After Twitter: Fragmentation,Platform Polities and Protective Sociality

Abstract

This article argues that a profound change has occurred in the spaces of social media, centring on the region formerly occupied by Twitter. More than Twitter rebranding as X, After Twitter refers to a historical punctuation point in the timeline of social media and an emerging social media reality. After Twitter registers the slow death of a set of ideals and related practices specific to platforms like Twitter, but also to the waning of ideals in relation to the communicative potentials of the open web more generally. We make three broad claims which characterise social media After Twitter: First, by way of an overview of alternatives and competitors including Bluesky, Mastodon, Threads, Truth Social and more, we observe a social media fragmentation. Such fragmentation is not solely driven by economic forces or technological development and instead is understood along explicitly political lines. Second, we observe the rise of polarised platform polities. These polities reflect divergent political positions, create distinct political realities and foster different modes of interaction and belonging. Third, we observe a general shift from connective to protective forms of sociality, where users approach social media as if they are constantly in the presence of adversaries, and the ‘weak ties’ that once defined a web of opportunities are replaced by an assumed toxicity of ties. We conclude by reflecting on the nostalgia for the Twitter-that-was, suggesting the need to foster a critical and reflective relationship with the Twitter of old.

Introduction

Something profound has changed in the spaces of social media. It centres on that region formerly occupied by Twitter. It’s hard to pin down exactly what that space was, since what Twitter was always contested and never singular. In their excellent ‘platform biography’, Jean Burgess and Nancy Baym (2022) remind us that Twitter was an ‘enigmatic platform – open to multiple uses’, though ‘there has always been a particular tension between the idea of a simple social technology for updating your friends on the one hand, and an informational public communication platform on the other’ (p. 12). Stressing this role in public discourse, they observe that it ‘has become the go-to for journalists seeking to follow the flow of news’ and that:

Many journalists, academics, and politicians are virtually dependent on it as a social listening, professional dialogue, and public relations tool, and it is widely considered an essential component of civic infrastructure for emergency communication. (Burgess & Baym, 2022, p. 4)

Whatever Twitter was for – updating friends, social listening, professional dialogue or PR – Twitter itself is no more. What is there instead? One answer is Twitter has become X. Twitter was rebranded as X on July 23, 2023. Indeed, as is well documented, Elon Musk acquired the platform in late October, 2022 and had already made many significant changes to the platform prior to the rebranding (Conger & Mac, 2024).

Twitter becoming X is only part of something larger, a change we call ‘After Twitter’. After Twitter is the slow death of a set of ideals and related practices specific to platforms like Twitter. It is the waning of idealism in relation to the communicative potentials of the web in general. It is a historical marker, a punctuation point in the timeline of social media – one with wide-ranging implications for users, third-party providers, platforms, governments, advertisers, researchers, educators and interest groups. After Twitter is a new social media reality.

Mapping this new reality is in part an empirical task: one can identify alternatives and competitors (as we do below) or detail Twitter’s profound changes as it morphed into X, for example. Our goal in this article is to add an interpretive layer to these empirical developments, to zoom out a little and outline what we see as the main contours of an emerging social media reality in the space formerly occupied by Twitter. In doing so, we see the article as a contribution to critical and cultural analyses of social media platforms and particularly those that seek to situate platforms historically. José van Dijck’s (2013) ‘critical history of the rise of social media’ (p. 5) is a notable early example in this regard, as is Nick Couldry’s (2024) more recent The Space of the World. Both seek to make sense of social media through reflecting critically on its development. We are also inspired by studies of platforms that work across technical, cultural, political and economic questions and are able to explore their interrelations (Bucher, 2021; Burgess & Baym, 2022; Poell et al., 2021).

This article makes three broad claims. First, through an overview of the pre-existing and emerging alternatives, we observe a social media fragmentation. While there is clearly more economic competition in this social media space, we suggest that fragmentation is not reducible to market logics, nor is it the result of purely technological developments. Second, we suggest a need to reframe platforms more explicitly in terms of politics, and to see them as distinct political entities which we refer to as platform polities. This is a context wherein a user’s choice of platform has become a knowingly political one. Platforms now articulate distinct political positions as part of their constitution; they aim to foster and support different kinds of discourse, communicative action, interaction and belonging. Understood as polities, we draw out questions of group identity and the capacity to mobilise for political ends. Third, we witness a broad shift in sociality. The connective ideal built into much social media and seen as a default good has been superseded by a strong desire for users to protect themselves and their like – we understand this as a shift from connective to protective sociality. Social media have always been filled with dangers, but what we suggest is a broader shift, whereby the very idea of the ‘open network’ no longer appeals. Instead, we now socialise under conditions of assumed potential antagonism, and where we once saw a world of potential friends, we now see potential enemies.

