π

Please Don't Misuse Content Warnings on Mastodon

Show Sidebar

If you are user of the social network Mastodon (part of the Fediverse), you may have stumbled over the feature called "content warning" (CW):

Mastodon features a Content Warning system. It’s an optional mask that covers the content of a post with an editable warning message.
It’s used to cover content that is admitted by your Instance policy but may still hurt people, like spoilers, nudity, depiction of violence or threads about sensitive topics.
For example, if you want to start a thread about the ending of a fresh new movie, you can do it using a CW like “Spoiler about the ending of...”
Every Instance has its own rules about CWs and therefore it’s common to see them used in different ways, like on selfies or depictions of food. That is because what on an Instance is considered a sensitive topic on another may be something commonly accepted. It’s possible that an Instance is blocked by others because of its misuse of CWs on certain kinds of topics.
This said, if you want you can always go in the Setting page and set to automatically uncover all the CWs.

While there are tons of valid reasons to use a CW, there is a growing number of posts that do seem to dilute this great idea of CWs.

Random Real-World Examples

I just took screenshots of some posts from my current personal and global timeline to give you an impression on what I mean. The examples are absolutely random and are only used here to demonstrate the real-world use of CW.

Mastodon post: "Facebook, help request, boosts appreciated".

"Facebook, help request, boosts appreciated" doesn't give me enough clues why it requires to click on "SHOW MORE" to read the content. After clicking on the CW button, I saw:

If you could visit this link: [...] and tell me what you see, I'd be super grateful. [...]

Well, I really don't get why this required a warning.

Mastodon post: "FOSS Rec? (Help Wanted)".

FOSS stands for Free/Libre Open Source Software. I don't know what "Rec" stands for in that context. Well, I had to click on "SHOW MORE" to see:

Hello all!
I'm looking for a FOSS journaling application for my phone or computer. I would like for it to be able to use text, images, url links - but dates are not necessary.
I'm having trouble finding much on FDroid.
If this helps the inquiry, I'm looking for something to catalogue my house plants in and put down information like their name, watering, light, soil, photo, etc.
Let me know if you have anything in mind!
#foss #help #recommendations

OK, now I know that "Rec" stood for "recommendations". But why on earth did I have to click on that button? Who would be offended or triggered by this message?

Mastodon post: "ph-".

This one's perfect. Two letters and a hyphen that doesn't mean anything to me at all. I can not have a clue about the content of the post or the motivation why I should expand the message.

I did anyway and could read a harmless personal rant about a person who was helping a peer to get rid of a mold although the person is allergic to mold.

Poor gal but I don't see any reason why this deserves a CW to click on.

Implications

Somebody could argue: "If you are annoyed to click on the button in order to see all the CW, you might enable following option in the settings of your Mastodon account":

Mastodon preferences with "Always expand posts marked with content warnings" (bottom). (click for a larger version)

Yes, this really would solve the issue for me.

However, this would be shortsighted for this great platform in general.

Imagine a person who is really triggered by certain topics such as suicide, war, death, or other clearly horrible topics. Such a person is not able to enable the option above because there are really good cases *where CWs are perfectly valid and in my opinion also necessary*.

Following that logic, you either have to decide to enable the global preference to basically disable CWs by showing their content without clicking on "SHOW MORE" all the time. Or you do need to mass-click posts of your time-line in order to read perfectly normal posts that do not need any warning at all.

This kills the CW feature for this platform and it harms the people who do rely on warnings.

If you do try to find as much reasons to use a CW, you're going to end up with a CW for every post. You can't win here. You can always find a poor soul who's offended by any post you do without CW.

What to Do About It?

In my opinion, there are great alternatives to mis-using CWs for non-warnings or warnings related to stuff that doesn't trigger most people.

Everybody should learn how to maintain a set of words or hashtags in his or her personal filter settings:

Mastodon preferences for filters. (click for a larger version)

Using that method, you can avoid contact with certain topics. This makes much more sense than forcing the sender to anticipate potential personal aversion of something for each potential reader.

