‘I can’t believe you said that!’: Social media, democracy and the public good

In the summer of 2022, I watched the battle over the future of Twitter, now named X, unroll in real time, both on the platform itself and in media commentary. No one fully understood the logic behind mega-billionaire South African tech magnate Elon Musk’s acquisition of the platform. Journalist Kara Swisher noted that, despite Twitter’s ‘outsize role’ as a ‘kaffeeklatch’ for politicians, influencers, activists, and business leaders around the globe, the Musk offer of $43 billion was a lowball offer. The company had almost never made a profit or, under founder Jack Dorsey’s leadership, produced a plan for profitability and global growth that other social media companies founded around the same time – Meta (since 2021 the parent of Facebook) and LinkedIn – had charted. A man of outsized ego, Musk had bragged that he was the only tech executive capable of ‘saving’ Twitter, not from its financial problems, but from reputational ones stemming from Twitter’s exclusion of prominent right-wing voices – including, in the aftermath of an unprecedented and violent attempt to prevent Joe Biden from becoming president of the United States on 6 January 2021, former American president Donald J. Trump (Swisher, 2022). Indeed, it had been such exclusions that led to the formation of explicitly right-wing social media platforms like Gab and Parler that played such a prominent role in organising the attempted coup d’etat – the consequences of which were still being litigated as Donald Trump mounted his third presidential campaign in 2024.

But mainstream platforms like Facebook and Twitter may have played an even bigger role in promoting populist, illiberal democracy in the United States (Davies, 2023). Extremist networks were able to spread propaganda and organise openly on these platforms without fear of censorship, or even surveillance. More important, however, was their effect on shifting the nature of political conversation. As Cas Mudde (2016) has observed about the rise of populism in Europe, a voter did not have to be a dedicated extremist to be affected by outlandish, false, statements, or conspiracist assertions. These forms of speech, which power illiberalism, mean the electorate is generally ‘more plugged in to political debates and more independent minded (although not necessarily better informed), making them more critical of and less deferential toward traditional elites’ (Mudde, 2016, p. 29).

Importantly, however, social media technology isn’t just a conduit for ideas: it has a past out of which certain uses of and ideas about speech have emerged. For example, the libertarianism of tech billionaires like Musk and PayPal co-founder Peter Thiel is not new; it rests on a long history of libertarian entrepreneurism and free market thinking in radio, television, and cable (Brownell, 2023; Pickard, 2014).

The United States has a strong legal tradition of protecting political speech and, perhaps unsurprisingly, free speech ideology in its most populist incarnation was baked into an American-born technology that was almost unregulated from its birth. But, the United States is virtually the only developed nation where this is still true. Russia and China exercise strict control over social media, and American countries defer to those standards when they operate there. The European Union (EU) has embraced more modest reforms, achieving substantial improvement of online content moderation through the 2022 Digital Services Act. In contrast, the United States has allowed Section 230 of the 1996 Communications Act, which provides limited federal immunity to internet companies, to remain in place. This makes it difficult, if not outright impossible, to force social media companies to regulate propaganda and extremist speech.

More significantly, it has left American users without a political analysis that might permit them to think critically about the takeover of a major platform like Twitter as a contest over speech. On the left, anxiety about the consequences of fully free speech prevails; in contrast, Musk’s acquisition represented a triumph for those on the extreme right who maintain that all moderation is censorship, and speech of all kinds is a legitimate tool for leveraging political power. And in an interesting twist, a platform that rose to prominence for its capacity to launch pro-democracy movements and link progressive organisers around the world (Tufekci, 2017), began to have the feel of a nation with hostile troops massing at the border. As Musk’s negotiations with Twitter plodded towards an eventual forced purchase, thousands of liberal and left-wing users logged on only to threaten to leave the platform for good. The hashtag #LeavingTwitter began to trend.

Having put together the acquisition price by mortgaging his Tesla car company, recruiting a few billionaire investors, and tossing in a sizeable chunk of his own personal wealth, Musk marched into Twitter’s San Francisco headquarters holding a kitchen sink and declaring that he had arrived to save the world. ‘The reason I acquired Twitter is because it is important to the future of civilization to have a common digital town square, where a wide range of beliefs can be debated in a healthy manner without resorting to violence’, he wrote in an October 27 post targeted to advertisers. Addressing the anxieties of millions of users, Musk promised that Twitter would not become a ‘free-for-all hellscape’, but would be ‘warm and welcoming to all’ (Musk, 2022). He then enacted his critics’ worst fears, firing half of the staff, gutting the safety committee, and reinstating banned users who had a track record of promoting violence: anonymous antisemites, white supremacists, troll accounts like Libs of TikTok (which exists only to spread right-wing disinformation), and, of course, former President Donald J. Trump (Mac & Browning, 2023).

Let’s be clear: social media is not journalism. Journalism is a practice where the distinction between fact and opinion is clear; social media is a highly diverse public square where people talk. Despite the descriptor ‘citizen journalist’ becoming attached to social media early on because of its capacity to allow knowledgeable observers to report in real time, and despite legitimate news outlets using it to promote edited content, social media wasn’t built to distribute news. This means that it is unlikely ever to self-regulate, particularly when propaganda and data collection are its profit centre (Vaidhyanathan, 2018). After 6 January 2021, Facebook dealt with its own disinformation problem by getting out of the news business and confining conspiracists to their own pages, much as Reddit has always done. But these were not solutions to the problem of extremist organising, propaganda, and disinformation driving illiberal politics. They were pragmatic dodges that allowed the platform to distance itself even further from what users, often paid professionals, do.

But lack of self-regulation does not equate to no regulation. The information any given user sees is determined by the affordances of the platform (for example, feeds, home pages, and hashtags) and, most importantly, the algorithm devised by the company itself. What is lacking in the United States, and what social media companies have successfully resisted, is government regulation of the kind that the European Union and Australia implemented between 2021 and 2023 – laws that required platforms to combat disinformation and to pay for news they distribute (Meese, 2023, pp. 14–32; McCarthy, 2021; Satariano, 2022). Under such laws, the EU launched an investigation of Meta in April 2024, charging that the company was failing to police disinformation on Facebook and Instagram in advance of the parliamentary elections to be held in June of the same year (Satariano, 2024).

But what is most important about Musk’s transformation of one of the most popular, and politically useful, apps, is that it marks the end stage of an important and politically fruitful period of digital history in which social media platforms made a thin, disputable, but plausible claim to enhancing the democratic public sphere, in part by elevating minority voices that are poorly represented in the mainstream. This honeymoon began in 2006 when Facebook opened to users beyond the .edu community and Twitter launched with its cheerful, distinctive baby blue bird icon. Like all successful honeymoons, user optimism in this period required characterising social media by its most democratic features and ignoring its potential for illiberalism. As both Kathleen Belew and I have argued, right-wing movements were early internet adopters in the 1980s, transitioning easily to political organising on Facebook and Twitter (Belew, 2018; Potter, 2020).

Still, by the twenty-first century, social media was most frequently associated with progressive and radical movements that sought to expand democracy. Facebook quickly became identified with Barack Obama’s campaign for President of the United States, and Twitter became associated with left populist insurgencies around the world: the Arab Spring in 2010, the rise of the Tea Party in 2010, Occupy Wall Street in 2011, the Movement for Black Lives in 2013, and the relaunching of the #MeToo movement in 2017.¹

These social movements pumped new life into the notion that the internet was an inherently progressive space, a concept that was probably well past its sell-by date in 2006 (Rheingold, 1993, p. 43). Yet, the honeymoon period is worth revisiting for the insights that can help us understand not only what social media had become by 2015, but what it might be today if it were re-engineered for democracy. Jeffrey Goldfarb’s book, The Politics of Small Things: The Power of the Powerless in Dark Times (2006) is one of the texts that helps us return to those optimistic years when social media had not been reduced to polished, data-hungry apps on the mobile phones of billions worldwide, a period when they instead foregrounded the conversations and disagreements between strangers that platform democracy. ‘When people freely meet and talk to each other as equals, reveal their differences, display their distinctions, and develop a capacity to act together, they create power’, as Goldfarb hypothesised about the rise of democracy in Eastern Europe. To understand mass-mediated politics, particularly in nations ruled by illiberal regimes, Goldfarb urged us to look ‘off the central political stage’ to see the small things happening in public spaces that populate the geography of our lives. One of these spaces, he concluded, was the internet (Goldfarb, 2006).

And yet, when social media is no longer the peripheral but the central stage, where do we look? That is, as yet, unclear. When Musk’s ownership of Twitter became inevitable in October 2023, the only thing everyone agreed on was that chaos on that stage, a state of nature in which new things inevitably occur, was imminent. Users on the left were horrified at the prospect that Twitter accounts that had been banned after the attack on the Capitol on January 6, 2021, would return to the platform. Users on the right naturally viewed this possibility as a triumph of disruption. One phrase – ‘I can't believe you said that’ – recurred in right-wing threads as users tested the limits of speech under Musk’s regime. It was an expression of delight: delight at speaking freely (some accounts posted things as juvenile as ‘penis, penis, penis’), and at the kind of boundary-crossing they hoped would send their opponents on ‘the left’ into disarray. Extremist pundits and ordinary MAGA foot soldiers flooded Twitter with images of ‘libs’ weeping inconsolably or ‘melting down’. Novel forms of humiliation and rudeness were rewarded with: ‘I can't believe you said that!’

What remains of the post-Musk Twitter public square is, if not entirely dominated, at least significantly shaped by a lack of conversation and a dramatic uptick in the trolling and disinformation that has become a critical tool for illiberal and authoritarian movements. This leads us to what Goldfarb did not predict: that the same technology that had helped grassroots movements overthrow authoritarianism would put the politics of small things to the purpose of undemocratic politics. Goldfarb was correct that public conversation leads to action and power. What we all missed, however, as the early, innovative phases of social media gave way to corporate platforms built to sell advertising and suck down data, was that conversation could turn a central tool of democracy, free and fair elections, to the business of bringing about an undemocratic world.

The easy story is that ‘good’ social media – and by this, I mean the virtual communities of the 1980s, the internet chatrooms, wikis, bulletin board services, the programmed blogs of the 1990s, and the content management services (LiveJournal, WordPress, and Blogger) of the early 2000s – gave way to ‘bad’ social media powered by smart phones. But in fact, the good is still with us, as anyone following a sporting event or trying to understand a breaking news event knows; and the bad has been there from the beginning. The WELL, a pioneering internet community launched in 1984, was both a magical model of radical free speech and community regulation; it too found personal attacks and sexism almost impossible to govern (Citron, 2014; Hafner, 2001). It’s easy to idealise how exciting blogs were and all the good they did – after all, the online magazine Goldfarb began as a result of his work on the internet and political change, Public Seminar, began its life as a blog.

But analysing the difficulties of blogging, and the dangers attendant to the outrageous, also give us insight on why today’s social media requires state, not corporate, regulation. To begin with, the verbal abuse and anonymous threatening behaviour of the early internet was a baked in feature of the blogging culture that flowered alongside early Facebook and Twitter. In both subgenres, communities of interest gathered around individual bloggers and could dissolve into conflict when invaded by trolls. The interconnections between social media platforms and political movements were not entirely clear, but the capacity to do good and harm in equal measure was (Potter, 2020, pp. 162–164).

Yet, many of us tried to do good. Blogs allowed us to reinvent real-life, or ‘meat world’ spaces – universities, domestic life, politics, and the military – by telling stories from the inside. I started an academic blog, Tenured Radical, in 2006, the same year that Goldfarb published The Politics of Small Things. If I imagined the potential transformation of academic life in the United States – characterised by debt-ridden graduate students, a permanent army of adjunct faculty, a deepening job crisis; collapsing structures of departmental and faculty authority – through internet storytelling, Goldfarb believed that the same technology could democratise nations, and perhaps entire continents. In both instances, social media conversations linked the like-minded, made us visible to each other, and allowed ideas and practices of resistance to flourish.

For some of these reasons Twitter, expressly built for the new smart phones, launched to immense optimism in 2006: not coincidentally, it was initially known as ‘microblogging’ (Burgess & Blaine, 2020). But in fact, Twitter was not like blogging. Instead, it reproduced the performative, and often contentious, quality of a blog’s comments section – but without the blogger, or anyone who had a face or an email address, to moderate and direct the conversation. And like Twitter personalities who acquire devoted followers, certain blogs (like conservative law professor Ann Althouse’s eponymous platform or Brooklyn College history professor K.C. Johnson’s Durham-In-Wonderland) assembled and cultivated legions of reactionary devotees who, in the name of free speech, would then sally forth to attack other bloggers.² In the comments section and in the posts themselves these blogs experimented with and refined a form of threat that was misogynist and racist but framed as reasonable discourse. Trolls bonded and found common ground over conversations that featured violence, homophobia, and meandering, rapey sexual fantasies (Jeong, 2018). In retrospect, blogs like these were an early phase of the reactionary populism that would emerge in MAGA Twitter and Facebook a decade later. In both cases, simply being outrageous counted as political activism and, as political journalist Adam Serwer describes MAGA culture, ‘the cruelty is the point’ (Serwer, 2021).

This long-lost blog world also gave birth to that gleeful phrase ‘I can't believe you said that’. Rather than signaling the perfection of cruelty or an epic dose of Schadenfreude as it does today, in the old blogging world that phrase recognised that the blogger had revealed a previously hidden truth and thus opened a new, or previously repressed, conversational space. This was particularly true in the blogs that chronicled university life, where hierarchies of power often cause people to self-silence and even the most radical faculty are small-c conservatives when it comes to structural change. For academic readers, ‘I can't believe you said that’ could be said as a prayer, as a paean for having delivered the perfect zinger to a troll in the comments section, or as praise for a post that trod all over the sacred cows of departmental life. It also signalled that the blogger had extended the range of what could be said in polite company about the real world. Such utterances became particularly important after 2008, when access to new tenure-track jobs plummeted and anxiety spiked.

As opposed to what I would say is Twitter’s essentially totalitarian character – elevating opinion and discouraging analysis – blogs were essentially democratic in the ways Goldfarb described. This was not because they promoted radically free speech, although they sometimes did, but because they made new conversations between strangers possible and created a discursive gateway between the academy and the non-academic world of politicians, mothers, journalists, and activists. Blogs also enabled exchanges across lines of status within university life. Academic bloggers, like me at Tenured Radical and an anonymous contingent faculty member who wrote as Invisible Adjunct, announced our perspectives openly in the pseudonyms we chose, and we engaged comments from anonymous readers about what they knew and believed. Whether securely or insecurely employed, we spent our days as obedient university citizens and our nights – even our office hours! – contesting the terms and conditions of our day jobs.

Bloggers were a people in search of a democracy and of the democratic freedoms we didn't experience in ordinary life. We were in flight from constraint and discipline, from an academic world that no longer fully made sense as the rewards for submitting to traditional university practices became increasingly obscure as job insecurity accelerated. In response we created a digital backstage for readers who were not actually ‘the public’ but in fact a ‘counterpublic’, a social formation based on conversation that emerges organically as a subaltern response to dominant forces in a democracy (Fraser, 1990, pp. 56–80; Warner, 2002).

In The Politics of Small Things, Goldfarb also continually brings us back to the importance not simply of the public sphere, but of the potential and transformative power of the conversations that occur there. Free speech may or may not promote democracy. It may, as we have seen it deployed by right-wing elements around the world, hinder it. But without creating space for conversation, we cannot have democracy either. Academic blogging, for example, made visible conversations that were already happening within higher education. It expanded what could be said out loud in universities, who it could be said to, and who might overhear it.³ Sexual harassment was a big topic, as were class and racism. Johnson’s Durham-in-Wonderland elevated a right-wing conversation about the loss of white, male privilege that needed to come out into the open. Whether on the left or the right, academic bloggers were interested in the various forms of intellectual and social terror that infect university hierarchies. That interest only intensified as the job market narrowed, employment became more tenuous, and conservative attacks on the university intensified.

In the policy world, speech that exceeds existing standards of tact, politesse, deference, and acknowledged reality is said to expand the Overton Window. Mostly associated with the libertarian right, the Overton Window describes a process by which performances and policies become available within the new reality that disruptive speech enables (Robertson, 2018). Traditionally, politicians, university presidents, corporate executives, and religious leaders have only been able to pursue ideas and policies for which there is widespread consent (which is notably different from agreement), or that are recognised as lying within the bounds of quotidien business. But the Overton Window creates the possibility of making minority positions not just visible, cracking that frame and making those ideas active players in a larger conversation. Returning to the phrase ‘I can’t believe you said that!’, we can understand that these words do not just signal a shift in affect. They alert us to the fact that the Overton Window has expanded, and new possibilities for speech, thought, and action have emerged.

Again, the good and the bad exist side by side. While the Overton Window can theoretically be generative of more progressive democratic norms, to a global, networked right wing it is a critical element in making illiberal democracy possible by breaking the rules with impunity. To return to academic social media as a microcosm of this impulse, at their origin point, academic bloggers spent a lot of time trying to rectify what we saw as unfair about the institutions that paid our salaries and supported our aspirations. But it was a paradoxical enterprise, because in essence bloggers weren’t solving problems but documenting an expanding Overton Window of university policies that replaced tenure-track jobs with post-docs, long-term contracts, and so-called ‘professors of practice’. What we were tracking was the totalitarian impulse immanent in all institutions under neoliberalism: as Nancy Fraser has argued, capitalist structures perpetuate themselves not by reproducing resources, but by consuming them (Fraser, 2023).

Democracy, Goldfarb argues, begins with individual, seemingly insignificant, conversations that connect, grow larger, and lead to collective problem-solving. As he argues in The Politics of Small Things, ‘When people talk to each other, defining a situation on their own terms and developing a capacity to act in concert, they constitute a democratic alternative to terror and force’ (Goldfarb, 2006, p. 8). For academic bloggers, then, the internet was our kitchen table, our clandestine bookstore, our basement theatre and living room literary salon – all the alternative public spaces where Goldfarb traced the social bonds, ideas, and energy amassing in Poland as a democratic society began to emerge through conversation after decades of Soviet-style dictatorship. If academic bloggers did not solve problems, they still maintained a democratic space where problem solving was possible.

Goldfarb’s insights about the power of conversation should also help us generate a bracing defense of our institutions as global social media companies, by simply doing what they normally do, facilitate equally global illiberal political movements. Autocrats and aspiring autocrats like Donald Trump, France’s Marine LePen, Russia’s Vladimir Putin, Hungary’s Viktor Orbán, and Brazil’s Jair Bolsonaro learn lessons from each other’s failures as well as successes, using internet-facilitated free speech to radicalise and reshape the political environment to their ends. They are capable of this because, although it is true that virtual space can nurture democratic interactions even under totalitarian restrictions, we now know that totalitarianism can carve out its own conversations in spaces built for democracy.

In 1927, Associate Justice Louis Brandeis of the United States Supreme Court issued a concurrence in a censorship case in which he proposed what has become a classic defense of free speech: that ‘the remedy’ for false or politically noxious statements issued on the American left is ‘more speech, not enforced silence’. With the speed and stealth that misinformation flies around the internet, Brandeis’s proposition no longer seems true. It is not the quantity of the speech that brings us closer to the truth but, as Goldfarb understood, the quality of the conversation.

Autocrats around the world recognise this. In 2021 Viktor Orbán, facing an election within the year, launched a pre-emptive attack on Twitter, Facebook, and YouTube, labelling them censors whose algorithms ‘shadow-banned’ his supporters. In 2022 Elon Musk’s acquisition of Twitter lifted bans, including that given to Donald Trump, imposed in the wake of January 6. And in that same year, Marine LePen, fighting to unseat French President Emanual Macron, vowed that if American social media continued to censor the French right then she, as President, would create a government-funded, sovereign French platform to free the choices of her movement.

The right to political speech is more contested than it has been since the nineteenth century, and thus it is crucial to understand that the debate over Twitter’s future is not really about the importance of a platform, but about who controls the pace of change as digital media evolves. As The Economist editorialised, when compared with rivals like Facebook, Instagram, and TikTok, Twitter (now X) is but a minnow. Its size belies it its importance, however, as a haunt of politicians, pundits, and wonks. A ‘digital public square’, as Musk himself put it, even in its anaemic form, still does much to set the political weather. And on Twitter, at least, that weather is shifting to the right.

Other platforms are now competing to create conversational spaces that might recreate the promise of the internet world that preceded the weaponisation of social media: Mastodon, a federated platform that links different communities in a ‘fediverse’ is one; Meta’s anodyne, but robustly populated Threads is another; and Twitter founder Jack Dorsey’s Blue Sky is a spicier, more Twitter-like competitor. But there is much to learn from the loss of Twitter to illiberalism, particularly about what kinds of regulation permit conversation, and a healthy political culture, to thrive. If the internet is to be a democratic public space, its users must take account of new circumstances, as well as the new forms illiberalism takes before it announces itself.

This is the project that is now before us, and it is not a small thing.

No potential conflict of interest was reported by the author(s).