In the concluding section, we reflect on how After Twitter is also conditioned by memories of the Twitter that was. Here, we draw on the concept of nostalgia, as it has been developed in memory studies to illuminate how future platforms are conditioned by conflicted memories of the past. Whether on Threads, Mastodon or Bluesky, there is a tendency to see the old in the new and to see these platforms in terms of what Twitter was once imagined to be. Given that Twitter is no more, what kind of relationship might we develop with its legacy? How should we think across the historical divide? We approach this discussion in terms of nostalgia and use it to advance our own position with regard to how Twitter should be remembered and how its history should inform the present and future of social media platforms.

Social Media Fragmentation: Twitter Alternatives and Competitors

It is generally assumed that platforms are monopoly tending (Lovink & Rasch, 2013), but we are living through a counter period defined by social media fragmentation (Chatterjee, 2024; Lewis, 2023). While the platform model hasn’t been superseded, in the After Twitter space, we witness variation and emerging alternatives and competitors – both noncapitalist and corporate (e.g. Chatterjee, 2024; Lewis, 2023). In this section, we provide an overview of the new social media landscape and consider the nature of its fragmentation.

Mastodon and the fediverse attracted a great deal of media attention as Musk completed his purchase of Twitter in 2022. Mastodon is a Twitter-like microblogging service, but unlike Twitter, it was designed to be noncentralised. Mastodon and the fediverse run on a shared technical protocol called ActivityPub. As free and open source software, Mastodon can be installed on a web server by anyone with a bit of technical knowledge. Whomever sets up a Mastodon server can invite people to join it, and those users can connect to users on other Mastodon (or ActivityPub-enabled) servers. The result is a large network of thousands of small social media servers.

Because Mastodon is noncentralised, those who were alarmed by Musk’s purchase of Twitter saw it as a safe haven – no single individual could simply take over the network of thousands of servers. Millions of people signed up for Mastodon accounts in late 2022 and early 2023. However, not long after, Mastodon was yet another victim of the ‘killer hype cycle’ (Zulli et al., 2020), with news reports arguing that it was ‘slumping’ and thus failing to ‘kill’ Twitter (Hoover, 2023). Bluesky (discussed below) has largely supplanted Mastodon as the Twitter killer du jour (Louise, 2024).

While it is no longer presented by journalists as a ‘Twitter killer’, Mastodon and the rest of the ActivityPub-based fediverse are carrying on. In part, this is due to the predominantly noncapitalist approach of the fediverse. As a nonprofit funded by donations, Mastodon has largely eschewed surveillance capitalist funding models – few if any Mastodon servers sell advertising space, and none of them are selling user data to marketers. As a result, Mastodon and the fediverse enjoy the benefit of not having to grow. In interviews with Mastodon administrators and moderators, one of the authors heard a common statement: ‘I want the fediverse to be as large as possible, but I don’t want my server to be large’. By this, these independent operators mean that they do not want to be responsible for large userbases on their servers, which would make those servers harder to moderate. Instead, the admins call for growth in the number of servers on the fediverse. Indeed, this includes servers running ActivityPub-compliant software other than Mastodon. This can include other Twitter-like microblogging systems (e.g. Misskey) or software that emulates other forms of social media. In fact, while hype over Mastodon has subsided, as of this writing there is a hype cycle around Pixelfed and Loops, which are ActivityPub-enabled alternatives to Instagram and TikTok, respectively (Perez, 2024; Ropek, 2025).

In contrast to federated, noncapitalist social media, we also see corporate social media Twitter alternatives emerging. Two companies, Hive and Post, attempted to position themselves as viable Twitter replacements. Both ran into issues. Hive was developed prior to the Musk purchase and received interest in the wake of that event. But a cybersecurity audit prompted Hive to temporarily shut down in late 2022, stunting the momentum of growth it had been experiencing up to that point (Ropek, 2022). Post was launched immediately after Musk’s purchase in late 2022. The company received venture capital funding but did not attract enough users, shutting down less than 2 years later (Roth, 2024).

This is not to say corporate Twitter competitors have failed. One corporation benefiting from Musk’s purchase of Twitter has been Substack, which launched a microblogging ‘Notes’ feature early in 2023 (Peters, 2023). Substack is a blogging platform for writers who can ask for and receive monetary subscriptions for their writing, with Substack taking a cut. Elon Musk even sought to purchase Substack at that time, reportedly because he wanted to integrate the Substack subscription model into X (Burman, 2024). Another corporate start-up, Spill, has had growth, particularly among Black Twitter users who left in the wake of the 2024 U.S. presidential election (Davis, 2024; Hale, 2024).

All of the corporate competitors cited above replicate the topological structure of Twitter/X – they are centralised systems, with data and power flowing to a logical centre. Two other corporate competitors are experimenting with a decentralized approach, somewhat similar to Mastodon and federated social media. Bluesky, a Public Benefit Corporation, operates predominantly in the for-profit model of social media: accepting venture capital, engaging in marketing campaigns and seeking a path to profitability (as of this writing, through the sale of custom handles or premium accounts). However, Bluesky also tapped into the desire for decentralised social media apparent from the rise of Mastodon: Bluesky’s main service, bsky.app, is a proof-of-concept for an underlying protocol, ATProto, which is being designed to decompose the hosting of social media into various parts, such as data hosting, indexing and content presentation. Critics have pointed out that Bluesky is acting in a contradictory manner: by building a relatively centralised, for-profit Twitter alternative, what incentive does it have to build a decentralised protocol that could allow for its users to ‘credibly exit’ (Lemmer-Webber, 2024)?

Bluesky isn’t the only corporate social media system exploring decentralisation in a contradictory manner. Meta is doing the same with Threads, a microblogging service it launched soon after Musk bought Twitter. Threads is implementing ActivityPub, the same protocol used by Mastodon and other federated systems. This allows Threads users to follow fediverse members and vice-versa – assuming fediverse members do not simply block Threads (Theophilos, 2024). Such a move by Meta – the quintessential walled garden corporation – is surprising, though there are a number of possible explanations, including convenience, the strategic positioning of Threads in relation to X, and European regulations regarding account portability (Hof, 2023).

There are also a number of notable dedicated ‘alt-tech’ – that is, right and far-right – alternatives to Twitter which emerged before its demise, including Gab, Gettr, Rumble, Parler and Truth Social. Of these, Gab and Truth Social are adaptations of Mastodon, though they are closed off from the wider fediverse. These platforms are typically commercial in orientation, though some have struggled to attract advertisers (e.g. Gab). Subscription, premium features, payment processing, selling data and merchandising are other notable alternatives sources of commercialisation. Significantly, platforms such as Parler were able to attract early investment as a direct result of their political stances (Yurieff & Benveniste, 2020). These right-wing platforms were often set up in response to Twitter’s handling of specific users or events (Gettr, Truth Social) or the (now historic) opinion that the larger social media platforms were left-leaning and too restrictive (Gab, Parler). Gab, Gettr, Parler, Rumble and Truth Social carve out a small but significant part of the After Twitter social media landscape. Some seem in perpetual struggle and re-launch mode (Gettr, Parler), while Truth Social and Gab seem to have found their respective niche user bases.

Taken as a whole, After Twitter is defined by platform fragmentation. Fragmentation is an established term in media studies and has been used to describe the effect the internet has had on more traditional broadcast media – often journalism specifically – and their audiences (Mancini, 2013). Media fragmentation poses challenges to the ideals of an informed democratic citizenry as there is no longer an informational common ground to underpin discussion. Updated variations on this theme are found in discussions of ‘echo chambers’ and ‘filter bubbles’ (for a critical treatment, see Bruns, 2019) and more recently, ‘sharded media’ (Merrin & Hoskins, 2025).

However, while fragmentation After Twitter is related to these other notions, it has its own specificity. The current social media fragmentation has a technological component to it, but it is not driven by technology in the sense that earlier uses of the term often imply. Indeed, the current fragmentation tends to use different underpinning technologies to (nevertheless) reproduce very similar interface experiences. This is not the ineluctable, breathtaking force of a new technology, but the conscious deployment, development and adaption of software to create social media with varying resemblances to the Twitter that was.

The recent social media fragmentation does have informational and political dimensions to it but with important qualifications. Fears of echo chambers and filter bubbles were largely about the open internet not living up to its potential. The background assumption was that users are on sites and platforms with significant diversity of views and opinions, but through their own habitual clicks, preferences, follower selections and further algorithmic personalisation, users are not exposed to meaningful difference. ‘Chambers’ and ‘bubbles’ served as critical responses to more celebratory accounts of the open web and its potential to foster public discourse. The current fragmentation is much more deliberate. It doesn’t begin on the open web, where the like-minded eventually band together. The moment of homophily happens a priori, at the point of platform selection (more on this below). And while Merrin and Hoskins’s (2025, p. 25) notion of ‘sharded media’ points to a ‘cambrian explosion of realities’ emerging after the downfall of mainstream news media, we would like to stress the reality building and consolidation work that is also clearly evident After Twitter. These fragmented platforms are reality making, not only in terms of shared meanings and agreements regarding the world beyond the platform but also in terms of their infrastructures, technical features, platform rules and guidelines, and modes of interactivity.

A final note on platforms and competition. Couldn’t this claim regarding fragmentation simply be understood as competition as usual? Indeed, haven’t there always been start-ups and entrepreneurs vying to create the next big social media service, with a long list of failed endeavours or early acquisitions, from ello and Vine to Google+? We think there is an important distinction to be made, namely, that what is happening After Twitter isn’t fully captured through a market logic of competition. Much of what populates the After Twitter space emerges from a profound negativity; it is the retreat from Twitter that has energised these other providers. The term alternatives speaks more directly to the platforms that do not want to compete with Twitter/X – less Schumpeterian ‘creative destruction’ and more seeking to create after experiencing destruction. There are economies underpinning the fragmented social media space, but After Twitter’s fragmentation cannot be explained solely in these terms. Indeed, as we now turn, we suggest the current fragmentation needs to be understood more directly through a political lens. If platforms once sought to compete economically while maintaining an aura of political neutrality (Gillespie, 2010, 2015), the post-Twitter space is characterised by the emergence of what we call platform polities.

Reframing Platforms: The Rise of Polarised Platform Polities

In the fragmented After Twitter space, there is a proliferation of functionally similar social media. X, Mastodon, Gab, Bluesky, Threads and Truth Social all offer what first appears as a similar user experience in terms of design, layout and basic features. Upon closer inspection, however, they differ in fundamental ways, including approaches to moderation and permissible content; business and ownership models; underlying protocols; features; how content travels and is filtered and follow/follower dynamics. After Twitter, we suggest, these differences are best understood through a political lens, and this entails a reframing of how we think critically about social media as platforms.

Few commentators were willing to openly declare Twitter as a global or digital ‘town square’, but the platform was routinely referred to in relation to this notion (e.g. Otto, 2013; Sacasas, 2019; Yeung, 2023). It never quite lived up to the town square standard, yet time and time again it seemed an obvious reference point for thinking about Twitter. Whatever one thought of Twitter as a digital town square, we are now witnessing a radically different state of affairs. Social media platforms now each support and express divergent political positions, and the choice to participate on any particular platform is itself already a political one. One does not, for example, go on Truth Social to participate in a debate about DEI; one goes on Truth Social because a position on DEI is already established – even if one joins Truth to discuss non-capital-P Political topics, such as sports or television. Any notion of the town square (recalling the Athenian agora) is definitively lost. What is emerging is something very different. On X, the prior agonistic dynamics have now tipped in favour of a more or less coherent political stance and where remaining disagreement is now framed in terms of the ‘antagonistic’ (Mouffe, 2013, 2019) acts of the Other. Rather than drawing idealised similarities with the agora, the dynamics on X are now much closer to Athenian pnyx.

In his discussion of the agora and the pnyx, Richard Sennett (2019) makes a series of relevant distinctions. The two had radically different spatialities, with the agora (the main square) being open and used for many purposes, often at the same time, and the pnyx (the main theatre) being bowl shaped and used more narrowly for theatre and political gatherings. Sennett (2019, p. 206) describes how events unfolded ‘synchronically’ in the agora and ‘sequentially’ in the pnyx. Both spaces had their modes of being, with citizens in the agora moving through the hustle and bustle of others with purpose, though taking care to make ‘eye-contact with strangers’ when stopped (Sennett, 2019, p. 206). In contrast, the pnyx ‘gathered together more submissive urban bodies’ (Sennett, 2019, p. 207), who could sit for hours at a time and focus their attention on a single speaker or the spectacle of the theatre (Sennett, 2019, p. 208).

Without suggesting Twitter was ever an agora, X has come to look much more like a pnyx. With X, we see a platform increasingly organised around the activities and political will of a sole individual. Musk acquired the platform in late 2022 and immediately went to work reshaping the company. He made many changes in those first months, though one event was especially revealing. In February 2023, Musk was annoyed that his tweets during the U.S. Super Bowl had been outperformed by then-President Joe Biden. In response, he had Twitter’s engineers boost his tweets, with reports claiming it was by a factor of 1000 (Schiffer, 2024, pp. 1–4; Taylor, 2023). Users awoke to their feeds filled with Musk’s tweets – an ‘unrelenting flow’ – wondering what had happened. While this engineering hack would be wound back, it showed that the new owner of the platform expected to be its main voice and was very willing to mould the platform around this desire. A month later, in March, Musk overtook former U.S. President Barack Obama as the most followed account on the platform, and in the years since the acquisition, Musk has consolidated his position at the top (Wikipedia Contributors, 2025). While follower numbers can be misleading (see Binder, 2023), they nevertheless give a strong indication of his growing influence.

