combination of secret rules and secret enforcement, combining all judicial and executive authority into a small group of people. The new version repeals "The Committee will hear or not hear disputes according to the wishes of the community, where there is a consensus" and replaces it with "jurisdiction over all disputes", meaning that the community would be stripped of the legislative power to override this. Holding all the powers, the resulting ArbCom could promulgate any policy without regard for existing Wikimedia principles - in fact, early in the WP:Pedophilia discussion, several people claimed it was responsible for creating the policy in that case, until it was clarified that ArbCom has no such power and was being promulgated by Jimbo Wales under WP:CONEXCEPT.

Now it is one thing to allow Jimbo Wales special "monarchist" powers over the English Wikipedia, because by founding it he showed himself generally trustworthy. But it is another thing entirely to risk that after the next election (or sooner) ArbCom might strike out radically on a campaign to purge Wikipedia of whatever it disagrees with, or placing it under the thrall of some nation, political party, or corporate entity which happens to gain control over a majority of its members.

Enacting this policy means abandoning the attempt at a broad democracy where everyone has a say on policies, and turning to a system where your vote in the ArbCom elections will be the only say you have on Wikipedia policy. Too bad you don't know who the candidates really represent.

Now one might argue that not all cases or policy decisions are "unsuitable for public discussion", but let's be realistic: there's always some argument against public discussion, and WHO is going to contradict ArbCom when it decides so? Being from U.S., which has been under a "state of emergency" since 1933, I know that things like this can be abused... and far worse has come from enabling acts in other countries.

Some additional issues from a quick reading:

• WP:Oversight currently describes a community election to ratify oversighters, which appears to be removed by this draft - a power that also greatly lends itself to secrecy and abuse, especially now that the protections of the old GFDL license have been repealed.

• Is there any guarantee a case would ever be closed? I see that "An arbitrator whose term expires while a case is pending may remain active on that case until its conclusion" and "At any time between the opening of a case and its closure, the Committee may issue temporary injunctions, restricting the conduct of the parties, or users generally, for the duration of the case." To me this suggests that an old ArbCom administration might try to hold on and maintain its policies even after losing an election, creating an (even greater) constitutional crisis.

Finally, let's bear in mind: if Wikipedia policies are really best created by a tiny group of "clued-in" people, rather than by a broad consensus of editors, then shouldn't Wikipedia articles be created the same way? If you really believe in this policy, then you should accept that Wikipedia is just a plain bad idea from beginning to end. Wnt (talk) 05:20, 1 August 2010 (UTC)

To my knowledge, there have been only two cases heard "in camera" in the history of the Arbitration Committee: One which involved real-world harassment of an editor and also involved email (either of which would have prevented public discussion), and the other revolving around checkuser data (which is covered by the WMF privacy policy). A third case, which involved a mailing list, was largely public, although the emails themselves (to which all of the parties had access) were not made public as they contained considerable personal non-public information; another recent case that included allegations of sockpuppetry was also public, with results of the checkuser investigation being described in as much detail as possible without violating the privacy policy. The WMF policy on Oversight and Checkuser explicitly places appointments to these two positions into the realm of the Arbitration Committee. The draft that is being discussed here is entirely based on existing practices that have simply not been documented.

I'm at a loss as to your suggestion that cases will never close. They all close eventually; the proviso that arbitrators can complete a case they've already started is so that when the new office-holders come in at the change of term, they don't have to immediately jump into cases that have already been running for several weeks, and render decisions; meanwhile, those who are already current on them finish those cases (if they wish), act only on those cases, and leave immediately after they close. Note that the "new" arbitrators may choose to become active on cases that carry over from one term to the next, and many of them do so. For the past few years, and into the foreseeable future, there will be at least as many new arbitrators appointed as there will be "old" ones remaining; the current two-year term means that half the committee will change every year, and appointments will also be made to fill any seats vacated during the term. Further, the Arbitration Committee has always had the power to issue injunctions during the course of a case; it is in the current policy, which is badly outdated.

The draft you are currently reviewing is the third one, written after broad community consultation on the first two drafts that were carried out last year, after an extensive and wide-ranging RFC on the Arbitration Committee carried out in 2008. It more accurately describes the practices of the Arbitration Committee than the current policy, which has been out of date since its last significant amendment in 2005. The incoming "class" of arbitrators in 2009 resolved to incorporate as many community-originated ideas from the 2008 RFC as was possible, including the revision of this policy, and this work has continued with the 2010 term. This particular draft of the Arbitration policy is the culmination of the input of a few hundred community members over time, and after this final review it will proceed to a ratification vote by any and all interested community members. There are no new powers here. Risker (talk) 06:18, 1 August 2010 (UTC)

Per Risker. Tony (talk) 07:12, 1 August 2010 (UTC)

Indeed. The current policy draft is really an attempt to codify existing policy and the community may well feel it is better to have it codified and limited rather than continue uncodified, with the risk of disputes about exactly how extensive the powers involved actually are. Arbcom is the only body which can take up 'matters unsuitable for public discussion because of privacy, legal, or similar concerns' and so once it is accepted that there are such matters, Arbcom has to deal with them. And in practice Arbcom is very reluctant to do anything in private which would involve the making of a new policy, because any changes have to be accepted by broad community consensus in the end. Sam Blacketer (talk) 10:27, 1 August 2010 (UTC)

I have to question Risker's statement that only two cases were heard in private, because the WP:Child protection policy describes a whole class of cases that must only be heard in private, and the site Wikisposure ([1]) claims a long list of people banned by this process, who are in fact listed as indefinitely banned at Wikipedia:List of banned users. How can this be?

More and more I am hearing that Wikipedia policies don't actually set the rules, they only describe them with varying degrees of accuracy. In that case, what is the point of having a vote to ratify them? Whether you vote for it or against it, it's equally valid as a description of current practices! And all the debates on all the policy pages would be meaningless.

But if the purpose of putting a draft up for a vote is not just to describe current practices but to set policy and limit ArbCom, then it should set good policy and create real limits. If ArbCom can't make policy, then it should say they can't make policy; or if they can't make policy unless a broad community consensus accepts it (which is about the same thing, since any editor can do as much) then the policy should say that. Reasonable what-if scenarios need to be anticipated, and vague terms like the "privacy, legal, or similar concerns" need to be nailed down very precisely. Even so, I don't see why these things have to be centralized under ArbCom rather than left to office actions, OTRS, or some other group. In "real life" I think people would recognize that merging the Supreme Court with the CIA would be a bad idea. Wnt (talk) 16:15, 1 August 2010 (UTC)

Wnt, why do you think it is better that unelected, unidentified people handle these concerns rather than community-elected, known people? Your arguments seem to be pressing for greater accountability, but I'm hard pressed to see how anything you're suggesting would provide that. Risker (talk) 16:34, 1 August 2010 (UTC)

Actions by the WMF are controlled by their elected board of directors, who (as far as I know) so far have not seemed to rule with a heavy hand, partly due (I think) to specific DMCA legal guarantees that make them unwilling to wade in unless they really have a reason. But avoiding centralization of power can be beneficial even when unelected officials are involved - for example, the U.S. Supreme Court is unelected, yet we're better off with that than allowing Congress to pass a law and rule it constitutional on the same day. Wnt (talk) 18:37, 1 August 2010 (UTC)

Well, that is the WMF Board, and it has little to do with the Arbitration Committee on any specific project. Arbitration Committees don't report to the Board, nor do the individual communities. Only WMF employees, through the Executive Director, have a reporting structure that ends at the Board of Directors. So, who exactly are you proposing to be the final step in dispute resolution on this project? The Board of Directors isn't going to do it, nor are WMF staff. And this is a website, not a nation, so comparisons with the US Supreme Court are specious at best, although I will point out that they are all known by name and go through a rigorous vetting process including vetting by elected representatives. Risker (talk) 18:53, 1 August 2010 (UTC)

I'm saying that if Wikipedia is actually doing something illegal, the WMF will step in to stop it. But if something is not worthy of an office action, it should not be left to a secret resolution by ArbCom for "privacy, legal, or other reasons", possibly according to secret or partially secret procedures. It should be handled in the open according to policies generated by the community (or by Jimbo Wales, and only Jimbo Wales, via CONEXCEPT). You've said that you've only seen two cases handled in camera - if that was really needed only in those two cases, couldn't you have called WMF and requested a formal office action each time? Due to unfortunate weaknesses in American free speech protections, Wikipedia might not be able to be fully as open as it should, but there doesn't have to be layer after layer of secret processes and redactions - one such process should be enough. And since I doubt WMF is going to give up office actions I think it's your new draft policy that ought to yield. Wnt (talk) 19:03, 1 August 2010 (UTC)

I think you're conflating unrelated matters. In both instances, the private proceedings were because of private information that was central to the case.You cannot have a public discussion of that information without doing more harm to the victim. What you're suggesting would be to tie the ArbCom's hands behind its back in those cases, and force them to remit to the WMF. Those cases did not require office actions, so you're basically wanting to give the WMF more authority over Wikipedia than it already has. Which really means you're just shifting the private proceedings from one group (ArbCom) to another (WMF). What does this gain us, really? You'll still have no access to the proceedings, it won't be open at all, so you won't gain any more openness by doing this. — The Hand That Feeds You:^Bite 16:52, 2 August 2010 (UTC)

The WMF runs Wikipedia. I think it's better to trust all secret cases to one group than to split them half and half and have two different bodies secretly removing information and giving orders without community discussion. I reiterate that WP:Child protection, originally described as an "ArbCom policy" and introduced without community support, currently calls for secret proceedings in many more cases - according to it, a Soccer Mom who complains at ANI that someone self-identified as a pedophile in a public edit can be blocked for not going directly to ArbCom, and her allegations can be oversighted. I know Jimbo Wales imposed an anti-pedophile policy by fiat but I don't think all that stuff should have to be part of it. Wnt (talk) 00:42, 3 August 2010 (UTC)

Given the appalling real-life consequences an accusation of pedophilia may have, why do you think it's appropriate to make them in public on one of the world's most visited websites?  Roger Davies ^talk 16:44, 3 August 2010 (UTC)

First, because all the edits that would either self-identify such a predilection or hint at a POV-pushing pattern are already public. Not only are they already public, they're doubtless already logged and archived by this mysterious organization "Wikisposure" which gives the impression that it has more involvement in the current process than the Wikimedia community. And when that organization lays false charges - for example, against a certain WMF employee a few months ago - it ends up being heard here anyway. Now if you hold a public debate, then all the diffs are gone over, the meaning is debated, and it should become clear how much is fact and how much is malicious interpretation. Wnt (talk) 18:15, 3 August 2010 (UTC)

Before you propose that the WMF take over this responsibility, perhaps you should ask them if they are interested and/or willing to do so at either the Board or staff level. This project does not have the right to demand that they accept responsibility for solving all of our problems. As you're the one proposing this idea, I will leave it to you to obtain their opinion on the matter, but will look forward to you sharing their response. Risker (talk) 01:09, 3 August 2010 (UTC)

I don't have to ask them. I've read meta:Office, which makes it very clear that even people who have never edited Wikipedia are welcome to call them up and ask for an office action. Those two extra cases you mentioned shouldn't burden them overly much. Wnt (talk) 04:08, 3 August 2010 (UTC)

Office actions are about deleting articles and edits, not the activities we are discussing here.  Roger Davies ^talk 16:44, 3 August 2010 (UTC)

What is there to Wikipedia besides articles (general sense) and edits? Office actions don't just delete those - they come with a solid promise of a block for people who reinsert the information. The only other things I can think of are the Wikipedia e-mails, which opens a can of worms, since you couldn't know an alleged statement in an e-mail were true unless you were going through logged copies of them... do you mean e-mails? Wnt (talk) 18:15, 3 August 2010 (UTC)

Meta - It is untrue that the Pedophillia policy has never been supported or approved by the community. This has come up 2 times fairly actively for discussion; the first time, a pseudo-vote on the policy was held, with significant (nearly 100) responses and 80% plus support for the policy. The second time there was much vocal opposition by about 5-6 die-hard opponents, who were offered a re-run of the poll but declined doing so.

That we have not had a formal policy approval for this does not mean that it's not what the community wants. It's exceptional in some ways - but it's not exceptional in the sense of being something that there is any evidence that the community writ large opposes and would strike down if given the chance.

Georgewilliamherbert (talk) 00:55, 7 August 2010 (UTC)

Note: I didn't say that the community would strike down the policy entirely. I said that if allowed to change the policy for themselves, they would write a better policy, more clearly linked to other Wikipedia policies against harassment and intimidation, and stripped of bizarre provisions like blocking people who publicly discuss that an editor has self-identified as a pedophile or seems to be pushing such an agenda. I understand the community is required to defer to Jimbo Wales on this point anyway (though I wonder if his "zero tolerance policy toward pedophilia" really extended to this latter point) but I don't think the community should ever defer to ArbCom on this sort of policy-making, because Jimbo will still be Jimbo next year, but who knows who will get elected to ArbCom and whether they will behave as well as the voters hope? Wnt (talk) 15:06, 8 August 2010 (UTC)

Three thoughts

• I think the statement that appeals can be made to Jimbo Wales needs to be struck. Frankly, I believe his history demonstrates that he is not very concerned with acting within policy or the community's wishes, and as such, I do not trust him to serve as the final appeal in the process. I would rather such appeals went to the WMF board, or if necessary, back to ArbCom itself.

