One of the absolute worst things that can happen to a civic/public community online is for moderators to be bullies or for moderators to take the side of the bullies. Once that happens, the community is at grave risk of ceasing to be a public community and instead embracing cliquism. If the moderators enforce the will of their friends rather than good discussion norms, the space is no longer going to be a space for good discussion but rather one for a certain friend group.

The most common way I've seen this happen goes something like this. A newcomer with locally unusual ideas joins the community. Conflict between their ideas and the more established norms arises. Because these ideas are unpopular, people push back against them, often in mean or uncharitable ways. If left unchecked, the newcomer may soon become a target of bullying and sniping. [1]

At this point, moderators need to intervene in favor of the newcomer, because mean and uncharitable behavior shouldn't be allowed to stand in a civic/public space, even if it's towards ideas that are locally unpopular. Moderation is needed to rein in the attacks and keep things civil and productive. However, in practice what often ends up happening is that the moderators intervene against the newcomer, enforcing the local social hierarchy rather than good discussion norms.

This is toxic to a civic/public space and, if left unchecked, drives out views or discussion styles other than those that are locally popular.

One potential antidote to this sort of behavior is holding moderators to significantly higher standards than users. If a moderator and a user are in an angry, insulting argument with one another, the moderator should be removed from moderation or at minimum recuse themselves. If a moderator posts insults against another user - especially someone who isn't popular - they are at fault and should apologize or be removed from moderation.

Yes, this is a harsh standard. Yes, this means that being a moderator limits what you can say in some circumstances. But that's what you need to do to keep the bullies at bay, and ultimately, being a moderator shouldn't be a position of power but rather a position of responsibility.

Lastly, I want to point out that it's totally fine for a space to exist for a friend group or for those who agree with certain perspectives - and for those sorts of spaces, it's entirely fine for moderators to enforce local social norms or locally popular opinions! However, there's a big difference between that and a civic/public space, and if you're going for civic/public norms a higher standard is needed of moderators.


[1] This obviously doesn't apply to Nazis and the like, which should IMO be banned outright.

[2] Note that footnote [1] should not be construed as an excuse to go around calling everyone you don't like a Nazi in hopes of getting them banned, and such rules should be clearly articulated beforehand - the intent is merely to point out that you can have a civic/public space that still prevents certain objectionable content.



New Comment
51 comments, sorted by Click to highlight new comments since:

https://www.lesswrong.com/posts/tscc3e5eujrsEeFN4/well-kept-gardens-die-by-pacifism

I think you should explain in substantially more detail why you think communities should do the opposite of following Eliezer's advice.

Eliezer and I grew up at different times and have seen different communities. He saw communities ruined by foolish chat drowning out the intelligent discussion; I see communities ruined by people misusing vote systems to downvote people who don't agree with them and upvote people who hold their beliefs regardless of the quality of those people's arguments, moderators abusing their power to help their friends instead of doing what helps overall discussion, and so on.

I don't know whether that's different "eras of the Internet", just different experiences, or what. But my experience has not much been one where communities refuse to defend themselves against bad norms, trolls, and the like - instead, it's been one where voting systems are often used to enforce groupthink and stifle legitimate criticisms, moderators protect their friends' interests rather than upholding norms, and so on.

This very site I think has been damaged by similar issues. For a long time, LW 1.0 was well known to have silly, nitpicky comments and voting patterns that drove away its best people - not because the community wasn't protecting its norms, but because the norms that were locally protected were bad and unpleasant to deal with! If the majority were promoting such behavior, action should have been taken to censure the majority - though in point of fact the extent to which this was a majority issue is unclear, because LW 1.0 voting was infamously subverted by someone who used multiple accounts to heavily influence voting towards his own preferred ideas.

Similarly, on LW 2.0 we lost our best poster (Duncan_Sabien) because moderation did not stand up for the deep, important values and Duncan wasn't willing to put up with it.

It is really very, very important that moderation not take the side of the bullies. That doesn't mean giving in to trolls, that doesn't mean letting people waste everyone's time - but it does mean that if there's someone making good, well-reasoned arguments who is getting hassled with bad comments because those arguments support locally unpopular conclusions, it is the job of the moderators to protect the person making good arguments, not to protect the local social order.

My memory of LW 1.0 is that it had a lot of mediocre content that made me not want to read it regularly.

This at least plausibly seems like it could be a clear second order effect of the thing Davis was pointing out.

Only if nitpicking (or the resulting lower posting volume, or something like that) demotivates good posters more strongly than it demotivates mediocre posters. If this is true, it requires an explanation. My naive guess would be it demotivates mediocre posters more strongly because they're wrong more often.

[-]dxu120
My naive guess would be it demotivates mediocre posters more strongly because they're wrong more often.

A lot of the time, "mediocre posters" tend to be the source of the nitpicking. This is because writing up a nuanced objection takes time and effort, and requires much of the same skills as writing a good top-level post; whereas posting low-effort nitpicks is easy, especially if other people reward you with karma when you do so. (And empirically, I observed a great deal of poorly reasoned comments receiving upvotes on LW 1.0 towards the end of its lifespan, although I will decline to speculate publicly as to the cause of this.)

It doesn't have to demotivate good posters more often, it just has to demotivate them enough to make them come less.

There are (at least) two types of mediocre posters – those that produce mediocre content, and those that like nitpicking and or picking fights. If you evenly drive away all content producers, and are left with nitpicking and drama, it doesn't matter if the mediocre content producers are wrong more often. (And the effect doesn't seem to be that strong to start an evaporative cooling process)

This seems like an argument for the hypothesis that nitpicking is net bad, but not for mr-hire's hypothesis in the great-grandparent comment that nitpicking caused LW 1.0 to have a lot of mediocre content as a second-order effect.

I'm not 100% sure what mr hire meant, but I saw my comments as being in line with this shorter comment by Davis. Not sure if that's the same or different from what you meant.

(That said, I get that this subthread was about your particular experience and if the issue was lack of good, positive content it makes sense for the above model to apply less.

I do think there's an alternate model which might or might not apply to your experience, which is:

'What matters is whether the good posters come faster than they leave, independent of how much mediocre there is.'

i.e. if a site has at least a core group of good content generators, it's easier to pick those people out of a crowd, even if there's a lot of mediocre content. And it doesn't take much nitpicking for good posters to feel like they'd rather be someplace else)

(I agree with this and made the uncle comment before seeing it. Also, my experience wasn't like that most of the time; I think it was mainly that way toward the end of LW 1.0.)

It's not necessary for good posters to be disproportionately effected. Good posters already HAVE a disproportionate effect on the health of a community, so a small impact on good posters is worse than a large impact on mediocre posters.

Yes, that's my view. My model of what went wrong with LW 1.0 culturally was something like:

1. Nitpicky standards get into the culture

2. Many of the strongest contributors dislike interacting with the nitpicky standards and move elsewhere

3. Many of the remaining contributors don't have as good content to contribute

4. LW is perceived as mediocre and no longer "the place to go", reinforcing migration away from the site

So I think there's a core of good advice somewhere here. Don't nitpick is different from don't bully, of course.

And, of course, whether pointing out a flaw in a post is bullying, or nitpicking, or assisting in finding the best expression of a valuable idea, is in the eye of the beholder.

eliezer's problem is what you have if your friend group is getting diluted. this problem is what you have if you're trying to dilute your friend group as much as you can.

I'm not sure I like the word "dilute"/"diluted" here, but in any case Eliezer and I are responding to rather different circumstances. Eliezer was writing after having experienced the SL4 mailing list being overrun by low-quality discussions and withering away; I'm writing after having experienced LW1.0 being overrun by overly high standards and withering away.

SL4 quite plausibly died thanks to pacifism; LW1.0, on the other hand, quite plausibly died to enforcement of the wrong standards. In other words, one might say SL4 was *too* welcoming, even to low-quality content; by contrast, I would say LW1.0 wasn't welcoming *enough*, and I believe my opinion on this matter is shared by many of its top contributors, who found it too annoying to deal with all the nitpicking and critical comments!

(Now, one might argue that the nitpicking and overly critical comments themselves represent LW1.0 dying by pacifism - but in my view it's still notable that SL4 and whatever other groups Eliezer is alluding to in his post seem to have died thanks to letting too much bad content in, while LW1.0 seems to me to have died thanks to screening too much good content out!)

I strongly agree with this point. This is the core reason I have mostly stopped using less wrong. I just made a post, and being able to set my own moderation standards is kind of cool. That might make less wrong worth of use as a blog, actually.

I suspect being nitpicked is only aversive if you feel the audience is using the nitpicks to dismiss you. People aren't going to leave the site over "I agree with your post, but China has a population of 1.4 billion, not 1.3 billion". They might leave the site over "Your post is nonsense: China has a population of 1.4 billion, not 1.3 billion. Downvoted!" But then the problem isn't that unimportant errors are being pointed out, but that they're being mistaken for important errors, and it's a special case of the problem of people being mistaken in general.

I agree with this classification, but want to note something that sometimes happens, that is seen incorrectly as an example of the latter scenario (i.e., of the “small errors mistaken for big ones”).

I am talking about a case like this:

Alice: Blah blah. For example, so-and-so.

Bob: Now, hang on there, Alice; so-and-so is actually not an example of blah blah (and possibly so-and-so does not even exist / so-and-so does not happen the way you say / etc.)!

Alice: Yes, sure, fine. That’s just an example, though, don’t nitpick.

Now, at this point Alice often simply ignores anything else Bob says, or gets frustrated and angry and stops reading the comments, etc., but if the conversation continues, what Bob says (or might properly say) would be—

Bob: But wait—what do you mean, “just” an example? It was the only example in your post! And now that you have (apparently) agreed with me that your purported example is actually not an example of your thesis, your post is left with no examples! That is a very serious flaw; you are now presenting a thesis with no empirical support or case studies whatsoever. And I note that you haven’t made any attempt, in your response to me, to replace the defeated example with others, which is odd, since you claim the thing you describe is commonplace… You say I’m nitpicking, yet as far as I can tell, I have inflicted a serious blow on the whole edifice of your essay!

(I once summarized this as: “that [example] was not an example of the thing you mention; but also and relatedly, maybe the thing you mention doesn’t really exist?”)

And in my experience, this never results in Alice seriously re-examining her thesis, because “dismantling an example” is, somehow, automatically classified as “irrelevant nitpicking”, even when that classification is completely nonsensical because without examples, the entire piece of writing is just empty noise.

It turns out that cranks and hucksters are indistinguishable from confused and vulnerable newbies. And protectors of conversational norms are indistinguishable from bullies. I think others have pointed out that your footnote hides the entire problem, because you don't actually have a nazi detector.

If you'd said "a given community should be transparent about the norms it'll enforce", I'd agree. Even saying "norm enforcement should start out gentle and only gradually ramp up if the participant appears to be working on it" would be totally reasonable. Saying "be nice to all participants, even if they're disruptive and not fitting in" is much harder for me to swallow as general advice. There are communities where it'd work (mostly small ones where there's time and energy to more gently bring someone up to speed), but at a certain size you have to decide if you value inclusiveness more than you value the actual stated purpose of the group.

For most such groups, there is always the actual important vote: participation. If you see norms being enforced that you disagree with (or in ways that you disagree with), definitely say something - people will either agree with you or defend themselves against what they see as trolling. If the latter, it's probably not the place you want to be.

(note: I don't see much on LW that I'd call bullying, or even incivility. If this is a complaint about a specific event that I didn't notice, I apologize for your bad experience, but I don't actually know what happened, so I can't advise on whether you're being oversensitive or a moderator was unnecessarily harsh about something. Both are possibilities to consider.)

It turns out that cranks and hucksters are indistinguishable from confused and vulnerable newbies. And protectors of conversational norms are indistinguishable from bullies. I think others have pointed out that your footnote hides the entire problem, because you don't actually have a nazi detector.

I think that's just false. If a moderator can't tell the difference between a Nazi and someone who's just locally unpopular, they have no business being a moderator. It's not actually hard to tell - I've moderated active communities before and never really had much trouble with it!

