“First they came for the intolerable harassers, and I did not protest, because getting rid of those people is necessary.

“Then they came for the irredeemable assholes, and I did not protest, because getting rid of those people is a good idea.

“Then they came for some people I wasn’t sure about, but I figured they knew what they were doing, because they’ve always done good work in the past.

“Then they came for some people I thought were right-on, and I was confused, because I didn’t see what good it did.”

— definitely not Martin Niemöller

There’s a problem with the Block Bot.

The problem isn’t that it exists. Discourse communities and political groups need some kind of internal discipline. We need rules for behavior that prevent the destruction of the group and the oppression of the more vulnerable members. This is true all the more for groups that specifically want vulnerable members to be accepted and to participate equally. The Block Bot is part of that effort.

What worries me about the Block Bot is that it’s getting too easy to get on the damn list, impossible to get off it, and the criteria whereby one is added are getting more and more subjective and opaque. It’s becoming a blackball.

This is bad for the Block Bot. It’s bad for the people who need it.

I know, I sound like a complete concern troll. But hear me out, and keep in mind that I think the Block Bot is a good idea.

The Block Bot works best when the decisions of the blockers are clear and sensible – obviously necessary. When it comes to serial harassers, obnoxious psychopaths, trolls and fakes, it’s impossible to disagree – these are the people who blow up discourse communities. You can’t get along with them, so they have to go. Show them the door.

In other words, at L1 and L2 (the higher levels of Block Bot, reserved for the truly damned), I think the Block Bot is working well. Consensus on these people is achievable.

L3 confuses me, though. It’s for “annoyances”. The Block Bot FAQ indeed defines presence on L3 list as relatively trivial – “someone on the internet thought you were an asshole,” is how it puts it. It’s not something to be sweated.

Except, it is, though. The Block Bot, precisely to the extent it has worked well and helped egalitarian discourse sub-communities on Twitter to thrive, has credibility. Being on the Block List, at any level, carries weight. L3 may not be the mark of Cain, but it’s still pretty much an instant disqualification from most of the social justice conversations on twitter.

And what gets you on it? “Someone on the internet thinks you are an asshole”. Do you see the problem with this definition? It’s essentially all-encompassing. Anyone can be on it, because it’s the nature of the internet that virtually everyone has at one time or another been an asshole.

“Speak for yourself.” Yeah, well, I am – but the number of people who have never had a high-handed or insulting word for another person on twitter is something like the number of people in the world who have never had a covetous thought.

Particularly given the nature of the conversation, which is to say, the social justice activism on twitter. It’s combative. That’s a feature, not a bug. Everything is at stake.

Are you sure each and every person to whom you have been vituperative is always and only on the side of the devil?

If not, you plausibly should be on L3.

Arguably I should. I can’t remember anything specific. No blocker has reported me. But I can’t imagine I’ve never had a sharp word with someone on the right side of the social justice divide.

Comrades quarrel.

(Nobody denies this is all partisan and ideological, do they? It’s not just bad behavior, it’s bad behavior directed toward people we like, people with good politics who say good things. To get on the Block Bot block lists, you have to be bad in a certain way to a person who is good in a certain way. It’s not merely about the nature of the behavior in general – the block bot isn’t a liberal policing mechanism. Okay, just so we’re clear. I don’t regard that, again, as a case against it. Just so “civility” as a general principle isn’t used to defend it, because clearly that’s not what’s at stake. It’s to protect certain groups from other groups.)

The reason I and thousands of others are not on the L3 list – that is the problematic thing. It’s because nobody knows exactly what qualifies you for membership on L3, because the definition is so broad it doesn’t serve to separate any particular person from any other person. It makes such specificity inconceivable.

But you can’t have transparency and trust without specificity. You can’t have those things without showing me the rules and then being able to show me, in each case, how the rule was broken. There is no such specificity or transparency for the L3 list. Individual blockers make the decisions and commit the members of the block list without the necessity for further review.

As I see it, categories like “asshole” and “annoyance” recognize patterns, dispositions to behave, not individual acts. The L3 list, if it should exist at all, should be for people the community can’t tolerate. Not people who once said a bad thing to a big wheel.

In other words, it’s not just a question about what you do to get on the list – it’s a question about whom you do it to. The bad tweets that get noticed and get people on L3 are directed at some people and not others.

For example, there is no attempt for the Block Bot blockers to protect people who merely happen to use the Block Bot.

Blockers notice bad tweets tweeted at their friends and at important people with large numbers of followers. The blockers notice bad tweets aimed at people they follow and know.

This is perfectly natural. It’s how social networks function. You notice the activity among your connections.

But this extra level of protection – a Block Blot blocker watching over one’s mentions – is not available to most. It’s available to people who are important nodes in the social justice conversation on twitter.

Structureless organizations have this particular problem – they always have structure; it’s just covert. There are people who dominate the social network. The reasons why are rarely purely about “merit”, whatever “merit” might mean in this circumstance. They accrue recognition and power by whatever means, and that recognition and power is perpetuated by the weighting of the network.

I hardly need to say that this can be problematic, especially in a community that is self-consciously egalitarian and accepting.

I submit that the Block Bot is, despite itself, making this phenomenon worse, precisely by hyper-regulating the discourse surrounding the powerful nodes of the network.

Combine these two effects: First, a widely respected blacklist, membership in which is imposed subjectively by unaccountable individuals acting quickly and perhaps arbitrarily. Second, an aura of hyper-regulation surrounding specifically the conversations concerning powerful individuals in the network.

A set of ultra-vague rules under which almost everyone is, in effect, a violator, does nothing but allow those powerful in the network to mute and expel anyone they want, anytime they want. If everyone is a criminal, everyone is present on the sufferance of the important and their friends.

Communities on sufferance, defined by the menace of their shoulder-hitters, is surely not what the social justice communities on twitter aspire to be.

The term for something like that in free-speech law is “chilling effect”, but that’s not the effect I want to focus on, because one of the (correct) assumptions behind the Block Bot is that not every utterance counts the same. I’m worried about the effect of the Block Bot on the always troublesome dynamics of in-group power.

I am arguing that the Block Bot can easily be a tool that comforts the powerful and afflicts the powerless – within the very social justice communities that it is intended to protect.

Putting people on a list in this manner is not a protective function. It’s called blackballing. That secret process whereby the entrenched protect themselves. Blackballing is always and everywhere a signal feature of closed communities that practice oppression against the weak. Blackballing reinforces hierarchy.

I know how that sounds. The Block Bot’s purpose is the exact opposite of that – it’s to remove the people who disable the very possibility of participation by the vulnerable and silenced. It sounds like I’m saying all the white boy chumps on the L3 list are the real oppressed.

Of course they’re not the real oppressed. What I’m saying is that blackballing is a bad tool. It’s almost impossible to use it for the good. It’s not self-protective, it’s retaliatory. It invariably turns inward on the community it protects.

The L3 block list is at that point. Its purpose needs to be more carefully considered. Its existence should be re-evaluated.

There should be clear definitions of infractions. There should be external input into the nature of these rules. There should be clear cases made against the people put on the list. (Retweeting one context-free tweet isn’t it.) There should be periodic review of the list. There should be external review of the list. There should be external appeal. (Appeal to “the admins” is insufficient. Their instinct and interest naturally leads them to support one another; also, their deliberations are every bit as closed and opaque as the considerations that get one committed to the list in the first place. And, naturally, they built this flawed system, and have an interest in defending its functioning.)

In my opinion, the L3 block list is itself a bad idea and should go. It’s by its nature too prone to crappy use, and to outright abuse. That the block bot admins themselves don’t treat anyone’s membership on L3 seriously already tells you what you need to know about it.

I know this will sound both naive and legalistic, in a way designed to disable the system – but a system like this, to be credible, effective, and useful, has to be transparent and to have serious limits. Otherwise it is potentially no more than ridiculous – the easy target of the lazy accusation that it’s a group of closed-minded friends batting aside whatever troubles their complacency.

I don’t believe that’s true of the Block Bot. That’s not what the people who run it have in mind. I think they’re sincere people doing a hard job.

But it’s becoming true. And, to repeat, that is bad for the Block Bot. It’s bad for the people who need it.

Look, the social justice conversation should by its nature be a broad one. People should be welcome, not terrified. Every community needs rules of good behavior. But at the same time communities are easily destroyed, whittled down to useless sects, by star chambers that expel people without excellent reasons. That is the real danger – that the Block Bot will damage the communities it exists to preserve. That it will destroy the village trying to save it.

What occasioned this rant? You have a right to know. A friend of mine got L3-blocked. Someone I know, who should be in on the social justice conversation. Not just abstractly, but concretely – this is a good person with good politics who is willing to go to some lengths to defend the side of the angels. The community needs people like that; it is weaker without them. The block was a snap judgment and an error, and it not only does wrong by this person – more importantly, it does wrong by the community.