Use CW only for things that typically triggers normal people, not snowflakes. If you triggered now, please do read this article how I mean that. The Internet is and never will be a safe space for everybody. Using Mastodon filters is a near-perfect tool of providing a solution for this issue.

Of course, this also requires the use of proper #hashtags which I also recommend for other reasons because on Mastodon you can search for content using hashtags only.

My personal approach is that I'll send out links to this article to people who mis-use CWs. If I get a response that doesn't explain why a warning was really necessary (I might change my mind as well) or if the person can't explain the reason to me and keeps using CWs like that, I add the account to my list of muted accounts so that I never see anything again from them.

Small Mastodon Survey

Update 2022-06-07: I should have done this before I wrote this article. I started a Mastodon survey which asked for the preference of using CWs. I tried to come up with neutral questions in order to minimize any bias.

Here are the results after 53 people took part:

alterantive-text for the image
Mastodon survey with results: 36% for CWs everywhere, 25% don't care, 40% only for important warnings.

To my surprise, it's a result with no clear winner.

Anyway, if the majority of people voting for a very broad definition of CWs, I'd still think that we should change that behavior as it is hurting the platform.

Back in the good old days of the Internet, we had a Netiquette which contained rules and suggestions so that services and platforms like email, the Usenet, instant messengers are working fine for its users.

I miss something like that on nowadays platforms in general. And it should be clear that this is not a matter of taste according to the majority of the users, this is a matter of providing workflows and habits so that the platform is working in a good way. Therefore, I don't care for the current majority, I care for the long-term aspects for the platform.

Update: Friendica posts in Mastodon

Friendica posts in Mastodon tend to come along with CWs where the title is shown and the rest is hidden. If you're interested in the technical details, please do read this thread that describes the issue at hand and how you can fix it in your settings.

Update 2024-03-04: I received a comment by @jupiter_rowland which I want to quote here:

Posts and comments coming into Mastodon from Friendica, Hubzilla and (streams) are special cases. That's because neither of them have Mastodon's dedicated CW field. Instead, they have reader-side content warnings that can optionally be created automatically. By default, a post is automatically hidden behind a content warning button if it contains "nsfw" and/or "sensitive".
Automatic reader-side content warnings have been part of the culture of both Friendica and Hubzilla since before Mastodon was even made as well as of fairly new (streams). Poster-side content warnings in a dedicated field have never been part of their culture, both because they've got a much better solution for content warnings and because they don't provide dedicated means for poster-side content warnings.
Technically speaking, Hubzilla and (streams) do have Mastodon's CW field. But unlike Mastodon, it isn't labelled "CW". It's labelled "summary". It has been a summary field originally, but since a hard-coded maximum of 500 characters doesn't require summaries, Mastodon repurposed that field for content warnings. Hubzilla and (streams) have no character limits at all, so using that field for summaries is still justified.
Hence, Hubzilla and (streams) users use this field for short summaries for very long posts, if at all, but not for content warnings.
Also, just like Friendica, both Hubzilla and (streams) have a conversation model like Facebook or Tumblr or blogs with only one post and many comments. All replies are comments, and there is a separate, dedicated entry form for comments under each post. This entry form does not, however, have a summary field because what sense does it make to give a summary for a blog comment?
Hence, Hubzilla and (streams) users can't put Mastodon-style content warnings on replies at all. Also, neither Hubzilla nor (streams) can re-use content warnings from posts or comments they reply to.
Friendica, the oldest one of the three, doesn't even have a dedicated summary field. The only way to add a summary, i.e. a Mastodon-style CW, is between the Friendica-specific BBcode tag pair [abstract][/abstract] . This works in both posts and comments, but the availability of this tag pair is neither advertised in the post editor nor in the comment editor.
Since Friendica doesn't have a dedicated UI element for Mastodon-style CWs, neither for posts nor for comments, most Friendica users don't give Mastodon-style CWs at all because they simply don't know that this is possible in the first place, much less how.
Also, users of Friendica, Hubzilla and (streams) tend to find the idea of writing a content warning between a pair of obscure BBcode tags (Friendica) or into the summary field (Hubzilla, (streams)) ridiculous. Trying to convince them otherwise would be like trying to convince Mastodon users to double their CWs with keywords or hashtags to trigger the NSFW content-warning generation on non-Mastodon projects.

Update: Popularity of This Article

With the 2022-11 mass migration from Twitter to Mastodon instances, this article gained momentum. At a certain time, it became even my most popular article:

With the latest #Mastodon boom, my article has over three times as many readers as the second most popular article on my blog. Over 10% of all page visits end up there. It does seem to hit a nerve. #CW #publicvoit
My message from 2022-11-08 about the popularity of this article.

It got even mentioned in media such as this daily dot article.

I got all sorts of comments on Mastodon. The vast majority was very positive and supportive. Only a few negative comments were written. Most of those people did not seem to have read this article because they simply objected to the idea in general without going into details about my concrete arguments listed here.

As I wrote: It does seem to hit a nerve.

Update: Eugen's Message on Consensus on Usage of Content Warnings

On 2022-11-11, Eugen (the founder of Mastodon) published an interesting message that relates to the subject at hand:

Having been here since 2016, I can tell you there is definitely no such thing as a consensus on usage of content warnings on the fediverse. It's a decentralized network that doesn't belong to any one party, so by definition there is no single culture on it. Different corners have different expectations and customs. [...]
Eugen's message about "consensus on usage of content warnings".

While some people may think that this is settled and everybody may use CWs as they wish, I do think that all of my arguments above are still perfectly valid.

Of course, CWs that can't describe properly what's hidden below are an issue in any case. It's useless like the popular but senseless email subject "a question".

For all the people who really do rely on content warnings (and I do mean warning when I write "warning"), any different use of CWs other than for warnings with a clear purpose is an issue.

We might start to use the CW feature for titles, summaries, abstracts, whatever, yes, of course. As Eugen is writing, the consensus of the community may change. However, this also means that there is no working "warning" feature any more. First, we would need to rename this feature as soon as possible in the UI and documentation in order to avoid misunderstanding and establish the new consensus.

At the same time, it is inevitable that people with legit issues with problematic topics (death, suicide, pornography, strong violence, ...) for whom this feature was made originally (I presume), are not able to profit from its use any more. You can never tell what an arbitrary sender thought about this feature when using it. There might be some very violent content hidden, there might be an informative message suited for everybody. You can never tell any more.

Neighbor Bob gives rat a list of sensitive topics and asks rat to add a trigger warning for all those topics. Rat explains that he is sensitive to the word "trigger". In the end, rat reports to swine that he and Bob can't no longer converse.
Pearls before swine about trigger warning topics. (click for a larger version)

To me, you can not weaken up CWs without accepting the consequences.

I'm lucky that I don't rely on CWs because I'm able to deal with most problematic content I've come across so far. The reason why I wrote this article in the first place was sympathy for the people who don't have that luxury.

Warnings and titles/subjects/summaries/abstracts are two very different use-cases to me. Each do have their legitimacy. Done properly, we would require to have separate features for explicitly defining both. I would not recommend Mastodon to go into that direction as the interface should keep its current simplicity and not overcomplicate things.

So far, I did not read or hear any single argument that proved the arguments of this article wrong. I got the impression that people tend to ignore the fact that using CWs for a different purpose than for warnings does cause issues for people who relied on CWs so far. Either way, we need to understand the consequences of any change of consensus in that direction.

Update: Github Issue on Renaming CWs

There is a Github issue which is discussing to rename the CW feature to something else.

Update: Research Proves CWs Fruitless

Here's a research study from 2023-08: A Meta-Analysis of the Efficacy of Trigger Warnings, Content Warnings, and Content Notes

It's free to read online. I'm just quoting the Conclusions:

Existing research on content warnings, content notes, and trigger warnings suggests that they are fruitless, although they do reliably induce a period of uncomfortable anticipation. Although many questions warrant further investigation, trigger warnings should not be used as a mental-health tool.

I guess that really supports my point here from a slightly different point of view.


Related articles that link to this one:

Comment via email (persistent) or via Disqus (ephemeral) comments below: