Black Twitter, quoting, and white views of toxicity on Mastodon

Last updated: January 12, 2023.  See the update log at the bottom.

For a detailed look at design considerations of a "quote boost" implementation, see the followon post Social threat modeling and quote boosts on Mastodon

"The platform is made up of people. Elon Musk bought Twitter, but he did not buy the communities that exist on the platform. And one in particular has defined the zeitgeist here: Black Twitter, which media studies scholar Meredith Clark defines as “a network of culturally connected communicators using the platform to draw attention to issues of concern to black communities.”"

– For Black Folks, Digital Migration Is Nothing New, Chris Gilliard and Kishonna Gray, Wired, December 13 2022

"Whether we were watching ABC’s Scandal, dissing actress Paula Patton’s fried chicken, or calling out hypocrites with a quote tweet and “This you?,” Twitter is where we come to find humor, complain, and seek solutions."

–What happens to social justice and culture on Twitter with Elon Musk at the helm? by Elizabeth Wellington in the Philadelphia Inquirer, April 2022

"As André Brock notes in his book, Distributed Blackness, Black Twitter as a phenomena emerged by means of the affordances of Twitter, through how things like quote tweets, retweets and hashtags allowed Black users to engage in Black digital practices."

– The Whiteness of Mastodon, Dr. Johnathan Flowers, Tech Policy Press

"Why can’t we “quote boost” at Mastodon, the way we can “quote tweet” at Twitter? That might sound like an innocuous question, but it’s a volatile one. Quote tweets let you tag someone else’s tweet with your own addition above it, as a new tweet from your account. Both the original and your add-on are in full display."

– Reflecting on Twitter, White Flight, & “Quote Tweet” Tensions at Mastodon, Hilda Bastien on Absolutely Maybe

It's an interesting time on Mastodon, the decentralized open-source source social network with Twitter-like functionality.  In contrast to Twitter, where conditions continue to deteriorate, there's a sense of possibility on Mastodon and other areas of the "fediverse",  a network of interconnected sites that's been around for more than a decade.  

At the same time though the rapid growth is putting a spotlight on issues that haven't been addressed ... like quote boosts for example.

Terminology note: "boost" is Mastodon's equivalent of Twitter's "retweet", so "quote boost" (QB) is often used for the not-yet-existent equivalent of Twitter's "quote tweet" (QT).  Sometimes, though, people refer to the hypothetical Mastodon functionality as a quote tweet because that's what we're used to.  And since Mastodon's equivalent of tweets were called "toots" until the latest release, sometimes people call them "quote toots".

Does quoting "inevitably" add toxicity?  Of course not.

Mastodon's Benevolent Dictator for Life (BDFL) Eugen Rochko gave his reason for not allowing a quoting feature back in March 2018: in his view, quoting "inevitably adds toxicity to people's behaviours." It's not clear why he thinks this.  There are plenty of examples of positive uses of quote tweets, so they clearly don't inevitably add toxicity.  

• Amplifying and endorsing opinions, mutual aid requests, actions, or important reporting
• References to tweets in another thread.
• Starting up a discussion on a related topic to avoid hijacking a thread.
• Calling attention to hypocrisy [Wellington] – or racism, anti-LGBTQ+ language or attitudes, or patterns of harassment and abuse
• For Black Twitter, participating in a call and response phenomena similarly to "playing the dozens" in the Black community [Flowers]
• For Autistic Twitter, presenting parallel experiences as a way to identify with and express solidarity with a given experience. [Flowers]
• For cross-language communication [@mawhrin]
• For citation and pedagogical purposes. [Carwil Bjork-James]

The #MissedQuoteBoost hashtag on Mastdon has other examples.  The list goes on ...

And while it's certainly true that quoting is sometimes used to harass and abuse people, it's easy for harassers to share links and take screenshots even when quote tweets aren't available ... so it's really not clear how much toxicity they actually add.   From a research perspective, as far as I know there's virtually no good analysis of whether QTs contribute to abuse[1] – and in particular, nothing specifically looking at abuse of specific identities like Black, indigenous, trans, queer, and disabled people.  

It's really a stark contrast to research highlighting the value of quote tweets in general (and very specifically for Black Twitter and activism purposes) from Feminista Jones, Clark, Sarah L. Jackson, Moya Bailey Brooke Foucault Welles, Clark, Flowers, Brock, and many others.

A frustrating conversation

"I've been very frustrated with this entire conversation because I think there are two big assumptions that are taken without examination:

1. Quote-boost was a significant *cause* of conflict on twitter that is less present here. Decisions about things like if people can use quote-boost can be effective protective measures.