• I would like to see a statement indicating that community consensus may overrule an ArbCom decision. ArbCom exists to serve the community, and given its intended role is to resolve disputes the community can not, I would recommend it be codified that the community ultimately stands above ArbCom.

• Related to above, I would also like to see an explicit statement that ArbCom does not exist to govern Wikipedia, and as such does not have power to create or modify policy by fiat. Nor does it have the power to retroactively justify actions taken against policy. Resolute 17:22, 1 August 2010 (UTC)

In response to your three points:

• Appeal back to the Committee is already noted in the policy; appeal to the WMF is almost certainly not an option for legal reasons. Whether the possibility of clemency from Jimmy is to be retained may be a slightly academic question; I cannot recall a single instance in the last five years when such an appeal was successfully made.
• While I generally support community primacy over the Committee in principle, I don't think we have a practical means available for gauging "community consensus" on a level necessary to overrule a Committee decision. I don't see any particular reason, for example, why the opinions of a dozen editors on AN/I—which are considered "community consensus" for something like the average ban discussion—should be given priority over a Committee elected by hundreds of editors. If we wanted a community process for overruling a Committee decision, it would need to be rigorous enough and broad enough to ensure a similar level of participation, and I'm not entirely sure of what a good structure for it might be; and instituting something like it would probably be a broader matter of democratizing community governance rather than something specific to arbitration in any case.
• The first part of your request is already present in the policy:

  The arbitration process is not a vehicle for creating new policy by fiat. The Committee's decisions may interpret existing policy and guidelines, recognize and call attention to standards of user conduct, or create procedures through which policy may be enforced.

  The second is rather more problematic, however, since it poses the question of how one would distinguish between "the power to retroactively justify actions taken against policy" (which you want to prohibit) and "the power to determine whether actions are or are not against policy" (which is inherent to the purpose of the arbitration process). This may be something only feasible through some broader community oversight mechanism (as in your second point).

Kirill ^{[talk] [prof]} 17:53, 1 August 2010 (UTC)

I don't think it is a fair comparison to make between the number of votes to elect an ArbCom member who later announces a decision, and the number of votes supporting that decision directly. I think that the opinions of the community often become rather clear on certain issues - by the time you have two or three archives of debate and have spawned RfC's and Village Pump discussions, a lot of people have heard and weighed in on them. This level of discussion is reached not just for major policies but even for such apparent minutiae as whether the LGBT WikiProject can place its tag on a skater's biography talk page. Wnt (talk) 18:52, 1 August 2010 (UTC)

Well, by the same token, the volume of material produced in a debate does not necessarily correlate to anything other than the ability of certain editors to keep talking at length. Even lengthy discussions on policy pages will usually revolve around a dozen core participants, and perhaps another dozen or two passerby that happen to chime in; only the largest and most heavily publicized debates (e.g. those listed here) garner levels of participation could possibly be considered participation by the community "as a whole", rather than by a small group of interested individuals.

So, yes, there could potentially be a process for the community to rule on such a matter directly; but it would need to be done on the level of those debates, and would need to be constructed so as to allow for an impartial presentation of the matter to the community. I'm not actually opposed to having a system of community referendums to decide issues of this sort, mind you; I simply don't think we're ready to institute such a process yet. Kirill ^{[talk] [prof]} 19:19, 1 August 2010 (UTC)

• • I have to agree with all three points. I think as a founding priciple the notion that the community can override ArbCom is important. We don't need to specify how, just that it can in theory be done. Hobit (talk) 23:50, 1 August 2010 (UTC)

But, Hobit, unless the "how" is specified, and specified in some detail, such a principle would be relentlessly and endlessly gamed on every forum on the site. It would remove the closure which an arbitration case brings. Worst still, it would provide an incentive for the warring parties to unite briefly to have the decision overturned (and if they're all topic-banned, for example, they may well have nothing better to do) before returning to their previous antics. Plus, there's the very real problem of making informed decisions: the evidence and submissions in the Scientology case last year involved dozens of editors and ran to the length of a novel. Who realistically has the time to review that in detail? And what would happen with deleted or suppressed edits/articles, which are not generally acessible? It's fine to talk about this in principle – and, in principle, I support it – but the practicalities probably make it unrealistic.  Roger Davies ^talk 00:25, 2 August 2010 (UTC)

If we continue the judicial metaphor, the community may not need to go over the facts of the case, just as appeals courts typically focus on procedural issues. I think that in most plausible scenarios, community discussion would determine that a logical argument or procedure used to reach the decision was flawed, and require ArbCom to go back and revise its decision on that basis. Wnt (talk) 01:11, 2 August 2010 (UTC)

But community discussion on what scale, and with what participants? Would the sanctioned parties be part of the consensus forming process? How do you deal with coordinated editing (cf. EEML and Scientology)? What about sockpuppets and suspiciously process-familiar newbies? What about block-voters? I have difficulty seeing any way of ironing out these wrinkles.  Roger Davies ^talk 01:58, 2 August 2010 (UTC)

According to the RfA I read the Scientology case looks like it had about 50 users involved total. I don't deny that I would worry that a well-organized group could influence either a well-attended policy vote with 200 support votes —— or an ArbCom election with about the same number of people voting. But here we're discussing a political dispute in which the majority of ArbCom is opposing the hypothetical vote-stackers. Such well-known editors, citing what they feel is sound logic, would probably canvass this site like it has never been canvassed before. There would be site notices, Village Pump and Jimbo Wales and RfC threads all going at the same time, if not before then certainly after the vote started to look like it was going against the ArbCom position. I really don't think ArbCom could end up losing unless a real majority of all Wikipedians had a strong gut feeling that they'd made a bad decision. Wnt (talk) 18:28, 3 August 2010 (UTC)

I agree that the how is important. And I especially agree that a vote by a few of the ANI All-Stars falls well below that bar. I would argue that an RFC or some other centralized discussion that produces a clear and broad consensus should have the ability to overrule an ArbCom decision. In fact, I expect that such reversals would be very rare. I argue it mostly in principle as the community should stand above anything other than the WMF itself, and would like to see this codified. It should not be necessary for the community to justify to ArbCom if it chooses to change the terms of a judgment, though defining the acceptable process would be problematic, I agree. Resolute 01:14, 2 August 2010 (UTC)

As I mentioned earlier, I agree entirely with the principle of the supremacy of the community: I simply cannot see a way of applying it to ArbCom decisions that would bring swift closure and/or not be susceptible to gaming.  Roger Davies ^talk 02:09, 2 August 2010 (UTC)

An example of this would be the BLP controversy. My comment on rewriting policy by fiat was referring directly to that mess (and I did miss that statement in the draft policy, thank you for highlighting it, Kirill). While the infamous ArbCom motion deemed the disruption the initial deletions caused to be a "reasonable exercise of administrative discretion", I would argue that the resulting RfC and it's overwhelming rejection of such deletions as a valid solution to this problem overrules the ArbCom motion. The community has deemed that this is not an acceptable solution, and as such any future disruption that uses that motion as a justification should be rejected as invalid. Resolute 01:21, 2 August 2010 (UTC)

Well, given that example, I'm no longer entirely certain what you mean by "overturn". The community decided to proceed in a different manner from the one approved by the motion, certainly; but that is not necessarily the same thing as saying that the reasoning of the motion was incorrect at the time it was made, nor as saying that the Committee was not permitted to pass the motion at all. I have no particular problem with having the community do something other than what the Committee has decided—provided that we have a reasonable process to determine whether that is, indeed, the will of the community—but I am quite loath, for obvious reasons, to set up a system where the community's decision to do something different retroactively makes the original judgment itself a violation of policy. Kirill ^{[talk] [prof]} 02:04, 2 August 2010 (UTC)

Well, I think the original judgment was outside policy, but we don't really need to rehash that over yet again as what's done is done. I used it as an example as it was the most obvious one I could think of where the community - and after a significant discussion at that - felt differently than ArbCom did. You are correct that this isn't an example of overturning an ArbCom decision itself, since the RfC was not phrased that way and certainly was not intended to do so. However, in this case, using the motion to justify future disruption along such lines would be invalidated by the subsequent community decision. Resolute 02:26, 2 August 2010 (UTC)

(e/c) That's one interpretation but not mine. The purpose of the motion was to kick the overall BLP problem, which had been a festering sore for some time, back to the comunity for the community to resolve. At the time the motion was written, it was accurate. Policy has since changed and is clearer. I don't see that as a slap in the face for ArbCom but as a good result. Your mileage clearly varies :)  Roger Davies ^talk 02:09, 2 August 2010 (UTC)

And to respond to the Jimmy Wales question, Kirill, whether it is academic or not, I am not in support of deifying Jimmy. He should be viewed in the context of being a WMF board member and/or an editor, and nothing more. If appeals to the WMF are problematic for legal reasons, then appeals to Jimmy should be avoided for the same reason. Resolute 01:25, 2 August 2010 (UTC)

I agree with the idea of rewriting the policy to exclude appeal to Jimbo. Frankly, I think the Godking concept has been outdated at Wikipedia for many years, and actually getting it acknowledged as such would be a good step for our community. Though this issue may be outside the scope of this policy rewrite and be something for the community to deal with as it comes to the role of Jimbo in general. Heimstern Läufer (talk) 07:13, 2 August 2010 (UTC)

• • • We either need to clearly and in gross detail spell out how the community can override ArbCom, or we just say it can and move along. If we need an RfC, what % is required? How many participants for how long? I think you're best off just saying they can and move along. I've no real objection to getting down to the nitty gritty, but I don't think it serves us here (or any set of details would have consensus). I think the idea that the community can override the Committee is important and needs to be here. Hobit (talk) 04:46, 2 August 2010 (UTC)

Uh, I think it's kind of the other way around. We either need to clearly or in gross detail explain it or we need to not say it can (which is not exactly the same as saying it can't). We can't just put that wording in without details; Roger's absolutely right that that would just open the situation to gaming. In theory, I approve of the idea of community override of ArbCom; in practice, I simply don't see it working. Even RFCs usually attract such a small portion of the community that it can hardly be considered a sufficient consensus to override something like this: I think the standard, if I dare bring numbers into the equation, would have to be something like at least a majority, and probably more, of all active editors. That just can't feasibly be attained. Heimstern Läufer (talk) 07:05, 2 August 2010 (UTC)

I think this would be an entirely unfair standard. Bear in mind that an ArbCom member can be elected to a three year term by a vote of 306 to 111, or a one year term by a vote of 154 to 87.[2] Since I and some other editors are now beginning to understand that these votes may be our one chance to have a say in Wikipedia policy, I suppose these numbers will go up, but they are not a majority of all editors. Wnt (talk) 11:59, 2 August 2010 (UTC)

I am adamantly opposed to a usual RFC being sufficient to overturn an AC ruling. The RFC crowd is a self-selecting group. A real, genuine sample of the community can probably never be achieved at the level needed for overrule binding arbitration, which is, after all, supposed to be binding. If not a majority of all active editors, it must be something really close. Heimstern Läufer (talk) 16:00, 2 August 2010 (UTC)

In my mind, I don't see a normal course RfC as being enough either. After all, we do appoint ArbCom to do a job that is mostly thankless and it is not my intent to undermine the committee. I am not sure at what point we as a community could safely say that we will set aside an ArbCom decision - but I would suggest that the times when it could happen are so rare that we don't need to spell out a formal policy on it. Chances are it would happen only as a result of a very public ... response ... to a decision such that it receives widespread input. More or less, my intent in my comments is to reverse the positions of Jimmy Wales and the community in this policy. As it reads now, ArbCom effectively stands subservient to Jimmy, but above the community. It should be the other way around, even if only in principle. Resolute 16:17, 2 August 2010 (UTC)

In regard to how to determine "community consensus" on a level necessary to overrule a Committee decision, given that it falls upon admins to enforce Committee decisions, why not simply ask the admin corps to vote on any contentious Committee decision. Given that those applying for admin membership is subjected to scutiny on admission, the likelihood of EEML style gaming is virtually nil, particularly if the vote threshold to overrule is set sufficiently high, say a simple majority of active admins. You would just need to work out how to certify a particular Committee decision for such a vote. That kind of governance works in the real world, what makes Wikipedia any different? --Martin (talk) 04:33, 2 August 2010 (UTC)

I really doubt that half of the active admins are going to bother to vote at all, and if they do, they're not likely to have read the rationales for the findings they're voting on that well. It's also just not really what the admin corps signed up for (despite all the admins who live in the drama, there are those who really did sign up basically as janitors). Heimstern Läufer (talk) 07:09, 2 August 2010 (UTC)

Martin, Hobit, et al.: It is seriously impractical to talk of the overriding of ArbCom decisions by some community process. We elect the arbitrators – more of them, more often, nowadays – and that election embodies the trust of the community. You are welcome to probe individual candidates on their talk pages about their attitudes to cases past and hypothetical. And there is ample scope for outside opinions to be expressed during specific cases. Let the arbitrators do their job, please.