Much like the political orators of the pnyx, Musk is the dominant voice on X. He commands attention sequentially, through his own political stunts or through his own form of charismatic narration of external events. He shapes discursive norms on the platform through his engagements, giving legitimacy to certain topics (and positions on topics) while rendering others illegitimate. While he does this through his own posting, sharing and engagements, it is equally done through his changes to X’s features (e.g. blocking), content moderation and the restoration of formerly banned accounts (including Trump, Andrew Tate and Kanye West). In these ways, Musk’s voice should be seen as taking on infrastructural and regulatory qualities well beyond his posting activity. His theatre is designed for him, by him. But whereas the pnyx produced its submissive observers, who could sit still for hours, Musk’s followers are not idle spectators. They are followers in a much more active sense, helping to circulate material or sourcing new events for him to comment upon. Blending Foucault’s (2002) notion of governance as ‘the conduct of conduct’ (p. 341) with the orchestral conductor, Musk is perhaps best understood as a conductor of conduct. He commands the attention of his followers and their actions follow his signals. While X is the largest platform organised around a clear leader/follower dynamic, Trump has a similar position with Truth Social. From the beginning, Twitter was defined in part by its follower feature – its lack of reciprocal friending – and with the emergence of leader/follower platforms we see the affordance of this feature taken to its extreme political ends.

Some years back, Paulo Gerbaudo (2018, p. 66) documented the rise of the ‘the digital party’ or ‘platform party’, referring to the fact that some political parties had come to ‘adopt the quality of Facebook’ and the ‘logic of platforms’ more generally. What we see After Twitter approaches a reversal of Gerbaudo’s then-astute observations. Social media platforms are not actively creating parties, per se, but they are creating distinct polities. Here, we take care not to associate a polity with a state, government or geographical territory. International Studies scholars Ferguson and Mansbach provide an account of polities precisely as a way to think of political groupings beyond the Westphalian state model. For them, a polity has a ‘distinct identity, a capacity to mobilize persons and resources for political ends, and a measure of institutionalization or hierarchy’ (Ferguson and Mansbach, 1996, p. 262). Such polities have a spatial dimension which does not necessarily map onto geographical territories, though it does include a ‘territorial “reach”’ (Ferguson and Mansbach, 1996, p. 262).

With the leader/follower platforms (namely X and Truth Social), the act of polity building is highly visible. It is, in some way, the telos of the platform. But the other platforms can also be seen as experiments in polity building. As part of the fediverse, Mastodon is structured around a federated vision of social media and has been described in terms of a covenantal mode of governance, borrowed from federalist political theory (Gehl & Zulli, 2022). Bluesky is perhaps best understood as aiming to embody what Johannes Thumfart (2024) has called the ‘liberal internet’, closely connected with liberalism and aligned with liberal values. Each platform with its distinct collective identity, with the ability to mobilise for political ends (and we include here everything from mobilising bodies to enact harm off-platform to the creation of politically desirable protocols), and with its own institutional form – this is what we mean when referring to platform polities.

A market-oriented view of the rise of Twitter competitors would see increased choice and thus a better social media environment for everyone as platforms compete for users. But this assumes social media are broadly interchangeable. Instead, if we understand platforms as polities, we can see clear limits to the economic framing of choice. Ferguson and Mansbach (1996, p. 264) note that polities can overlap, and one polity can be nested within another, and one can easily imagine users participating in more than one platform polity. However, just as the Green Party in the United Kingdom cannot be easily substituted for Reform, or Buddhism easily substituted with Catholicism, not all platforms are interchangeable because they address matters far beyond the economic frame. One cannot substitute Gab for Bluesky, for example, without the user undergoing a profound political transformation. This extends to non-Political interactions on these platforms: sports or television discussions, for example, operate against the backdrop of the more explicit political positions developed by the platforms.

We speak instead of a polarisation of platform polities. Such polarisation shares significant continuities with the political polarisation widely observed in national and international politics, where far-right movements and parties have become increasingly popular and helped empty out the ground of the centre-right. Platform polities certainly overlap with and help facilitate this broader polarisation. For example, different platforms now often have wildly different takes on the political issues that once could be analysed from within a single platform through the notion of ‘issue networks’ (Marres, 2015). Basic search queries for currently contested political topics on the different platforms – Ukraine, DOGE, Gaza, DEI and so on – starkly show the new divide. However, when it comes to platforms, this is only part of the picture.

Mastodon was not made as a Twitter clone but as an attempt to build an alternative. Twitter was never the goal; it was the point of departure. Truth Social was similarly created in response to Trump being kicked off Twitter, and Bluesky, born from within Twitter, was designed as an improvement on it. Even X became X as a way to signal its shift away from Twitter. The specificity of these various movements is evidence of polarisation. Each platform performs this distance in unique ways, and rather than seeing a single pole or spectrum, one can identify multiple poles of alignment: features, moderation, regulations, ownership and content.

Reframing social media more explicitly around politics through the notion of platform polities also helps us to understand the changing nature of sociality on these platforms, including how people engage with more than one at once. Future studies could map out the political positions of platform users – very interesting studies could be developed based on users’ thinking as they maintain accounts across different platform polities. Are they doing so to participate in different polities, and what distinctions do they draw across those different contexts?

In addition, we observe a shift away from social-expansionist and outward tending social connectivity to concerns about protection. The latter not only refers to safety, including for marginalised groups, but equally the protection of users wanting to engage in hate speech. It is a protectivity driven by the values of the polity. These values obviously conflict across polities, within the fragmented space After Twitter.

Changing Sociality: From Connective to Protective Media

The explicit politicisation of social media, their becoming fragmented polities, is equally reflected in the changing conditions of sociality. In her excellent critical history of the first decade of social media, Jose van Dijck (2013) wrote about the rise of what she called ‘an ecosystem of connective media’ (p. 21). She uses the term connective media to describe how social media platforms turned users’ desire for ‘connectedness’ into a viable business, in part through the promotion and fostering of a particular kind of sociality, characterised by sociotechnical instantiations of practices such as liking, sharing, friending and following. Constitutive of this new connective sociality was the existence and latent potential of ‘weak ties’ (Van Dijck, 2013, pp. 8, 13), a notion borrowed from the sociologist Mark Granovetter (1973) and widely adopted by commentators writing about social media during this period. In distinction to ‘strong ties’, generally thought of as family and close friends, ‘weak ties’ are that much larger network of acquaintances – with Van Dijck (2013, p. 13) extending this to ‘total strangers’ in her account. Because weak ties stretch beyond close family and friends, and act as bridges to expanded networks, they have often been considered valuable as sources of information or for new opportunities. The concept of ‘weak ties’ was informed by a network ontology, where society was approached as a network. This translated easily to understandings of social networking sites built on the internet. Connective media offered a sociality based on formalising connectivity to these expanded networks of weak ties. Van Dijck’s critical account showed how the assumed widespread will to connect was captured by platforms and made to serve their own largely economic ends.

The Culture of Connectivity was written at the peak of the excitement over the potential of social media, when companies such as Facebook were pushing their mission to ‘make the world more open and connected’ as an assumed good. The following decade of social media – roughly 2013–2023 on Van Dijck’s timeline – would see a whole range of features or dedicated apps that aimed to reclaim more of a space within the open world of connectivity for those stronger ties. Most social media providers introduced more nuanced privacy settings; blocking, muting or visibility features; the introduction of groups and more general control over connectedness. This period also saw the rise of digital detox and disconnect initiatives and related research (Fast, 2021; Hesselberth, 2018; Jorge, 2019; Karppi, 2018; Nassen et al., 2023; Syvertsen, 2020). Detox and disconnect initiatives respond to the sense that the platform life is too much, that connection can be overwhelming.

After Twitter, it is not so much recognised that connectivity has its dangers, which has long been assumed. Rather, for many, connectivity itself has lost its appeal in a more fundamental sense. The desire for connectivity is superseded by protectivity – a will to protect. Protective sociality sees openness less as an assumed good than a condition of ever-present potential antagonism. It is much closer to a cryptographic worldview, where communication occurs ‘in the presence of adversaries’, as Jean-François Blanchette (2012) once put it. Rather than see the open network as one of possible weak ties in waiting – Van Dijck’s ‘total strangers’ – the ties are recoded more explicitly around a political logic of something approaching a Schmittian (2007, p. 26) friend-enemy distinction. From a belief in the ‘strength in weak ties’, underpinned by an assumption that positive opportunities lie with acquaintances and strangers, the protective impulse is guided by an assumed toxicity of weak ties. We are, of course, aware that marginalised and minority groups have always had different experiences of the open web, and with the question of adversaries ever-present. The fact that Black Twitter remained after the Musk takeover is a stark reminder that the Twitter of old was always seen as a hostile environment for this group and others (Parham, 2024b).

Protective sociality is reflected in a number of recent developments. The first is the increase in user migrations reportedly ‘flocking’ away from Twitter/X en masse (Zia et al., 2023). These include the highly reported Twitter to Mastodon migration of 2022, following Musk’s acquisition of Twitter, as well as the 2024 migration from X to Bluesky (Boyd, 2024). However, it is important not to discount rightward migrations to Gab, Parler and Truth Social. In each case, a user’s flocking ‘away from’ X (or splitting time across X and another system) contains within it the hope of finding a new home with like-minded others on a platform that more explicitly aligns with their values. Second, we see the presence of alternatives that explicitly respond to a desire for protectivity. We might not talk of Mastodon and Bluesky purely as ‘protective media,’ but both offer a range of protective features beyond Twitter. Bluesky has, for example, its ‘anti-toxicity features’ (Frazee, 2024), which includes the ability to detach quote posts to prevent ‘dog piling’ or the ability to hide replies to posts, among other things. Since its founding in 2016, Mastodon is further along the protective path, with its instance-level blocklists, enhanced privacy, reduced ‘searchability’ and community-level moderation. The point here is not whether such features are effective; it is that they reflect user’s changing desires at the level of feature design.