This obviously doesn’t apply to Nazis and the like, which should IMO be banned outright.

You understand, of course, that these four words are doing all of the work in your post, yes?

Nope. My post is primarily about something else - this is just a footnote that serves to point out that it's totally fine to have a civic/public space that nevertheless holds that the most extreme content and ideologies are over the line.

For whatever it's worth, it definitely had a really big impact on my experience of this post in a way that felt to me like it invalidated most of its intention.

How so? The main point of the post is quite unrelated to this footnote.

My most charitable interpretation of footnote 1 is this: It's possible to imagine a profile picture, bio or first post so beyond the pale that the best course of action is to ban that person outright. And if you cannot imagine such a profile picture, bio or first post, then you have a poor imagination.

That would be quite a high bar for me, though. There would have to be overwhelming evidence that this person is going to be a net-negative influence. "They are a self-professed Nazi" would not clear that bar.

Oh, I don't think there's a disagreement here. I strong-upvoted the comment I responded to. "We can ban a Nazi because they're a Nazi." is a bad rule.

What I'm trying to add to the conversation (apart from an attempted steel-man of that footnote) is that the actual reason we ban people from communities is not because of what they've done in the past, but what they're likely to do in the future if they stay.

Usually we need to observe someone's actions before we can make such a determination, so it almost always makes sense to give people a fair chance; even a second and third. But I can imagine scenarios where a utility maximizer can be confident much earlier. Even if those scenarios are contrived, it seems important to keep an eye on our terminal values (e.g., keeping the community healthy and prospering), and recognize that our instrumental values may admit of exceptions, lest we become prisoners of our own rules.

When I say Nazis I am, in fact, referring to actual Nazis. I'm not validating "denunciation", I'm saying you have no obligation to provide a space for certain forms of objectionable content, and indeed you shouldn't. I do not consider such restrictions to compromise the rule of law - part of the rule of law involves establishing clear boundaries for what content is and isn't out of bounds, and Nazi stuff is on the wrong side of those bounds.

I think I do not know what an "actual Nazi" is. It is obviously an extremely fuzzy boundary that could range from including over 100 million people over humanity's history, or barely 5000, and I do not know which you mean.

I mean people who literally, actually support the Nazi party.

There exists no literal Nazi party. Do you mean anyone who has ever said anything good about the original german Nazi party? What does "supporting" mean?

Do you mean people who self-identity as a member of the Nazi party?

These are good questions that would need to be answered if it weren't for "and the like", which makes the rule fuzzy again no matter how unambiguously we define "Nazi".

FWIW, I also find this comment kinda nitpicky. I realize there's a few different things it could mean, and I think the phrase 'and the like' is doing some weird, ambiguous work.

But I do think it's fairly straightforward and reasonable to clump "people who literally think [former] Nazi party should have won the war, and people who nowadays call themselves Nazis or Neo-Nazis (whether that's associated with a formal party or not, it seems to me like an obvious endorsement of the original party)".

[Don't want to argue further about this, but wanted to register the disagreement]

[This comment is no longer endorsed by its author]Reply

But you realize this isn't just random unmotivated nitpicking, because it's also fairly straightforward and reasonable to clump "Nazi" with "HBD", and from there to ban someone like Gwern for his GWAS and embryo selection research, right?

(I regret having posted the original comment since my opinion depends on a lot of related points and I don't feel good about how people are likely to interpret the comment without all those related points. Will offer this one clarification)

I think there are lots of bad and dumb ways to conflate Nazi's with other things, and a lot of bad policies I have seen people endorse re: Nazis. I *do* have a moderately strong opposition to the phrase "Nazis or the like", as a concept, because that is opening all kinds of room for unprincipled slippery slopes like the one you mention.

But I don't think it's particularly weird, if you are running a private space, to say "in this space it is not acceptable to openly self-identify as a Nazi." (This is in part because I think it's generally essential for private spaces to have pretty strong leeway to define their culture pretty arbitrarily)

[This comment is no longer endorsed by its author]Reply

But I don’t think it’s particularly weird, if you are running a private space, to say “in this space it is not acceptable to openly self-identify as a Nazi.” (This is in part because I think it’s generally essential for private spaces to have pretty strong leeway to define their culture pretty arbitrarily)

[Emphasis mine]

Note that the OP is explicitly and specifically about public (a.k.a. “civic”) spaces.

