@realTimo @tuxom @wion
As far as I know this attitude to flag "difficult" content came up in the heydays of quitter.se
with its main admin Hannes clearly opting for banning unpleasant and politically inconvenient views not of his taste. It generated a huge debate on censorship back then, with the usual rebuttal that "being free to speak doesn't entail being free to be heard".
Technically and legally, Hannes and others had a point though. As their servers were based in Sweden, and Sweden has certain laws prohibiting the public posting of content like Nazi propaganda, the provider/admin of such a server would have been legally liable for the content appearing on the Public Timeline of his server.
Imagine you were on that GNU Social instance in Sweden (even if it prohibits Nazi stuff just for legal reasons) and subscribe to somebody on a server that is located in a different country in which the posting of Nazi stuff is legal (the U.S., e.g.). Like consenting to access adult sites, you may still subscribe to this account posting Nazi stuff. The Nazi stuff enters your Home Timeline (and by that the Whole Known Network of the instance). Problem now is that not just a repeat but also a (critical) reply could now heave the original Nazi post (via conversation) into the Public Timeline of the Swedish instance, thereby violating the Swedish laws and put the provider of the Swedish instance at legal risks. That was the technical (or legal) reason quitter started blocking instances and accounts.
Which was perfectly fine. In a sense. (To protect the provider of an instance from legal liabilities). But it had two impacts:
1) You were restricted in to whom you could reply to. Thus the sensible ban of Nazi stuff got mixed with the nonsensical intrusion in your freedom to even critically reply to such stuff. Restricting federation entailed restricting freedom of speech.
2) In the end, the Fediverse would divide into mutually sealed off sub-diverses of instances that interact with each others and closing off those they don't.
Then the blocking was expanded to content that was deemed "inadequate", that is, of different political or "cultural" positions. SJW and safe space attitude creeped into the criteria of moderating the instance. Instances became topical instances. The admin's saying was: "It is my house, it is my rules." (Thereby deciding that an instance is a house, not a conversation.)
In this blend of legally justifable and subjectively felt necessary blocking the idea of CWs only added its own flavour. Given a now established "culture" of banning content, hypersensitivites and the wish for "safe spaces" put pressure on admins to introduce means that the Public Timeline was "safe" to watch for anybody. Controversial content needed to be flagged in order not to expose people to stuff they deemed insensitive or even traumatic.
This seems to have been the way with GNU Social and Ostatus. I don't know how Mastodon and Pleroma and their protocols add to or deviate from these difficulties. Perhaps their protocols now allow interacting with problematic content without heaving it into the Whole Known Network and then Public Timelines of an instance that cannot allow for such to appear.
Anyway, muting and blocking of accounts by accounts is available, and the rest evades my understanding.