From the Bourgeois Public Sphere to the Social Media Public Sphere

To understand the social media public sphere and the forms of democracy prevalent there, it is necessary to start from the general idea of the public sphere as it has been traditionally conceived in debates at the crossroads between political science, sociology, and the study of media and communications. The Structural Transformation of the Public Sphere, by Jürgen Habermas, originally published in German in 1962 but translated into English only in 1989, has within a few decades become one of the most influential theories at the crossroads of political science and media studies (Reference HabermasHabermas 1989). Its theoretical interest lies in its ability to capture the structural, institutional, and spatial dimensions of public communication systems, rendering how different media channels and sites of discussion are embedded in a common space organized around a common logic.

In his historical analysis, Habermas famously tracked the historical emergence of the bourgeois public sphere between the seventeenth century and the nineteenth. He argued that the emergence of news media and places of public meeting, such as tea houses, coffee houses, and theatres, opened a rational space for discussion and deliberation, informed by the information and opinion available through the emerging modern press. For Habermas, the rise of the modern capitalist press, which turned information into a commodity available on the market, progressively extricated access to news from aristocratic privilege, making it available to anyone who was able to purchase it. In Habermas’ analysis, this historical reconstruction went with the attempt to develop a normative and idealized model of public communication.

Habermas found much to praise in some of the key “criteria” that had emerged from the public sphere in the era of Enlightenment. These included: a separation between private interests and public issues; a degree of inclusivity in discussions, overcoming the exclusionary tendencies of aristocratic privilege; and a commitment to rational-critical argumentation aimed at finding a consensus among interlocutors (Reference HabermasHabermas 1989: 90-93). Further, in the bourgeoise public sphere, Habermas also retrieved an important mechanism of social coordination, through which emerging social consensus would go on to influence political decisions. Political rulers could not completely ignore the emerging points of consensus in the public sphere. Yet, Habermas was also aware that this positive content of the bourgeoise public sphere had progressively become imperiled, and already in that seminal work he decried what he saw as a progressive involution of the public sphere in mass democracies. Widespread commercialization of the news media, as well as the development of advertising and techniques of influence and manipulation, had fundamentally subverted the normative principles of the liberal public sphere (ibid.: 164).

Since its publication, and more so since its translation into English in 1989, Habermas's work became the target of intense criticism. Feminist theorists highlighted that what Habermas saw as an inclusive place was highly exclusive, as it barred women from participation (Reference FraserFraser 1990). Others pointed to the class exclusion inherent in Habermas's public sphere, as made explicit in its very qualifier (which is often forgotten), namely the bourgeois public sphere. In the early 1970s, Alexander Kluge and Oskar Negt (2016) pointed to the presence of an alternative public sphere of the working class, which crisscrossed the space of production in the factory and the city, including economic “penny cafés,” providing cheap access to arenas of discussion. The organizing principle of this public sphere was neither information nor opinion (Ahnung) but rather experience (Erfahrung) (Reference Kluge and OskarKluge and Negt 2016). The interest of this work lies in its awareness that different public spheres could co-exist at any given historical time (Reference BreeseBreese 2011), raising interesting questions as to the nature of the social media public sphere and its relationship with other public spheres.

At the inception of the Internet, authors discussed whether it should be considered a public sphere in its own right: a “virtual public sphere” (Reference PapacharissiPapacharissi 2002). Many of the interactive features of digital communication were seen as offering new forms of participation for citizens that could have rehashed some of the features identified by Habermas in his normative model, such as the ability for horizontal communication, reasoned discussion, and participatory deliberation. However, the development of social media, sometimes described as a second wave of the Internet, or “Web 2.0,” has comprised tendencies that appear to contradict these expectations. Social media have absorbed many features we used to associate with mass media, such as the logic of one-to-many communication and broadcasting (Reference CastellsCastells 2010). Far from being a de-centralized and networked system, this is often an environment dominated by strong power dynamics like those typical of scale-free networks (Barabasi and Bonabeau 2003), in which the majority of users consume messages produced by a small number of new media outlets and “power users.”

