4 Rules
4.1 Introduction
While the Fediverse promises more autonomy in social media operations, most servers on the Fediverse have actually adopted very similar rules, even directly copying them in many cases. Why, despite the opportunity to innovate, do Fediverse servers tend to adopt the same rules?
We consider this puzzle through the persepctive of institutional isomorphism. Through interviews with 17 server moderators and administrators combined with longitudinal records from thousands of servers, we find evidence of isomorphism between Fediverse servers and describe some of the mechanisms through which this occurs.
Our study suggests that rules on the Fediverse perform an isometric role in addition to a practical role: as a signal to other servers and to people on the server. Rules encode not only what behaviors are permissible, but also community values.
4.2 Background
4.2.1 Institutional Isomorphism
Meyer and Rowan (1977) consider rules as myths which insitutions incorporate to bolster their own legitimatacy, sometimes at the expense of internal efficiency. Under this neo-institutionalist view, individual actions are constrained by societal expectations, which in turn lead to the development of formal and informal rules within insitutions.
DiMaggio and Powell (1983) consider the problem of why orginizations tend to look similar to each other and theorize that orginizations tend to adopt similar norms through a process of insitutional isomorphism, which has three types: normative, coersive, and mimetic. Normative isomorphic change is a byproduct of professionalization. Coersive isomporpohic change comes from pressures from other organizations. Mimetic isomorphism describes how orgainizations tend to imitate each other.
4.2.2 Empirical Setting
The Fediverse is a set of independently operated servers which send data with the ActivityPub protocol. The most popular servers on the Fediverse use Mastodon, though alternatives like Pleroma and Misskey are also common. servers can pass messages to each other regardless of the actual software being used.
As the servers are autonomous, content moderation is handled differently by administrators and moderators on each server.
4.2.3 Content Moderation
Significant scholarship has investigated content moderation in the context of commercial, centralized social media platforms (e.g. Gillespie 2018; Roberts 2019). The function of content moderation in this context is shaped by economic factors: platforms often use contract labor and the people making case-by-case decisions are often divorced from the context of the communities they moderate (Gray and Suri 2019). However, a significant amount of the Web also relies on volunteer moderation: service work from unpaid people who are usually active members of their own communities. These volunteers might have different motivations for their actions.
4.2.3.1 Content Moderation on the Fediverse
Due to its technical design, the Fediverse faces unique content moderation challenges. Administrators may be responsible not only for posts on their own servers, but they may also have to deal with posts and conflicts from external servers.
Like the servers themselves, the costs of content moderation is distributed among many servers.
Servers must also consider their relationship to other servers from which they may receive posts. Hassan et al. (2021) explored the use of server-to-server federation policies and suggested that some server-level blocks may have collatoral damage on some users. Colglazier, TeBlunthuis, and Shaw (2024) measured the effects of server-to-server de-federation events and found asymmetric effects on activity for affected accounts on the blocked servers.
4.2.4 Rules in Online Communities
Some commercial websites like Reddit allow volunteer moderators for sub-communities to create their own rules; however, moderators must also consider the site-wide rules (Fiesler et al. 2018). The Fediverse, by contrast, requires no base layer of rules for each community and account: there are no site-wide rules (Nicholson, Keegan, and Fiesler 2023). Rules on the Fediverse therefore do more than on sites like Reddit.
4.3 Data
My analysis of the use for and function of rules on the Fediverse uses data from interviews with N people on the Fediverse and trace data from the metadata on M Fediverse servers.