2.  Less conflict is *always* objectively good and this should guide decisions about UI."

– @futurebird@sauropods.win

Yeah really.  One reason quote tweets are so effective on Twitter at calling out hypocrisy (and racist language or assumptions) is the visual aspects, with more prominence on the quote ("this you?") providing context.  That often leads to conflict.

Do the quote tweets cause conflict?  

Are conflicts caused by calling out hypocrites always a bad thing?  

What about conflicts caused by calling out racists – or talking about Mastodon's whiteness?

Quoting happens anyway. So does toxicity.

I'll just say that the lack of quote toots did not stop me from being harassed to the point of blackmail Patreon pledges my first time here.

Harassers will harass regardless of tools.

– Creatrix Tiara

Even without quoting functionality, Mastodon's got a long history of toxicity. Creatrix Tiara, for example, was dogpiled within hours of arriving on Mastodon back in 2017, after commenting about the lack of people of color.

And it's worth mentioning that while arguing that quoting shouldn't be implemented because it "inevitably" leads to toxcisity, Mastodon's BDFL has also blocked functionality like local-only posts that would clearly reduce harassment, abuse, and other toxic behaviors.[2]  It's a great example of a point futurebird made in the Mastdon discussion: "Preventing abuse isn't the same thing as preventing conflict and often these can work in direct opposition to each other."

And just as happened on Twitter before quote tweets were officially introduced in 2015, people also find ways to quote on Mastodon even though the software doesn't have the functionality. A screenshot-and-link takes a lot more steps than a quote tweet (see the analysis below if you want the details) but if somebody really wants to call somebody out (or direct abuse or harassment at somebody who's called them out) they'll make the effort.

Sure enough, sometimes it leads to conflicts.

"As one of the Mastodonians that I have come to rely on for the history of users of color on Mastodon has said, trying to do “this you”– that is the Black Twitter call and response methodology where we point out a problem, the contradictions in a person’s statement by bringing up another statement that they’ve made and saying “this you,” right– that’s a narrow way of defining it. But doing something like “this you” on Mastodon invites a torrent of racist and sexist kinds of violence from other Mastodonians."

– Dr. Johnathan Flowers in The Whiteness of Mastodon

Huh.  It's almost like confronting people with hypocrisy (or racism) triggers the conflict ... not whether or not the platform supports quoting.  Who could have predicted?

"What strikes me with the mansplaining here when we bring up quote tweets, is how many of us who are experts on being harassed on and off line, people do this to. You don’t need to explain to me what leads to harassment when there have been articles written about the harassment I’ve gotten, and that doesn’t even cover 5% of it."

– Timnit Gebru

Affordances, norms, and inheriting whiteness

"Mastodon is a very white space. It is not unlike other tech spaces where whiteness is predominant. Insofar as this is the case, the norms, the habits, the affordances of the platform will inherit whiteness."

– Dr. Johnathan Flowers in The Whiteness of Mastodon

"Affordances" are the qualities and properties of software that defines its possible uses or encourages people to use it in one way or another.  Mastodon's lack of support for quote boosting is a great example of an affordance inheriting whiteness.  The unexamined racism of viewing a key component of Black digital practice as "inevitably toxic" is an example of whiteness as a norm.

One of the habits I see in discourse about quoting is the suggestion that since Mastodon was founded by marginalized people who wanted to get away from toxicity on Twitter, we should defer to the trans and queer people who have been here for years and don't want quoting. Don't get me wrong, as a queer person who's been on Mastodon since 2017, it's good that people want to take our opinions into account – as Rapid queer-led community innovation discusses, trans and queer people drove a lot of Mastodon's improvements in anti-abuse technology in 2016-2017.  And I really really really wish that site admins and Mastodon's BDFL would listen to the trans and queer people who have been advocating for local-only posts[2] ever since the trans-led GlitchSoc fork first implemented them in 2017!

When it comes to quoting, though, while it's certainly true that some trans and queer people feel that we'd be better off without it, the opinion is far from unanimous; quite a few of us think there are also good arguments in favor of quote tweets.  And trans and queer opinions aren't the only ones that matter.  As Flowers says:

"[W]hiteness is a large problem within the LGBTQ community. Being queer does not insulate one cell from the inheritance of whiteness"

So when cis white people cite white trans and queer people's perspectives on quoting without also bringing up Black perspectives, it's a great example of the dynamic Flowers points out: Mastodon's history often gets used as a shield from critiques of anti-Blackness.

I really don't care if QTs happen on this site, or don't happen. What's much more informative and important to me, is how this conversation is unfolding, and what is prioritized. I've heard everything from "The Black perspective isn't real" to "You Twitter refugees!" from folks with account create dates in November, talking to Black folk here since 2017

– mekka okereke

I don't really think mastodon needs quote-boost to thrive. It would be better, helpful, positive, but users are creative and will work around it. We will find new ways to communicate. What I'm concerned about now is what this restriction has come to represent.