This process of ever-growing concentration has been facilitated by the rise of large corporate platforms such as Facebook and Twitter progressively acquiring a position of dominance over online discussions. Over the last five years, the landscape of the digital public sphere has also been transformed by the rise of instant messaging apps such as WhatsApp and Telegram and the rise of video-sharing platforms such as TikTok. But the overall dynamic proceeds along similar lines, in the direction of an ever-growing hierarchy in terms of the communication power of different users. In a recent intervention on the consequences of social media for the public sphere, Habermas has highlighted how, while social media lead to the elision of communication borders (Engretzung), they also pose a risk of fragmentation of the public sphere and deformation in the perception of publicness (Reference Habermas, Martin and SebastianHabermas 2021: 471). This builds on previous reflections by Habermas, where he has expressed preoccupation about the consequences of the transformation of the media system for the epistemic dimension of democracy (Reference HabermasHabermas 2006).

The digital transformation of the public sphere has important implications for contemporary democracy. Many scholars have discussed the possibility of a “digital democracy,” a new type of democracy, built after the image of digital media (Reference Hague, Brian D., Barry N. and Brian D.Hague and Loader 2005). Different models of digital democracy have been proposed by scholars, with an influential typology by Lincoln Dahlberg listing liberal-individualist, deliberative, counter-publics, and autonomist Marxist positions on this issue (Reference DahlbergDahlberg 2011). Further, some scholars, such as Stephen Coleman, have proposed to look at digital media more as a means to correct representative democracy than to overcome it (Reference ColemanColeman 2005). However, the debate on digital democracy has mostly been informed by the ambition of overcoming a stale representative democracy and achieving a greater degree of directness and disintermediation, as proposed by pundits such Reference TofflerAlvin Toffler (1991). The hope was that, due to its network-like structure, the Internet would favor a more decentralized society, in which power would be taken away from central power and be made available to every citizen (Reference BarlowBarlow 2019; Reference ShirkyShirky 2010). Cyberspace seemed to offer an opportunity for a new Jeffersonian democracy, with as much power as possible in the hands of the individual (Reference TurnerTurner 2010) in an Internet of small producers, self-entrepreneurs, and “amateurs” who would have found online not only a place to thrive economically but also to make decisions democratically.

The technological and commercial development of the Internet, however, has gone in a direction that is opposite to the one predicted, and hoped for, by both liberal advocates of rational argumentation and libertarian techno-utopians. The Internet that has emerged from the so-called Web 2.0 era is one of enormous concentration, with a few platforms, services, and companies controlling large shares of the market (Reference McChesneyMcChesney 2014). Companies like the GAFA (Google, Apple, Facebook, Amazon) have come to control much of the everyday online traffic. This is similar to the situation in China, where companies such as Baidu, Alibaba, Tencent, and Xiaomi (or BATX) control the local Internet. These companies and their services have become hosts to billions of people who every day go about their customs and intervene in a myriad of discussions and interactions.

Already in Habermas's public sphere, the media of the bourgeois public sphere were mostly private and commercial ventures (1989: 166). While he initially considered that commodification of information acted as a means of inclusivity, he later decried mass commercialization as a factor leading to the vilification of public communication. With social media, this process of commercialization that worried Habermas has only been pushed to the extreme, as large-scale commodification now engulfs not only the media producing information and opinion but also the places of discussion that decide which content is permissible—in line with their “community guidelines.”

This popularization of social media driven by commercialization has had ambiguous consequences for democracy. It is undeniable that there has also been an element of democratization, whereby the diffusion of social media has allowed ordinary people, who in previous eras were mere media consumers, to access its self-publishing and instantaneous communication affordances and to post their ideas or favorite contents at the touch of a button, depriving the news media system of their exclusive role as institutional mediator or gate-keeper. Yet, the erosion of media gate-keeping is also at the source of many of the negative phenomena that are often associated with social media—from fake news to trolling and flames. All in all, self-publishing is an innovation that has significantly re-defined the nature of public communication, and it is also been enabling factor for the emergence of the collective subject that dominates the social media public sphere: digital crowds.

The Digital Return of the Crowd

Rather than a bourgeois public sphere, the social media public sphere resembles a plebeian public sphere, namely a massified one dominated by the action of online crowds. “Plebeian public sphere” is a term Habermas himself uses in the preface to The Structural Transformation of the Public Sphere, referring to a variant of the public sphere that was visible at different moments in history, such as the French Revolution, in which “a public sphere stripped of its literary garb began to function—its subject was no longer the ‘educated strata’ but the uneducated ‘people’,” and other phenomena, such as the Chartist Movement and the “anarchist traditions of the workers’ movement on the continent” (Reference HabermasHabermas 1989: xviii).

The term originates from the plebs in ancient Rome and the public spaces through which they acted politically. These were not limited to official meetings, such as the contiones, in which the people of Rome were addressed by public officials and popular assemblies (Reference MouritsenMouritsen 2001); it also encompassed more informal occasions of popular public opinion such as in “Circuli, rings of people who gathered for conversation or as an audience” (Reference Rosillo-LópezRosillo-Lopez 2017: 184). Further, it became visible around major spaces and occasions of everyday economic activity, consumption, and entertainment, from the market in the forum to festivals and circuses; while these normally acted as a means of domination and persuasion of the masses, they involved a gathering of large crowds that sometimes propitiated the spreading of dissent in the form of rumors (ibid.: 203) and even riots. More generally, the term “plebs” has also been used to describe the social subject behind major popular movements in history (Reference BreaughBreaugh 2013).

In this article, I use the term “plebeian public sphere,” rather than “proletarian public sphere” used by Reference Kluge and OskarKluge and Negt (2016), to express that, on the one hand, the subject intervening in the social media public sphere is more a case of “mass” than “class,” given that those coming together in online crowds do not necessarily share the same class condition, and on the other, that the digital crowd is not a subject that gathers around spaces of production, as Negt and Kluge proposed, but more around spaces of socialization, entertainment, and informal discussion, as with the ancient plebeian public sphere. My suggestion is that, to capture the nature of the subject in the social media public sphere, we need instead to fully account for its crowd-like behavior.

US Marxist scholar Stanley Aronowitz already highlighted some years ago that Habermas's theory of the public sphere is imbued with a “fear of crowds” and street protest (Reference Aronowitz, Mike and WarrenAronowitz 2000). These days we find references to crowds in discussions about social media and in the coining of terms such as smart-mobs, crowdsourcing, and so on (Reference LeeLee 2017). Howard Rheingold introduced the term “smart-mobs” to describe early forms of collective action via mobile communications (Reference RheingoldRheingold 2002). Andrew McAfee and Erik Brynjolfsson have argued that the platformization of the Internet has been accompanied by the rise of more dispersed and bottom-up organized crowds rather than by the traditional institutional “cores” (Reference McAfee and ErikMcAfee and Brynjolfsson 2017). This has made it possible for various forms of mass behavior to emerge, which have become visible in many different contexts: from the production of knowledge on Wikipedia to online trading and various forms of social and political action.

The sense behind the use of these terms is that, unlike pre-digital media, social media afford a virtual space in which a number of micro-behaviors can be aggregated to form collective behavior. As Dolata and Schrape highlight (2016), “crowds” stands alongside “masses” and “swarms” as terms used to describe non-organized collectives that manifest themselves online. Further, this notion chides with growing debates about the affective nature of contemporary politics (see, for example, Reference Slaby and ChristianSlaby and von Scheve 2019) and of social media (Reference Bösel, Berndt and SerjoschaBösel 2020; Papacharissi 2015) and the way digital communication involves a variety of emotional processes that are involved in gathering people and creating bonds of solidarity among them similar to those seen in traditional crowds (Reference CanettiCanetti 1984). As seen in many different movements, online crowds can quickly materialize into physical crowds (Reference GerbaudoGerbaudo 2012).

