Racial slurs that should never be written

July 7, 2020 9:44 AM

A new feature has been rolled out site wide to block the ability to make posts or comments with a specific egregious racial slur and variants. Mods currently delete comments with other slurs as fast as they can after the comment is flagged but the community has decided we'd rather not have those words appear even for a short time (previously). Currently if your text includes a prohibited word your ability to post/comment is disabled. This thread is for discussing which other words fall into the category of racial slurs which should similarly be prevented from appearing in the site under any circumstances.

Discussion of this new feature has the potential to encompass every bad thing that can possibly be written. And while many of those topics are worth having a discussion about I'm asking we narrow the focus of this meta to only racial slurs. Discussion of other potential classes of always inappropriate words would be better served by their own Meta where they'd be the topic and not drowned out by the focused discussion here.

It should also be taken as a giving that trying to game the prohibited list by being cutesy with unicode or other tricks is going to be an instaban so I don't think we need to spend any time on pursuing those what-ifs. We can not possibly cover every possibility and our system is set up to rapidly deal with things we don't enumerate.

To not have this thread nothing but wall to wall slurs please obfuscate your proposals and if they aren't clear link to an explanation. EG: the two words I want on the list are the H-word and I-word. The mods are also going to set up a publicly viewable list of words with numbers so we eventually can have a number that we can point to that isn't subject to link rot.

posted by Mitheral to Feature Requests at 9:44 AM (131 comments total)

Thanks for putting this together, Mitheral.

As Mitheral's post says, I've set up a list of candidates for the filter in a publically-viewable document, with room for annotations. Content warning for racial slurs, obviously: here is the list, and it should remain available and readable there indefinitely. The initial order is arbitrary but static; a numeric item will always be the same item, so folks can reference individual terms that way.

For proposed additions, I'll echo Mitheral in asking folks to avoid mentioning them in plain text in the thread; a link to a reference page for a term with link text that avoids the literal string would be better so folks don't have to run into stuff directly in the thread here if possible, and we can add the list as we go.
posted by cortex (staff) at 9:44 AM on July 7, 2020 [1 favorite]

A certain word that means bilking, flimflamming, or bamboozling, derived from a (possibly questionable) term for the Roma/Romani people. More here.
posted by beagle at 9:56 AM on July 7, 2020 [8 favorites]

The british use a word beginning with c to refer to Indians that is the same as the N word in the US. It has also been applied to the Chinese.
posted by infini at 10:12 AM on July 7, 2020 [3 favorites]

I would like to add another c-word as well as a J-word.
posted by Comrade_robot at 10:17 AM on July 7, 2020 [1 favorite]

Also British, and thus to US ears "old-timey" but still irksome words for Portuguese, Spanish and Italian.
posted by aramaic at 10:17 AM on July 7, 2020

Due to how Trump is a racial slur-generator, I am wondering how rapidly the list can be updated after Trump weaponizes a term, because the impact is severe and extensive (Stop AAPI Hate, Press Release, Jul. 1, 2020, See Also Press Release, Jun. 18, 2020, for more racist slurs promoted by Trump).
posted by katra at 10:17 AM on July 7, 2020

Number 20 is the last name of a beloved actor, and part of the name of a town in Minnesota. Maybe it should be banned anyway, I honestly don't know the right answer.
posted by paper chromatographologist at 10:24 AM on July 7, 2020 [5 favorites]

A similar conversation has come up in recent bans of cards in the game Vampire: the Eternal Struggle.
Some cards were banned for including multiple offensive words (G and T) for the Romani people.

And another card for a T word commonly associated with Disney's Song of the South.
posted by jozxyqk at 10:24 AM on July 7, 2020

Besides saying the town should probably change its name.
posted by paper chromatographologist at 10:24 AM on July 7, 2020 [2 favorites]

Sincere question - what is the preferred procedure for if someone is quoting someone else having used one of these words?

A. Find a way to rephrase, or
B. Use some kind of placeholder like "Then they said [Q-word]" or "then they said 'q***k'"?
posted by EmpressCallipygos at 10:31 AM on July 7, 2020 [1 favorite]

Number 20 is the last name of a beloved actor, and part of the name of a town in Minnesota. Maybe it should be banned anyway, I honestly don't know the right answer.

My opinion is that it is much better to ban these words and then have to do a little extra work when referring to cases like this (so, for example, abbreviating their names and linking to their Wikipedia page) than to let words like this appear on the site.

In other words, let's prioritize the experience and feelings of people of color over the convenience of being able to type out rarely occuring exceptions.
posted by mcduff at 10:36 AM on July 7, 2020 [38 favorites]

I would like to see the (hopefully soon-to-be-former) name of Washington's NFL team added to the list.
posted by Ampersand692 at 10:38 AM on July 7, 2020 [13 favorites]

not racial slurs, but the following should be considered

I agree, but I think Mitheral's suggestion above is a sensible one.

And while many of those topics are worth having a discussion about I'm asking we narrow the focus of this meta to only racial slurs. Discussion of other potential classes of always inappropriate words would be better served by their own Meta where they'd be the topic and not drowned out by the focused discussion here.

posted by zamboni at 10:39 AM on July 7, 2020 [12 favorites]

Some quick running notes:

I'm updating the spreadsheet as we go, annotating briefly. If I'm getting an annotation wrong or could make it more accurate while staying concise, please feel free to let me know. And if I miss one, let me know.

I'm vetting words against a standard UNIX dictionary file (just "grep xxxx words") to identify common vocab collisions for the "false match" column. This is unlikely to be thorough but helps for identifying major issues with substring matching problems.

I want to keep these focused on Mitheral's original structure here; anem0ne, I agree with you that those deserve attention and we will want to look at things outside the specific territory of racial slurs as we move forward, but I'd prefer to not expand the scope of this thread so I'd like to hold off on that for now.
posted by cortex (staff) at 10:39 AM on July 7, 2020

For finding "false match words", this word match tool is fairly useful. Search for something like *[slur]* to see words that contain the slur in them.
posted by saeculorum at 10:39 AM on July 7, 2020

Thanks for making this thread Mitheral, and the running document, cortex.

I'd add the word starting with Un- formerly used to refer to Dalits. I was particularly dismayed to see it written out in that thread (as a quote) just a few comments below my original comment calling attention to it as a slur.
posted by basalganglia at 10:45 AM on July 7, 2020 [7 favorites]

I support everything on the list thus far (it's up to 32 as I type this), and in particular as a person of Jewish heritage I'd be particularly grateful to see the k-word go away.
posted by Stacey at 10:53 AM on July 7, 2020 [4 favorites]

The one I would suggest is actually a racial slur for another ethnic group, so it's covered. Most of the other words I'd want on a list fit other non-racial categories so I'll save them for a different discussion. I would like to add to Mitheral's list generation by including this s-word which has become one of my go-to's for searching for bad faith assholes on Twitter.
posted by jessamyn (retired) at 10:55 AM on July 7, 2020 [5 favorites]

Also speaking as a person of Jewish heritage, I am not looking forward to the argument over whether a slur against Jewish people is a racial slur, and thus germane to this thread, or not.
posted by Faint of Butt at 10:56 AM on July 7, 2020 [5 favorites]

The entire category for List_of_ethnic_slurs_by_ethnicity#Indigenous_Australian.
posted by zamboni at 10:57 AM on July 7, 2020 [8 favorites]

There is a term that could be included, because it can blend a racist aspect of anti-semitism into something that many people have experienced as not funny at all, e.g. harassment, abuse, discrimination, etc.
posted by katra at 11:08 AM on July 7, 2020 [2 favorites]

I'm happy not to use the Jewish hair term that katra references, though the link isn't right that it's a gentile term. I have only ever heard Jewish people use it, and then only in reference to the hair on their own head (as opposed to anyone else's). It's still a bit problematic: it's not a positive term and implies that hair that is curly like Black hair is bad.
posted by jb at 11:13 AM on July 7, 2020 [2 favorites]

I've used that term to refer to my own hair plenty of times, but if Black people find it offensive I'll gladly stop.
posted by Faint of Butt at 11:21 AM on July 7, 2020

I felt the link rounded up a variety of perspectives on how it may be used by some people referring to their own hair, but it also has implications that when used widely could harass, abuse, and discriminate against people. I have tried to call attention to the implications embedded in how people (including non-Jews) use the term, and what is essentially the difference between intent and impact.
posted by katra at 11:33 AM on July 7, 2020

For entries that generate a lot of false hits (I am mainly thinking about #20, which has a lot of non-racial uses including as a not unusual surname), I wonder if it might make sense to also check whether the term has a history of being used in an OK (or not OK) way on the site. #20 seems to have a lot of OK usage, though I haven't studied it past the first couple of pages of results.
posted by Mid at 11:48 AM on July 7, 2020 [1 favorite]

This will be problematic.
posted by Foci for Analysis at 11:51 AM on July 7, 2020 [5 favorites]

The words I would have suggested from a Canadian perspective that aren't already on the list are used by different peoples to refer to themselves, so I won't put them forward. I will stick to flagging and/or sending a mod note as appropriate though.
posted by ODiV at 11:54 AM on July 7, 2020 [1 favorite]

This will be problematic.

Not if frimble knows their regexp.
posted by jessamyn (retired) at 11:54 AM on July 7, 2020 [24 favorites]

False match words for #29 should include references to the Toni Morrison novel.
posted by MiraK at 12:00 PM on July 7, 2020 [1 favorite]

This [Scunthorpe problem] will be problematic.

We have the twin advantages that a) we have good live modding and b) nothing we do is mission critical. If an overly aggressive filter becomes a problematic it can be addressed essentially immediately and the only damage is a comment/post gets delayed for a few hours. Generally I'd be on the side of caution on this sort of technological solution to a social problem but in this case the good is far outweighed by the (temporary) harm of a Scunthorpe error.
posted by Mitheral at 12:05 PM on July 7, 2020 [18 favorites]

This will be problematic.

Not if frimble knows their regexp.

As someone who has done text analysis in the past, even if you are good at regexes there will be many corner cases. It will be hard, and there will be metatalks, and we will need to be cool with it. G-- as a beverage is or verb is going to be hell.
posted by Going To Maine at 12:07 PM on July 7, 2020 [7 favorites]

Didn't we already have the this can't possibly be done, it's too hard part of the conversation?
posted by mittens at 12:10 PM on July 7, 2020 [36 favorites]

I want to ask folks to hold off on arguing the specific implementation ease/difficulty for filtering in here; ultimately it will come down to frimble to actually handle that and I will trust them to provide feedback on what is and is not a headache in practice.

There are some complexities. I can lay out some specific categories of difficulty in a bit, am in a meeting right now. But please let's not do a volunteer version of that in here.
posted by cortex (staff) at 12:12 PM on July 7, 2020 [11 favorites]

Didn’t we already have the this can't possibly be done, it's too hard part of the conversation?

Much like going to the moon, “hard” isn’t the same as impossible.
posted by Going To Maine at 12:16 PM on July 7, 2020 [3 favorites]

There's another H-word used for multiracial people.
posted by any portmanteau in a storm at 12:26 PM on July 7, 2020 [1 favorite]

We're making a list of words for the new filter, which, at present, isn't an advisory hey, are you sure you want to use that word in this context, but a simple you can't post that comment until you take out that word. Anything that's on that list should be stuff that's egregious and always unacceptable, and there's going to be a host of things that are contextual, and maybe should go on a different list. Figuring out where those lines are, and how they differ for different members of the community, is going to be a bumpy ride. Let's be kind and thoughtful, hear from frimble, and figure it out as we go along.

(Re the Australian ones - I think 36, 38 & 39 are historical spellings that I've never seen used colloquially, and I think would now be regarded as misspellings for 35 and 37.)
posted by zamboni at 12:28 PM on July 7, 2020 [5 favorites]

Obviously you're going ahead with this feature so maybe use this thorough Wikipedia page as a data source. Most Mefites never seen these MeTas and it makes sense to represent all people and not those of us who are regulars around here.
posted by Foci for Analysis at 1:15 PM on July 7, 2020 [4 favorites]

Currently if your text includes a prohibited word your ability to post/comment is disabled.

Does it / will it highlight or somehow let you know which word is preventing you from posting? I could see that being frustrating if you have a long comment and no idea which word is the problem.
posted by ODiV at 1:35 PM on July 7, 2020 [6 favorites]

The word G***y might prove to be a problematic one. Certainly here in the UK it's not considered to be a slur when applied to people who identify as such, as opposed to being used to label a whole range of minorities. In fact, last month was G***y, Roma and Traveller History Month.
posted by pipeski at 1:40 PM on July 7, 2020 [8 favorites]

What about non-English slurs? Here might be a couple of pejorative terms in Spanish.
posted by zabuni at 1:43 PM on July 7, 2020 [1 favorite]

Does it / will it highlight or somehow let you know which word is preventing you from posting?

Frimble's announcement post explains the details. At present, it's more the rather than a prohibited word.
posted by zamboni at 1:44 PM on July 7, 2020

G-- as a beverage is or verb is going to be hell.

It's like racists ruin everything.

I generally favor a more expansive list and see where the problems come in. I can see G***y being difficult, because it's been used casually for so long, but that's... part... of the... problem... right?
posted by GenjiandProust at 1:44 PM on July 7, 2020 [6 favorites]

Okay, meeting over, following up on a couple things:

First up, one comment deleted (and a late reply just now while assembling this comment). We've specifically said please keep plain text suggestions out of the thread; I'm not sure mild terms of offense toward white people need to be part of this conversation right now in any case.

Does it / will it highlight or somehow let you know which word is preventing you from posting? I could see that being frustrating if you have a long comment and no idea which word is the problem.

The current version we built doesn't because we were aiming to get it in place for one specific word to start. We'll need to do some sort of specific highlighting or quoting for a broader set of distinct terms, so that the commenter will know what the specific issue is. The details of that is something frimble will need to scope out before we know what approach will work well and what the timeframe for implementing it will be.

What about non-English slurs?

I think including non-English slurs can make sense (there's a couple up on the list already). I think it's more a question of whether it comes up in English-language discussion on MetaFilter, rather than what the language of origin is.
posted by cortex (staff) at 1:49 PM on July 7, 2020 [1 favorite]

I'll talk a little more about some of the string matching and category issues hinted at above.

I want to say first off that I'm generally leaning toward the position a couple folks have expressed that we prioritize addressing the core racist problem, and then iron out the details later. So I'm looking at this through a lens of, if there's no major complications, including something on the filter and sorting out the edge cases after is a good goal. But we'll need to account for the major complications.

Categories of matching:

A. The easiest case we have is a term that has an offensive meaning and is not a common substring or homograph of any relatively common non-offensive words or phrases.

Items 9 and 15 are (among others) a couple of examples on the list that fulfill that category. Blocking just those specific substrings won't create any confusion or difficulty.

B. Cases where a term has an overtly offensive meaning in its exact form but it is also a substring in some common words or phrases with no offensive content.

Items 17 and 25 are standout examples of this; if we blocked comments containing either of those substrings at all, it'd match lots of words that have no objectionable content. For these we'd need to match only the exact substring (and stems of it for likely variants of it, primarily plural -s forms), and ignore any other substring matches.

C. Cases where a term has an overtly offensive meaning and there are also common exact homograph matches with no offensive meaning (and typically no etymological relationship).

Items 8 and 33 are examples. Both are offensive within a specific semantic context as a slur; both are non-offensive terms in other common contexts. The spelling is the same between cases; there is no practical way to detect which usage is in play. Handling these would require either (a) choosing to prevent people from using the non-offensive terms entirely, or (b) provide a softer warning/heads-up function to give the commenter the opportunity to assess whether the meaning they're using falls into the offensive definition.

- Exceptions, edge cases, etc.

Those are three fairly absolutist categorical definitions to illustrate the logic that'd need to go into implementation and finalizing a list or lists. Some items on the current list fall peripherally between categories or mostly fit one particular category, and will require a judgement call about whether it should be a full-on stop word or would be better as a warning.

For example, we made the call with items 5 and 6 for the initial one-term feature that the collateral damage of blocking the one non-offensive word each variant is a substring of was a worthwhile tradeoff. Because those non-offensive strings exist, 5 & 6 aren't a pure case of category A, but it's close enough and the words are offensive enough that it's not a reason to not act there. That is I think true for at least a couple new items on the list as well. If we implement exact matches for some terms, we can restore the use of those non-offensive strings after the fact.
posted by cortex (staff) at 2:31 PM on July 7, 2020 [2 favorites]

So that's the major logic issues we'll need to sort through. That's my own framework there, not any official set of categories; if folks have specific additions or elaborations I'm open to that. But I'm laying that out mostly just informationally; I would ask folks to continue to avoid getting into too much wide-open brainstorming and "what if..." stuff in here, so I'd suggest if you have specific proposals for organizing stuff consider putting it in an email to me or saving it for a followup discussion.

As far as implementation: I'll need to work with frimble on what is practical, and in what timeframe. Category A (and things close enough to it justify the collateral damage for now) are things we could I believe add to the filter without needing significant additional logic in the code. Categories B and C will both require some amount of additional work besides that. Any expansion of the list will require us to rewrite and expand the message folks see on a hit; the writing part of that can be mod work, but it will need additional coding to display stuff conditionally. Displaying info about what words matched and where will also require additional work from frimble.

I would like to aim to get bang for our buck with this, and figure out which bits we can implement most easily to capture the most likely points of concern with the expectation that building this out farther will take some time. I want to make sure we're (a) using frimble's time well and (b) not roadblocking other current site development needs that folks have asked us to prioritize, so we'll scope this out some this week and follow up from there. I can't provide any useful timeline estimates before we do that.
posted by cortex (staff) at 2:35 PM on July 7, 2020

Using your categories from above, could the tool do this:

For A: prevent the post from happening (as it does now), along with whatever message is passed to the user trying to post it (as I assume it does now).

For B & C: Prevent initial post, give them a warning about the potential slur, noting if they post their comment with the word in place and in context it is a slur they're going to be ______ (timeout, ban, whatever), and ping a mod if the post is made with the word in place, so it can be fairly quickly reviewed.
posted by maxwelton at 3:05 PM on July 7, 2020 [2 favorites]

Mod note: One comment removed. Again, please don't post candidate offensive terms into the thread directly in plain text, and please keep this specific thread focused on racial slurs rather than more general slurs and offensive terms.
posted by cortex (staff) at 3:13 PM on July 7, 2020

Several slurs used against Mincéirí (Irish Travellers) that should be avoided.
posted by scorbet at 3:23 PM on July 7, 2020 [2 favorites]

i think we should avoid the word starting with e, that people use to talk about arctic first peoples
posted by PinkMoose at 3:24 PM on July 7, 2020 [5 favorites]

Yeah, that's one of the ones I was referring to above and didn't want to suggest it because it's actually used by some of those groups in a "this is what you should call us" sort of way.
posted by ODiV at 3:34 PM on July 7, 2020 [6 favorites]

i think we should avoid the word starting with e

That is a serious edge case, because, while the Inuit in Canada and Greenland find it a slur, the people in Alaska are mostly not Inuit (being mainly Yupik and Aleut) and seem to prefer the e-word over being misidentified (although this seems really complicated and probably should be left to the people in question to decide). I certainly have no idea how to resolve that issue on the internet, when a term is a slur in one location and the preferred term in another....
posted by GenjiandProust at 3:43 PM on July 7, 2020 [11 favorites]

The block will tell the poster clearly what term was blocked so they're not ending up searching and guessing what they typed in to trigger the block?

What will happen to quotes that include those words - are we expected to bowlderise them appropriately? I'm not thinking of current usage quotes but older quotes where the words are used age-appropriately. I'd be deeply weirded out by someone using the C-word for Chinese labourers to refer to someone Chinese now, but it's nothing I would notice in a book set in the 19th c. Ditto for the word m- referring to someone of mixed race (frankly most mixed race terms end up being horrible slurs)
posted by dorothyisunderwood at 3:50 PM on July 7, 2020

I also think it's a different case than some of the others, because it's a word that some groups use specifically about their own ethnicity. Warning someone that it might be a slur would steer people away from using it where it would be correct, plus it would be shitty to issue such a warning to someone posting about their own ethnic group.
posted by ODiV at 3:52 PM on July 7, 2020 [2 favorites]

By the way, I want to reiterate that, while legitimate edge cases exist, that is not a reason for the vast majority of unambiguous slurs be added to the list and blocked expeditiously. “Difficult cases make bad law,” so let’s initially act against the easy cases.
posted by GenjiandProust at 4:10 PM on July 7, 2020 [14 favorites]

Thank you to Mitheral for compiling the list, and cortex and frimble for working on the technical aspects. And thanks to everyone in the thread adding more words.

Here is another c-word I think should be included, a Canadian slur used against Aboriginal people. It is a category C word, because there’s also a completely unrelated meaning, but is definitely still used as a slur (one that I’ve heard recently).
posted by hurdy gurdy girl at 4:18 PM on July 7, 2020 [2 favorites]

What about ableist slurs?
posted by 80 Cats in a Dog Suit at 5:05 PM on July 7, 2020 [2 favorites]

Those can go in another thread once the mechanism is working, is my understanding.
posted by jessamyn (retired) at 5:07 PM on July 7, 2020 [2 favorites]

There's a word beginning with "k" which is a vicious slur in South Africa that's often used to refer to the plant/culinary ingredient otherwise known as Thai lime or Makrut lime.
posted by Lexica at 5:14 PM on July 7, 2020 [7 favorites]

1000% what stoneweaver said.
posted by TwoStride at 6:38 PM on July 7, 2020 [7 favorites]

Yeah, what stoneweaver said. This whole meta makes me very uncomfortable. Putting the list of words in a separate document and having people use hyperlinks to wikipedia etc seems like a weird kind of figleaf. To participate in or understand this metatalk, a reader kind of has to mouse over or click through and see a whole bunch of slurs, so this doesn't stop anyone seeing them. To be able to refer to "#34" or whatever, you have to click through to the spreadsheet. Maybe that's better than having the slurs in plain text on this page, but only just.