Whose concerns matter?
Who is welcomed?
Who is supported?
Who isn't?

– futurebird on GitHub

An opportunity for progress

"I'm not sure why there is any debate on QTs.

There are multiple issues open on the project to implement it that were rejected. Forks have already been made to implement it as it's primarily a client side thing.

It's clear there will not be a consensus on adding it. So, either work with those that want it to implement it, or work with those who don't to filter it out."

– rose

Quote Toots are coming. The choice is to make them benign, or not to.

– Isabelle Moreton (@epistemophagy)

Mastodon forks like Fedibird have quote-boosts and so does other Mastodon-compatible fediverse [3] software including  Friendica, Misskey, and Akkoma. Just a few hours after I first posted this, treehouse.systems admin (and longtime fediverse developer) Ariadne Conill released a Mastodon patch adding quote tweets.  Even Mastodon's BDFL (Benevolent Dictator for Life) Eugen Rochko has said he doesn't feel as strongly about it as he used to.

The good news is that there's a lot of energy in the discussion trying to make them benign. There are multiple Github issues (1, 2, 3) discussing potential design implementation, as well as hashtags including #MissedQuoteBoost, #QuoteBoost, #QuoteTweet, #QuoteToot, and #QT – and lots of interesting threads, some of which are linked to from the QuoteToot page on the toot.cat MeW wiki.

If done right, there's a great opportunity for Mastodon here. There's certainly plenty of room for improvement over how Twitter does quote tweets. Are there are ways to design a quoting mechanism that make it less effective for harassers? Are there ways to make it less likely to trigger accidental dogpiles?

As Afsenah Rigot discusses in Design From the Margins, centering the marginalized people directly impacted by design decisions leads to products that are better for everybody. With quoting, this means working with, listening to, and following the lead of:

• marginalized communities who make extensive use of quoting, especially Black Twitter and disability Twitter
• people and communities who are the target of harasssment using quoting along with other tactics, especially women of color, trans and queer people and disabled people

As Bastian says in Reflecting on Twitter, White Flight, & “Quote Tweet” Tensions at Mastodon,

"The principles from the Design Justice Network are one of the good places to start to move the quote boost question forward. Those principles begin with the need to seek liberation from oppressive systems, and include prioritizing “impact on the community over the intentions of the designer.”"

Appendix: Comparing screenshot-and-link to Twitter's QTs

A common response to the original version of this post from  people who oppose adding quoting support to Mastodon make is that there's an existing alternative: take a screenshot and include a link.  Of course it adds "friction" by requiring several extra steps, but putting that aside for the moment, isn't the existing functionality good enough?  It's a good question, and worth addressing – the issues it highlights are important to consider in the design of a quote boost for Mastodon or other systems. Snce the analysis is fairly lengthy, I decided to put it at the end of the post in an appendix.

To get a handle on this, let's look at some of the key properties of Twitter's QT's as they're currently implemented, have several important properties.  To be clear, I'm not saying that Mastodon should just clone Twitter's implementation – there's clearly room for improvement! But, it's a concrete existing implementation, so a useful basis for comparison.

Screenshot-and-link can be a decent approximation to QTs in several ways

• It's easy to navigate from the quote to the original tweet (although if the original tweet had a link to an article you want to read it takes an extra click)
• People can see the new commentary as well as the original tweet and any embedded images (although not GIFs or videos)
• As long as the quoter has taken the time to put in alt-text, it's can be a good experience for people who use screenreaders (although it's a lot harder if the original tweet had an images, GIFs, or videos)
• As long as the quoter has tagged the person who made the original tweet, there's a notification when you're quoted (although it's got less information than a notification for a QT)

On the other hand, Twitter's QT functionality also does things that screenshot-and-link doesn't approximate:

• It's easy to see who's quoted the original tweet and what they've said – and to navigate to the quote tweets.  For playing the dozens, splitting off a new thread to have a related discussion, citing a thread as part of an academic discussion, or calling out hypocrisy, these makes a big difference.
• If somebody deletes the original tweet, or changes their account to protected, you can no longer see a copy in the QT. The screenshot, by contrast, still shows the original now-deleted tweet – good news for harassers, because it prevents the target from being able to stop the quote from spreading.
• A QT always shows the original tweet accurately. Screenshots, by contrast, can be edited, as harassers and disinfo spreaders frequently do.
• QT's of polls show the current results.  Screenshots, by contrast, show the results of the poll at the time they were taken.
• You can't QT a tweet from a protected account.  This is very important from a privacy perspective – although in practice, it only stops accidental quoting of protected tweets, because harassers can still screenshot-and-link.
• You can't do a "private QT" – quote tweet using protected account.  Twitter originally allowed private QTs, and it was a very common attack vector: the private account's followers could see the QT, but the person being targeted couldn't.