Along the same lines, on the matter of appeals to Jimbo, I do believe this wording should be removed: "Remedies and enforcement actions by the Arbitration Committee may be appealed to, and amended by, Jimbo Wales, unless the case involves Jimbo Wales's own actions." The fact that this arcane provision has not been used since the Jurrasic Period, and that it theoretically casts community-elected arbitrators as second-fiddle, is reason enough to modernise by dispensing with this arrangement. Jimmy plays a huge role, and does not need such symbolic tips of the hat. Let him get on with his role, and ArbCom with its role. Tony (talk) 07:25, 2 August 2010 (UTC)

I don't think it's necessary to have a quorum of the entire community, or the entire admin community, to see a community consensus form. In everyday news polling, a small sample will give a fairly good idea of how an election will turn out - it may be off by five percent, but it won't be off by fifty percent. True, discussions here involve self-selection (and potentially stealth-canvassing), but a proper random sample could be collected if needed. But if opinion runs strongly enough in one direction isn't even needed. Also, there is some distinction between consensus and voting, in that consensus puts somewhat more weight on strong opinions, as evidenced by some degree of explanation by the editor, whereas voting can hinge almost entirely on the opinions of a large group of "undecided" people who really don't know or care much about the issue at all. This also favors a greater acceptance for the results of large but not Wiki-wide discussions and votes. Wnt (talk) 11:49, 2 August 2010 (UTC)

• On getting Jimbo out of the policy, as a matter of principle I doubt this page and its editors have the authority to make that change (much as it is needed) unilaterally. Further, a discussion of Jimbo's final appeal authority should also consider removing him from formally appointing ArbCom and from participating on the ArbCom mailing list.
• While we are on matters of principle, it is disappointing how often both arbitrators and mere mortal editors disregard principle in the so-called interest of pragmatism. For example, that it is hard to devise a suitable mechanism for the community to over-rule ArbCom is not a good reason to abandon the principle that the community is sovereign and supreme. ArbCom's terrible record in issues of recusal also owes a frightening amount to disregarded principles of propriety and ethics. There is an announcement at WT:ACN at this moment where most if not all arbitrators should know there should have been a recusal, yet there is no mechanism whereby the community can even force ArbCom to openly discuss recusal concerns.
• The 'ArbCom not GovCom' issue needs to be addressed. In the recent retroactive approval of BLP deletions by fiat, ArbCom acted as a governing body, ignoring established content and conduct policy, its own policy and principles, and once again provoked outrage in the community. The skills for which arbitrators are chosen are not the skills needed for governance.
• As a matter of accountability, an arbitrator recall mechanism is years overdue. Problematic arbitrator actions / behaviour have led to both sand-bagging and resignations in the past, and unfortunately the community has lost good arbitrators and kept some... well, let's not get specific there.
• On acountability, the present statement "Any arbitrator who repeatedly or grossly fails to meet the expectations outlined above may be suspended or removed by a motion supported by two-thirds of the Committee." needs to go. I know it says may be suspended etc but I can think of no case where arbitrators hae publicly done anything more than mildly rebuke / criticise a colleague. Perhaps there have been harsh words behind the scenes - I don't know. What I do know is that arbitrator solidarity makes ArbCom appear unwilling to admit to errors in its actions and even more inflexible when issues of arbitrator behaviour arise. EdChem (talk) 13:58, 2 August 2010 (UTC)
• In principle the idea of a perfect democracy where we are all equal is wonderful, but in practice I would want to see a very carefully written (by lawyers) constitution before any real attempt to make the "community" pre-eminent. As Wikipedia becomes more important (rising in Google ranking, and surpassing other information sources), there will be ever-increasing attempts to game the system and "take over" either a part or a whole of this site. Interminable POV disagreements like WP:Arbitration/Requests/Case/Race and intelligence show that community consensus cannot achieve desirable results without an arbitration system that is reasonably final. Given that this is an encyclopedia and not a place to exercise free speech or right wrongs, some discussions simply have to finish (until resumed in a few months), and an ArbCom that delivers a result (even if "wrong" in the eyes of some) is better than the never-ending struggles that would ensue if a handful of people could indefinitely push their case by arguing that the community has primacy. The community has primacy: we could change the ArbCom system if we wanted, but we don't. ArbCom is the worst form of governance, except for all the others. Johnuniq (talk) 03:42, 3 August 2010 (UTC)
• As others have noted, this is ArbCom, not GovCom and it has limits. Those limits are listed in this document. But what mechanism is there if they go off the rails? In particular what if they go off the rails in a direction Jimbo likes? A good constitution would have recall mechanisms and the like. We aren't going to get there I don't think, nor perhaps should we. But a founding principle that ArbCom's decisions can be overridden by the community in principle is enough for me. If the issue arises we can find a fair mechanism. Should they go that far off the rails the community can reign them in. It would have to be a clear case of the vast majority objecting to have a shot of actually working, but the principle is important. Hobit (talk) 04:15, 3 August 2010 (UTC)

• I'm very curious to know exactly how ArbCom could go off the rails.  Roger Davies ^talk 15:55, 3 August 2010 (UTC)

Even Nazi groups have more than 300 members - more than the number it takes to win an ArbCom election. Assuming one single political group somewhere is not completely incompetent... Wnt (talk) 17:08, 3 August 2010 (UTC)

[Chuckle]  Roger Davies ^talk 17:28, 3 August 2010 (UTC)

I think that the subjugation of ArbCom by some political group can become a more serious possibility, harder to detect, and even harder to define, if ArbCom takes on general "government" type powers. Individual cases for arbitration are hard to predict or classify, and people want good well-known arbiters for them. But if ArbCom takes on a role of pushing policies, then editors won't be voting for people any more, they'll be voting for platforms. You could see a group of obscure candidates come out and assume they're no threat because they're nobodies who nobody knows, but then on account of a censorship debate or allegations of ethnic, religious, or political bias, they could suddenly turn up with 500 votes apiece. You might suspect something irregular happened but you wouldn't be able to prove it. Wnt (talk) 19:19, 4 August 2010 (UTC)

Awesome job Godwinning the thread.  ;) I think you are building a strawman though. We saw the anger over the perceived policy by fiat motion created in the BLP mess. There is no way the community would stand for ArbCom becoming a government type body. Resolute 20:46, 4 August 2010 (UTC)

Oddly, I can't seem to find anything in WP:BLPBAN or WP:BLPLOG or WP:BLPSE to suggest that they are not still current policies. The deletion of Inge Lynn Collins Bongo a few months back, shortly after it was mentioned on Jimbo Wales' talk page, despite her having been mentioned in a Congressional subcommittee report and various mass media sources, was an example of the "BLP crackdown". So I don't see the evidence that ArbCom ever recognized that it might have overreached by announcing a BLP crackdown in an arbitration case about inline citations in the biography of a dead person. As for Godwin, I never suggested anyone was a Nazi - I just want Wikipedia to design itself with things like separation of powers designed to protect us from them in the future. Wnt (talk) 22:45, 4 August 2010 (UTC)

• Sure, for example they could A) start creating policy B) endorse the actions of people who are clearly doing wrong things. They've done B recently for some definition of "clearly" (mass out-of-process deletions somehow are acceptable?) and frankly I get the sense of them heading toward A. It's not too far of a walk to "off the rails" in my opinion. Hobit (talk) 17:43, 4 August 2010 (UTC)
• To be more clear, we are setting up boundaries for ArbCom, but not providing any way to judge if they've crossed the line. I don't see Jimbo jumping in as I get the sense he'd prefer a stronger ArbCom. Hobit (talk) 17:48, 4 August 2010 (UTC)

"To resolve matters unsuitable for public discussion because of privacy, legal, or similar concerns"

I would oppose this line of text. These concerns should be placed out in the open for all to see and cases too extreme for public participation shouldn't be handled by ArbCom at all. They should be brought up to a higher power able to handle sensitive information without being tied down to the consensus-based model we use on-wiki. People like Mike Godwin, the Foundation, etc. should be the proper authorities for such extreme cases. I am extremely wary of letting ArbCom deal with any material that the community isn't privy to. ArbCom is a community-elected body and as such its actions should be available for all to see and judge. We are an open, transparent encyclopedia that operates under the model of consensus. As such we should prohibit any policy which allows binding decisions to be discussed offwiki. ThemFromSpace 15:30, 3 August 2010 (UTC)

It's not entirely true that Wikipedia is a transparent organisation. For years, administrators have been able to delete material, which significantly reduces its transparency. They can also now apply suppress. It is probably neither possible nor practical now to create a truly transparent organisation.  Roger Davies ^talk 16:24, 3 August 2010 (UTC)

(edit conflict) Themfromspace—thanks for your query. I'm sure Mike Godwin and the Foundation Board would be biting their collective lips reading this. It would be quite inappropriate to foist onto them responsibility for dealing with the large number of sensitive issues—probably best described as a continual stream, actually—that the arbs have to dispatch on a daily basis out of the glare of the spotlights. These matters are "under the radar" for very good reasons: to protect all parties, the project, and the Foundation—you, me, and ultimately the readers we serve.

Perhaps I could quote Miss Scrimshaw, a notable character in Patrick White's A fringe of leaves (1976): "I do not believe one will ever arrive at the end of people's frightfulness" (p. 10). The arbs work incredibly hard to deal with the kind of problematic behaviour—extreme attacks, sad or dangerous cases—that would bore most of us to tears. It must be unpleasant work (we've lost a few arbs this year already), but it just has to be done, with skill, sensitivity, and confidentiality. I cannot go along with the claim that because ArbCom is an elected body, "its actions should be available for all to see and judge".

I have felt the rough end of an ArbCom case, so perhaps I might be quick to indulge in conspiracy theories. However, I have come to see how easy it is to take for granted the work done out of the public arena to keep this huge and complex wiki running smoothly. Please remember that we put the candidates through the hoops, very publicly, during the annual elections. Every voter has the opportunity to probe candidates deeply on their talk pages or election pages. Tony (talk) 16:30, 3 August 2010 (UTC)

I keep getting mixed messages from the posters holding high office in Wikipedia. I am at once told that resolving cases in secret is so extraordinarily rare that an arbitrator knows of only two instances, and that it is nothing to worry about - and that it is so common, so thick and burdensome a stream of cases, that it is far too much for the front door of the WMF to handle — the same group that handles every crank e-mail the site gets from the general public! This is an awfully large contradiction, isn't it?

Now Roger Davies seems to speak common sense, as people seem to know it on the Internet: everyone knows that every internet service can and does claim the right to censor any and all content without any public admission of it, ban any and all editors, without explanation; that every site provides a Code of Conduct and every code of conduct ends in "Etcetera" and "other reasons as we see fit". But this "common sense" isn't common sense as we know it from the real world, but reflects some weird Internet combination of slaveowner-libertarian ideology and a thick gooey layer of C.Y.A., and it's not a precedent for a large volunteer organization such as the world has never seen. Wikipedia doesn't have to protect the editors from democracy. At many an unruly demonstration, people have chanted, "This is what democracy looks like", and we should take that to heart. Wnt (talk) 17:04, 3 August 2010 (UTC)

Thanks Tony. I still have issues with this provision, but I can understand where it's coming from now. Rest assured I'm not accusing ArbCom of any conspiracies, I'm just trying to uphold our commitment to transparancy, which is something I value as a member of the community. ThemFromSpace 18:27, 3 August 2010 (UTC)

Just as a point of information, complete transparency is not policy. In fact, WP:NOT says "Wikipedia is free and open, but restricts both freedom and openness where they interfere with creating an encyclopedia", which has been policy since February 2005.  Roger Davies ^talk 11:06, 4 August 2010 (UTC)

Incidentally, the foundation has been alerted to this part of the discussion and if they have anything else to add to ArbCom's position then I'm sure they'll say something.  Roger Davies ^talk 17:30, 3 August 2010 (UTC)

"To consider appeals from blocked, banned, or otherwise restricted users"

Why does that need to be the case? It seems a bit iffy to me to have the same group that bans people also be the one that reviews a ban. Of course people should always be able to appeal to the Arbitration Committee in hopes that it will reverse its action, but for cases where the Arbitration Committee has clearly made a mistake (Matthew Hoffman, some editors in the Scientology case who hadn't edited for over a year), there should be another body that has the power to overturn ArbCom decisions, though not make its own. NW (Talk) 15:07, 5 August 2010 (UTC)

a) It's not always ArbCom that bans people. b) We don't have a "higher court" to appeal to. The main reason this function has always been there is that blocked users are technically incapable of appealing to AN/I or RfC or wherever, since they are blocked. ArbCom can handle cases brought to it by email in a way that the community can't. (Yes, there is the unblock template, but that only really works for getting a second opinion from another admin, rather than any kind of detailed appeal.) --Tango (talk) 15:11, 5 August 2010 (UTC)

This might not be the best time or place to be proposing new institutions, better to focus on whether ArbCom ought/ought not be responsible for x or empowered to do y. Skomorokh 15:18, 5 August 2010 (UTC)

Four Net Votes

Has there been any talk of getting rid of this? I remember that Newyorkbrad raised the issue around the time of EEML, but I don't think there was any follow up with that. NW (Talk) 15:15, 5 August 2010 (UTC)

• What was the issue raised? Hobit (talk) 15:10, 6 August 2010 (UTC)

No, not really. It gets very occasionally (and inconclusively discussed) on the list. It's a procedural device for accepting and closing cases: it's more expedient and less cumbersome than a full committee vote.  Roger Davies ^talk 19:48, 6 August 2010 (UTC)

Preamble

I think we need something of a preamble. What the purpose and guiding principles are for ArbCom. I'd personally want to include something about being an elected body there to resolve otherwise intractable disputes, not to make policy. That a small group of people are needed which the community has expressed faith in so that highly detailed and time consuming cases which the community can't possibly deal with get addressed in a thoughtful and fair way, etc. Thoughts? Hobit (talk) 15:09, 6 August 2010 (UTC)

• Isn't an executive summary putting undue weight on some aspects and not others? To return to the policy thing, the current draft says:

  The arbitration process is not a vehicle for creating new policy by fiat. The Committee's decisions may interpret existing policy and guidelines, recognize and call attention to standards of user conduct, or create procedures through which policy may be enforced.

  Previous Arbitration Committee decisions are considered useful and informative, but are not binding on future proceedings.

  Which seems pretty categoric to me, particularly in tandem with the "ArbCom decisions don't create precedent" statement. Would adding a ===Policy and precedent=== heading help emphasise it more?  Roger Davies ^talk 20:01, 6 August 2010 (UTC)

I'm actually fairly mellow about what is in the Preamble, but A) I dislike having a document without a lede, and B) we really should be able to summarize the point of this thing. Hobit (talk) 00:27, 7 August 2010 (UTC)

How about having a very simple one-sentence lead, along the lines of These are the policies of the Arbitration Committee? --Tryptofish (talk) 00:32, 7 August 2010 (UTC)

That seems reasonable; we'd presumably want some text before the TOC on the page, in any case. I'd suggest something along the lines of:

This policy governs the Arbitration Committee and the Wikipedia arbitration process. It was ratified by the community on [date].

with links to the ArbCom and arbitration pages as appropriate. Kirill ^{[talk] [prof]} 02:22, 7 August 2010 (UTC)

Indeed. I've tweaked it slightly ("This policy governs the Arbitration Committee, arbitration proceedings and arbitration processes. It was ratified by the community on [date].") and added it in a new Preamble section.  Roger Davies ^talk 11:07, 7 August 2010 (UTC)

I've also just inserted a header ===Policy and precedent=== to bring out the text beneath it.  Roger Davies ^talk 13:33, 11 August 2010 (UTC)

Arbitrator recall

Reading back through the noticeboard archives, I see that the prospect of recalling arbitrators has been on the Committee's agenda in the past. Might this be an opportune time and place to address it? Skomorokh 19:00, 6 August 2010 (UTC)

Despite discussion in the past, the Committee hasn't really formed a consensus view on this. The obvious major drawback with a formal recall process is that it has great potential for harrying and neutralising arbitrators. Another is that it would probably take a month of Sundays to write a procedure and establish parameters that everyone could agree on. Despite the meme to the contrary, in my experience, the committee has been very effective in suggesting that it might be a good time to go. In extreme cases, a steward could desysop as indeed could the committee using existing procedures: which would certainly have the effect of securing a resignation by making the abitrator's position untenable.  Roger Davies ^talk 20:24, 6 August 2010 (UTC)

This a classic pragmatism over principle answer. It's too hard so we can't / won't do it, never mind what the principles of accountability and good governance demand. By the way, we aren't talking about forcing an arbitrator to be accountable to ArbCom (on the occasions where ArbCom decides not to stonewall), we are talking accountability to the community. As an aside, I'm presently speculating that a dislike for the idea of a principle-driven approach is the reason my earlier comments seem to have been basically ignored. The principle-driven approach here recognises that the accountability mechanism is necessary, and so ignores all the arguments about it being hard when they're advanced as reasons to do nothing, instead recognising those are problems that must be solved. Start by deciding what needs doing, then find a way to do it. Don't get talked out of doing what is necessary by concern about how difficult it will be, that's the reason we don't have an administrator recall process. EdChem (talk) 20:51, 6 August 2010 (UTC)

What would the bar for recall be? It would have to be substantially higher than "s/he made a decision that I don't agree with." I can only think of two instances where an Arbitrator did something that I'd consider recallable - in one case, the Arbitrator had failed to disclose alternate accounts with significant disciplinary histories, and in another an Arbitrator concocted an entire ArbCom case from soup to nuts by himself and presented it as a fait accompli, without a clear rationale for the need for secrecy and (apparently) without the consensus of the Committee itself.

In both cases, the Arbitrator in question resigned, although in the second case the resignation occurred long after the event in question and its motivations were presumably complex. I know many people felt that the failure of an Arbitrator to disclose their knowledge of an active sockpuppet was a recallable offense; while I didn't necessarily share that view, the Arbitrator in question did resign in response to the community outcry.

So that's three events, and in only one case was the lack of a recall mechanism truly an issue (although ultimately, presumably due in part to internal pressure, the Arbitrator in question did resign). I see a lot of drama ensuing from any recall process, no matter how carefully constructed, and given the history I'm not sure that there's a pressing need for the process which would outweigh that drawback. MastCell ^Talk 21:51, 6 August 2010 (UTC)

Indeed, the main problem is that a recall mechanism with a low barrier to entry will be used simply to exert political pressure on arbitrators voting in a way that displeases some influential group within the community, while one with a high barrier will be functionally indistinguishable from the current provisions for having an arbitrator removed by the Committee itself.

I'm not disputing the principle itself, of course; but would having the policy contain a statement like "Arbitrators are subject to community recall in principle, but no process for such a recall exists" really be helpful? In particular, I think it's reasonable to say that establishing any such process would require community agreement on a level commensurate with that required for amending the policy itself; and, given that, it would be just as simple to amend the policy with a recall provision if or when a method actually gains acceptance as it would be to insert a principle that cannot be invoked in practice and then amend the policy again to add the method for invoking it.

Or, in slightly more general terms: do people really want to see more statements of principle in the policy even if there is no provision for enforcing them? Kirill ^{[talk] [prof]} 02:19, 7 August 2010 (UTC)

Yes, yes I do. I'd prefer mechanisms, but I'll take statements of principle. Hobit (talk) 02:50, 7 August 2010 (UTC)

In spite of controversy, or how frequently/occasionally it's used or needed, some principles have enabled practices to emerge that works in the way that is intended. Either way, it might be an election issue this year. Ncmvocalist (talk) 07:30, 8 August 2010 (UTC)

Election issue? Well, that doesn't inspire confidence that a sober, objective, rational debate will take place... :P MastCell ^Talk 17:08, 9 August 2010 (UTC)

Agreed. Administrators, and especially Arbitrators must be free to make unpopular but necessary decisions. I do understand why recall for administrators can be a good thing if done voluntarily (note, I have my placed myself under those terms, and twice have had recall petitions filed against me, neither time came close to reaching the low bar I had set). No one has come up with a system that A) would be useful in situations where an Arbitrator has not already resigned due to public pressure and B) would not be open to being used as just another battleground tool. Personally, I don't think such a system is possible, but I've been proven wrong before. SirFozzie (talk) 23:26, 10 August 2010 (UTC)

What is the enforcement mechanism?

Having rules and guidelines for ArbCom is all well and good and I'm pleased we are having this discussion. That said, we seem unwilling to create a mechanism, or even hint at one, for dealing with the breaking of these rules. I'd say ArbCom has created policy in the past (which would seem contrary to this document). What mechanism do we have if they do so again? We've ruled out having a !vote on the issue and ruled out arbitrator recall. What does that leave? I've never seen a document like this without some reasonable mechanism for enforcement. If there is no mechanism, what's the point of this document? Just to tell them what we'd like them to do? Hobit (talk) 00:26, 7 August 2010 (UTC)

Where has ArbCom made policy? Can I have an example please?  Roger Davies ^talk 01:28, 7 August 2010 (UTC)

Sure. [3] endorsing folks ignoring policy is pretty much identical to making policy. [4] is clearly the creation of policy. Perhaps it was a good policy to create to allow all administrators to take such measures at their own discretion, but it wasn't allowed before and after this it was. Hobit (talk) 02:49, 7 August 2010 (UTC)

I disagree I'm afraid with your analysis regarding the motion. There is (and was) wide, probably universal, consensus that BLPs need to accurate and sourced.

Through instruction creep, and against policy, notability guidelines were being used to trump the WP:V and WP:N policies.

The committee declined to accept a case against various arbitrators (it is entitled under longstanding policy to do so) and urged the community to sort out the problem.

In both the motion and the Badly-drawn Jeff case, the committee simply reiterated existing policy. I can't see what new policy was created.  Roger Davies ^talk 11:25, 7 August 2010 (UTC)

I certainly can accept that others might not see it as policy creation. I will point out that they didn't decline the case, they endorsed the out-of-process action. Those are quite different things. In the WP:BLPBAN case (hey look, it has a shortcut to it--makes you think at least some folks believe it will need to be referred to on occasion) I think that FT2 did see it as policy creation and clearly a large number of people felt the same way in the aftermath. I'm asking what we do if a large % of everyone thinks they are overreaching. Hobit (talk) 12:16, 7 August 2010 (UTC)

The fact remains that deletions probably weren't out of process as there was, to put it mildly, a tension between the guidelines and the core policy. (See WP:BURO, for an explanation.)

The solution is not to try to beat ArbCom over the head, which is best a distraction, but to use existing community processes (RFC, mostly) to address the core issue. This incidentally nearly always what ArbCom recommends, and did so in this instance.  Roger Davies ^talk 12:28, 7 August 2010 (UTC)

The other documents where you have seen enforcement procedures would have been dealing with constituted bodies where there is a clear line of legal authority, ownership, and liability. In those bodies, the consequences for failing to follow a directive would include loss of pay, loss of employment, and state-enforced sanctions such as fines and prison. Wikipedia is different in many respects, and a semblance of control here is maintained by delegation: contributors delegate powers to admins and arbcom to manage the more egregious trolls and POV pushers, while admins/arbcom delegate maintenance of encyclopedic content to the contributors. Meanwhile, those with legal ownership and liability delegate all they can to the community. Arbcom may make mistakes, but they don't go crazy because they are good people and because they would get massive pushback from the community if they did. Ultimately, Wikipedia works because the participants recognize that the three groups (WMF, admins/arbcom, contributors) are all essential. In short, there is no need, nor is there any feasible mechanism, to maintain enforcement procedures to control arbcom. Johnuniq (talk) 01:36, 7 August 2010 (UTC)

I'll disagree. As noted, I feel they've gone off the rails before and I very much worry they will again. If they do something 90% of the community thinks is wrong there really should be a way to address the issue. Hobit (talk) 02:49, 7 August 2010 (UTC)

But there already are ways. An RFC on the issue for a start. The focus on ArbCom rather than on the issue is likely to be more of a distraction than anything else.  Roger Davies ^talk 11:31, 7 August 2010 (UTC)

I agree with Roger. The BLP issue here did go to an RfC and workable processes emerged from that. In the end, ArbCom's role was relatively minor, serving more to calm things down and help the process along. It should also be noted that there was not unanimous agreement among ArbCom on the motion passed. A better link is here, where you can see that the motion passed 9-4-3 and there were extensive comments by arbitrators. That alone should indicate that ArbCom wasn't "going off the rails" as some have put it. To my mind, an ArbCom "going off the rails" would be one where more than half the committee resigns en masse due to disagreements with other arbitrators (rather than individually for other reason), and the remaining arbitrators pass a series of hardline motions with unanimous voting and refuses to engage in on-wiki discussion of the motions (this scenerio has happened on other language Wikipedias, IIRC). At that point, what would likely happen is administrators would refuse to enforce the motions. Of all the problems that may arise, I don't think ArbCom going off the rails is really one to worry about. There are enough checks and balances in place to prevent this happening, or to make it clear that it has happened if it does happen. Carcharoth (talk) 11:03, 9 August 2010 (UTC)

Appeals, amendments and clarifications

There is mention of an appeal process through Jimbo here, but I wonder if something should be mentioned in regard to amendment and clarification requests? --Martin (talk) 01:26, 7 August 2010 (UTC)

Both flow, I suppose, from remedies but it would easy enough to broaden if that's now adequate.  Roger Davies ^talk 01:30, 7 August 2010 (UTC)

Okay, on re-reading I found this: "Where the meaning of any provision is unclear to any arbitrator, the parties, or other interested editors, it will be clarified upon request"[5], that would cover off clarification requests, but what about amendment requests? There is nothing really stated about that. --Martin (talk) 01:33, 7 August 2010 (UTC)

The second half of the appeal section covers normal amendment requests: "A party or other editor affected by remedies may request reconsideration or amendment of the applicable ruling." Kirill ^{[talk] [prof]} 02:06, 7 August 2010 (UTC)

I thought that sentence was related to appeals to Jimbo Wales. Perhaps something like "A party or other editor affected by remedies may appeal to the Committee to reconsider or amend the applicable ruling" would make it more clear, then I would move the part about Jimbo Wales after, since appeals to him would be considered as the last resort. --Martin (talk) 12:39, 7 August 2010 (UTC)

That makes sense though it is probably worth taking out the bit about Enforcement appeals going to Jimmy. They already have quite a complex appeals structure and adding Jimmy to it is probably overkill.  Roger Davies ^talk 13:20, 7 August 2010 (UTC)

I've just tweaked the text more or less along the lines you recommend. Enforcement reversals/appeals are already covered by in the procedures document so I've taken the reference out here. If the AE policy goes to ArbCom, then the Jimmy appeal provisions would kick in.  Roger Davies ^talk 13:38, 11 August 2010 (UTC)

I must say that this is the fly in the ointment for me. It is anachronistic to retain the QEII role, especially when it's not even used. Didn't he say in public he intended to use that about as often as the Queen sacks a government, or some such? I think that gives the imprimatur for gently excising that bit. Tony (talk) 12:35, 10 August 2010 (UTC)

If the provision is there, it's intended to be used, and the chances are that they will be used. Then, we will be catapulted into a scenario not unlike the 1975 Australian constitutional crisis. I think it best that we did not have such a provision. Ohconfucius ^¡digame! 12:57, 10 August 2010 (UTC)

I agree. Jimmy is almost certain not to do anything regarding ArbCom beyond offer it advice and if he does do more than that it is likely to cause more trouble than good (based on past experience with Jimmy's interventions), so it really doesn't make sense to preserve the power in writing. If we need some kind of safety valve to protect against a runaway ArbCom it should be some form of community recall. --Tango (talk) 12:59, 10 August 2010 (UTC)

The commonest types of appeal are about lengths of topic ban; sockpuppetry determinations; and requests for unbanning from banned users. The membership of the committee significantly changes every year so there's no shortage of new eyes to look at appeals of old decisions. Jimmy did make known, very publicly, his opinions about "secret trials" affair about eighteen months ago and I am not aware of any criticism at all of his doing so.  Roger Davies ^talk 14:05, 10 August 2010 (UTC)

I suppose Jimbo's reserve powers to overturn ArbCom decisions, while rarely used, can bring an element of "adult supervision" to the deliberations of the Committee. Given that the membership of the Committee does change significantly over time, this probably also adds an element of stability to the structure. The scenario that has occurred in another language wiki, where half the committee resigns and the other half starts drafting extreme rulings, I just couldn't imagine happening here, Jimbo's involvement provides the necessary circuit breaker to prevent such a scenario occurring. --Martin (talk) 19:06, 10 August 2010 (UTC)

Scope

In looking over the two "scope" sections (Scope and responsibilities, and Format and scope of decisions), the latter says "The Committee's decisions may interpret existing policy and guidelines, recognize and call attention to standards of user conduct, or create procedures through which policy may be enforced.", but the former gives the impression that the scope is limited to behavioral issues alone.

I don't contest this, I merely would like to see the former section clarified, to prevent future haggling over policy and what the actual "scope" is. : ) - jc37 03:41, 10 August 2010 (UTC)

Good catch :) I've added "primarily" to the first scope section and text which Casliber suggested the other day ("Although the arbitration committee does not rule on content, it can propose a pathway by means of which the resolving of a content dispute can be facilitated.") to the second.  Roger Davies ^talk 13:28, 11 August 2010 (UTC)

This wording is found here: :“Although the Committee does not rule on content, it can propose a pathway by means of which the resolution of a content dispute can be facilitated.”

I have no idea what a "pathway" consists of. It does seem to me that this statement means that WP:OR cannot be entertained by ArbCom. I suppose a "pathway" might be for ArbCom to suggest to those disputing content that they should observe WP:OR, although ArbCom cannot, apparently, enforce its observation. Or can they? For example, would it be possible for ArbCom to suggest sanctions would be imposed upon those who fail to address sourced counterarguments on Talk pages? Or upon those who make unsourced assertions on Talk pages? ArbCom needn't determine whether a source is "reliable" but only insist that the appearance of genuine discourse be observed. Brews ohare (talk) 16:28, 17 August 2010 (UTC)

I've tweaked this a bit to read: "The Committee does not rule on content but may propose means by which community resolution of a content dispute can be facilitated".  Roger Davies ^talk 03:59, 20 August 2010 (UTC)

ArbCom might also be charged with enforcing that the use of WP:OR on Talk pages always be accompanied by an explicit statement of what text exactly is being labeled by the appellation WP:OR and why, as often this and other guidelines are bandied about in isolation as though self-explanatory. Brews ohare (talk) 16:37, 17 August 2010 (UTC)

More comments

Overall echoing others' words. A nice job overall, not much to suggest. A few quick comments though:

(1) "Format of decisions" - I wouldn't tie the format of a ruling (which could vary) in as a policy point. No need. Replace by a more general wording:

"Decisions are written in clear standard English. They should explain concisely and at a minimum the salient facts and principles, the remedies (or binding rulings) which will apply, and any enforcement provisions."

(2) "Selection and appointment" - add a bullet to cover the situation where new positions urgently need filling. Arbitrator burnout rates are high and this could be a realistic situation.

(3) Arbitration pages - there is no mention of pages related to the Arbitration Committee except that clerks ensure decorum on them. Some aspects of Arbitration practice and norms are specified on other subpages, so their editing is not completely open to arbitrary communal changes. Consider adding something like:

"The Committee manages the content of Arbitration subpages, which may expand upon arbitration norms and processes."

(4) "Appeal of decisions" - the clause "directly affected by remedies" is redundant and the title too narrow. Any user may seek amendment of a ruling, not just those "directly affected", and they may seek loosening or indeed tightening, not just appeal. Suggested retitle and 1st sentence:

Appeal and amendment of decisions: - Any editor may appeal or ask the Committee to reconsider or amend past decisions, usually when a remedy is either too restrictive or else needs updating as a result of subsequent events. [...]

(5) Enforcement - in the current text, specifying enforcement provisions is the Committee's responsibility, but the enforcement process generally, is not. Enforcement is the backstop for arbitrated decisions and therefore ultimately for all dispute resolution in the community. If the Committee doesn't have the formal right to ensure enforcement of its own decisions then arbitration achieves nothing and fails to be a "final resort" because as we have had in the past, cases will have no protection against unilateral dismissal of AE. Arbcom needs the ability to regulate enforcement of its rulings. Consider this fix:

The Committee retains jurisdiction over all matters heard by it, including over their enforcement processes, and may, at its sole discretion, revisit any proceeding at any time.

(6) "Ratification and amendment" - proposed changes with adequate support may be put to the community without Arbcom approval. In a few areas the community probably shouldn't amend the policy without Committee endorsement. For example if there was a sufficient vote that privacy based issues should be heard on-wiki it could be a problem. I would add that:

"Changes which may affect the scope, responsibility, internal procedures, or jurisdiction of the Committee require the approval of a majority of the Committee".

(7) "Policy and precedent" - the Committee at times:

• Proposed how policies may be reconciled in cases where they conflict (as opposed to merely interpreting them)
• Considers the spirit and norms of policies and communal practice (different from the wording of "policies and guidelines")
• Creates processes by which policy divisions and other issues may be temporarily addressed to reduce their disruptive effect, and
• In cases of extreme upheaval may set temporary measures in place to reduce conflict pending a proper resolution by community consensus.

Quick recheck needed that the draft policy clearly allows these.

FT2 ^{(Talk | email)} 18:30, 13 August 2010 (UTC)

Added numbers to pars for ease of reference.  Roger Davies ^talk 12:47, 25 September 2010 (UTC)

You've raised some interesting points and made some good catches :)

(1) Tweaked accordingly.

(2) Yes, I suppose this is theoretically possible so I've reinstated the earlier text about it. The long leadtimes for elections probably means this will never happen.

(3) I've broadly agree and have tweaked this. Suggestions about better, conciser wording, and better places to insert the text welcome.

(4) Tweaked this too, in line with your suggestion, adding a bit about the committee accepting or declining at its discretion.

(5) Yes, good catch.

(6) Hmmm, I don't agree with this, as framed.

(7) Yes, it probably does.

Thanks for the feedback,  Roger Davies ^talk 14:06, 25 September 2010 (UTC)

Parties

We are still lacking a definition of "parties". This matters, as being forcibly made a "party" to a case, or being "expected" to take part in the mud-slinging process, is extremely stressful to some editors. "Parties" should mean - only - the complainant(s) plus those of whom the Committee finds that there is reasonable evidence of possible serious misconduct. This would also help avoid the appalling situations that have occurred, where editors who have done nothing exceptionally wrong end up getting massively disproportionate sanctions - Kotniski

I agree this needs a defintion and some tidying up. Let me think it over ... 10:31, 14 August 2010 (UTC) — Preceding unsigned comment added by Roger Davies (talk • contribs)

The other approach would be to make it much easier to ask to be removed as a party - that might be easier to word:

"Parties are generally users whose conduct was central to the case or greatly affected the central issues involved, users who present or are listed in the request for arbitration, and users who deem themselves parties out of a wish to participate in the arbitration or resolve the dispute. Any user whose involvement was not central to the disputed matters or to the request for arbitration may ask to be removed as a party.

FT2 ^{(Talk | email)} 14:10, 14 August 2010 (UTC)

That seems more a point for the case management procedures than for the policy itself, since we would then need to specify (a) how parties are added and removed and (b) what the location and significance of the "official" list of parties might be. Kirill ^{[talk] [prof]} 14:29, 14 August 2010 (UTC)

Well the significance (according to the current wording of the draft) is that the people on the list are "expected" to do various things, which they may have no wish to do. If we're going to make this kind of imposition on people, we'd better make it very clear on whom we are permitting such an imposition to be made.--Kotniski (talk) 14:41, 14 August 2010 (UTC)

My initial inclination would be to say that anyone determined by the Committee to be a party would be expected to respond, and leave the actual process by which the Committee makes that determination and announces it as a procedural matter, as it's likely to be somewhat involved (notionally, it would probably be the list of initially named parties, plus any additions and minus any removals, with provisions for editors to request addition and removal, as well as for addition of editors being named in draft or proposed findings). Kirill ^{[talk] [prof]} 14:48, 14 August 2010 (UTC)

All 3 of the above comments make sense - both Kirill's about leaving it to case management and the committee, Kotniski's about stress on people who are (perhaps tenuously) named as parties and don't have an easy way to step out of the role. That is why I suggested have a one sentence summary that sets out the principle of being a party and that people with tenuous connections can ask to be dropped as parties, then leave the rest to case management. If it's all left to case management that would work - except that the draft also considers edits to case management so it just moves the whole thing a few lines down. FT2 ^{(Talk | email)} 15:25, 14 August 2010 (UTC)

Arbitrators should not hear cases themselves

I think it would be better if after accepting a case, the Arbitrators appoint a panel of Admins and experienced editors to hear the case and make decisions. It is then possible for far more cases to be heard, so one can intervene effectively in disputes far sooner with effective arbitration than we do now.

After a case has been heard and the decisions have been made, one Admin writes up a report explaining how the decisions were arrived at and sends it to ArbCom. In case of an appeal, that report can be used to see if it has any merits to let it go forward. If the Arbitrators decide to let it go forward, it will be heard by a new panel of Admins and editors.

If after one appeal the problems persist, e.g. Arbitration Enforcement doesn't seem to yield good results, the Arbitrators themselves get involved in the case. They then have two detailed reports to go by. Because they have not been involved in making any decisions in the case, it is then easier for them to see faults in any decisions that have been made and correct them. Count Iblis (talk) 23:28, 14 August 2010 (UTC)

Experimentation's good. Delegating the case handling to a panel of Arbcom appointed admins (for cases without privacy or similar issues), in effect placing the case more in (Arbcom trusted) communal hands and positioning Arbcom itself in a supervisory / review role for reasonableness and issues, might well have merits and be worth exploring.

As far as policy goes, the draft already states "The Committee may create or modify its procedures, providing they are consistent with its scope; and form subcommittees or designate individuals for particular tasks or roles". If some future committee wanted to try new approaches of the kind you suggest, they would not be prevented from doing so by it. FT2 ^{(Talk | email)} 23:55, 14 August 2010 (UTC)

Going a bit off topic, but I've said before that ArbCom should always be the second step, after admins (except in special cases where privacy/harassment/urgency/etc. are involved). If ArbCom were a simple court of appeal, to review and possibly amend specific actions or inactions of administrators, it would greatly simplify the issues and ensure that everyone gets a fair hearing (as we've seen in practice, ArbCom - consisting of fallible human beings - can get things disastrously wrong with respect to individual editors, and such editors deserve a second instance to consider their cases, which they don't have when ArbCom takes on matters in the first instance).--Kotniski (talk) 06:56, 15 August 2010 (UTC)

Arbcom is already the "second step after admins". The Committee will not usually consider a case until other routes have failed - which includes attempts at admin handling. FT2 ^{(Talk | email)} 09:03, 15 August 2010 (UTC)

That may be true, but what a "case" is is so nebulous and changeable, that in effect editors can get punished by ArbCom for things that have never been brought to admins, and in ways that no reasonable admins would ever consider (this is from my and others' experiences a year or two ago - things may have changed, but policy should set such things in stone to prevent a drift back).--Kotniski (talk) 09:16, 15 August 2010 (UTC)

It may not be true in all cases (but certainly in many of them, especially the bigger ones) that there has already been very significant administrator involvement before the case comes to ArbCom. What also sometimes happens, and I'm not sure what can be done about this, is that admins who try to intervene are overwhelmed by force of numbers, volume of complaints etc. It usually only takes a handful of vocal editors to stall community processes, especially if they are civil and articulate. Happily, though, fewer cases are coming to ArbCom as the community finds ways to tackle problems. Unhappily, these that come are the most difficult ones.  Roger Davies ^talk 10:14, 15 August 2010 (UTC)

Yes, it would be good (though perhaps not here) to develop some more effective ways of resolving these recurring types of problem. My observations are that the system we have doesn't really work all that well - in Central/Eastern Europe editing, for example, despite several ArbCom cases and hundreds of admin interventions, we still have a situation where various groups of obviously tendentious editors are constantly sniping at each other, burdening admins with complaints, harming article quality, making life intolerable for ordinary editors, etc. I'm not claiming to have the answers, but probably more rapid, hands-on, enforced guidance by admins, with more emphasis on principles than individuals, is the way to go (sorry, just thinking out loud here; this has nothing to do with this page any more).--Kotniski (talk) 10:37, 15 August 2010 (UTC)

This suggestion for treating cases using a board that is appointed is closely related to a suggestion I made long ago: the notion of filtering the cases that are taken to ArbCom. I'd like to draw attention to this proposal. Brews ohare (talk) 15:34, 17 August 2010 (UTC)

Addition to "Scope and responsibilities"

I would like to see the section on Scope and responsibilities extended to include activities to catalyze genuine discussion on Talk pages. Enforcement of guidelines such as WP:CIVIL are presently clear responsibilities of ArbCom. However, WP:CIVIL, for example, often is enforced in a lopsided manner to stifle conflict, rather than to bring about sensible engagement. Therefore, I'd propose as a responsibility the enforcement of discussion on Talk pages using something like the following language:

6. To catalyze discussion on Talk pages by even-handed enforcement of behavioral guidelines such as WP:CIVIL, by enforcement of content-specific Talk page discussion, and by discouragement of interruptions of discussion threads with diversionary tactics.

By "diversionary tactics" I mean sneers, sniffs, catcalls that should be taken as violations of WP:CIVIL and misuse of polls to stop discussion that should be stopped under WP:POLL. It may be that some guidelines need clarification in this regard.

In general, much more could be done to improve Talk pages if enforcement clearly had the objective of catalyzing discussion instead of punitive squelching aimed strictly at clamping things down. Brews ohare (talk) 14:52, 18 August 2010 (UTC)

Concerns and Suggestions

While I agree on the overall purpose of this draft, there are a few areas in which I have serious concerns and suggestions

• (1) Scope of ArbCom authority.

• ArbCom was designed to be limited to matters of user conduct such that they would not erode or supersede the role of the community in making decisions regarding content and policy. This draft introduces wording that, whether intended or not, makes ArbCom capable of binding content and policy, which significantly oversteps their mandate.

• (2) Role of consensus in determining whether a case should go before ArbCom

• The previous policy included the language "The Committee will hear or not hear disputes according to the wishes of the community, where there is a consensus.". I'm not sure if this was ever invoked in an RFAr, but this was also a check on the power of ArbCom - the ability for the community to keep a matter outside of ArbCom jurisdiction.

• (3) Role of Jimbo Wales

• This seems to be a hotly debated topic, and one that we might be well to actually separate out into it's own policy, as Jimbo's role, and willingness to act in it have diminished significantly in recent years. As founder of the project and a de-facto "monarch", he is highly respected and trusted by the community, but at the very least, we should probably start changing that language to give the authority held by Jimbo to the WMF board, either as an alternative to Jimbo or with "and/or" language, as well as move the "royal assent" for the appointment of Arbitrators to the foundation board with the expectation that they will act in a manner that insures the long-term integrity of Wikipedia's judiciary functions.

• (4) Role in appointment of oversight and checkuser.

• The Arbitration Committee should be empowered only to deny these advanced permissions (as they may be aware of circumstances not generally able to be known to the community due to privacy reasons), not grant them against the wishes of the community, however, as those rights are inherently required to perform the duties of arbitrator, those rights should be implicitly granted to current, duly elected arbitrators by virtue of their election, and to those under interim appointment with concurrence of the community.

• (5) Removal of arbitrators.

• As arbitrators are elected by the community at large for limited terms, and their power checked by that of other arbitrators, the removal of an arbitrator should rarely if ever be necessary, and as a general rule, arbitrators have been of sufficient integrity that they have resigned when it was demonstrated that the community had lost trust in their actions. Should an arbitrator need to be removed, it can be expected that other arbitrators would privately ask for their resignation, and that failing, an appeal to Jimbo or to the WMF board seem to be acceptable remedies. A general ban on editing of any significant length by the Arbitration Committee against one of it's members would also create a de-facto removal.

• (6) Hearing of cases involving only private evidence, and similarly sensitive matters.

• So long as the committee is explicitly limited to matters of conduct, I don't have a problem with it. As soon as the scope goes beyond user conduct, I don't believe there is a place for secrecy. and the only cases which might in my view demand such level of secrecy without being matters of content are already handled adequately by the Foundation as office actions.

• (7) Long term project governance.

• As Jimbo's role as defacto benevolent dictator for life, founder, monarch, godking, or whatever label you wish to use has diminished, there has been a sort of power vacuum created which is not filled by any existing process. An alternative executive authority is needed, but I don't feel comfortable with that authority vested in the same instrument as our judiciary functions. Whatever form that executive authority takes, the purpose needs to be to make the sort of binding decisions and injunctions that were once commonly done by Jimbo himself. While consensus is and continues to be the primary governance of Wikipedia, there have been times where consensus was not forthcoming on a matter of pressing importance to the community, where the community drifts from fundamental guiding principles of both Wikipedia and the Foundation, or where consensus was not clear. The guiding principles of such authority should be to maintain the fundamental principles of the project and Foundation, to intervene in matters where lack of consensus poses an ongoing disruption to the project, and to act as a final check on all other processes on Wikipedia. If we fail to do something in this regard, we will continue to see matters of content and policy escalate into matters of conduct and conflict - which will in many cases exhaust the other avenues of dispute resolution and ultimately be adjudicated by ArbCom, thus causing policy and conduct matters to fall into their scope. As noted in other threads, matters of policy and content fall under a different set of skills than which we expect from arbitrators, so this is not a desired outcome.
• It should also be noted that the foundation itself can't be expected to involve itself in day to day governance of any of it's projects, nor can it directly appoint anyone to do so as they would be seen to be an agent of the foundation itself, therefore increasing the legal exposure of the foundation for actions of editors. This means that the foundation itself can't act in the way Jimbo historically has - Jimbo has done so primarily in his largely ceremonial role as "leader", rather than any official role as a member of the foundation staff - the most the foundation can likely do is give it's assent to some form of community-established process for appointing an executive authority, and possibly step in if that process violates a founding principle as recognized by the Foundation

I apologize for the length of this post, but I hope that it helps to explore the issues involved in adoption of a new arbitration policy, as well as larger issues which need to be addressed. Triona (talk) 05:13, 5 September 2010 (UTC)

No need to apologise for the length: it isn't tl;dr at all :)

(1) I've been through this again carefully and the language to which you refer isn't jumping out. Could you clarify this please, with a quote from the text?

(2) Indeed and I can't think of an instance were it has been mentioned, let alone been a key part of an RfAr either. In fact, the opposite is usually applies; with the community sometimes putting great pressure to accept a case. I suppose the conclusion is that it isn't a major community concern.

(3) Jimbo's role is unique and his role doesn't automatically devolve to the Board. Apart from which, all the indications are that the Board and the Foundation want less rather than more to do with day-to-day management of the individual wikis.

(4) Previously, CU/OS appointments were made by the committee a week or so after asking for community feedback. At the time, no one batted an eyelid and the whole business was very low key. I personally felt that the community should be more involved, was one of the arbitrators who pushed hard to implement elections; and played a major part in setting them up. Overall, I've been very disappointed by the lack of interest shown by the community. The core problem is that, with or without community input, we need to regularly replenish the ranks of CU/OS operators to keep up with the volume of work.

(5) Does this need spelling out? If the committee removed an arbitrator's tools, that arbitrator would be obliged by peer/community pressure to resign as no one would take him/her seriously.

(6) Without going into too much detail, there's a whole load of stuff which involves conduct and privacy which the Foundation is not remotely interested in dealing with.

(7) I'm not actually convinced that a specific executive authority is needed as we had the RFC process which usually gets there in the end. Perhaps what does need addressing is what one editor referred to as "demogoguery", where a charismatic and articulate editor is able to exert undue influence or, put more graphically, punch above their weight.

Thank you again for the interesting and thought-provoking comments,  Roger Davies ^talk 14:50, 25 September 2010 (UTC)

Is this on?

Hello, is anyone still working on this? Can we expect any more revisions, on the road towards presenting a version for "ratification"? --Kotniski (talk) 08:37, 13 September 2010 (UTC)

Indeed you can and within the next week or so, I expect :) Sorry about the delay in getting back to people but I've been away for nearly three weeks.  Roger Davies ^talk 09:42, 13 September 2010 (UTC)

A suggestion

Currently the Arbitration Committee functions as the governing body of the English Wikipedia. I believe that ArbCom and MedCom were set up by Jimbo to aid in the process of dispute resolution, MedCom has stayed true to that but ArbCom has evolved into the governing body of Wikipedia, it has become a case of "What we say is final". The Committee itself has too much power. I am proposing the following changes:

1. Decentralisation and removal of the power currently vested upon the Committee.
2. Creation of a new committee that will be responsible for looking after Policy and putting forward the remedies proposed by ArbCom to the greater community in order to gain consensus on these remedies, the final decision then lies with the Stewards who will determine the consensus established and make the final decision.
3. Removal of Jimbo's authority, he should not be the person to make the final decision. He should be allowed a say in discussion but should have no part in enacting these decisions, such decisions are for the community and the community alone.
4. Allowing the community a greater say, the community should be allowed to have a greater say in decisions. The Arbitration Committee should then evaluate the opinions of the community and take them to heart. They should not override community consensus and should not have the final say, the final say always lies with the community.
5. User rights, sanctions and editing restrictions should be decided by the community through a simple Request for Comment, the Arbitration Committee should not be the ones to decide who does what, the community are capable of deciding for themselves. Whether or not a user should keep their admin/crat bits and whether or not a user should be made a Checkuser and given Oversight permissions should be decided by the community in the same manner as a request for admin/bureaucratship, with a requirement of 80% Support for promotion.

In setting up a new hierarchy we can eliminate the gap of power and give the community greater say. This suggestion was copied from the Policy proposals section of the Village Pump. —Ғяіᴆaз'§Đøøм • Champagne? • 07:37, 17 September 2010 (UTC)

I hope you don't mind but I've changed the stars to hashes for ease of reference. Please revert if you hate it :)

re: #1. This is essentially abolition of the committee. Can you provide please examples of the excessive power that ArbCom has and why its abolition is crucial?

re: #2. Something similar was tried in the Climate change case, with a community sanctions noticeboard. Through no fault of the participants, it proved ineffective. This rather assumes that ArbCom decisions are about policy/content: they're not; they're about conduct.

re: #3. Jimbo plays almost no role in arbitration.

re: #4. The community is already handling more and more of the straightforward stuff. Cases these days coming to ArbCom tend to be of enormous complexity (cf. Climate change).

re: #5. In the context of nasty longterm disputes, there is no such thing as a "simple RFC".

More later, I expect.  Roger Davies ^talk 08:18, 17 September 2010 (UTC)

I'm saying that the Arbitration Committee should not have any real power, they should not determine who gets to keep their privileges, impose sanctions etc. but rather should make these recommendations and allow the community greater say. —Ғяіᴆaз'§Đøøм • Champagne? • 08:35, 17 September 2010 (UTC)

The Committee should allow the community to decide the changes needed and where needed, they should not override community consensus and do what is good from the ArbCom's perspective. The community currently has little or no say, what is the point in gathering the community's opinion if in the end all that time was wasted. The revocation of privileges and appointment of new CheckUsers and Oversighters should lie with the community, there should be no internal discussion, if a candidate meets a set criteria then by golly let's get on with the elections, if a candidate has been blocked let the community decide we don't need all this internal hubbub from the ArbCom.