However, there are platforms that explicitly roll back safety and moderation initiatives that were previously taken for granted. This might appear to be the opposite of protectively sociality, but it is not. Removing forms of content moderation and significantly weakening rules around safety and abuse, for example, offers a form of protection for those who want to speak ‘openly’ about themes that others would find offensive or worse. Knowing that a platform openly welcomes racist, ableist, sexist, trans-phobic and other discriminatory and hateful content is a strong signal that those wishing to post such material will be protected, while those wishing to defend against such content will not. These initiatives are best understood within the context of platform polities and friend-enemy distinctions. After Musk’s Twitter acquisition, for example, it was noted that formerly banned accounts had come ‘roaring back’, with one commentator noting ‘Elon Musk sent up the Bat Signal to every kind of racist, misogynist and homophobe that Twitter was open for business . . . They have reacted accordingly’ (Frenkel & Conger, 2022). The authors further detail what they describe as an unprecedented rise in slurs and hate speech. Meanwhile, on the platform, quotes and video footage of Musk’s declared attack on the ‘woke mind virus’ circulate widely.

Such an understanding of protective media is a platform-centric version of Paulo Gerbaudo’s (2021) discussion of protection in The Great Recoil. Like many others, Gerbaudo is trying to understand the shift away from a global neoliberal consensus and a strong return of the state. His Great Recoil captures ‘the moment when societies turn backward and inward: when globalisation goes into retreat, . . . when people have to withdraw to their homes due to lockdowns, quarantines and confinement measures, and must shrink from contact with others’ (Gerbaudo, 2021, p. 14). There are clear parallels to be drawn with the Great Recoil and the emergence of platform polities and the protective turn in social media. What is especially relevant, though, is how Gerbaudo (2021, p. 30) links the polity with protection and how he describes protection as having contrasting variants that cross the political spectrum. He notes how ‘protection is central to politics’, how it is ‘a good in itself, but also a necessary condition for social cohesion’ (Gerbaudo, 2021, p. 172). And, in dialogue with Plato’s Republic, ‘Protection is what keeps the city together, what gives it a sense of purpose, what commits its citizens to the pursuit of a common mission. Protection, in this sense, is not just defensive; it is productive of the community’ (Gerbaudo, 2021, p. 173). We can see that how a platform polity enacts its protectivity in part defines that polity as a community. That is, the more generalised protective turn in social media, the recoil from the open web, has as many instantiations of protection as it does polities.

Conclusion: Nostalgia and Polity Building

In this article, we’ve suggested the space formerly occupied by Twitter is a fragmented one. It is a fragmentation that runs counter to typical monopoly-tending platform dynamics and it cannot be understood purely in terms of economic competition. We’ve suggested, instead, that these platforms need a more explicitly political framing to make sense of the fragmentation and to understand its variations. We used the notion of platform polities to do this work. Alongside and partly constitutive of the rise of polities, we observed a protective social media turn – an eschewing of the open network and a turning inward. As Gerbaudo writes of the recoil, there is also a turning ‘backward’. After Twitter, a key frame of reference for users is the Twitter-that-was, whether real or imagined. Here, we take up nostalgia, particularly as it appears in the field of memory studies (Atia & Davies, 2010; Radstone, 2010), to consider this question of how to reach across the historical break, of what to take and what to leave behind as platforms develop After Twitter.

How people react to the new fragmentation is conditioned in large part around their relationship to what Twitter was. One common comment about Mastodon is that it is like the Twitter of old. Early coverage of Mastodon often framed it as ‘Twitter without the Nazis’ – a gesture towards Twitter pre-Gamergate (Glaser & Oremus, 2018; Jeong, 2017). More recently, Meta’s Threads has been embraced as fun – just as Twitter used to be. Kari Paul (2023) writes in The Guardian that ‘with Twitter getting clunkier and progressively less usable since Musk took it over, opening an app and actually being able to see and engage with content smoothly felt like a breath of fresh air. . . I actually enjoyed using Threads’. Writing in Medium, Winston Ford (2023) says, ‘spending an hour on Threads is like a time machine to 2008. It’s your friends cracking corny jokes, begging for follows, and asking “is this thing on?”’. Similar nostalgic assessments can be found in commentary on Bluesky. Particularly after Musk bought Twitter, many people have turned to Bluesky to recapture the old spirit of Twitter. Political science professor David Karpf (2023) argues that ‘Bluesky feels like a party that I don’t know how I got invited to, but I see a bunch of people I know and they seem happy to see me’. Kate Knibbs (2023) declares in Wired that ‘the most impressive feat Bluesky has managed isn’t technical, but cultural. It has recreated an older, better era of the internet, one that’s actually fun’.