Interestingly, screenshot-and-link is in some ways more effective than QTs for harassers: the person being attacked doesn't get any information about where the attack is coming from – and they can't protect themself by deleting the tweet or protecting their account.

For most other purposes, though, QTs are noticeably more useful.

"Friction"

Which brings us back to the question of "friction", the extra work that needs to be done by the person doing the quoting and by the people looking at the quote. There are two ways to do a QT on Twitter, both fairly straightforward:

• when looking at a tweet, click on the "retweet" and then choose the "quote tweet" option and type in your commentary
• copy the link to the tweet you want to quote, then compose a tweet as you normally do and paste the link to the tweet you want to quote into it.

Doing a screenshot-and-link is much harder, especially if you want to include alt text, and even more so if the original tweet has an iamge.  Here's the steps I went through

1. copy the link to the tweet I want to quote, then compose a tweet as I normally would, paste the link to the tweet I want to quote into it, and make sure to tag the person whose tweet I'm quoting
2. go back to the original tweet and make a screenshot of it.
3. return to the tweet I'm composing, compose a tweet, add the screenshot as an image.
4. go back to the original tweet, select the text (including the tweeter's account and name)
5. return to the tweet I'm composing, edit the image, paste in the alt text (but don't click save yet)
6. go back to the original tweet, bring up the alt-text, and copy it
7. return to the tweet I'm composing, and continue to add alt text by typing the words. "With an image of " and pasting the alt-text from the original image

That's kind of a lot!  It's really worth trying it out yourself, either on Twitter or Mastodon, especially on the phone.  

When quoting is being used for positive purposes, all this additional friction's a significantly usability issue – especially for those who care about accessibility.  And note that selecting and copying text can be very challenging for people with impaired vision or motor skills, so the friction's noticeably higher for them.

When quoting is being used for harassment, additional friction is a good thing.  On the other hand, harassers are likely to skip adding the alt-text, so they can stop afer step 3 – the friction's a lot lower for them.

Notes

[1]  Hilda Bastian's Quote Tweeting: Over 30 Studies Dispel Some Myths (from January 12, 2023) is an outstanding look at research relating to quote tweets – but it didn't exist when I originally wrote this, so here's the references I originally included.

Magdalena Wojcieszak et al's Echo chambers revisited: The (overwhelming) sharing of in-group politicians, pundits and media on Twitter notes that quote-tweets of posts by "elites" are more likely to be used to provide negative commentary ... but negativity isn't the same as abuse, and the work equates anti-racists calling out racist posts by conservative politicians and racists directing harassment at progressive politicians of color.  

A few other interesting papers:

• Kiran Garimella et al's Quote rts on twitter: usage of the new feature for political discourse (from 2016) found that quote tweets had fewer insults than replies.
• Haesoo Kim et. al.'s November 2022 When Does it Become Harassment?: An Investigation of Online Criticism and Calling Out in Twitter, based on interviews with 32 Twitter users, suggest that "QTs could be perceived to be more aggressive when used for calling outpurpose."
• Martin Riedl. et. al.'s Platformed antisemitism onTwitter: Anti-Jewish rhetoric in political discourse surrounding the 2018 US midterm election notes that quote-tweets were used both to add antisemitic references and context in their quote-tweets, and to contest anti-semitic tweets.  Note however that this work appears to use the IHRA definition of anti-Semitism, which many scholars argue has been “hijacked” to protect the Israeli government from international criticism.

[2] Local-only posts can't get shared with other instances, so if you're on a well-moderated instance this significantly cuts down the risk of having your post targeted by harassers, terfs, or nazis. They're also valuable from a privacy perspective, since they prevent posts from being shared with admins or software that doesn't respect privacy).  The GlitchSoc Mastodon fork, which at the time was led by trans and non-binary women, implemented local-only posts in 2017; Darius Kazemi's Hometown fork has supported local-only posts for years.  Alas, Rochko's antipathy to them means that they're not available on most instances.  

[3] EFF's The Fediverse Could Be Awesome (If We Don’t Screw It Up) defines the fediverse as "an expanding ecosystem of interconnected social media sites and services that let people interact with each other no matter which one of these sites and services they have an account with."  Mastodon is the most popular software in the fediverse today but there are lots of other options as well.  Per Axbom's The many branches of the Fediverse has more.

Updates

• December 25: add link to treehouse implementation
• December 26-27: typo fixes
• December 28: add poll, rework list of positive purposes, include github quote from futurebird, add appendix
• January 12, 2023: update footnote 1 with info from Hilda Bastian's Quote Tweeting: Over 30 Studies Dispel Some Myths

Image credit: philroc, via Wikipedia, published under a Creative Commons CC0 1.0 Universal Public Domain Dedication.