Linking instead of typing also doesn't mean they aren't printed on the site in case it's like, the possibility that metafilter might turn up in searches for these terms that you are concerned about - they are right there in the hyperlinks for anyone (or every machine) that looks at html.

And I honestly don't see how any debate or discussion isn't going to end up feeling like people pushing back against anti-racism, just like the last meta along these lines.
posted by lollusc at 7:54 PM on July 7, 2020 [2 favorites]

Nthing what stoneweaver said. This kind of feels like that episode of The Office (US) where Michael tries to get everyone to act out negative stereotypes of different ethnicities as part of his "diversity training." I love that this filter is happening and am very excited to see it put into place (with the understanding that there will probably be some bumps along the way), but coming up with the list itself really doesn't seem like something that is well-suited to a metatalk thread.
posted by DingoMutt at 8:18 PM on July 7, 2020 [7 favorites]

Mod note: Comment removed. This is thread has a pretty specific and narrow scope and it is explicitly not for general theorizing about free speech and censorship and so on. If you want to talk about that, save it for a thread where that's within the scope of the intended conversation.
posted by cortex (staff) at 10:20 PM on July 7, 2020

I'm not saying the criticism is invalid, but the alternative courses of action are also flawed. If they came up with a list with a consultant, they'd get criticized for a lack of transparency and for not getting community input. If they waited until they had established the anti-racism advisory board they'd get criticized for not acting fast enough.

There's not going to be a comfortable way to generate this list. The discussion seemed to be going about as well as could be expected; I know stoneweaver said they didn't read the preceding thread but there wasn't any bad faith "no that's not really a slur" debate happening.

I don't know, y'all might be right that this isn't the best way to handle this. But given the heated criticism they were getting for the time it took to establish the filter as it is now I can see why they wouldn't want to wait to make it more robust.
posted by JDHarper at 10:49 PM on July 7, 2020 [18 favorites]

Sometimes there's no perfect way to handle things, but there are a lot of different flawed ways and some are more flawed than others. Sometimes you get criticism for not moving fast enough, which is going to happen no matter how fast you move if you slow-walked on things for the past 20 years. But that isn't a reason to do things the fastest possible way no matter the consequences.

As for transparency, it doesn't always mean discussions have to happen in real time in a completely open forum. It can mean people are informed about how the discussions did happen after the fact.

From my perspective, a better approach would have been to put this energy into fast tracking the advisory committee, and then see if they are willing to help generate a list of words. Then maybe have a metatalk that links to that list in one place only, and requests anyone who has feedback on it to contact the mods by email. Finish up by providing the community with a summary of the feedback and the decision-making process in broad brush strokes, along with a process for people to give further feedback in the future if they want to.

But that's just me and some thoughts I had off the top of my head. No doubt the advisory committee would be best placed to decide on a process for this kind of decision-making too.
posted by lollusc at 10:58 PM on July 7, 2020 [6 favorites]

I despise this s-word and would be happy never to see or hear it again.
posted by h00py at 1:21 AM on July 8, 2020

Following stoneweaver’s lead, could we take the list as it exists and try implementing it? People could use the contact form to add terms, and it should be a priority for the Advisory Board (which should also be a priority). This would avoid pushback and possible meltdowns here. As I said above, there are edge cases, in terms that will be very difficult to filter, but that is all the more reason why we should try to expeditiously get a solid core of unambiguous words in place.
posted by GenjiandProust at 4:18 AM on July 8, 2020 [3 favorites]

"there are consultants and best practices"...

I think we're getting this done very quickly, efficiently; and avoiding the expense of hiring a $3.99 a minute consultant.
posted by Afghan Stan at 4:34 AM on July 8, 2020 [1 favorite]

Speaking for myself, personally, I am not aware of any terms applied to Jews that IMO need the same treatment as some of the words above. There absolutely are slurs, but I don't think there any that could be used by someone accidentally or by someone pretending to be innocent. The only thing that comes close IMO would be someone using "jew", lowercased, as a verb. I don't imagine anyone would be hurt by a filter that ensured the word was always spelled with a capital letter.
posted by Joe in Australia at 4:52 AM on July 8, 2020

w-word, another c-word, a potentially difficult to implement n-word.
posted by Comrade_robot at 5:40 AM on July 8, 2020

Numbered Sequential World
posted by infini at 6:43 AM on July 8, 2020 [2 favorites]

deprecated term for USian

Huh? Whatever happened to calling people what they want to be called?
posted by emelenjr at 7:46 AM on July 8, 2020 [5 favorites]

iirc the argument was that USian was the unwanted term but memory does dim wiht the passing of years - 7 years for each second of pumpkinhead
posted by infini at 7:52 AM on July 8, 2020

(Let's not get sidetracked onto USian/American please; it's not the same thing.)
posted by LobsterMitten (staff) at 8:39 AM on July 8, 2020 [15 favorites]

+1 on Comrade_robot's w-word. Their c-word(s) should also go, I agree.

Here is an additional H-word suggestion.

# 45 on the list is a bit complicated. Definitely not used in Canada and considered offensive, though still somewhat in use in Alaska, by the people themselves. I think, as my link indicates, it is on the way out in Alaska, partly because of the Canadian influence. In 2016, President Obama signed legislation that replaced that term with "Alaska Native" in federal laws.
posted by gudrun at 8:41 AM on July 8, 2020

hurdy gurdy girl: Here is another c-word I think should be included, a Canadian slur used against Aboriginal people. It is a category C word, because there’s also a completely unrelated meaning, but is definitely still used as a slur (one that I’ve heard recently).

I actually think the two words (#44 on the list) are closely related and both should be banned -- while I'm not sure which came first, the coupling of an ethnic slur to sloppy and rapid alcohol consumption is not a coincidence. The OED says the first use of the drinking meaning is not until 1958 though so I wouldn't be surprised if the slur came first.

Use of this word to describe the sound of a steam train or a generator is perhaps a better false match confound to consider.
posted by Rumple at 9:56 AM on July 8, 2020

About No. 44 -- I don't know what we are going to do about false positives, but the dictionary refers to that as drinking in large gulps in general, with no reference to alcohol. I think of it broadly as that also. And according to Google, the word is also used to describe dogs that are a mix of Chihuahua and Pug.
posted by NotLost at 10:11 AM on July 8, 2020 [1 favorite]

More on No. 44, from the Word Detective.
posted by NotLost at 10:14 AM on July 8, 2020

I think it might be useful for a lot of these words if we had an algorithm that automatically flagged posts containing words that are sometimes used slurs as "potentially offensive" as soon as they're posted, so the mod on duty can take a look, rather than have the algorithm automatically delete the post and take action against the poster. If almost all flags for a particular word turn out to be false positives, the mods could consider taking it off the list.
posted by nangar at 10:28 AM on July 8, 2020 [8 favorites]

A note about #21: "[G-word] jazz" is a mistranslation of the French term jazz manouche. The word manouche is the term used by the specific Romani group Django Reinhardt was part of, who mostly live in eastern France, to refer to themselves. The word is derived from the Romani word for "human" and isn't considered a slur. I think most fans of Django Reinhardt and this style of jazz are familiar with the term jazz manouche. So if you want a non-offensive term to refer to this style of music, there you go.
posted by nangar at 11:45 AM on July 8, 2020 [13 favorites]

I'm quite alarmed that when poc say they disagree with how action is taken on this matter when it comes to racially based slurs, there is pushback from what i assume to be white people. If mefi wants to move the needle towards poc being heard and being taken seriously on matters that affects them the most, they should listen.
posted by xtine at 11:53 AM on July 8, 2020 [9 favorites]

Before I write this, as an FYI I am black. I do not speak for all black folks.
I feel someone should point out that #2 is a widely used phrase in the black community, and not as a slur. I'm thinking of (warning if you click my links, you'll see the word) when Larry Wilmore said "Barry you did it my [#2]" or multiple uses in hip hop. I use it affectionately with my friends. I'm not here to debate the origins of the word, that's pretty clear, but I do think it should be noted that this is part of many of our vernaculars and I don't think policing POC is what this filter is intended for. How that nuance works, I dunno.
posted by Pretty Good Talker at 12:05 PM on July 8, 2020 [6 favorites]

Would you clarify what you mean by "pushback," xtine?
posted by NotLost at 12:14 PM on July 8, 2020 [1 favorite]

Would you clarify what you mean by "pushback," xtine?

Stoneweaver wrote a comment which so far has 45+ favorites regarding that the work in the racial slur list would be better suited to a committee than worked out live on this metatalk. There are followup comments on this thread then sidetracking saying that a consultant would be expensive. Then further comments downstream completely ignoring that poc have said "this talk makes me uncomfortable" to continue to discuss what to add/not add to the list.
posted by xtine at 1:21 PM on July 8, 2020 [10 favorites]

I think that if there is a possible question of listening to people of color or not, people need to identify themselves as such. Otherwise, the rest of us can't be expected to know.
posted by NotLost at 1:33 PM on July 8, 2020 [1 favorite]

I think that if there is a possible question of listening to people of color or not, people need to identify themselves as such. Otherwise, the rest of us can't be expected to know.

I am going to hand hold you just this one time, I am not being facetious. One does not need to explicitly say I AM A POC when writing a comment. Yes, sometimes it may not be obvious to realize a commenter on mefi identifies as a poc. HOWEVER, in stoneweaver's case, you can simply click on on their username to see they have even tagged themselves with "poctakeover." This is really minimal effort.

Also, I think you need to realize that this is creating more work for us poc here. Rather than white people going HERE IDENTIFY YOURSELF, can you take a a few minutes effort to see who they are?
posted by xtine at 1:45 PM on July 8, 2020 [12 favorites]

Then further comments downstream completely ignoring that poc have said "this talk makes me uncomfortable" to continue to discuss what to add/not add to the list.

Some poc have said this, while others have not said it and have been making suggestions in the thread. Not a monolith.
posted by betweenthebars at 1:48 PM on July 8, 2020 [5 favorites]

I do understand that this is making POC uncomfortable, and an open metatalk as the way to come up with this list has some definite cons.

However, I also think that if we want metafilter to be a place where POC feel they are a part of, then all mefites need to be part of the solution to stop people feeling marginalized. I don't want the fact that there are new mods, and that there will be an Advisory Committee, to mean that everybody else thinks that now the work is done and that the problems are taken care of. I have uncomfortable feelings about an Advisory Committee made up of people who have probably had some of the slurs hurled at them being the ones who have to make this list, as one of the first things they may wind up doing (also there are some slurs on the list, correctly included I think, that are not directed at POC). It is with that in mind that I have participated in this metatalk: my feeling that we all have to start doing the work.

All that being said, I think GenjiandProust's suggestion is a possible compromise going forward:
could we take the list as it exists and try implementing it? People could use the contact form to add terms, and it should be a priority for the Advisory Board (which should also be a priority). This would avoid pushback and possible meltdowns here. As I said above, there are edge cases, in terms that will be very difficult to filter, but that is all the more reason why we should try to expeditiously get a solid core of unambiguous words in place.
posted by gudrun at 2:43 PM on July 8, 2020 [1 favorite]

I was too obtuse above. I agree with xtine (and stoneweaver) that discussion of slurs in this thread is... not great. Can we take what we have, implement it, and see how it goes? If people have slurs they really want to add, they can contact the mods, or the advisory board, when it is formed rather than discuss them here. I'm white, but these coy allusions, even well-meant, are upsetting, and not just to my fragile sensibilities.
posted by GenjiandProust at 2:52 PM on July 8, 2020 [4 favorites]

I don't think that 21 is a slur everywhere, or rather in particular in the UK it is a term that a specific group of people use for themselves in preference to alternatives. It is also used routinely for example in the UK census and when collecting demographic information; it is not a reclaiming just the normal word.

I think it would be a better candidate for a 'potentially problematic' list, than a 'cannot be used' list.
posted by plonkee at 4:38 PM on July 8, 2020

Catching up this afternoon; I've expanded the list with items suggested since I clocked out last night and have amended some of the annotations based on folks feedback. Thanks for helping out with this, everybody, and again if I missed something please let me know.

I'll follow up on a few things that folks have brought up.

I feel someone should point out that #2 is a widely used phrase in the black community, and not as a slur.

I think this is a real difficulty with this project and I appreciate you broaching it; I think it's going to similarly be a difficulty with some non-race-related terms too, wherever there's an issue of offense/slur vs. in-group reclamation of a word. There's a balance we'll need to continue to figure out how to strike between (a) preventing clueless/thoughtless harm by non-group members mentioning/quoting harmful language and (b) empowering in-group members to use the language that is meaningful to them, and part of that is not all folks within an in-group are going to agree about whether it's safe/okay to use that language in a more mixed-group context.

Like you, I don't know what the right way to resolve that nuance is, and that's something I'd rather continue to hear from folks on the site about than try to make a call on myself. I think it's a good example of something that a structured discussion with e.g. the advisory board could help set some boundaries on in the long term.

I think that if there is a possible question of listening to people of color or not, people need to identify themselves as such. Otherwise, the rest of us can't be expected to know.

People on MeFi get to decide whether and when and how they want to disclose details about their identity. That's it, end of sentence. As much as there are challenges involved in navigating intersections of social and cultural and racial issues in a pseudonymous text context, nobody gets to demand that other people post bona fides in order to be heard.

We need to be responsible to each other, and thoughtful about whether we're predicating a comment on personal details that aren't visible, and in general just generous in our assumptions. That's all true and sometimes it's complicated. But we need to trust one another to be representing ourselves fairly and thoughtfully and otherwise leave people's identity in their own hands. Anything that veers into "prove you are/aren't x" is super duper problematic and not how this place can work.

It is with that in mind that I have participated in this metatalk: my feeling that we all have to start doing the work.

That's the spirit in which I think this new thread was suggested last week in the initial slur thread, and in which Mitheral pursued putting it together. I appreciate the work they did there and the care they took in clearly outlining the content of and expectations for the thread to prevent folks from being surprised, and I appreciate folks in here by and large working to take similar care in how they forwarded suggestions and discussed some of the ambiguous or complicated items. It's been useful to get some input and feedback and expand the list according to MeFites' specific concerns and personal experiences, and again I appreciate the effort folks have put in here.

Can we take what we have, implement it, and see how it goes? If people have slurs they really want to add, they can contact the mods, or the advisory board, when it is formed rather than discuss them here.

I'd be alright with that. I think folks did a good job yesterday of adding content to the list and helping identify some of the complexities with some items. I can take the list as it exists now and bucket it out by relative ease or difficulty to implement, and work with frimble to get the more straightforward stuff in place first. There are a couple new moving parts required to accommodate multiple distinct words even without more complicated matching, so it'll take some time for frimble to nail down a new implementation. I'm hopeful that keeping it simple we can get a basic version of that working within a week, but that's my own estimate and frimble would be better able to scope it out.

As the advisory board is still in formation (travelingthyme is working actively on getting an initial board and meeting organized but it takes time), for now I'd say if folks have additional suggestions it'd be good to send them to the contact form or to me directly at [e-mail].

I can leave this thread itself open for the time being if folks have thoughts specifically about useful additional processes or considerations moving forward on this, but if it's seeming like the thread being open is counter-productive we'll go ahead and close it up.
posted by cortex (staff) at 6:04 PM on July 8, 2020 [5 favorites]

I am going to hand hold you just this one time, I am not being facetious. One does not need to explicitly say I AM A POC when writing a comment. Yes, sometimes it may not be obvious to realize a commenter on mefi identifies as a poc. HOWEVER, in stoneweaver's case, you can simply click on on their username to see they have even tagged themselves with "poctakeover." This is really minimal effort.


I disagree and really object to the idea that if someone does not explicitly identify themselves as a POC, you should assume they are white. We are not a monolith--there are a multiplicity of perspectives represented here. One thing that bothers me about MetaFilter's extreme white normativity is that there's assumed to be some sort of standard POC line on every issue. It leads to things like a white poster in the previous racial slur thread "educating" a POC on racism.

I'm quite alarmed that when poc say they disagree with how action is taken on this matter when it comes to racially based slurs, there is pushback from what i assume to be white people.

Please stop assuming.
posted by armadillo1224 at 6:42 PM on July 8, 2020 [5 favorites]

My comments have been taken out of context.

My comment about people of color identifying themselves were not isolated; they did not come out of the blue.

My comment was in response to this comment by xtine: “I'm quite alarmed that when poc say they disagree with how action is taken on this matter when it comes to racially based slurs, there is pushback from what i assume to be white people. If mefi wants to move the needle towards poc being heard and being taken seriously on matters that affects them the most, they should listen.”

In response, I wrote, “I think that if there is a possible question of listening to people of color or not, people need to identify themselves as such. Otherwise, the rest of us can't be expected to know.” The last sentence is key.

Tell me if you have a different interpretation, but I read that as xtine saying that white people should listen more to people of color about racial slurs. I read that as xtine saying that people of color should be given more weight than white people in any discussion about racial slurs.

Now, how are white people to do that, if the white people do not know who the people of color are? How are white people to know who to listen to more? How is anyone to know who is pushing against whom?

Those are not rhetorical questions.

If I wanted more credence in a discussion based on any aspect of my identity, I would not expect that credence to come out of thin air.

Nor would I expect all readers of a thread to click on the profiles of all commenters, to read and correctly interpret all of those profiles.
posted by NotLost at 6:58 PM on July 8, 2020 [2 favorites]

Plenty of the POC participating in this discussion are long-time members who are completely open about their identity. Is it really asking so much that if you want to do the work that you pay attention to who you're talking to? What do you lose by waiting to find out what the power dynamics are before you decide how to act?
posted by zeusianfog at 7:18 PM on July 8, 2020 [5 favorites]

@armadillo1224 Hi. I don't like to argue. However I would like to give you my viewpoint and I would like to be more understanding of yours. I am a POC. I see other open POC here say they are uncomfortable. I understand that not all POC will automatically agree on issues. However I do want to say that I would like for us to work together rather than to encounter "we are not a monolith" statements. It does alarm me when several long time POC mefites get pushback here. I only say "assume" because when I do try to click through to posters profiles to try to understand who they are, some of them present as white in either a picture or what they say. I say "assume" because there may be white passing people. I don't know for certain. However there is a history of white people automatically overtaking the narrative when POC open up to their vulnerabilities.

My wording of alarmed means just that, I am worried. I do not want efforts to help POC (as this thread is exclusively titled for race) be stymied by white people arguing. I do want to understand your perspective if you feel otherwise.
posted by xtine at 8:52 PM on July 8, 2020 [3 favorites]

NotLost, we don't need to talk about this in terms of hypotheticals and generalisations. Stoneweaver is the person who made the original comment. Stoneweaver is active in almost all race-related threads and is open about their identity. I would think that most people who have read more than a couple of race-related threads on this site would know that Stoneweaver is a POC. If you didn't know, then first of all, it might be worth reading and listening more in these threads before starting to actively participate in them. Secondly, the comments replying to Stoneweaver's are the kind where the user could think to themselves, hey, the result of pushing back on this would be pretty different depending on whether this is a white user or a POC, so maybe I'll check their profile before doing so.

That's not the same as saying you have to check the profile of every person you engage with on metafilter. It's a pretty specific and probably (hopefully) rare, use-case, when you want to argue with a person about something they say about race and it will come across differently depending on whether their statement is about their own race or someone else's.

No one is saying you need to magically be aware of people's race if they choose not to disclose it anywhere. That's a pretty big strawman.

(Just to be clear about where I'm coming from in my own comment above and this one: I am white.)
posted by lollusc at 10:12 PM on July 8, 2020 [10 favorites]

(And I'm not saying that if a person doesn't explicitly identify themselves as a POC, we should assume they are white. I agree with Armadillo1224 that that is harmful and unfortunately happens far too often here too. I don't think that was the case in the other thread they refer to though, where a white person was trying to "educate" a POC about racism. If it's the one I think it was, the race of the POC was very clear from their original comment and the white user was not paying attention.)
posted by lollusc at 10:18 PM on July 8, 2020 [5 favorites]

I think that if there is a possible question of listening to people of color or not, people need to identify themselves as such. Otherwise, the rest of us can't be expected to know.

Unsurprising to me that the first suggestion is to put the onus on a POC to self identify to be listened to when a person makes a recommendation that is intended to make this place more hospitable to a POC.

Also unsurprising to me that it never occurred to suggest instead a white person self identify when they're pushing back against such a recommendation. Such is the privilege of white being the default.
posted by Karaage at 8:08 AM on July 9, 2020 [20 favorites]

I can't emphasize enough what Karaage said. I mostly stay out of race discussions on Metafilter (and honestly, in real life too) because of this constant dynamic where white people get to be the default and "can't be expected to know" how to behave kindly to others (of any skin tone) who are literally saying "This word is harmful."

At least with the straight-up white supremacists you don't get this Good Place Committee nonsense.
posted by basalganglia at 9:34 AM on July 9, 2020 [11 favorites]

History lesson on the t-word.
posted by Ideefixe at 3:01 PM on July 9, 2020 [1 favorite]

The North American Scrabble Players Association (NASPA) moves toward banning 200+ slurs: arstechnica story; NPR story; NASPA webpage.
posted by gudrun at 7:56 PM on July 9, 2020 [4 favorites]

Number 20 is problematic; capitalized, it is the surname of some beloved relatives of mine. Apparently it's an abbreviation of an Irish or Scottish surname with the "Mc" or "Mac" prefix. Perhaps it would be possible to allow the version that begins with a capital letter?

I am white, for what it's worth.
posted by SereneStorm at 9:54 PM on July 9, 2020

^SereneStorm, here's one in-thread response, a couple of days ago, to the same concern:

Number 20 is the last name of a beloved actor, and part of the name of a town in Minnesota. Maybe it should be banned anyway, I honestly don't know the right answer.

"My opinion is that it is much better to ban these words and then have to do a little extra work when referring to cases like this (so, for example, abbreviating their names and linking to their Wikipedia page) than to let words like this appear on the site.

"In other words, let's prioritize the experience and feelings of people of color over the convenience of being able to type out rarely occuring exceptions." posted by mcduff at 1:36 PM on July 7
posted by Iris Gambol at 11:58 PM on July 9, 2020 [8 favorites]

Here's another list of ethnic slurs.
posted by chavenet at 1:29 AM on July 10, 2020

On the G word, a particular cowtown puppet show that makes fun of bad movies that I make Fanfare posts about and kinda-sorta run showings of (let's be specific, it's Mystery Science Theater 3000) might, or might not, be considered a false match. In the KTMA age the female robot character was sometimes called Gypsum, but that connotation seems to have been abandoned in the cable era.

In the most recent version of MST (produced for the stage), it's worth noting that the character has received a redesign to make her easier to transport, perform on-stage, and even participate in theater sequences. This character has been rechristened GPC, pronounced Star Wars style ("Gee-Pee-See"), due to fan requests that the name be changed to be less offensive.
posted by JHarris at 1:30 AM on July 10, 2020

(Re the Australian ones - I think 36, 38 & 39 are historical spellings that I've never seen used colloquially, and I think would now be regarded as misspellings for 35 and 37.)

Seconding zamboni. Never seen 36 used as a slur, neither 38 which is an Anglicization of an Aboriginal word that is used on the north coast. For example, bong bong can mean many watercourses. The town name Woodenbong is a Githabal word related to ducks on the water. I hope 38 is removed from the not-to-be-used list.
posted by Thella at 3:39 AM on July 10, 2020

Just a quick comment to say that I have read the list of words that cortex posted and would be happy to have them all auto-banned, with people needing to engage privately with the mods for an exemption if they feel they really need to be able to post one of them. (I would also be fine with the mods denying all such requests.) I am also fine with new words being added routinely upon request. Basically I would like us to bias on the side of preventing harm, because the cost of needing to rewrite to avoid harmful words is so extremely low.
posted by Susan PG at 5:24 AM on July 10, 2020 [2 favorites]

Using the numbers from Cortex's list I see three main categories, the last of which divides into a spectrum:

1. Words that can be treated as offensive in every context, like the infamous N-word. Metafilter isn't the place for these words even if they are used internally without offense by particular communities.
2. Words that are used self-descriptively by some communities but are offensive to others, like #45 (the E-word that describes North American indigenous groups.) This category will be tricky! I don't know what to suggest.
3. Words that have innocent meanings but are potentially offensive. There's a spectrum here:
   1. Words that are offensive even when not applied to people, such as #13, the K-word for makrut lime.
   2. Uncommon words that are potentially used by people being jerks. E.g., the older English word for "stingy". It sounds like the N-word.
   3. Less common words that are offensive when applied to people, like #33, the "Un-" word used for some subordinated Indian persons.
   4. Common English words that are offensive when applied to people, like #40, the G-word that also means a common alcoholic distillation.

I think the third category would ideally be flagged, and the user offered an alternative. E.g., like "makrut lime" instead of #13. That would relieve pressure on the moderators and allow a wider variety of words to be placed on the list without great inconvenience. Otherwise we're going to end up by enforcing discrimination ourselves. E.g., I don't think #40, the G-word, is in common use by racists today but regardless: do we want to say "sorry, our need to discuss alcoholic beverages outweighs your offense, so we're not going to list it"? We're going to have to use good judgment on at least some words, so maybe rely on warnings for at least the last two subgroups of this category. If someone is warned but then goes ahead with using a word in an offensive context it would at least clarify things: they weren't doing it accidentally or innocently.
posted by Joe in Australia at 4:56 AM on July 11, 2020

Words that are used self-descriptively by some communities but are offensive to others, like #45 (the E-word that describes North American indigenous groups.) This category will be tricky! I don't know what to suggest.

For these, would it be possible to arrange it so that you get an option to rewrite the comment, or alternatively, if you are sure that it's the right word to use in this context, your post can be reviewed by mods and maybe gets a note about terms/language? I feel uncomfortable with the idea that a marginalised group cannot be referred to by the word they would like to be used for themselves.
posted by plonkee at 10:53 AM on July 12, 2020

I don’t want to make a ridiculous assumption here, but CF or not there’s not any obvious webdev reason the language would preclude matching these with regular expressions?

Right. The point of identifying that stuff is not to mark it out as impossible to deal with, just to bucket stuff into groups of what's easiest to deal with first prior to additional work.

We talked over the details some in today's team meeting and will push forward on the low-hanging fruit on this; it sounds like much of the substring exclusion stuff is already at least halfway in place so I'm not anticipating any great difficulty there. frimble and I are talking through implementation details today with the hope of being able to roll out some updated functionality for hard-stop stuff in the next few days.

Warning-level stuff will be some additional work and a secondary goal so that we don't get roadblocked on that.
posted by cortex (staff) at 11:26 AM on July 12, 2020 [1 favorite]

In the middle of the night last night I was looking through all my comments using the word 'fire', and way down in the pile, one up from the bottom, was this old MetaTalk thread, which devolved into a discussion of whether it was appropriate to use ethnic slurs in a "neutral context" on Metafilter.

I had forgotten it existed, but people might be interested in checking it out to get a sense of just how far Metafilter has come in the last 14 years.
posted by jamjam at 6:31 PM on July 12, 2020 [1 favorite]

As far as implementation goes, would it be possible to make the slur list fairly comprehensive, but then temporarily "shadowban" matched comments instead of blocking them outright or requiring users to petition the mods? I.e., make it so it looks like it posted from the user's end, but hide it from everybody else in a private queue. That way the on-duty mod can look at it fairly quickly and approve it if it's not harmful (or contact the user if it is). This also avoids scandalizing folks with an unexpected "IT APPEARS YOU ARE TRYING TO POST A RACIAL SLUR ಠ_ಠ" message when they're just trying to talk about pedagogy (#27) or their magnificent new housecat (#20).

The only potential issue I see is where to place newly-approved comments chronologically -- at the time they were posted, or as the newest comment in the thread? But then, most threads aren't real-time or chatty enough that this would make a huge difference.
posted by Rhaomi at 9:13 AM on July 14, 2020

I get your thinking there, but I don't have any interest in shadowban stuff. MeFi works best when we can make the guidelines and expectations explicit in the documentation and function of the site. If we're going to build something to signal that something is unacceptable or needs care or caution, I want that to be transparent to everybody involved.
posted by cortex (staff) at 12:32 PM on July 14, 2020 [2 favorites]

Just to clarify, it would only act as a silent filter for innocent comments like the ones I mentioned, to avoid unnecessary friction from false positives. If a term is actually being used as a slur, the mod could then dash off a note to the user letting them know what the issue is, with a copy of their comment for easy editing. This way we could have a much broader wordlist without worrying about making a lot of people feel frustrated over being wrongfully blocked by an overzealous filter.
posted by Rhaomi at 12:52 PM on July 14, 2020

Re #20, again, and edge case worries, again: It's not like Maine is churning out such a varied assortment of felines. They can be "Maine cats," without much effort.
posted by Iris Gambol at 1:32 PM on July 14, 2020

Okay, I sat on this a bit but it really hurts my feelings that when I ask about ableist slurs I'm basically told wait your turn we're dealing with racism first since for a lot of disabled people it's an insection of discrimination.

