MeatballWiki | RecentChanges | Random Page | Indices | Categories

An EditWar is what happens when two (or more) users repeatedly overwrite each others' edits of a wiki page.

Most people would find such editing pointless, and would use their brains to find a way out of the conflict. The most natural way is to take a step back and see the world from a different perspective, what WikiPedia calls a NeutralPointOfView (NPOV). Instead of writing "Darwin was right" (or wrong), the result of taking this step could be to write "Many people believe Darwin was right, but far from everybody".

Sometimes, however, this doesn't happen naturally. WikiPedia has taken the measure to raise NPOV to an official policy, a rule of conduct, to which new users are pointed. Ultimately, users who fail to follow such rules might be blocked from editing the page or the entire site.

On SusningNu, two users, one ultra-right wing, the other leftist-anarchist, both hiding behind pen names, constantly fall into edit wars over controversial topics. In order to be liberal and to be able to study the phenomenon, I, the BenevolentDictator of SusningNu, have not tried to stop them. It is clear that the state is not contagious--other users are not becoming infected with the disease. It also doesn't seem to heal by itself, but has been going on for months already. After the November 20, 2002, "fork" by some SusningNu users to the SwedishWikipedia, these two users are now fighting the same edit wars on both sites. The SwedishWikipedia is still very young, and hasn't started to enforce any of the rules set forth by the other WikiPedia websites. I will continue to study this case. -- LarsAronsson

One problem with wikis is the tendency of amateurish contributors to steamroll over thoughtful contributors. This problem is a special case of the EditWar enigma. What makes amateurish steamrolling particularly problematic is that it discourages thoughtful contributors from participating in wikis while the amateurs are encouraged by the success of their steamrolling strategy. This happens in other aspects of life too: idiots win because they give more of a damn.

Closely related to the problem of steamrolling is the problem of arbitrary censorship by GodKings. To wit, most wikis are hosted by private parties who can impose their will on contributors. Assuming that a particular wiki's content is freely licensed, aggrieved parties can simply exercise their RightToLeave when their contributions are arbitrarily censored, or even exercise their RightToFork. Even so, forking is a necessary evil, not a desirable outcome. -- DavidPrenatt

See also: WikiPedia:Wikipedia:Edit_wars_in_progress

On WikiPedia, RK (not to be confused with RichardKulisz), a pro-Israeli user, has been active since December 2001. Doug, an advocate of a fringe scientfic theory, has been with the project since before October 2001. Michael continues to vandalise, and he has been active since March 2003 and banned since April. Lir, albeit with some degree of reform, has been trolling since September 2002. What happens when everyone gives up arguing and lets them do what they want? Observe these changes [1] [2] made by Doug, unopposed since August 2003. They don't stop caring, they check back periodically, ready to start a fight with anyone who opposes them. -- TimStarling (December 2003)

Remember, you do not have the ability to punch them in the face, or more to the point, the ability to shove them out the door and call the cops. The only thing you can do is control whether you emotionally bond with someone or not. If the CommunityDoesNotAgree you should not, someone will bond with them, and then they will be there forever. Emotional bonds do not have to be positive. They may negative, such as being very angry at someone. Then you get into cosmic FlameWars where people will not leave. They may return just because they know they can get a rise out of you and they are seriously deprived of any social recognition (cf. WhatIsaTroll). They may return to protect their reputations (cf. PunishReputation).

Wikipedia has a lot of problems that contribute to it attracting such painful people. First, it has grown faster than it should have. Thus, it has attracted a lot of people who have not been acclimated to the community norms nor do they feel unified with others on the project. Compare here where we only accept people who feel comfortable working with us. This is because SoftSecurity does not function well with too much internal disagreement.

Second, WikiPedia allows pseudonyms, so you get a masquerade ball effect where people feel free to do crazy things. This is for two reasons. One, they feel free of social pressure because they are acting through a proxy (their avatar) that they can discard. Two, they do not feel the same levels of remorse because they feel they are only hurting other people's proxies--they cannot put a face to the name or more like a name to the nym, so to speak.

Third, Wikipedia uses HardSecurity, which is a guaranteed way to say to a wanker, "Hi, we noticed you, and that you were so important that we're going to ban you from an open system." Plus, it really agitates people, and that makes them form an emotional bond.

Fourth, people often feed the trolls by arguing with them or supporting them or transmuting their posts. After a certain point you have to give up on the person and forego any valuable text they may write. You've got to start to DissuadeReputation by canning all their writing in one place. You can do this under the auspices of waiting until things cool down before sorting through their ideas. However, if the CommunityDoesNotAgree, you will be attacked. Note that it only takes one oddball to undermine you, not the mainstream of the community. You have to bounce dissenters as they come or they pile up on you.

Fifth, a lot of people argue vehemently about banning an individual. Some people are idealists in that they insist on RadicalInclusiveness, whereas others are hardliners and wish they could nuclear bomb vandals. All of that attention bolsters the person's ego so they will feel confident they can return unimpeded. After all, half the community is behind them, right? (Even if everyone would agree they are incredibly annoying.) You'll often hear requests to prove your authority in the middle of this as they think they see the community withdrawing support. Often, it does look like the community is withdrawing its support. Not good.

All of these are debatable points, but they at least all point to an inability to control emotions and more so, the inability to control forming emotional bonds. Once you do form a bond with someone, you are in for a multiyear fight until you can eliminate the bond. This means in your mind, you have to stop caring about the person, and then act accordingly. However, on a project like WikiPedia, that means everyone has to stop caring, which is impossible to force or convince people to do, but as long as you have enough people who will clean up after those who do care so that it appears as a whole the project does not care, then maybe that is sufficient. In a way, you are working against both ends: the attacker and the friend, both who have negative emotions. You have to keep them separated.

But it's Wiki:FixBrokenWindows. The more tiresome people you attract, the less valuable people will remain, and then you won't have enough people around to defend the project. It's easier to destroy than to create, and it's certainly more fulfilling to destroy someone else's work than to defend your own (hold the gains). This is why, in some respects, you have to be merciless in your inability to care about people. If they aren't into BarnRaising, get rid of them. Life's too short. If it isn't fun, it isn't worth it. If they aren't fun, they aren't worth it. There is no ideal greater than the project's survival; anything else is suicide. (That sounds overly jingoistic to me, though.) -- SunirShah

Thank you for your reply Sunir, I'll address a few points.

I'm very grateful for Wikipedia's newbies -- most of them seem to pick up the basic principles of Wikipedia culture within a few months. A constant stream of newbies is necessary to replace the oldtimers who either leave or become averse to any kind of argument. Those who become averse to argument generally recover after a few months, but after a few repetitions, they often consider leaving.

The particular flavour of anti-social activity which has caused trouble of late is the so-called "POV-pusher" -- in other words people with ExcessBaggage. A POV-pusher might be classified as a troll, but unlike most trolls, they don't exist merely for the sake of making people angry. Hence, if everyone stopped caring, ignored the troll and went back to their business, the POV-pusher would write what they want to write, and check back periodically, making "improvements" to their own prose, and making sure no-one has changed it.

Hence I have trouble seeing an EditWar on Wikipedia as merely interpersonal. An EditWar, in the worst cases, is a clash of ideologies -- NeutralPointOfView versus an agenda. POV-pushers do not learn NPOV by simply hanging around long enough. They never accept the community norms.

Wikipedians are not prepared to accept a compromise of NPOV, they feel a responsibility to the readers (who far outnumber editors) to keep the content maximally accurate at all times. What is at stake is truth -- education of the general public in whatever the issue may be. For that goal, Wikipedians (even newbies) are prepared to sacrifice anything.

Wikipedia continues to grow extremely quickly, and its controversial articles remain roughly neutral. But it does so at the cost of a great deal of stress and contributor turnover. -- TimStarling

Ignoring a person is more active than letting them build a reputation. You just continuously evaporate anything they have done without a word to them. For Wikipedia, you might just want to quietly move controversial text to the /Talk pages with a summary like, "Doesn't seem sufficiently charitable; moving out-of-line pending further meditation." Just keep doing this over and over again. The trick is that you will actually have to refactor that text into something inoccuous and boring. Being boring is critical. If you do this with great aplomb and wisdom, you may bore the person to death with your stodginess.

A good way is to turn the invective into a point of view.

Ice cream is immoral. In Xavier 3:16, the Bible says, "Let us stone those who eat French Vanilla."

Turns into

Some particular Christians feel that ice cream is immoral, occasionally quoting from Xavier 3:16, "Let us stone those who eat French Vanilla."

Thus you achieve your neutral point of view with a small amount of effort. Eventually someone will disagree and fix it, but at least the reader is forewarned the following text is likely crazy.

Of course, you need many more people than the one person writing this content to actually do this, as it is tiresome and painful and it wastes their time which could be better applied to writing something they care about (which is the fun part). BarnStars are a good way of at least thanking people for taking time out of their lives for something so inane. -- SunirShah

We do remove content to talk pages, or just delete it altogether. The removal will be reverted, so an EditWar breaks out. The page is alternately protected and unprotected by sysops, in an attempt to force discussion. The warring parties discuss the issue on the talk page, but the POV-pusher fails to see reason. From there people get stressed, leading to editor turnover as I mentioned previously. An excellent description of the evolution of WikiPedia edit wars is at MetaWikiPedia:More_heat_than_light. The version without the profuse comments is [here].

People rarely write "ice cream is immoral", and when they do, it is easy to rephrase it like you suggest. The divisive issues occur on another level. For example, it may be that even stating what some people think is indicative of bias. People will claim that too much space is being given to what Fred thinks, when Fred is clearly wrong and in any case only represents a very small portion of the community. See for example WikiPedia:Mother_Teresa and WikiPedia:Rick_Santorum (also WikiPedia:Talk:Santorum). Arguments may also centre around page naming, for instance the AKFD case. Similarly, the Israel-Palestine issue is a minefield of conflicting interpretations, emotional anecdotes, and wild conspiracy theories. Again and again, the argument turns to inclusion, not phrasing. -- TimStarling

I think the problem with DissuadeReputation is that it assumes that all the trolls are the same. It's true that the strategy is still equivalent: convince the troll that they are wasting their own time. Alternatively, evaporate their frame of mind that they are fighting the mainstream culture, because they will then entrench their position to prepare for the onslaught. However, the tactics change depending on the person. Gather as much information as you can about the person, and then try to put yourself into the person's shoes. What is motivating them?

Some may feel the need to fight evil, in which case you present yourself as friendly and they will stop fighting you. Some may want to show that they are smarter than you, in which case you "give in" and let them feel smarter than you. Some may just be lashing out at someone to release anger, in which case you let them vent and yet empathize with them to help them through their anger. Some immature people may want attention (e.g. vandals), in which case you don't give them attention (that is where DissuadeReputation comes in).

Whatever you do, never descend the their level: do not fight (FightingIsBoring). This is easier said than done, especially if you get into an AngryCloud.

Note that there are also violent ways to exploit someone's flaws. That is, while NonViolent tactics attempt to bore the person away, violent strategies attempt to anger them away.

One way to take down a rationalist is to learn more about their argument than they know, or at least enough to complain they aren't representing the position correctly. For instance, if someone is arguing for libertarianism, provide quotes from AtlasShrugged?, interviews with AynRand, speeches from the Libertarian Party, and so on to demonstrate they don't understand libertarianism. This will catch them off guard as they are prepared to fight the mainstream, but not their rear flank. If you use a lot of ad hominem attacks to derail them, they will eventually defer to the "expert," after which you can end the discussion.

Another less violent method is play possum; ask to be taught everything the person knows. Keep asking questions. Build a ForestFire. You will tire the person out too much to refactor the conversation. (If they try, ask questions about the refactoring.) Then you will have an ugly blob you cannot put into the mainstream corpus. Further, if he tries to add more text, you have a gravity well to drag the text back into and thus end the discussion.

Violent strategies cost social and emotional capital. They are rarely worth it. Further, angering someone is a guaranteed way to keep someone around. Consider that the ForestFire is a giant PunishReputation sink that he or she would like to clean up. Until you do, you'll have to deal with them. It's better to stick to boredom.

One major problem here is that we do not yet have a general theory of trolling. This is partially because it is really painful to think about these annoying people; I already have a headache thinking about this problem. It is also partially because it is really disgusting to think we should psychoanalyze people and come up with ways to use their character flaws against them; however, it is necessary because we only have psychology rather than physicality. UniversalAccess is something to work towards, but we're nowhere close. -- SunirShah

Perhaps this discussion should move to DissuadeReputation.

There is also a Wikisite with the name editwar.com (http://editwar.com) . They seem to like editwars and maybe that's a good thing, to concentrate those wars there.

CategoryConflict CategoryJargon


MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions