Wikipedia:Headcount alone does not constitute consensus

A group of well-versed editors preparing to launch a well-aimed snowball into an otherwise unanimous !vote

The process of building consensus, whether through discussions, silent consensus, or the bold, revert, discuss cycle, is a vital part to the curation of content on Wikipedia. Disputes that pop up are almost always solved through community consensus instead of office action by the Wikimedia Foundation, and in this way, Wikipedia has a unique community of editors curating an encyclopedia of content written in a much more neutral and methodical way compared to many other places across the internet. Arguments about content are often settled using the !vote process on talk pages, using bolded support and oppose votes. When an uncontroversial change appears, and a desiring editor chooses not to be bold and change it themselves, a discussion opens. Many times, these end up being stacked with supporting !votes, or end up ignored - in both cases, consensus is unanimous or assumed to be through silent consensus, and, to avoid holding up uncontroversial changes through bureaucratic means, discussions are often closed using WP:SNOW - the snowball clause, which states some variant of "the argument has a snowball's chance in hell of failing" to close discussions much earlier than they would have otherwise.

Some dissenting arguments made in the later stages of otherwise unanimous discussions can raise good points, arguments with solid foundations that should not be ignored. Discussions closed prematurely through the snowball clause should avoid not considering the arguments of these "well-aimed snowballs" when closing through the snowball clause, as doing so can create a shaky consensus that is ripe to be challenged later down the line. There is no deadline for discussion to end and you need not be hasty to "save other editors' time" if the tide of consensus could reasonably change with the newer arguments that appear.

Well-aimed snowballs

How would you handle this discussion? Would you close it prematurely through the snowball clause, or would you read through each argument and weigh them before leaving a !vote of your own?

The snowball clause is one of Wikipedia's most infamous and useful concepts; after all, who doesn't want to close discussions early when a clear consensus has emerged? This is the romanticized version of the snowball clause: a field of bolded support votes and no opposition in sight. In practice, most discussions that would end up in this way don't end up making it to a !vote as editors, whether consciously or not, are bold and make the change themselves, and, in many cases the snowball clause is cited in the close of, the discussion ends up having some opposition.

The humble snowball may seem to be a poor contender to survive the scorching flames of hell on its face, but in practice, many discussions with a unanimous or nearly unanimous consensus have some reason to scrutinize the support of. New arguments can be added at any point during an open discussion, and simply being added late is not a reason to ignore an argument outright. There is no rule or policy against dissenting against an argument with unanimous consensus (or, in other words, a snowball's chance in hell of passing). These dissenting arguments, which will hereafter be referred to as "well-aimed snowballs", should not be ignored for the purposes of invoking the snowball clause.

For instance, in discussions in which a majority of !votes in support have no explanation or arguments behind them, an argument of opposition should be given extra weight, especially if it's a more concrete argument to begin with. If these arguments are added late in an argument, discussion should not be closed early, as 10 support !votes to 1 late oppose !vote can shift the consensus over time and can eventually end up as 10 supports to 15 opposes; when closing discussions per the snowball clause or through strong consensus, keep an eye out for these well-aimed snowballs and consider allowing discussion to continue longer for more editors to consider the dissenting viewpoint. If a well-aimed snowball ends up lasting in hell, recognize that closing the discussion will require acknowledging significantly more nuance than most other discussions, as you will need to justify why a single argument is able to override the head count of numerous other editors, even if it may seem entirely clear to you the well-aimed snowball will last.

Specific cases to doubt unaniminity

Sockpuppetry

Sockpuppet accounts are those created by an editor, whether in good or bad faith, that are separate from their primary account. In the context of discussions, sometimes these accounts are used in bad faith to give more support to a proposal. Sockpuppetry does not aid any side in a discussion, as it does not add anything that the single editor themselves couldn't bring alone, and could be argued to prove that at least one editor on one side is willing to act in bad faith to strengthen the headcount of an argument.

Behavior in discussion is often used as evidence in sockpuppet investigations. If you are certain sockpuppetry is occurring in a discussion, you can strikethrough the comments of a sockpuppet by wrapping the <s> and </s> markup tags around their comment, and make it clear the removal was due to sockpuppetry.

Meatpuppetry and canvassing

Meatpuppetry and canvassing are functionally identical concepts in discussions, where an editor increases participation in a biased way. These almost always have the goal and result of boosting one side of an argument. Having an unbiased group of editors to discuss is a vital part of consensus building and if that process is disrupted by a surge of biased editors there is much reason to doubt apparent unaniminity.

Temporal changes

Changes in time can influence the way editors generally feel about a topic. This can come about for many reasons, including the facts of the argument changes, and new arguments that convince new participants in discussions. Nobody is required to participate in a consensus-building experience, and those that do are by no means mandated to come back later on when the facts change, so consider asking yourself whether people making one argument relatively early on in an argument would feel the same way later on. This does not mean that old voices should just be ignored or that the substance of opinions should be interpreted differently, but arguments should be weighed accordingly when either passing to participate in discussion or closing it. Wikipedia has no deadlines so there is good reason to not close a discussion early if there is reason to believe the consensus being built will change drastically over time when considering more recent arguments.

Good aim can't save all snowballs

See also: Wikipedia:One against many

While it may seem that any well-aimed snowball is an unstoppable force in its own right, consider that many times the snowball clause is invoked, many editors simply don't want to overcomplicate their own writing, so if an apparent well-aimed snowball shows up, don't be surprised if editors on the same side as the majority headcount respond with equally concrete arguments that had not been made before. Alternatively, a consensus discussion that may seem a prime target to a well-aimed snowball may in fact be built on a concrete argument that no snowball, regardless of aim, will end up preventing the consensus of. Not all discussions have or even need a well-aimed snowball - in many cases, the snowball clause can be used to institute consensus on the foundation of concrete arguments.

See also

• Wikipedia:Polling is not a substitute for discussion
• Wikipedia:Wikipedia is not a democracy