Wikipedia:Governance reform

This page is currently inactive and is retained for historical reference. For more and current information on Wikipedia reforms please go to Wikipedia:Reforms.

Shortcut

WP:GVR

Failure of the traditional governance model

Has Wikipedia's traditional governance failed to scale adequately? With the project's growth, has the model adopted on a much smaller wiki now become ineffective?

Some major internally driven policy proposals, such as the attribution policy, have failed to result in anything. A "no consensus" outcome perhaps might have been predicted. Internally motivated policy formation, it seems, has effectively stagnated. The last major changes to the main body of policy—the BLP policy and the non-free content policy—were both essentially imposed from the outside. The impression is that external pressure on the project means more than internal debate.

Policy debate becomes, in most cases, nothing more than an endurance contest between those who wish to effect some change and those who wish to retain the status quo. This is the land of the filibuster: so long as those opposed to any proposal are sufficiently dedicated and sufficiently vocal, they can keep the debate going without any effective means being available to force a decision. The few attempts to do so by means of a general referendum have proven ineffective. Is this in some sense, inevitable in a project with a perpetually open set of participants?

Policy formation thus essentially devolves to three processes. These cannot be regarded as well-constructed means of governing a project of this scale.

Policy changes by edit-war: Objecting editors simply burn out from the effort of opposing a group bent on forcing through change. Small-scale changes to existing policy are typically made by one or several editors. They quietly make changes to the policy, or more actively edit-war to force in certain changes over the objections of other editors. Here, as in more general debates, sheer persistence will often overcome opposition.

Policy changes by The Death of a Thousand Cuts, or Shifting sands: Persistent editors make small but frequent edits which seem innocuous individually, but over time accumulate to a radical change in policy.

Policy changes by the Arbitration Committee: The Committee is not actually authorized to change policy. Changes are typically presented as "interpretations" or "clarifications" to what is presumably already agreed upon. The Committee affects policy through the rulings it adopts in its decisions. Its approach is unduly colored by the specifics of disputes before it. This can lead to sporadic and even contradictory attempts at changing policy. The Committee is not, therefore, well-suited to play the role of policy-maker. In short, the Arbitration Committee does what it suggests, arbitrate. An analogy would be the judicial branch of a government.

Where is the feasible way to change outdated and unscalable processes? The lack of an effective, community-oriented policy-making process has caused a variety of other unfortunate consequences. Among others:

The dispute-resolution process has largely collapsed into a single arbitration step. The various preliminary and community-driven methods are either no longer effective or have been abandoned.
Policy pages are often dominated by small groups of editors, who actively resist any change away from their favored position.
Attempts to develop new policy sometimes degenerate into outright conflict among groups of editors. Consider, for example, the various proposals regarding "attack sites".

Adopting a new model

What are the methods used by other groups of a comparable size in the "real world"? The obvious solution to the lack of an effective policy-making process would be to adopt something already known. The most obvious way—and perhaps the one most suited to easy adoption within the constraints of Wikipedia—is the elected decision-making group.

One option would be to push policy-making responsibilities onto an existing elected group. The Arbitration Committee is perhaps the closest thing Wikipedia has to a "governing body", and could potentially be used in such a fashion. ArbCom is not, however, well-suited to exercising both a judicial role and a legislative one; and it is, in many ways, simply too small and too overworked to function effectively in the latter area.

An alternative would be to convene a dedicated policy-making body (the "Wikipedia Assembly", perhaps?):

The body would contain a sufficient number of editors (at least twenty) to serve both as a reasonable cross-section of editors, and as an effective deliberative forum.
The body would be freely elected, eventually via a tranche system that would allow overlap between successive terms.
When the tranche system is in place, terms would be for two years, as for the Wikimedia Foundation board. Initially, half the seats would be for one year.
The electorate would be determined in the same way as for elections to the WMF board, except based solely on contributions to the English Wikipedia.
The body would be given essentially unlimited authority to set project policy, so long as it is in compliance with Wikimedia Foundation policies and resolutions, the relevant laws, and so forth.
The body would set its own agenda. Any member would be allowed to introduce discussion of any specific policy page.
The body would make decisions by vote, initially using simple majority.
After initial discussion, the body could vote to reserve any policy page to itself. Once so reserved the page would be protected and only changes authorised by the body would be allowed, until such time as the body voted to release the page. The current consensus policy-making would continue on all policy pages not currently reserved.
The body would have authority on par with the authority Jimbo Wales has held historically, and would replace him in this role (but not in his role as the appointer of Arbcom members).
Only something truly massive such as a site-wide no-confidence action or intervention by the Wikimedia Foundation (not any one Wikipedia-user or Board member, regardless of who they are) would have veto power over the body.

