There are very different philosophies on when moderators should remove any content. Most Stack Exchange sites are on one side of the extremes with rather low barriers for deletion. Especially comments are often removed at the first sign of trouble.

Some communities fall on the other end of the spectrum and almost never remove content unless it is extremely offensive or spam. The users in such communities often feel strongly about this, and removal of borderline content is viewed as censorship.

My impression as a moderator on one of the most strictly moderated SE sites (Skeptics) is that the ability to simply remove comments is extremely valuable in preventing conflicts from escalating. If a conversation is getting personal I tend to nuke comments immediately. This often prevents any escalation as at that point both sides don't really care enough to continue, at least as long as there is no previous bad history between the users.

What I've observed in other communities that have strong opinions against removal of content is that small-scale conflicts between users can escalate and lead to very long threads of aggressive posts and a rather toxic atmosphere.

What preventative measures can be taken to avoid such escalations if simply removing non-constructive posts is not possible?

4 Answers 4


If you really can't delete old or out of date content but you want to prevent arguments over it then your only action is to put the posts into such a state that they can't be edited, voted on, commented on or replied to.

Such a lock must be visible and have a simple explanation of why the post is locked and possibly a link to where the lock can be discussed.

This preserves the content but prevents the arguments from continuing.


Disemvoweling, i.e. editing a message to remove vowels, is a compromise between deletion and preservation. Unlike deletion, it satisfies all but the most hard-line anti-censorship voices, because the content is preserved publicly and can be read by anyone who cares. It nonetheless offers many of the advantages of deletion: the content will be skipped by casual readers, cutting down on the risk of readers being offended or jumping into the discussion with their own retorts.

Disemvoweling is a form of public shaming, which is a double-edged sword. Other people involved in the discussion can see that a user misbehaved and get the signal that they may want to try and keep away from that user. Disemvoweling can help turn the attitude towards the rude or non-constructive user towards slight ridicule rather than antagonism. On the other hand, disemvoweling leaves a record highlighting the user's misbehavior; deletion has the advantage that it lets bygones be bygones. If you implement disemvoweling, make sure that affected users can remove their disemvoweled posts at their own discretion, even if they would not otherwise be allowed to remove their contributions.

Disemvoweling has been practiced since the days of Usenet in an earlier form of splatting out (replacing vowels by *). It has been used in various other communities, notably by Teresa Nielsen-Hayden on her blog Making Light since 2002; this is where both the name and the formalized practice originated. By the way, you'll find a lot of good advice on moderation on her blog.

Although disemvoweling is less strong than deletion, it is not benign. Using this technique doesn't preclude the need to be fair (as much as possible) and perceived as fair.

The concept may not apply to languages written in scripts where vowels and consonants are not separate letters.

  • 2
    -1, mdrtrs tnd t b n ngh ht wtr s t s wtht pblcly shmng mmbrs f thr sts.
    – Caleb
    Commented Aug 13, 2014 at 9:40

This is a very wide question. First of all we need to talk a bit about the concept of escalation. There are different ways in which a back-and-forth between a few users can escalate:

  1. On our network, comments should be used to improve, or criticize answers and not their authors. Predictably, when this happens, the authors get... irked.

  2. Some other times, comments are used to add speculative or "pseudo" answers. This can lead to bickering because there are no down votes on comments - people can say whatever they want without risking to lose rep.

  3. Fanboi-ism or its opposite: some people downvote and just add the equivalent of "-1: your answer is wrong." ot "+1: I totally agree", which don't add any information besides disclosing the source of a vote. These comments should be ignored, but sometimes they spark off discussions because they are unqualified

  4. Passive-aggression, flame-baiting and trolling: "And of course only a [person who likes something I dislike] would say something like this" or "Anonymous downvoter, don't be a coward and leave a comment"

Depending on which case, there are alternatives to deletion. In case number 1. you can try suggesting moving the convo to chat, or try to stop it via a lock or a strong comment. While this might work, the damage is done - I know we lost contributors on Skeptics because of personal attacks.

To defend against pseudo-answers, we've tried some community education in meta. This works in the sense that the community monitors the situation via flags, but there's little we can do besides deletion or suggestion to move to chat. These conversations are not generally that problematic elsewhere, but on Skeptics they are endemic since they are a free for all "I'll tell you what I think without providing any evidence".

Eliminating fanboi-ism is a matter of community culture. Meta discussions should try to educate the community, and I don't see a defensible reason to leave these comments on any SE site.

There's little we can do to prevent point 4. Ultimately we have a suspension reason for "excessive discussion in comments". If the users cannot be educated with kind guidance, banning them for a little bit does the trick -- or in some cases, a perma-ban is the only option. We've had one such case on Skeptics.

  • 4
    This answer sounds highly SE-oriented, can you confirm that you would recommend this answer for general communities?
    – Talisman
    Commented Jul 30, 2014 at 9:55
  • 1
    It really depends on the community. The question is specific to SE sites.
    – Sklivvz
    Commented Jul 30, 2014 at 9:56

You could edit parts of the post, not necessarily removing parts of it, but wording those alternatively. If you believe that the post is essentially "flamebait" (where a user is just asking for a conflict to happen) then you may be left with no alternative but to delete. If using the likes of vBulletin, you may want to use something like a soft delete however, to keep the post present if needs be.

Being a moderator on a site like that is not easy but sometimes the community will have to accept that you are doing a job and are only looking out for the best interests of the majority. You ideally want to have the control there to be able to remove posts using your best judgement without being subject to abuse from all quarters.

If these users are prone to becoming aggressive you may wish to consider ejecting them from your community. It's not a light decision and it shouldn't be made alone, but you have a job to do and if you think a post should be edited/deleted then you have to make that call.

  • 3
    If a community is strongly against deletion, it's likely to be even more opposed to imposed rewording. Commented Jul 30, 2014 at 12:39
  • 1
    Yeah, that's why I mention that short of really having the content visible you will have to carry out your duty as a mod.
    – Talisman
    Commented Jul 30, 2014 at 15:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.