Recently there have been some discussions about the political stances of the Lemmy developers and site admins. To clear up some misconceptions: Lemmy is run by a team of people with different ideologies, including anti-capitalist, communist, anarchist, and others. While @dessalines and I are communists, we take decisions collectively, and don’t demand that anyone adopt our views or convert to our ideologies. We wouldn’t devote so much time to building a federated site otherwise.
What’s important to us is that you follow the site rules and Code of Conduct. Meaning primarily, no-bigotry, and being respectful towards others. As long as that is the case, we can get along perfectly fine.
In general we are open for constructive feedback, so please contact any member of the admin team if you have an idea how to improve Lemmy.
Slur Filter
We also noticed a consistent criticism of the built-in slur filter in Lemmy. Not so much on lemmy.ml itself, but whenever Lemmy is recommended elsewhere, a few usual suspects keep bringing it up. To these people we say the following: we are using the slur filter as a tool to keep a friendly atmosphere, and prevent racists, sexists and other bigots from using Lemmy. Its existence alone has lead many of them to not make an account, or run an instance: a clear net positive.
You can see for yourself the words which are blocked (content warning, link here). Note that it doesn’t include any simple swear words, but only slurs which are used to insult and attack other people. If you want to use any of these words, then please stay on one of the many platforms that permit them. Lemmy is not for you, and we don’t want you here.
We are fully aware that the slur filter is not perfect. It is made for American English, and can give false positives in other languages or dialects. We are totally willing to fix such problems on a case by case basis, simply open an issue in our repo with a description of the problem.
I’m clearly “left-leaning”, so I might be biased, but I don’t agree with your criticism toward the slur filter : the project is open source, and as such people wanting to use these slur can work they way to another version. The devs explain here a clear intention to make this change difficult enough to prevent at least partially the migration of some communities they don’t want to support and/or give a platform to. I think that’s an honest way to do things ?
It also open up the debate on free speech and how saying some things actively attacks fundamental rights of others. In those cases, defending free speech as a “right” becomes irrelevant since both sides of the debate can use this logic to defend opposing actions. Trying to be short here, hope you understand what I mean !
I’m happy it’s becoming harder for neonazis to find a home online, however i’m not happy that this makes lemmy english-centric, and i’m not happy that honest discussion about some topics (including thoughtful criticism) will be made harder.
Related example: on another message board a few weeks back i couldn’t post a message containing my criticism of “bitcoin” because bitcoin was part of the slur filter to filter out the crypto-capitalist clique… i understand and appreciate why it was put in place, but i felt really powerless as a user that a machine who lacks understanding of the context of me using this word, decided i had no right to post it. I appreciate strong moderation, but i don’t trust machine to police/judge our activities.
I quite agree with you that moderation is hardly a machine job, and not saying it is the perfect solution. It sure as it’s drawback. I am just arguing that the benefits outweigh them. I would prefer to be in a world where there are not needed, be as of the world today, I admit I prefer having this filter rather than not having it, mostly because of the systemic effects I explained.
I agree that the relevance of he content of the filter can be discussed too, and that banning some words can make it difficult to discuss certain topics. But I think some words are almost always meant to harm, and can be easily replace by more positive or neutral term.
As a direct example : I can talk in this post about homosexuality, and I can event paraphrase to talk about the way some f word is used as a slur for it and how I think allowing it here isn’t a good idea in my opinion. See, I can talk about it, be respectful about it. I just prevent to call you a [insert here whatever banned slur] pretending to use my free speech.
That’s also the case for me, in case that was not clear :)
I don’t think it’s that easy, because of the context. Should all usage of the n***** word by black people be prevented? Should all usage of w****/b**** words by queer/femmes folks in a sex-positive context be prevented? etc… I agree with you using these words is most times inappropriate and we can find better words for that, however white male technologists have a long history of dictating how the software can be used (and who it’s for) and i believe there’s something wrong in that power dynamic in and of itself. It’s not uncommon that measures of control introduced “to protect the oppressed” turn into serious popular repression.
Still, like i said i like this filter in practice, and it’s part of the reason i’m here (no fascism policy). As a militant antifascist AFK, i need to reflect on this and ponder whether automatic censorship is ok in the name of antifascism: it seems pretty efficient so far, if only as a psychological barrier. And i strongly believe we should moderate speech and advertise why we consider certain words/concepts to be mental barriers, but i’m really bothered on an ethical level to just dismiss content without human interaction. Isn’t that precisely what we critique in Youtube/Facebook/etc? I’m not exactly placing these examples on the same level as a slur filter though ;)
As often in cool debate, I think in the end we mostly agree. I especially agree with you on the point that reclaiming a word is a valid way of using some slur, and that it should not be to a privileged group to choose when a word is ok or not. On this point I have to point out that this is still the case with manual moderation, if most moderator are privileged. So I agree that diversity should be push in all places of power, and all decision are better made (and more legitimate) with a diversity in the group that make them.
But on the automated part, I really think the psychological aspect is strong and should be questioned. You talk about “human interaction” but this definition is really hard non only to define, but also to defend as an efficient way of reaching you goals. I am quite sure that when the devs made their filter, there was quite a lot of human interaction and debate around it, and the simple fact the put one show that they interacted with other people around them. And is a “manual” moderation a human interaction when you don’t see or know the person, don’t know their culture, the context, their tone, etc. Moderation will never be perfect, will always involve bad decisions, errors. When errors are mades “directly” by humans, compassion and empathy help us to try and understand before judging (but judging nonetheless in the end don’t get me wrong). Why is it so different when an automated system (created by an imperfect human) ? Why is an automated error worse than a human one if the consequences are the same ?
Long story short, I don’t like thinking along great principles like “automated moderation is dangerous”, but rather try analyze the situation and think : would this place be better if there was not this automated moderation ? I agree that this is a wide and difficult debate one what is “better” of course, but the focus should always be this one : how to make things better.
Thank you so much for your answer, i’m not used to debate online because I didn’t feel at ease anywhere else before, but I love it and it is thanks to people like you and all the other interesting answers I get that I can enjoy that and think about it so much ! Thank you thank you <3 !!
(edit : typo)
Thanks for your comment, I’m really happy to read something like this. I’m glad that people can really get along here :)
Sure, but given a /c/blackfolks community, a white admin would probably think twice before getting involved in internal matters over there. Which an algorithm will have no clue about.
The latter is true, but i believe the former isn’t. Having some kind of filter shows great concern for people experiencing harassment/bullying online, but using a word-based filter is a known anti-pattern since about the end of the 90s. I remember i used to go to this library, and from there you couldn’t access the library’s own website because the name of the library contained a french slur inside (though the whole was not a slur really) and the library-wide MITM proxy had a slur list like the one lemmy implemented. That’s how clueless such systems are.
For the reason you mentioned: lack of context and empathy.
Certainly not. I’m not advocating for removing the slur filter on this specific instance. I’m arguing having it hardcoded into the source is a strong political posture and we don’t really measure the variety of consequences it may have on the ecosystem as a whole.
Thanks to you too <3! I strongly appreciate online debate in such settings. Are you by any chance too young to remember when (before Facebook) forums/BBS were the craze? We really lost something (on a human/political level) when everyone moved to these centralized platforms where interactions were turned uniform and bland, and real-name policies have led to real-life crisis (bullying, suicides…).
Yes, at least a sensible one ! But the term you used is great : “thinking twice”. I really do think that the admins though twice before choosing this filter. This is a human choice, just made with automated tools. As in your example, the moderation will be questioned in either case, and that’s great ! I’m here for it. And i promote a way of organizing the critique against moderation around the question “what are the concrete consequences of either choices”, kind of an utilitarian point of view I have to admit. I think that initializing this new platform with a quite strong political stance on these issue will help this place have a positive impact. It is “hardcoded” yes, but in an open source project. If the platform grows, forks will appear for sure, especially if strong opinion arises on this kind of “hardcoded” issues. So I think about it more as a launch measure than a definitive stance.
I’m too young yes, just missed it ! (I’m born in 1994). (I am very social AFK, but never used Facebook/Twitter/… so I’m not as used to online interaction, that I think have specific codes.) But things go back and forth, let’s hope projects like Lemmy are the sign of a new era of progress !
That’s the defence of the “slur filter” that everyone can agree on. It’s harmless because it does almost nothing. It has no real benefit or cost.
The people who say it deters fascists - it just doesn’t hold water.#
I don’t know, if I believe some comments around here, there are clearly some of them that explicitly explain they would not come here because they feel “hated”, in public, so clearly to deter anyone close to them to come here for these reason. If so, it means it has some positive effect, and it seems plausible to me.
I don’t understand a lot of your message.
But if i get the gist, that might not be so positive. People who feel hated, isolated, afraid to express themselves in public, they are the people we should welcome.
It sounds like they are teenagers who are just figuring out their views. They all have strange and offensive ideas at times, but with help most people figure out a sensible worldview in the end.
As long as they can stop themselves from insulting or attacking other people, they can come to Lemmy with no problem.
You’re just the ones who can’t express themselves properly yet. They have to learn that by chatting with adults and getting negative feedback.
But that doesn’t have to be our problem, at least not while lemmy is still so small.
Sorry, it was not clear at all ! I was talking about fascist publicly denouncing lemmy as a platform suppressing their free speech, and that, as such, it should be avoided.
Okay i didn’t get that at all.
Bit this new idea sounds like a paradox - someone pro free speech would mind object to that denouncement at all.
I’m not sure I understand what you mean here sorry.
Perhaps I was a bit too harsh on the filter. After all, you’re right that someone could just modify the code. Even so, it doesn’t really seem like it lines up with the philosophy of federated platforms. It makes it more difficult to customize moderation on the instance level. I also feel like the problem of platforming nasty people could be solved by moderation on the instance level and blocking instances which don’t have adequate moderation. That’s what it’s going to need to be in the end anyway if Lemmy grows enough and people customize the code.
It does bring up the free speech debate, but I find those usually aren’t very productive in these sorts of contexts. It’s not really a legal question since the government isn’t involved, and they usually just end up being each side stating their presuppositions.
It’s not terribly important in this case anyways, I just thought I’d share my thoughts on it.
I mostly agree. And I agree that if the platform really grows, it will come down to per instance moderation and instance admins choosing wisely the instances they choose to federate with. But I think the choice is to make sure to give a head start to the people they want to welcome here. With the recent events in the US, imagine lemmy being the next tool used by “some people” the devs wich didn’t come. Then the platform as a whole would be much less attractive to some other people the devs are more interested in helping and interacting with.
So I think we agree, on the long term, if Lemmy grow, someone will come up with a modified version without thoses filter. It will just take more time. Meanwhile, Lemmur gets to be at peace as much as it can ?
Thanks for your answer !