Even when we want to end an unproductive relationship, it is generally not a good idea to "burn your bridges." We want to maintain the possibility of recreating that relationship in the future, or transmuting it into a new form of a relationship. When leaving a job, for instance, we may need those people as references or we may want to recommend co-workers as future hires to our new places of employment. Even quiescent relationships maintain potential. Mature adults prefer a "clean break" approach to ending a relationship as messy loose ends create conflict that will degrade the quality of the relationship in the future. To help mitigate a potential clean break in the future, many people enter a relationship with an ExitStrategy planned out. Sometimes this exit strategy is not explicit, but a learned behaviour operating from our bank of social skills; nonetheless, we remain capable of leaving when necessary.
Proper social etiquette requires one to respect other individuals desire to exit in order to avoid conflict, such letting people out of a boring conversation. People typically use ParaLanguage to signal their desire to exit, such as staring blankly, squirming, or coughing; however, ParaLanguage is not available online. One may have to resort to either catching the more subtle cues such as noticing attention wavering or positive reinforcement waning or to using more overt signals such as respect an outright claim to end the conversation without taking grave offence.
However, sometimes you find yourself needing to manage a relationship with people you do not value, and you want to engineer their departure. For OnlineCommunities, this often means some undesirable like a troll (cf. WhatIsaTroll) or a schizoid has disrupted your community. While in a physical community social cues are much more effective at demonstrating that someone is not wanted, it is easier to control access to an online community overtly through HardSecurity. CommunityExile is an easy way of ending an interaction, but it hardly ends a relationship as there are at least two parties in any relationship and you have only controlled yourself. It does not serve well to terminate the relationship "cold turkey", as they must go from highly valuing communication with you (which may be either positively or negatively patterned) to not valuing communication with you instantaneously. While you have made that decision, they might not by themselves. Then, the mismatch in desire to communicate will result in frustration and anger, increasing the probability of an attack, and destroying your ability to create a new relationship in the future.
Therefore, DissuadeInteraction rather than terminating it. Bring the other party to the direct personal conclusion that the relationship is not worth her time, rather than hoping they will make the indirect conclusion that because it is not worth your time is not worth her time. Provide a path for them to follow towards this conclusion.
There are many ways to DissuadeInteraction. One can use the person's already learned social skills as a lever. Demonstrating boredom or other discomfort, say through squirming, staring blankly, avoiding eye contact, etc. will normally end a dialogue with normal people. They will see that you are not a receptive target for their ideas, and thus not worth the cost of interaction. Don't do this frequently with people you know as that is immature, but certainly with strangers whom you incidentally meet it can be an effective psychological wall, although of some concern for urban social theorists.
You can be more structured in your approach. As a SurgeProtector works against a robot, a SurgeProtector can work against an over-zealous contributor. The more this person tries to interact with you, the more painful and tedious it becomes for that individual. As the length of time it takes for a TransactionBasedCommunity to asychronously respond increases, the more costly the interaction becomes for that individual. TomCoates? also [describes] a SurgeProtector that increasingly generates false errors; individuals will hopefully begin to think the underlying software running the OnlineCommunity is too fragile to justify discourse in that medium. However, when other good community members already aware of this practice experience actual (unintentional) system failures, they might think they were banned for some arbitrary reason (cf. GodKing), demonstrating the social problems of quietly screwing with reality--a la OpenProcess#badvogato. A CommunitySolution (or rather social solution) to the SurgeProtector family, AlexSchroeder's gives us his rule of DiminishingReplies, whereby he responds to annoying people with increasingly smaller statements than their previous statements. As the discussion continues, the conversation becomes smaller and smaller until it ends.
After a while, interaction with the community becomes more costly than any perceived benefit, and the individual will leave. As SurgeProtectors are normally gradual rather than a discrete, the individual will hopefully become acclimated to the costs. A SurgeProtector that can turn itself off dynamically if/when the individual becomes less annoying provides good tolerance and forgiveness (cf. BuildInTolerance, ForgiveAndForget). Conversely, the more overt the SurgeProtector is--particularly if it is socially controlled rather than automatic--the more obvious it is to the individual that your party is unilaterally closing the CommunicationChannel. This will result in a more frustrated and angered individual as she will not feel like she received FairProcess. That will not function if the individual perceives benefit from other aspects of the community (maybe other people who don't DissuadeInteraction with them, or article editing in ContentOverCommunity communities).
A final, much more concerted approach to dissuading interaction is to DissuadeReputation. One must be mindful not to PunishReputation, but certainly only valuing a person's valuable contributions will make unvaluable people reconsider the cost/benefit ratio of their relationship with you. After all, they will not receive much benefit (emotional validation) for their efforts. Following the metaphor of the AttentionEconomy?, this staunches the flow of attention capital; i.e. normally one dissuades reputation by ignoring the person away.
One can also DissuadeReputation by erasing that person's reputation. Quietly removing any trace of a person's interaction with you is only violent insomuch as any effort they put towards the interaction will be nullified. Even this small amount of violence should be carefully managed and not abused. It should be clear that erasing reputation is much more extreme than simply ending the relationship smoothly. You are in effect retroactively incurring costs to previous interactions, not creating new costs to future interactions. Thus, this is not dissuading interaction. Also remember that while digital media allows one to act in the past, people are not similarly pliable.
But, before you DissuadeInteraction, consider that passive aggressive behaviour is unhealthy: this is a sort of SnipingCriticism. It's better to be direct with someone than indirect. Be honest rather than subtle. Perhaps the underlying problem in the relationship is a miscommunication. Perhaps the person also feels the relationship is unvaluable, but is powerless to end it, and would himself like a reasonable ExitStrategy. Perhaps the person has a strong bond with you and will feel gravely hurt by being ignored, say like your spouse. Only DissuadeInteraction when you either have not bonded with the individual or the individual is too emotionally unstable or unaware to handle a direct dialogue. The individual can not understand what's wrong about his behaviour because nobody talks to him: that will prevent him from correcting himself, while FairCriticism can be more useful.
See also EncourageInteraction?.
Contrast DissuadeReputation, PunishReputation, RewardReputation.
It may seem contradictory to DissuadeInteraction when you value RadicalInclusiveness, but it's important to be cogniscent of the people you are really talking about when you say you are radically inclusive. There are two camps: the minorities who are traditionally oppressed and the anti-social who would normally be excluded. While most mature people will exercise their RightToLeave (i.e. manage the relationship) if they do not like the group, the anti-social people whom you are radically including do not necessarily do this. They may begin to incur costs not only to themselves but other people in the group. This is when to DissuadeInteraction. This is not to permanently exclude them but to make them realize the cost of their interaction with your group so they can decide for themselves whether it is time to go. One must not punish interaction if you are radically inclusive because you have agreed to incur personal costs in your relationships; further, you hope that a future relationship may be possible on better terms. The aim is always to find the least violent solution, keeping in mind violence to not only the individual but to the group members as well. -- SunirShah
I realized after reading Bayle's write up the problem with RewardReputation, DissuadeReputation, PunishReputation: they should only really talk about dissuade, punish, reward reputation, not people. After reading some more mitigating solutions, I wrote this "higher-level" Pattern, which also hopefully addresses some more of the ethical nuances of how to bounce people. Now I would kindly ask someone with more time on their hands to alter DissuadeReputation, etc. to talk about reputation, and more the appropriate other material here or elsewhere. This is not an easy task as the entire description of DissuadeReputation for instance is muddled with the two ideas of DissuadeInteraction and the more specific case of DissuadeReputation. -- SunirShah
OK, so first we DissuadeInteraction to create a CommunityExile. Then we can think about DissuadeReputation. The former gets them to leave, the latter demotivates the return, and reduces the echo effects of their participation. But if you try to DissuadeReputation while they're still around, then you're encouraging interaction. Just like a ForestFire, ReworkLater, once passions have cooled. --MartinHarper
Don't forget dissuading reputation before a bond is created and exile is necessary, such as the UseRealNames policy preventing PenNames.
A very interesting conclusion, Martin. That begs the question though: although CommunityExile is currently listed as HardSecurity only, is it also SoftSecurity? It seems like a rather hard choice; sure we make it more soft by adding tolerance and gradient, but the initial decision is still dichotomizing. -- SunirShah
Hard vs soft - see new para in CommunityExile
You know, I really hate this theory. I find DissuadeInteraction in practice is condescending, arrogant, and mean. As powerful as the theory of controlling emotional bonds may be in practice, but it certainly doesn't feel good as you have to constantly judge each person and ascribe either retribution or reward depending. It's also rather artificial, as no one can control whom they relate to forever. Suppose one of our ex-lovers started writing on MeatballWiki? Or worse, a professor or a boss?
I figure if the person is emotionally mature enough to be controlled through DissuadeInteraction, they are probably capable of wanting to learn to behave appropriately. The real trouble happens when you attract someone not this mature, and some are downright crazy. Early on in the WikiLifeCycle you can control for this since the only people writing on your wiki are those whom you invited personally because you know them. Later on, as your reputation grows, you attract people whom you don't know and may be people who you don't want to know.
Still most people are still highly susceptible to DissuadeInteraction, so I'm reluctant to give up on this just yet. But perhaps it should be a concept less in the foreground, and more as an undercurrent in a process that slowly tapers the person's emotional involvement in the community depending on how annoying they are. Currently it's a zero or one process, and that's not very wise; or actually, it's not very convincing. After all, the point is to convince the person they are wasting their own time, and that means diplomacy--not fighting.
However, once they are convinced they are wasting their own time, then that doesn't mean that they have to leave. They may be then primed for the better answer and a mutually rewarding BarnRaising experience. It may just be a matter of pointing out they aren't in Kansas any more, and the level of behaviour we expect is much better. (cf. PygmalionEffect, BeProfessional) I have found in the past that being stern about expecting better behaviour with the presumption this is possible from the person is more effective than assuming they are horribly broken beyond repair. -- SunirShah