Questions of what is, and what is not, appropriate for private spaces, are thus not applicable.

Ah, thank you, I missed that. (I'm not sure whether this changes my opinion about how fair it is to cluster all self-identified-Nazis together but it makes it complicated enough that I'm not prepared to make a strong claim on the subject)

[+]Elo-50

Either you’re not referring to actual Nazis, or your entire “Nazis and the like” argument is nonsense, because (to my knowledge) all those who were members of the NSDAP are dead.

So what do you really mean?

What's included in "and the like"?

That sentence doesn't propose banning people for extreme content but for who they happen to be regardless of how they behave in the forum.

Saying that certain people who belong to a certain social group that's generally accepted in good company to be the outgroup should be banned outright, seems to advocate something that's very similar to bullying.

I do think it makes sense to ban users who only join a forum like LessWrong to push one particular political perspective. Those people might be "Nazis" but they could also have many different causes.

As Scott writes:

The moral of the story is: if you’re against witch-hunts, and you promise to found your own little utopian community where witch-hunts will never happen, your new society will end up consisting of approximately three principled civil libertarians and seven zillion witches. It will be a terrible place to live even if witch-hunts are genuinely wrong.

You have to ban some things to implement basic standards. For instance, here are the list of banned topics from a space I used to moderate (well, still do but it's mostly quiet now):

-Racism, virulent political ideologies, etc. - even if they're clothed in "scientific" guises. "HBD" and the like are explicitly not welcome.
-Harsh, insulting language. Telling someone you think they're wrong is fine, cursing them out isn't.
-Pornography or any other sexually explicit or highly suggestive content.
-Any form of "doxxing", offsite harassment, etc. except in cases of preventing serious crimes - and if we ever get to that point things will have gone deeply wrong here!

These sorts of restrictions have not in my view led to bullying - instead, in many respects they've led to there being a safer space, where people don't have to worry about certain types of bad content that can be prevalent online.

LessWrong is a forum that does have basic standards without banning people for their political views. It has different dynamics of how content standards are enforced then banning.

To bring this back to the concrete - what is your evaluation of the discussion in the comments for this post? There has been no moderation applied (though members are participating who happen to be site moderators, they're not acting to moderate the post of the comments AFAIK). There has been nothing that I'd call bullying. There has been nitpicking, as part of attempts (some of which I think are valid), to point out ambiguity and problems with your position.

Are you inspired to (choose one or more):

1) edit your post to be more popular

1a) in agreement that the new version is clearer about your thesis and recommendation

1b) reluctantly, feeling bullied into it

2) decide it's fulfilled it's purpose and spurred good conversation in the comments

3) leave LW and condemn us all as jerks

4) continue to participate and get value from (if not enjoy) the style of discussion we often have

Agreed that there hasn't been bullying but has been nitpicking. I think the comments here have gone seriously off-track from the main intent of my post and I intend to write another post that deals with the "is it OK to ban Nazis" issue more directly; I'm unsure whether I'll write a new version of this post.

This situation has made me less likely to want to write on LW in the future, but it's not to the point where I'm quitting or whatever.

I suspect most of the challenge with this as a policy is that it's hard and requires empathy. It's fun and easy to keep people out, call them names, say they are dumb, and elevate your own status by pushing others down. It seems to be the default human behavior. Doing otherwise is hard, requires patience and training, and sometimes you still get it wrong.

I agree with the proposal, though, and when I have time I try to help those along who post here who seem new and confused with gentle words and encouragement towards our norms. It's not that we have to let low quality content pass so much as we can respond to it in a compassionate and loving way that fosters growth and encouragement rather than in a spiteful and exclusionary way that discourages growth, learning, and engagement.

In short, be nice first, keep things nice second.

Since people seem to have issues with footnote 1, I've added a second footnote clarifying it a bit. I should stress that this remains an aside and not the main thrust of the post.