It wouldn't be too hard to add some of the worst and most often used.

Anyway, yes. That was hurtful.
posted by 80 Cats in a Dog Suit at 5:16 PM on July 15, 2020 [1 favorite]

You have every right to feel and be hurt by us prioritizing dealing with racial slurs over ableist slurs. I can understand the rationale for wanting to have a relatively narrow focus for a thread like this but I don't like the idea of telling someone to wait their turn in order to receive a measure of respect and support from the rest of the community either. The mods should get in touch with you to see if you want them to put a post like this put up for ableist slurs or ask you for contributions that they can then add to the list directly.
posted by any portmanteau in a storm at 5:41 PM on July 15, 2020 [1 favorite]

I don't know if we get much of this on MeFi, but is it worth trying to keep up with the sewer of terms spawned as racist subreddits?
posted by benzenedream at 8:28 PM on July 15, 2020

Cortex' comment reads to me as he doesn't want other kinds of slurs to be dealt with in this post. That means there should be other posts up for those slurs, or if we've decided that maybe people chiming in with suggestions in a thread like this isn't the way to go then there should be another way for people to add slurs that shouldn't be allowed to get posted. We're in the process of changing the site's code for the better, and yes it's kind of a small step but it is still something, and we should be harnessing people's willingness to contribute towards it instead of asking them to wait.
posted by any portmanteau in a storm at 10:02 PM on July 15, 2020

Right, I didn't want to expand this post beyond Mitheral's original narrow focus, to support not having it wander and create a lot of extra friction and bad feelings the way the preceding did with multiple different lines of conversation colliding. 80 Cats, I hear you on the frustration of having it be a "wait" thing for other categories of harmful language, and I'm sorry about the hurt there.

We're limited in how much work we can do at once as far as implementation, which all of this hangs on, but I'm fine with the idea of doing organizational prep-work for other categories of offensive language in the mean time. My takeaway from this thread was there was sufficient discomfort on some folks' parts about crowdsourcing the list-making even with clear warning and careful elision of the suggested terms in thread that this approach probably isn't one we should repeat, so aiming for a more back-channel list-making process would be worth trying. I'm open to processes there, and it'd work fine for me to accommodate other categories in that same working spreadsheet on extra tabs, if that works for folks.

The good news at this point is frimble's been able to do significant implementation work on the expanded filter feature; we need to do some testing before making it live but the framework we've got now should make it easy to add additional terms to the hard stop process, and most of the stuff on the candidate list that came together in here will be able to be put into action as soon as that's live. We've had to make frimble change tack suddenly for top priority stuff the last couple days with recent site stuff, but I'm hoping to be able to test this and make it live in the next few days still as we get our feet back under us. I appreciate everybody's patience as we sort it out.
posted by cortex (staff) at 10:39 PM on July 15, 2020

cortex, any change in your view of the queue idea? That is, silently screen matches, then a mod can quietly approve false positives, or reject them -and- notify the user if it's actually harmful. (I think I made it sound like harmful ones should be silently removed, too, but that wasn't my intent.)

To make it easier for mods and users, the notification for mod-rejected comments could be an automated message with boilerplate about offensive language, a link to the new guidelines, and then a copy of the original comment with the matched word(s) highlighted so the user can easily edit it for resubmission. This avoids both frustration from being incorrectly auto-blocked and anger at losing a lengthy comment due to a matched word from a quote or a URL, and would help educate users about the guidelines at the same time.
posted by Rhaomi at 11:21 AM on July 16, 2020

That would be doing a lot more engineering to accomplish more or less what we're already implementing as a comment post/preview warning. Just for logistical reasons it's not on the table currently, we need to just get our straightforward expansion of the existing function tested and rolled out.
posted by cortex (staff) at 11:29 AM on July 16, 2020 [1 favorite]

Understood. I hope something like this is still a possibility down the road, because IMHO a filter with less nuance will be a lot more frustrating in practice and tend to discourage people from participating, especially new users.
posted by Rhaomi at 11:48 AM on July 16, 2020

FYI, I put something in the MeTa queue days ago about how to address other types of slurs.
posted by NotLost at 9:25 AM on July 21, 2020

NotLost, did you see this about posts needing more mod attention being delayed this week? Thinking that might be why your post isn’t up yet. (That post does say you should have been notified about the delay, I apologize if I’m telling you something you already know and you left your comment just an FYI.)
posted by the thorn bushes have roses at 11:58 AM on July 21, 2020

Oh, right, thanks, I knew about that. Mine went into the queue before they went up, and a mod told me.

I just wanted to let people know that something was pending.
posted by NotLost at 1:51 PM on July 21, 2020

I've asked Cortex to reconsider some of the words listed under 'Indigenous Australians' on the list. It appears that the Australian entries came directly from Wikipedia and not an actual Australian. The sources are not reliable and the title 'Indigenous Australians' is not the preferred term as it lumps many nations and languages together and does not recognise the diversity of cultures of First Nations people in our country.
posted by Thella at 2:22 PM on July 29, 2020

Yeah, I'm happy to amend the text there; if you have either a specific suggestion of better resources on those usages, I'd welcome that.

I created a second tab on the linked spreadsheet, "stop strings and messages", that is what we're using as a reference for the actual implementation; if you look there, you'll see there are some entries I've marked up as being either unlikely/unused (like the variant spellings in 36, 38 and 39 you mentioned previously) and so excluded, or complicated because of common homomorph collisions (e.g. 33 and 40) and so likewise excluded. Everything else is what's in place on the version of the feature in testing, and the "stop message text" is what will actually displayed on the site in the comment preview box when a given term is matched.

The "offensive use" field in the spreadsheet is intended more for organizational reference than anything; I'd still prefer to make sure it's accurate and so feedback there is totally welcome and I'll incorporate it, but it's the "stop message text" field that will be visible. Currently for 35 and 37 (and most of the other things on the list, that text is

Your comment includes the word [x], the history of which as a racial slur makes it unacceptable to post on MetaFilter, regardless of intent.

There are a few terms where the specific nature of the problem is described a bit differently, and there may be other terms using the above default descriptor where customizing it would make sense. But as a starting point it gives folks a starting point should they find themselves using one of these words without being aware of the issue, and we can iterate from there as needed.
posted by cortex (staff) at 3:14 PM on July 29, 2020

Here's a list that is more extensive but has less contentious entries (some contextual of course).
posted by Thella at 2:22 PM on July 30, 2020

Missed edit window: I meant 'fewer contentious entries' not entries that are less contentious.
posted by Thella at 2:48 PM on July 30, 2020

Mod note: One deleted - Thella, can you either send that list by email or allow me to edit the terms so they aren't in plain text in the thread?
posted by LobsterMitten (staff) at 3:08 PM on July 30, 2020

Mod note: Ok, edited to remove the in-thread list and un-deleted the comment, people can check the link if they want. Thanks.
posted by LobsterMitten (staff) at 8:53 AM on July 31, 2020