This discussion of online crowds seems to contradict a famous prediction of French sociologist Reference TardeGabriel Tarde (2010). Writing at the end of the nineteenth century, Tarde proposed that crowds were a pre-modern phenomenon that would eventually be substituted by publics, thanks to the possibility of increasing mediation offered by new communication technologies. While some authors have some reservations about the notion of digital crowds (2017), preferring alternate notions, such as “swarms,” Dolata and Schrape argue that the term “online crowds” can be aptly used to describe non-organized collectives emerging online that, while lacking the element of physical proximity characteristic of traditional crowds, nevertheless seem engaged in an experience of gathering or “crowding” (Reference Dolata and Jan-FelixDolata and Schrape 2016). This was a trend that had already been noted some years ago by the cultural theorist Mark Poster (1997), when he said that the Internet was becoming the new place of public crowding, taking the place of physical spaces like street corners, squares, taverns, and so on as organizers of social interactions.

These acts of crowding bespeak how the Internet and social media act as sites of gathering and aggregation from a distance. Partly this derives from the “economy of attention” of online spaces, the role of various algorithms involved in shaping visibility (Reference Van Dijck and ThomasVan Dijck and Poell 2013), and the connected filter bubble effect of social media (Reference PariserPariser 2011). There has been an intense discussion about how these mechanisms lead to polarization and the limits of the filter bubble theory (see, for example, Reference Jones-Jang and MyojungJones-Jang and Chung 2022). What is clear, however, is that social media platforms, in offering “expanded possibilities for the formation, coordination and control of collective behavior” (Reference Dolata and Jan-FelixDolata and Schrape 2016: 14), bestow heightened opportunities for the emergence of crowd-like phenomena. On these platforms, “like-minded” individuals are brought together around online content of common interest and connected conversations, gathering at short notice to engage in mediated forms of collective action and soon disappear, only to reappear in a changed form a few hours or days later.

In contemporary politics, the examples of this return to the logic of the crowd are manifold: from hackers coordinating at a distance to attack websites for political purposes, to online campaigns with mass support, to online traders’ communities, such as the Reddit group Wall Street Bets, which in January 2021 bought en masse the shares of the videogames company GameStop. In the field of protest, we have seen various social movements, such as Occupy Wall Street and Black Lives Matter, emerging on the back of popular Internet hashtags and viral memes circulated by social media crowds. Similarly, political insurgencies—such as the Sanders campaign in the 2016 Democratic primaries and the Trump election victory the same year—were fueled by mercurial “digital armies” of committed supporters, who shared content and mobilized people to take part in campaign events. Furthermore, on a day-by-day basis, online crowds battle on Twitter, Facebook, and YouTube in the comment sections, in what often take the form of fiery confrontations between different groups.

The most visible consequence of this crowd logic is the “turbulence” of the social media public sphere (Reference Margetts, Peter, Scott and TahaMargetts, Hale, and Yasseri 2015): new actors can rapidly emerge and then often fade away at the same speed. The social media public sphere is akin to a volcano, in which new phenomena bubble up like fluid lava. In turn, this volcanic or turbulent element of the social media public sphere needs to be understood as a consequence of the high degree of individualization or “networked atomization” that social media have contributed to fostering. Social media are “personal media” (Papacharissi 2003) whose accounts are by default defined as belonging to specific individuals rather than collectives in a context, in which the individual, rather than a group, is the basic unit of social organization (Reference WellmanWellman 2001). The logic of democracy that dominates the social media public sphere needs to be understood in light of this social polarity between individualized users and effervescent crowds, and the role played by social media reactions as a key mechanism through which the emergence of crowds becomes publicly visible.

Reactions as the New Mechanism of Democracy

The term “reaction” implies a response to a given action. In the context of physics, it is used to indicate the action in response to a previous action: for example, the movement of an object. In Newton's Third Law, it is famously stated: “To every action there is always opposed an equal reaction.” In other areas of knowledge, such as psychology or sociology, “reaction” implies a response to stimuli, a behavior triggered by an action that precedes it both temporally and logically. One example would be a fight-or-flight response when faced with danger, or a response to any given media content, eliciting laughter, surprise, or sadness. For example, a cinema audience may react to the sudden appearance of a ghost in a horror movie by shrieking in fear. Indeed, Hollywood and commercial cinema more generally have often been geared to elicit “effects,” or reactions, in the public, such as awe, sadness, joy, calm, thrills, and so on. Media scholars have studied the effects of different media contents, from television programs to video games, and the kinds of reactions they elicit in the public (Reference StaigerStaiger 2005).

We could thus see reactions as a universal aspect of all forms of communication. Yet, my contention is that in the social media public sphere, reactions have acquired a particularly eminent role that is decisive for understanding their overall logic. Social media reactions can be seen as the innovation introduced by social media vis-a-vis what Lev Manovich described as the “language of new media” (Reference Manovich2002). Reactions have become the most common form of “interactivity”—one of the common features attributed to new media, alongside hypertextuality and multimediality (Reference KiousisKiousis 2002). But it is a very minimal, and in some respects disappointing, form of interactivity. Whereas in the new media discourse of the 1990s and early 2000s the notion of interactivity conjured a view of horizontal exchange, where there would be some degree of reciprocity among different interlocutors, reaction manifests a version of interactivity as a far more asymmetrical and low-intensity mechanism.

Reactions may appear trivial, and the logic behind them has often been ridiculed as a “clicktivism,” a low-intensity activity such as signing a petition or liking a post that is ultimately inconclusive (Reference HalupkaHalupka 2014). But they are the most frequent micro-behaviors seen on social media (Reference MacafeeMacafee 2013; Reference Heiss, Desiree and JörgHeiss, Schmuck, and Matthes 2019). Many more users merely “react” to a content which is served to them than produce a new content. Reactions have also become a form of “content” or message in their own right. While in previous media conditions people were already reacting to content in various ways, these reactions mostly remained invisible to the message producers. Now, however, the reaction to a message is instantaneously seen by the sender, who is “notified” about them and can take cues from them.

Further, the word “reaction” has now also been used to describe posts that are created in response to other posts, or videos, as is the case with TikTok, where they have become a popular “genre.” Besides emotional reactions, such as those seen on Facebook, and other familiar interactions, such as liking and retweeting on Twitter, comments are also a key form of reaction behavior on social media. Tim O'Reilly famously contrasted Web 1.0 and its logic of publishing to Web 2.0 and its logic of participation (Reference O'ReillyO'Reilly 2005): what now mattered was not how many people read a piece of content, but how much they participated. More generally, much of the discussion occurring online is built around the expectation that people will react to content, either by endorsing it or disapproving of it.

While reactions are an affordance inherent in social media platforms, they are not merely “technologically determined”: their use and meaning are shaped by a number of symbolic processes, both on the designer side and on the user side, where different communities attribute to them different meanings and social uses (Reference Tian, Thiago, Giulio, Emilia and ChaoTian et al. 2017). Further, on some occasions they have been used as a means of political expression and contestation. One famous example was a BBC Facebook live interview with Theresa May during the 2017 election campaign; the interview was overwhelmed by a wave of angry reactions from online users, visualized in the form of a stream of angry faces during the event. On Facebook, it is well known how, on other occasions, users have reacted en masse with the Haha reaction to ridicule a given content. Further, this political use of reactions is also seen in the practice of “ratioing” (Reference Minot, Michael V., Thayer, Christopher M. and Peter SheridanMinot et al. 2021), with users “piling” on tweets they strongly disagree with by writing negative comments and “quote tweets.” However, conversely, reactions can also be used to convey approval of a given content: by liking, loving, sharing a given Facebook post and writing approving comments. by retweeting, loving, or commenting positively on Twitter posts, and by using similar reactions on other platforms.

