User:Dan Polansky/Consensus

What is consensus and what is it in reference to a wiki?

The dictionary definitions are as follows, as per “consensus”, in OneLook Dictionary Search.:

• "Agreement; accord; consent." Webster 1913
• "A general agreement or concord: as a consensus of opinion." Century 1911 (the current jpeg)
• General agreement (Along the lines of this.) MWO, 1a, Collins, AHD 2

When applying the government by consensus to a wiki dictionary, I see the following principles as acceptable by me:

• 1. In order to add a translation, an etymology or a definition to a particular page, the editor does not need to file a request for edit in a dedicated process page. Instead, unless there already was contention about the content added somewhere on the wiki such as on the talk page or in the edit history, they can directly edit the page without talking to anyone else before. They can do so even if they are an anonymous IP editor.
• 2. In order to start editing, an editor does not need to do any proving about who they are, nor any declarations of who they are. They do not need to disclose their age, sex, lack of education, lack of work experience, and the like. They can even edit using an anonymous IP. However, this is not really a necessary precondition for government by consensus on the wiki, since the consensus (general agreement) could really be among the eligible editors. But it is a typical requirement made as part of an extreme openness of Wikipedia-like wikis.
• 3. In forming of decisions, differences between editors are largely not taken into account. Thus, being an admin, being a senior editor, being an old-aged editor, having a putative wiki-designed rank (violet belt, whatever) or having an important social role in real life does not play a role. In particular, there are no wiki ranks allocated, not even for fun.
• 4. Ideally, when a disagreement about a particular page arises, the decision about the page can be made by discussion which might lead to a near unanimity. In that discussion, people are required to make substantive arguments or support particular substantive arguments made by others rather than voting bare "support" or "oppose". If near unanimity cannot be reached, something like 2/3-supermajority has to be considered good enough, especially if fairly many editors take part on the discussion.
• 5. Consensus is not unanimity. It is unrealistic to expect that every single editor taking part on a discussion will agree. Something like 2/3-supermajority has to be good enough, with caveats. One caveat is that if people vote and provide no rationale or reasoning, going by 2/3-supermajority does not look like a proper consensus-forming exercise. That being the case does not really depend on 2/3 or 8/10 being the threshold. On the other hand, 8/10 support created by voting with bare votes is still a "general agreement". Another caveat is that 2:1:0 (support:oppose:abstain) supermajority is not much of evidence of general agreement; 2:1:13 is better but still questionable.
• 6. Making large-scale changes requires express discussion unless there is a high probability that a supermajority agrees with them. For instance, if the overwhelming majority of pages uses "from" rather than "<" in etymology chains, an editor can mass convert "<" to "from". However, in that case, the requirement that things be explained and reasoned about is violated, as long as the person doing the changes does not post a rationale somewhere.
• 7. Making large-scale removals requires a vote lasting a month or so, since that is the most reliable method (or the sole one, actually) of finding out whether there is a general agreement among editors--even those who are temporarily absent from the wiki.
• 8. Policy pages are created to avoid repeated explaining and arguing the same thing that was already agreed on by general agreement of editors. Ideally, they can only be enacted and updated via a vote. Even without policy pages but an enablement of, say, RFD process, editors in RFD can present their arguments, and in doing so, start formulating pieces of their personal policy. Such a RFD could work acceptably well.
• 9. At least one policy probably needs to be enacted by a superpower, such as the policy that the government of the wiki should be consensus based. The superpower needs to be silently present across the wiki, presenting an implicit threat that if the decision making would quit following the consensus principle in a major way, there would be an intervention that is not open to discussion and consensus-forming exercise. Even if there is general agreement among the editors that the decision making should no longer proceed by the principle of seeking general agreement, that general agreement is irrelevant and invalid.