The following is a suggested practice, not confirmed good practice.
Intellectual arguments must adhere to certain standards, or descend into politics. A good example of an intellectually dishonest tactic is an ad hominem attack: one poster attacks, not the meat of another poster's argument, but the poster himself. For instance, they might make accusations of deception or ulterior motives, or bring in racial or sexual attacks. Such an approach does not AssumeGoodFaith, and can create significant conflict within a community if allowed to continue unchecked.
Therefore, demand good faith. If someone uses an offensive ad hominem attack on you, DemandApology. Further, on a wiki, demand a rewrite of the offending text. Cease participating the discussion until it returns to an acceptable standard. A short message to communicate this is ideal: DampenEmotions. If the author does not apologise, simply exercise your right to do something better with your time.
This by itself is vulnerable: by being silent, you give a voice only to the less rational. So, the entire community must move to support you, clearly and quickly disengaging with the offending author, except for ConflictResolution, until such time as the offence is rectified. This also lessens the risk of DemandGoodFaith becoming a TrollingTactic, since the community as a whole must decide on a case-by-case basis whether the demand is proportional, or simply a gambit to curry sympathy. (A reliable PageHistory is important for this to allow the community to determine the facts of the conflict.)
If an author refuses to apologise for an offence despite public ostracism, they are behaving sociopathically and are moving rapidly into the "DifficultPerson" category of the CommunityRoles. If, after a period of review, and after conflict resolution has been attempted, the situation remains unresolved, it could be sensible to escalate it into a question of community division (i.e. CommunityExile). If the author apologises at any point in the process, the community should fall back on ForgiveAndForget, and the discussion can resume on a more careful note.
But, an AdHominem attack is not the same as an honest criticism (cf. HealthyConflict) of a member's behaviour outside of the context of a rational argument. Any member of the community must be free to criticise the behaviour of others in an appropriate location, such as the user's HomePage.
Frequent resorts to DemandGoodFaith could be abuse, but it is always correct behaviour to AssumeGoodFaith, apologise, and retract. Absent the indicators of face-to-face conversation, it is impossible to determine how offensive a remark may have been except by such clear feedback.
Just some thoughts I had on viewing an exciting trolling session somewhere on the web, which would have been much less effective if this pattern had been employed. I think this works even if one values ContentOverCommunity, as in the very short term, it makes sense to place CommunityOverContent: you can't get interesting content out of trolling tactics, only interesting reactions. I've used the immediate-backoff policy myself, and have also relied on how people handle demands for apologies to separate troll from contributor in a couple of MB tussles. What do people think? -- ChrisPurcell
After thinking a bit about this, it is insufficient to demand anything unless you are in a position of power over the accused. If you stand at a peer or subordinate level, you risk escalating the conflict as making such a demand will ExpandScope. It is also, however, insufficient to simply hope the accused comes to their senses, or through the power of DissuadeInteraction, they will either leave or become motivated to curry favour again. What's needed is an objective standard that all parties, including third party bystanders, can agree to. Thus, the demand is not so much part of the inter-personal conflict, but rather the CommunityExpectation of what constitude suitable BehavioralNorms in this space. You'll have a much greater chance of garnering community support to DefendEachOther if you appeal to the common cause.
It's best if the objective standard is this so-called 'rational process' you refer to, which is merely a vague way of stating some process that leads all parties regardless of differences towards an outcome that satisfices their SuperordinateGoal. That is, a FairProcess to reconcile arguments. In academia, there are appropriate methods and forms of argument that are considered publishable in journals; others are rejected through PeerReview. On Wikipedia, they have the NeutralPointOfView. A StyleGuide would be helpful in this regard, provided it is consistently adhered to. -- SunirShah
I'm not sure how to relate to this concept. I agree with Sunir that it is difficult to demand something in online commnuities, because then there is always the issue of "who gives you the right to" in the background. As a general strategy I much more tend to suggest DefendEachOther, so that if some norms of the community are violated it is not the job of the one attacked or hurt to defend himself, but it is the job of the members of the community. It's perhaps one of the nicest ways (beside ThankYou) for anyone to get an informal member status. Self-restricting actions ("don't participate/communicate until..") by a rule also seems against tactical thinking - Why should someone make himself predictable in a conflict? I also think that "ignore a person X" which seems related, is often used in real world and in forums to signal difference in status. I know communities where the founders almost can't communicate any more, because to communicate with X means to accept him as something like a peer (this end in "oh, really B talked to you, he is number three in the hierarchy! now you are somebody in the community who will be taken seriously..." so it's not about the quality of what you communicate, but the status of who talks with you). It's like the king not talking to normal people, which is quite un-wiki-like and unproductive. ... On the other hand I think that this is good advice and it may actually be successful in certain classes of situations. Maybe especially with newcomers or people that expect real-world-like behaviour, e. g. in the situation of CommunityOnline. -- HelmutLeitner