Comments

This would, admittedly, be a significant departure from the traditional "Wikipedia is not a democracy" philosophy. The alternative to democratization, however, is policy stagnation of the sort that has been the norm for the past few years; and that is increasingly untenable as the project continues to grow and grapple with ever more sophisticated concerns.

One source of the anti-democratic tendency on Wikipedia is the open nature of the community. As a result, core democratic activities such as canvassing and get-out-the-vote drives are explicitly banned, since they may recruit people who have no real commitment to the wikipedia community or its values. But this contributes to policy stagnation by undermining the claim of any given group of editors to represent the community. The proposed Assembly overcomes this democratic deficit in two ways:

The explicit rules for elections ensure that the electorate are members in good standing of the community, and represent a fair sample of it. This allows proper campaigning, canvassing and get-out-the-vote activity.
The Assembly itself constitutes a complete and well-defined electorate for policy decisions.

Wikipedians would, of course, be able to canvass the Assembly or individual assembly members on any policy issue, including suggesting draft policies for consideration. The Assembly need not, of course, take any notice.

As defined above, the Assembly would have authority over its own constitution.

The assembly might choose to require super-majority voting on some or all policy changes to encourage stability. Examples might be

Its own constitution (including voting policy)
Other elements of the Wikipedia constitution, such as the rules governing ArbCom.
Foundational policies.

By allowing consensus policy-making to go ahead whenever it seems to be working, the Assembly could avoid getting bogged down by routine details and focus on the major issues.

Other options for change

Another option is to simply institute polling on a wider basis on policy changes. Protect policy and guidance, and through the watchlist poll on proposals.

Another option is to simplify our guidance, policies and procedures so that existing methods work better.

Yet another option is for a user who lacks time to get involved in complex policy discussions to appoint a representative to participate for him (aka liquid democracy).

A different opinion: Wikipedia governance is thriving.

A number of people believe that Wikipedia policy pages are binding rules, and that it must be fought over as in a democratic state.

These people forget that there is actually a policy page that states the opposite, to wit Wikipedia:Ignore all rules. This is quite literal, no one is required to follow any rule, as strictly speaking, there are none.

Instead, people are strongly advised to adhere to the current Wikipedia:Consensus on any particular matter. You could theoretically even ignore consensus, but you typically end up getting a friendly tap on the shoulder after a little while.

The pages in the project namespace document the current consensus on what best practices are.

Some people don't like this, and perceive this as a problem. Others perceive the resulting behavior of those people to be a problem in and of itself.

The dispute-resolution process is thriving. Powerful, scalable processes such as third opinion, The mediation cabal and Editor Assistance are shouldering the burden. Despite that fact, and in contrast, the arbitration committee has always had trouble shouldering its burden, because having a single committee-bottleneck simply does not scale.

Policy pages are typically maintained by a large and diverse community of editors who visit and help maintain the pages. New editors join in and help out. There are situations where people think it is appropriate to filibuster. Currently these situations are slowly being corrected. This process is thriving, with hundreds of policies, guidelines and essays to provide advice to people on every single subject from how to convince the community to change course to how many spaces should follow a full stop.

The policy proposal process has long been deprecated, though several die-hards continue to try to (ab)use this system. As long as these attempts don't harm the wiki, they are still tolerated, however. Almost all such attempts rightly fail.

One of Wikipedia's great assets is the inherently local nature of most editing decisions. The difficulty in passing new policy serves as a check on global instruction creep, and keeps policy decisions as local as possible.

Specific proposals

Please feel free to review and comment on any of the particular proposals listed below: