It’s been very fun and heartening in the last week to see a big wave of people I followed on Twitter moving over to open source federated alternatives, namely Mastodon (or its fork Hometown, in my case). There are a lot of good blog posts making the rounds where people offer tips on getting oriented in the fediverse and reflections on its differences from Twitter; this is not one of those. Rather, I want to reflect on this present and probably brief period where such a huge portion of what people are posting on Mastodon is about Mastodon – this moment of lively medium metadiscourse as people help each other explore a new kind of communication channel.

As a radio historian, it’s fun to imagine this kind of moment as a small and very distant window into how radio operators in the first two decades of the twentieth century would have related to each other and to the medium they were trying to define. As Susan Douglas documented in Inventing American Broadcasting, amateur operators held club meetings over the airwaves where they would tackle technical problems and share aspirations for what widespread radio adoption in the United States might look like.

This is not to say that we should glamorize medium metadiscourse. Douglas’s book shows how a specific complex of white, middle-class masculinity, bolstered by hero-inventor stories, propelled these amateur operators. Today we can look to “web3” for similar forms of entrepreneurship fetishism in an aggressive and grift-pervaded medium metadiscourse whose medium, in most cases I’ve seen, doesn’t even exist yet.

My corners of the fediverse have mercifully little incursion from web3 or its driving interests. Even so, the present levels of medium metadiscourse there are unsustainable – people can only stay interested in the nuances of cross-instance hashtag propagation for so long. Mastodon’s test will be (or rather has been; at small scales relative to Twitter, it has been succeeding for years) how many people ride each platform migration wave to a place where they feel settled and comfortable with a group of people they can and want to talk with. If, when placing a phone call in 1900, you still thought more about the operation of automatic switchboards than about what you planned to say to the person you were calling, then the telephone would not have felt like a stable or friendly medium.

The upside, in my view, is that these moments are when media are at their least transparent, in Lisa Gitelman’s terms. “The technology and all its supporting protocols” (both technical standards and pieces of etiquette like answering the phone with “hello”) have not yet “become self-evident as the result of social processes” (Always Already New, p. 6). The “success” of the medium will depend on the eventual capacity for its users to become inattentive to those protocols in favor of the “content” that they help transmit. But in these more opaque moments, infrastructural curiosity is closer to hand, as are exciting potentials for differentiation from other media (as can be seen in expressions about how Mastodon isn’t and shouldn’t become just like Twitter).

The twist I’d add, following from a few years of researching American radio since 1950, is that metadiscursive moments are not unique to a medium’s early years. Radio today features a lot more emphasis on radio itself than it seems to have featured during what’s often called its heyday. This emphasis today tends to be nostalgic or defensive, assuring listeners of the medium’s advantages over newer kinds of channels, rather than exploratory; but I don’t think it’s incompatible with a kind of return to that sense of openness, as a 21st-century flourishing in experimental radio arts has shown.

One narrative frame I’m testing out as a generalization from radio automation’s history goes something like this:

  • A medium is emergent; its socio-technical protocols are open to and in need of definition
  • Protocols stabilize to a point that they become “transparent;” they succeed as a medium
  • Protocols and routines become so stable that many of them can be automated
  • A feedback loop of automation and capitalist predation can begin to strip out the internal work that differentiates meaningful channels within the medium, increasing its homogeneity
  • Users/audiences and capitalists increasingly abandon the medium
  • The medium’s protocols and resources are left under-used and once again appear available for exploratory uses


This process has taken a century in broadcast radio, but I believe we can map it more or less cleanly onto internet media at much shorter time scales. Anxieties about automation (“bots” and “the algorithm”) have been pervasive on platforms like Twitter for some time; one way to read them is that the social protocols for expression in these venues have become so standardized that they can be convincingly automated, and it grows less easy to tell an automated speaker apart from a “real” one. If I am right that high levels of medium metadiscourse characterize either end of the above cycle, then it will be interesting to see how Twitter users continue to talk about its transformation under Elon Musk’s private ownership; but it will be much more exciting to watch what comes of the current flurry of general eagerness to understand and adjust protocols for the federated social web.