Clearly, there is a degree of nostalgia for a Twitter of old. If the Twitter-that-was is no more, there is desire for a smooth transition to an After Twitter that is somehow a return to Twitter. What comes After Twitter is in part conditioned by what form of nostalgia prevails: A party you weren’t invited to? A space for cracking corny jokes? A smooth content experience? Fun? Within these imaginings, we argue there’s also potential for a more radically transformative social network configuration. Otherwise, we be more or less stuck with a conservative continuation of the old. Here we can turn, as many have, to Svetlana Boym’s (2001) two conceptions of nostalgia, restorative nostalgia and reflective nostalgia.

Restorative nostalgia is ‘a return to the original stasis, to the prelapsarian moment’ (Boym, 2001, p. 49). It is a desire, per Boym, to restore nostos, the home or dominion. Restorative nostalgia elides the problems of the past in favour of a gauzy, idealised vision. As Hassler-Forest (2020, p. 184) puts it, ‘instead of depicting the virulent racism, sexism and homophobia of its historical period’, restorative nostalgic media present an unproblematic version of the past. Restorative nostalgia is easily exploited by politicians or industry. As Rose (2018) argues, Donald Trump’s first presidential campaign was marked by appeals to the ‘toxic nostalgia’ of coal miners in the Appalachian region of the United States – a longing for a past before widespread condemnation of fossil fuels as well as a past of White cisgender male supremacy. Similarly, as Mejía Estévez (2009) suggests, digital versions of restorative nostalgia are about avoiding ‘disruption, trauma and multiplicity’ (Hassler-Forest, 2020, p. 195) in favour of smooth continuity – it is a desire to keep the system the same as it was before.

Reflective nostalgia, in contrast, is focused on algia, or longing and reflection:

Where restorative nostalgia helps sustain cultural conservatism by transforming the past – and its cultural artefacts – into sacred shrines that we slavishly reproduce and consume, reflective nostalgia gives us occasion to consider the passage of time as something that is irreversible, and therefore marked by tragedy and trauma more than reassurance and comfort. (Hassler-Forest, 2020, p. 193)

This form of nostalgia is ‘ironic, inconclusive and fragmentary’ and can be critical (Boym, 2001, pp. 49–50). As Boym (2001) argues,

reflective nostalgia does not follow a single plot but explores ways of inhabiting many places at once and imagining different time zones; it loves details, not symbols. At best, reflective nostalgia can present an ethical and creative challenge, not merely a pretext for midnight melancholias. (p. xviii)

Overall, then, the Twitter-that-was that is reflected in the After Twitter alternatives may either be an uncritical continuation or it could be a critical reflection on how social media has dealt with ruptures and breaks – in short, history – over the past two decades. From this perspective, it is hard to see Threads as anything but restorative nostalgia: the smooth shift of Instagram accounts to Threads lends itself to this idea, as does the aesthetics of microblogging – including advertising and brands. The emergence of Bluesky – spun out of Twitter itself and replicating much of Twitter’s features as well as aesthetics – could easily satisfy those with restorative nostalgia, as well. Most importantly, the longing for a singular platform to replace Twitter is digital nostalgia: a longing for continuity rather than embracing the fragmentation of social media.

However, it is possible Bluesky’s commitment to open protocols and ‘credible exit’ shifts it away from being a mere continuation of Twitter, evincing some degree of reflection on how what comes After Twitter must be better than before. Both Bluesky and Threads have also at least gestured towards decentralisation – a far cry from the old ‘walled garden’ approach of corporate social media. Independent personal data servers, such as the Black-focused algorithmic feed Blacksky, are emerging in relation to Bluesky’s ATProto (Parham, 2024a). For its part, Mastodon is more in line with reflective nostalgia: it began not as a Twitter alternative, but as an alternative to yet another federated alternative, only becoming a Twitter alternative through surprise growth and Twitter’s spectacular decline. Mastodon’s more reflective nostalgia might be a resource for critical explorations of the past, more akin to ‘critical reverse engineering’ (Gehl, 2017). Critical reverse engineering has temporal aspects – it’s forward engineering in reverse. The result of critical reverse engineering bears some resemblance to the object that was reverse engineered, and yet the new object must have critical differences that stave off the problems of the old. This is more clearly in line with Boym’s reflective nostalgia. But restorative nostalgia also affects Mastodon, where calls for Twitter features such as ‘quote posts’ have been constantly happening.

In all cases, any nostalgia for the Twitter-that-was must have reflective elements if something better – not something merely continuous – is to emerge. What Twitter is being recovered? The Twitter of Black Lives Matter or the Twitter of Gamergate? Is it the Twitter of brands and public relations or the Twitter of democratic debate and journalism? Of course, upon reflection, Twitter was all of these things at one time or another – all the more reason for what comes After Twitter to be a critical engagement instead of merely a return to the old fun times. We hope that After Twitter, as a concept, serves not only as a historical marker that points to a new social media reality but also as a set of coordinates for our nostalgic impulses. Infrastructure, politics, sociality and the (re)making of social media history are all in play.