Editing restrictions and sanctions should be determined through an RfA or RfC like format in that the community decides through a vote with each user expressing his/her own individual opinion. I'm suggesting we retain the ArbCom but separate its power, the ArbCom should have no power, they should only be allowed to make recommendations. The community are capable of deciding for themselves and should be allowed to express their opinion without it getting ignored and overrided. —Ғяіᴆaз'§Đøøм • Champagne? • 08:51, 17 September 2010 (UTC)

May I comment on the first four of the numbered points above? I have been through an unpleasant ArbCom case myself as a party, so you'd think I'd be antagonistic; but I've since seen how the Committee functions and reflected on it, and I largely support it way it works now.

1. "Decentralisation and removal of the power currently vested upon the Committee". That would be a disaster. Most of the service ArbCom provides to the community is a hard daily grind and would bore us silly. I believe it can be very stressful; believe me, our lives are easier for the work it does. ArbCom is part of the cost of doing business on a free wiki: unfettered freedom would be chaotic and would be more than just an inconvenience for editors like us.
2. Policy committee: it just won't work. While ArbCom involves itself in content and policy only insofar it is absolutely necessary to sort out behavioural issues in a case, the idea of a WP government has been tried and seen to fail miserably. We, the people, are the government, and consensus is the mechanism.
3. "Removal of Jimbo's authority"—I used to be strongly against his authority as the ultimate appeal, but I've softened. He never uses it, and frankly, there's a bunch of reasons for not bothering to change this symbolic status. However, I believe Jimbo should not interfere in the democratic process of ArbCom elections. Tony (talk) 09:20, 17 September 2010 (UTC)
4. "Allowing the community a greater say". I am not at one with the Committee on the amount of bloated text involved in its more complex cases. I believe there should be far more strict evidentiary limits for the sake of all parties, the arbs, and the clerks. You appear to be advocating the opposite, and I can only warn that no system of justice allows the town square / peanut gallery to add whatever it wants. Tony (talk) 09:20, 17 September 2010 (UTC)

Common sense should be used if these motions are passed, the community is courrently underrepresented by the ArbCom's agenda. The Committee itself was setup to aid in WP:DR not to have absolute power. My suggestion is that its power be separated into separate committees and where it would not hinder the progress of DR the community should be allowed to voice their opinion.

When deciding on the elections of CUs and OS's the system of candidacy should be similar to that of RfB, the ArbCom doesn't need to have their internal discussions on whether a possible candidate is fit for the job. When enacting decisions community consensus should be required where possible and the plausibility of requiring consensus should be evaluated. —Ғяіᴆaз'§Đøøм • Champagne? • 09:28, 17 September 2010 (UTC)

The community already gets great say on evidence and proposed decisions in case pages, even though this often gets unmanageably long as old disputes are re-fought (the CC case PD talk page case, for example, currently has about ten pages of archives, with more on the way). Community input on one of the Race and Intelligence article talk pages ran to nearly 90 pages of archives. The community pays ArbCom to sift through all this stuff, separate the wheat from the chaff, and make considered findings.

Furthermore, ArbCom is increasingly using discretionary sanctions. These remand individual sanctions back to admin, with appeal to community noticeboards.

The current CU/OS set up is determined by WMF policy.

 Roger Davies ^talk 09:47, 17 September 2010 (UTC)

(ec)However, the concept of "community consensus" is also a troubled one - it is quite undefined, and disputes usually arise as to what the consensus is (or rather, how to proceed in the face of obvious lack of consensus). At least with ArbCom, a motion is either passed or it isn't, and you have some idea of where you are. However I think a rethink of the whole disagreement-resolving process is in order - I'm sure we're all aware of the faults of the present "system", how much time, effort and good faith gets wasted in a very inefficient and sometimes unpleasant process for deciding whether the encyclopedia should say A or B. ArbCom generally say they see their role as eliminating the troublemakers from the consensus-forming process so as to allow the process to proceed smoothly - but that shouldn't be how it happens - the troublemakers need to be eliminated (or better - shown the nature of their trouble and persuaded to stop making it) quickly at an early' stage in the process, not months after things have erupted. I think we should have enough experience by now of how trouble develops to be able to work out ways of nipping it in the bud, of steering the process in the right direction the whole time.--Kotniski (talk) 09:58, 17 September 2010 (UTC)

Well isn't that what consensus is? I mean look at elections, they're chaotic in the lead up to election day. I'm proposing that new committees be set up to take over some of the duties of ArbCom. —Ғяіᴆaз'§Đøøм • Champagne? • 11:48, 17 September 2010 (UTC)

Actually, you're not. You're suggesting scrapping ArCom altogether and transferring many of its responsibilities to the stewards, who do not seem to have been consulted over whether they actually want a ten- to twenty-fold increase in their workload ;)  Roger Davies ^talk 12:01, 17 September 2010 (UTC)

My whole proposal is that ArbCom gets less power, with some of that power being focused back into the community. Also the establishment of new committees that will take care of some of ArbCom's duties will lessen the workload. Ok the Stewards idea wasn't that good, but who else will be able to fairly analyse and evaluate consensus if it isn't someone related? —Ғяіᴆaз'§Đøøм • Champagne? • 20:47, 17 September 2010 (UTC)

Why would new committees be any better than the old committee? Why should ArbCom - which, for better or worse, the community has at least elected - not be trusted (or be trusted less than any other body) to fairly analyse consensus? (Though that said, perhaps there's a case for separating two roles, based on two kinds of trust - there are some people we can trust to have access to confidential information, and there are some (not necessarily the same people) we can trust to oversee and enforce the decision-making-by-consensus process. In particular, the number of people holding the first role ought to be limited to a practical minimum, while with the second, it may well be helpful to have a larger number sharing the load.)--Kotniski (talk) 10:53, 18 September 2010 (UTC)

It seems my arguments ARE WAY too old in relation to ArbCom today. I just thought that if ArbCom's lesser powers were separated into smaller sub-committees it'd lighten the workload of Arbitrators themselves to focus primarily on WP:DR. —Ғяіᴆaз'§Đøøм • Champagne? • 05:52, 19 September 2010 (UTC)

I've made comments over at Village Pump, I still believe that separating ArbCom's lesser powers and giving them to a sub-committee or a new committee altogether would be a net benefit since ArbCom will be able to focus on the dispute resolution itself and make suggestions for the remedies after the dispute has been resolved and all parties have come to a compromise. —Ғяіᴆaз'§Đøøм • Champagne? • 8:33pm • 10:33, 21 September 2010 (UTC)

In the stuff we deal with, the parties rarely compromise and sometimes the disputes have to be resolved by obliging them to disengage.  Roger Davies ^talk 11:13, 21 September 2010 (UTC)

Yes; as Roger says, the sorts of disputes that actually reach us are by definition ones where the parties are unwilling or unable to compromise. The (voluntary) mediation process exists to handle disputes where the parties are willing to cooperate despite their disagreements; the arbitration process, on the other hand, involves the imposition of binding remedies precisely because some people simply refuse to coexist peacefully. Kirill ^{[talk] [prof]} 01:41, 22 September 2010 (UTC)

I understand that but currently people are made to FEAR ArbCom with the prospect of edit restrictions and the like, a one-month trial of the Policy Committee and a feasibility study wouldn't hurt would it? —Ғяіᴆaз'§Đøøм • Champagne? • 7:05pm • 09:05, 22 September 2010 (UTC)

A few mostly copy-editing suggestions

Please consider these copy-editing suggestions; some of them are very fussy, but the policy is not changed often, so there's reason to rake over the text before the referendum. The only real qualm I have is the retention of the Workshop page, which I believe is very damaging to the administration of justice and what should be the goal of reining in the bloat.

Will the "Supplementary documents" be included in the referendum?

(Later: I just noticed you're supposed to make such suggestions as ref tags on the page; sorry, now I've written them here, I'll insert them under a collapsible banner.)

Copy-editing suggestions

• Scope and responsibilities
  • Comma after "serious"?
  • "And/or" could be just "or".
  • Numbered-list formatting: The semicolon endings followed by an upper-case start (lower-case would be ideal for a sense of single-sentence flow, IMO, or no punctuation at all? But this is no big deal, although I see lower-case is used, inconsistently, in Prior dispute resolution). Putting that aside, there's an inconsistency in that every item starts with "To ...", but just below, in "Conduct of arbitrators", the "to" appears just once, at the end of the lead (much neater, IMO). In "Selection and appointment", the boundary is a comma rather than a semicolon ("... data, and (2) ..."); could it be consistent throughout?
• Selection and appointment
  • "The election format is decided by the community." The running of the election, as well as merely its format, has always been done by the community, too. Is this clear?
• Recusal of arbitrators
  • "with or without giving a reason"—this makes me realise that "with" and "without" have quite different grammatical functions. "with giving a reason" doesn't work, even though the phrase looks nice and neat at first. "whether or not a reason is given" resorts to the passive, but I can't think of anything better at the moment.
  • "Arbitrators must recuse in any case, or from any aspect of any case,"—"in and then "from". You might consider "recuse from", which avoids the other meaning of "in any case", too. Perhaps this: "Arbitrators must recuse from a case, or from any aspect of a case," to avoid the repetition of "any", as well.
  • "arbitratorial"—it's even hard to pronounce (go on, say it!), and is not in my dictionary. Not the sweetest English word. You wouldn't consider just "Previous arbitration or administrative interactions are not usually grounds for recusal"? Here, you may decide that "arbitration" is OK as a qualifier of "interactions", even though it's usually a noun.
  • "An editor believing an arbitrator should recuse ..."—you might consider "An editor who believes an arbitrator should recuse".
  • "Should the arbitrator not respond, or not recuse, the user may ..."—The first comma could possibly be dropped.
• Procedures and roles
  • "The Committee may create or modify its procedures, providing they are consistent ..."—Could I suggest "provided", which removes fleeting ambiguity that the Committee might be going to provide something in the next clause.
  • "Where appropriate, the Committee will invite community comment on intended changes prior to implementing them." I'd have gone for "may invite", since the Committee is charged with determining appropriateness in the first place.
• Jurisdiction
  • "The Committee has jurisdiction over all disputes on the English Wikipedia concerning user conduct or users who hold advanced permissions"—niggle niggle ... so it has to be a dispute first? A member of the community has to complain first? Also, "user conduct" is paired with "users who hold advanced permissions"; I'm unsure about what that really means in relation to both groups. Perhaps, "The Committee has jurisdiction over the conduct of users, including those who hold advanced permissions, on the English Wikipedia". Did I understand correctly?
• Forms of proceeding
  • Under Summary proceedings, can't the second sentence ("To pass, the motion requires the support of a majority of all active non-recused arbitrators.") be binned as repeating the overarching statement below? ("Decisions are reached by a majority vote of active, non-recused arbitrators.")
  • "may resolve the matter in private"—please consider "a matter", which removes a slight tension with the preceding "exceptional circumstances".
• Participation
  • The opening sentence might deserve to be the first para per se, not joined to the next bit about new and retiring arbs.
  • "Editors who are parties to proceedings are expected to respond to allegations, and failure to do so may result in a decision being made without their participation." Is the second clause not redundant? It seems to weaken the first clause. And this is more fully covered in "Participation" under Supplementary documents". Do we need both sections?
• Policy and precedent
  • Perhaps a comma after the first "content".
• Ratification and amendment
  • This policy will undergo formal ratification through a community referendum, whose format will be decided by the community." But it does seem that the format has already been decided, in the very same paragraph.

Supplementary documents

• Arbitrator activity and voting
  • "An inactive arbitrator may become active by voting on any aspect of a proceeding." Users have expressed frustration that the number of votes required for a majority is not updated by the clerks when an arbitrator who listed as "inactive" suddenly votes in a proceeding. It's a little messy. Can it be cleared up?
  • Is the use of preferential voting precluded? IMO, this is a long-neglected possibility for allowing the more nuanced preferences of arbs to be expressed, for example, for lengths of topic bans, rather than the rigid choice of two or three possibilities, yes or no, with preferences often expressed in words by individual arbs?
• Prior dispute resolution
  • The "or" should be at the end of the second point after a semicolon, not the start of the third point.
• Opening
  • "More than 24 hours must have elapsed ..."—please remove "must" (as in the third bullet. I wonder why these are bullets rather than numbered points, as elsewhere.
  • The Workshop page is an invitation to bloat. I can't imagine why the Committee is interested in anything but evidence on the formal Evidence page, unless the Workshop page is true to its original design as a place for arbitrators alone to workshop. Once the parties are let loose on a Workshop page (no guidelines for it, either), it is likely to be very damaging to the administration of justice—to encourage bad blood, not cooperation. What is the relationship between the Evidence talk page and the Workshop page? This is the only aspect of this draft I have strong misgivings about.
• Participation
  • Possible comma after "bureaucrat"?
  • I think "decided" would be better than "directed".

The policy draft seems pretty much ready to go, apart from these little points. I wonder whether users might be invited to make any final comments, since it would be good if the referendum were not held too close to the ArbCom elections, which will probably start early November as they did last year. An RfC on the General Questions for the election will probably be necessary before they start, too.

• Evidence
  • I'd remove "about", or some parties will game the system. Very good section!
• Voting
  • "on on-wiki"—Perhaps the first "on" is unnecessary.
  • "with a separate vote for each substantive clause", then "Where several substantive items are combined in a single clause, they may at the volition of any arbitrator be split into separate clauses for voting." These points seem to contradict each other. Does the drafter need to be tied down so explicitly?

Apart from this, the policy draft seems pretty much ready to go. It would be good if the referendum were not held too close to the ArbCom elections, which will probably start early November as they did last year. An RfC on the General Questions for the election will probably be necessary before they start, too. It would be nice if the RfC and the referendum did not overlap.

Thanks; I think many editors appreciate the work that has gone into this and previous drafts.

Tony (talk) 12:06, 25 September 2010 (UTC)

Break to reduce scrolling

Thanks, Tony. I've given these a once-through and, as some need a bit of thought, I'll respond tomorrow when I'm fresh.  Roger Davies ^talk 15:33, 25 September 2010 (UTC)

Now done, in one hit, so a single diff will show the changes. I won't comment on individual changes at this stage but if you have any observations could you please put them in a new collapsed section, preferably numbered for ease of response? Many thanks,  Roger Davies ^talk 05:58, 26 September 2010 (UTC)

Why did you remove the comments regarding the "jurisdiction" sentence, saying you agreed with them, but leaving the objected-to sentence exactly as it was before? Was this intentional?--Kotniski (talk) 09:45, 26 September 2010 (UTC)

Yes, it was intentional, but on re-reading your reffed comment in the light of your comment above, perhaps you were suggesting that the scope and jurisdiction sections be consolidated to remove ambiguities and duplication? If so, that seems like a good plan.  Roger Davies ^talk 09:59, 26 September 2010 (UTC)

Yes, for me "Scope and responsibilities" and "Jurisdiction" are such similar concepts in this context that they ought to be covered by a single section, or at least by two consecutive sections.--Kotniski (talk) 10:08, 26 September 2010 (UTC)

I'll work on that later. Thanks,  Roger Davies ^talk 10:24, 26 September 2010 (UTC)

I've given this some thought and can't think of a way to make the two sections into happy bedmates. Easiest to delete the reference to permissions in the jurisdiction section as it's a bit of a red herring there.  Roger Davies ^talk 21:14, 30 September 2010 (UTC)

Recusal

Roger Davies has suggested this text to define when administrators are "involved":[6]

Examples of conduct that often lead to the appearance of administrator bias include (but are not limited to) significant or recent:

1. participation in an editorial dispute with one of the editors;
2. participation in an editorial capacity in a content dispute affecting the same topic or similar topics or the broader topic;
3. personal interaction with one or more of the editors.

That is markedly different from the recusal language in this draft and in the existing arbitration policy. Are there substantial reasons why the policies for recusal by admins and arbitrators should be significantly different? If so, perhaps those should be explained in the policy.   Will Beback  talk  20:28, 12 October 2010 (UTC)

I believe those types of incidents are already covered by the current proposal (but feel free to propose a better wording if you disagree). For example, "significant personal involvement with one of the parties" would be either editorial disputes or personal interactions while "prior personal involvement in the substance of the dispute" is the participation in the dispute (editorial or sometimes otherwise). Shell ^babelfish 04:12, 14 October 2010 (UTC)

Thanks for your reply. Unless there's a reason for using different language it seems logical to use the same or similar text. The (draft) standard for admins is "appearance of bias", while the (draft) standard for arbitrators is "significant conflict of interest". That seems like a much higher threshold. Would there be a problem with having it say something like, Arbitrators must recuse from any case, or any aspect of a case, in which they have a significant conflict of interest or there could be an appearance of bias.   Will Beback  talk  04:55, 14 October 2010 (UTC)

I'm concerned that simply stating "appearance of bias" would be an invitation for wikilawyering, especially with no qualifications as to who makes that decision or what kinds of things would constitute an appearance of bias. I'd be concerned about the same thing in the administrator policy - I've seen many circumstances in which repeatedly problematic editors have tried to remove admins using similar terms. Shell ^babelfish 05:30, 14 October 2010 (UTC)

Fully agree with Shell here; I've seen this attempt to be gamed before, as an attempt to try and select the "jury", with the explicit purpose of trying to avoid certain arbitrators with a reputation for being more (or less) supportive of certain types of sanctions, or being more (or less) tolerant of certain behaviours, for two examples. Risker (talk) 05:46, 14 October 2010 (UTC)

(od) No, there isn't a different threshold because there's no threshold at all. We're talking about two entirely different things. The draft "involved admin" text is in great part about advising people how to minimise problems in the future by avoiding actions that may create an "appearance of bias". Your suggestion here turns that into past actions which arguably created an "appearance of bias". 06:16, 14 October 2010 (UTC)

I may have misunderstood that clause of admin policy proposal. My general point is that if the standards are similar then using consistent language in the policies would make them clearer for everyone involved. I don't think anyone would want the involvement of either admins or arbs who have an appearance of bias.   Will Beback  talk  07:16, 14 October 2010 (UTC)

This is a really knotty one. Who is to declare the appearance of bias? What does it mean? That a couple of vociferous users are shouting bias from the rooftops, with claims that the admin/arb must have been biased to arrive at the decision they did? Admins working in contentious areas or with tendentious editors or with feuding factions get this all the time. Should they be driven off? The strong consensus at WT:ADMIN is "certainly not" and I don't think anyone would say the contrary.  Roger Davies ^talk 08:06, 14 October 2010 (UTC)

@Will: the fact that there are many other admins one can ask for a second or third opinion, and to perform or decide not to perform an admin action in question, changes the matter considerably. ArbCom does not have infinite human resources to mix and match, subsitute, recuse and the like—and nor should it. I do not think it is wise to consider WP:UNINVOLVED as a benchmark. <Aside: I would prefer that once an arb announced their recusal, they did not participate in discussions on the matter—in or out, not halfway house—but it would almost certainly be clunky to include in the policy, so I'm not proposing it.> Tony (talk) 08:25, 14 October 2010 (UTC)

The pool of active admins is not infinite, though certainly larger than the pool of active arbs. (Especially in the later months of the year.) Everyone on the arbitration committee has been involved in many topics and with many other editors, as have all of the most experienced admins. Rather than size of pools, the more important distinction is that admins work mainly as solitary "circuit court judges" and sheriffs, rather like Judge Roy Bean dispensing frontier justice, while the arbs are more like an appellate court justices who only act as a body and make decisions without enforcing them. For that reason there should be a somewhat lower standard of "appearance of bias"/"conflict of interest" than the admins who carry more personal responsibility. How much lower? And can we express it in similar language to the admin policy, or vice versa?   Will Beback  talk  08:59, 14 October 2010 (UTC)

Actually, I'm a bit confused with the reasoning here. Admins, who work alone, often go unsupervised and unscrutised for long periods, and can (and sometimes do) get up to all sorts of mischief. In contrast, arbitrators have no individual powers, work in one of the most scrutinised areas of the encyclopedia, and make decisions publicly en banc where any bias is diluted by the votes of their colleagues. Why does therefore does the bar need to be lower for arbitrators? Why does there need to be an avoidance of "apparent bias" provision at all, given the impossibility of predicting what might be seen by whom as an appearance of bias?  Roger Davies ^talk 11:14, 14 October 2010 (UTC)

I agree that the dynamics involved are different between arbitrators who vote as one among many, and admins who work alone. The best way to address entrenched Community conflicts is for a large number of arbitrators to vote in the case. Lowering the threshold for recusal as the potential to eliminate all arbs in some situation because of ArbComs contact with long term established users, particularly those who have special permissions such as oversight and checkuser, or who are heavily involved with areas that overlap with ArbCom's work. The ability to appeal to AUSC helps since it has Community seats. But overall we depend on ArbCom members to be well chosen by the Community at large and to represent the broad consensus of the Community that elects them. FloNight♥♥♥♥ 13:00, 14 October 2010 (UTC)

Don't we want to avoid the appearance of bias by arbitrators? In a recent case, just 5 votes were needed to pass remedies, so if there is bias it might not be diluted very much, especially if it's shared by more than one arbitrator. Getting back to the point of this thread, can we make the language in this policy more similar (but not the same, for the reasons given above) language as the admin policy? It just seems like they cover essentially the same issue and we should use comparable language for both. It could include an explicit statement that appearance of bias is less of an issue for arbitrators than for admins.   Will Beback  talk  01:17, 15 October 2010 (UTC)

Yes, we do wish to "avoid the appearance of bias in arbitrators" but as we can't control what the perceptions of others are I'm not sure it's possible. In the recent case you refer to, by the way, such a provision in policy would have been gamed to Kingdom Come as various players tried to neutralise arbitrators or, once they saw the way the case was going, neutralise ArbCom altogether. But, I repeat, the sentiment in the Involved Admin draft is guidance about types of behaviour to avoid as the conduct might give raise to the appearance of bias. It does not say that admins are involved simply because there's an appearance of bias, which is the language you are trying to introduce here.  Roger Davies ^talk 07:37, 15 October 2010 (UTC)

Will Beback, you would be introducing wording that is unpractical and would stop ArbCom from functioning in situations where the Community is less able to resolve an entrenched problem or does not have the information to solve the problem. Instead, we need to elect arbitrators who we trust to know when to recuse because they are too involved to make an impartial decision, or who remind other Arbitrators to bow out if they are unduly influencing a decision. FloNight♥♥♥♥ 09:50, 15 October 2010 (UTC)

Comments on October 2010 update

Since this is a cleanup of previous versions there isn't much to add. The following points were not incorporated in this draft but I think ther should be re-visited carefully, as this policy must be capable of forming the basis of Wikipedia Arbitration for many years to come.

1. Procedures and roles - The word "supervision" in this sentence could be improved - they don't just supervise them, the Committee actively manages the content. Worth clarifying given past cases centering on disputes over "who can edit what page on-wiki". Do not want to see dispute over AC's right to manage its own pages. Suggest replace by "the operation of..." or "the management of the Committee's noticeboard pages" which avoids doubt. Ditto suggest "subpages" not "procedural pages" which also avoids doubt (which pages are "procedural" anyway?). The sentence can then be simplified: - "These include the management of arbitration cases, and of the Committee's noticeboard pages and the content of other Committee subpages". Simpler and no ambiguity. (Note also missing word "the").
2. Ratification and amendment - although Roger stated above that he did not agree, I would urge reconsideration of this point. In a few limited areas the community probably shouldn't amend the policy without much stronger endorsement. For example if there was a sufficient vote following some cause célèbre, that privacy issues should be explained in full on-wiki or disclosed to all parties, it could be a problem. I would still urge to add this, or some wording with the same effect: "Changes which affect the scope, responsibility, internal procedures, or jurisdiction of the Committee require (some kind of larger majority or higher criterion)". Rationale is that these are crucial headers which together make Arbitration effective in helping the community, and safeguard participants' and third parties privacy rights. The jurisdiction and areas ArbCom can be involved, its responsibilities, and the way it works internally, are all fundamental to its operation. Some higher hurdle is needed to ensure these are almost impossible to "pull the teeth of" without some better process arising. Not convinced the current criteria are sufficiently preventative to future whim.
3. Policy and precedent - "interpret" should be "interpret and reconcile". More seriously the following isn't included, as a routine kind of action in some very divisive community issues. "In cases of exceptionally divisive issues, the Committee may create temporary processes and measures to address their disruptive effect or to reduce conflict pending community resolution". This needs explicit statement to avoid future drama over whether ArbCom is "making policy". The policy as worded excludes this - it merely allows that AC may "propose" for "content disputes", may "call attention" to policies, or may "enforce policies". Some of the most divisive issues with strong feelings on all sides have been ones that these would arguably not apply to. ArbCom really needs this ability in its remedies toolkit.

That's it, and good work! FT2 ^{(Talk | email)} 07:10, 16 October 2010 (UTC)

"with a separate vote for each clause"—Most sentences contain several clauses. I think a better word would be "point". (Even though I disagree with even that much pointillism in the statement of principles). Tony (talk) 01:55, 17 October 2010 (UTC) PS "Once voting on clauses appears to have ended"—Can "on clauses" be removed? Is it not clear what "voting" refers to? If absolutely necessary, it could be "if voting on [A, B, C, and D sections] appears to have ended". Tony (talk) 01:58, 17 October 2010 (UTC)

Adoption

Will this be adopted or ratified? It's a big improvement over the current policy, and months of effort have gone into it. It'd be nice to put it into place before the next ArbCom is seated.   Will Beback  talk  21:09, 2 November 2010 (UTC)

Given the schedule constraints, we're looking at putting it up for ratification at the beginning of next year, once the new arbitrators are seated. Kirill ^{[talk] [prof]} 05:57, 25 November 2010 (UTC)