Reactions often have a qualitative element (as seen in Facebook's emotional reactions, and most evidently with comments), but their impact is mostly quantitative: they are aggregated as a measure of popularity expressed in the numerical form of social media metrics. In general, social media communication aims to gather as many positive reactions as it can, while minimizing negative reactions. Indeed, the influence of Facebook pages, Twitter, and YouTube channels or TikTok accounts is often measured in terms of the number of people who follow and “like” their online content, their “user base,” and their “user engagement.” Further, reactions have a direct bearing on the visibility of social media content. Social media algorithms prioritize contents that have already received a high number of reactions, reading them as a marker of interest. It is thus only by attracting a high number of reactions that social media content can go “viral”; in this sense, reactions are both a measure and a fuel of social media communication.

The role played by reactions on social media has important consequences for the nature of “really existing digital democracy” in the social media public sphere: reactions can be understood as a vote. Given that the number of people reacting to content becomes an implicit measurement of its “popularity” (Van Dijck and Poell 2016), every reaction becomes a ballot of sorts: one “like” equals one vote. This implicit vote almost invariably takes a simplified form, which approximates the binary one of a referendum (Yes/No, For/Against, reacting or not reacting) or of selecting among a limited number of options (as in the case of Facebook's emotional reactions). This nature of reacting as voting is made explicit with Reddit, where people can “upvote” or “downvote” a post or a comment depending on whether they agree or disagree with it, and whether they think it is valuable and useful to the community.

This implicit online voting mechanism is also at work in dedicated recommendation websites or recommendation features now available on many websites, from Google Maps, where one can evaluate the quality of service of different shops or restaurants, to Airbnb or Uber, where one can vote for or against a host or a driver. Obviously, in most cases, this is a “vote” only figuratively, as people are not expressing a political preference, nor electing representatives to decide on their behalf. But the use of reactions has contributed to shaping the idea of the web as a sort of people's tribunal, the stage of a permanent referendum in which people constantly judge everything that is on offer. In this sense, it is significant that the number of reactions to a post is often cited by leaders and observers as proof of the level of popularity of a given discourse, sometimes even boasting it in dedicated posts.

Democracy or Ochlocracy?

Reactions embody the new logic of democracy that has become prevalent in the social media public sphere, one which—as the foregoing reference to a permanent referendum highlights—tends to have a strong plebiscitary character. It is plebiscitary because it approximates a binary vote around all sorts of issues in which what ultimately matters is the power of numbers and the size of the digital crowds that are mobilized. This points to a logic of democracy drawn from the idealized vision of a liberal model of reasoned conversation discussed by Habermas and other theorists (Reference HabermasHabermas 1989). Furthermore, it is also different from visions of deliberative democracy that have been supported by political theorists in recent decades (Gutman and Thompson 2009), sometimes drawing on Habermas’ public sphere. According to deliberative democracy theory, the legitimacy of decision making derives from the active participation of citizens in the shaping and discussion of issues. In this context, deliberation involves a qualitative aspect, as it is expected that participants will actively shape the actual content of the decisions being taken. With the plebiscitary democracy that is manifested in the social media public sphere, it is instead the quantitative aspect of preference aggregation, and the Yes/No, For/Against referendum that ultimately matters.

This, however, does not mean that this plebiscitary democracy precludes any condition for voice or that it only leads to collective manipulation. My suggestion is that, to capture the logic of these practices, it is necessary to adopt a different normative model of democracy and the public sphere, such as the notion of a pluralist and antagonistic democracy articulated by Chantal Mouffe, in which the logic of the public sphere is more one of conflict than of consensus (Reference MouffeMouffe 1999). If we follow this different model, we can see that online crowds, and individuals within them, have some voice in its possibility to lend support for one of the different antagonistic options that are on offer at any point in time. Yet, it is evident that this influence is mostly quantitative rather than qualitative. It involves choosing among “options” that have by and large already been “pre-packaged” by political leaders and influencers. But this participation remains “reactive” in the basic sense that it involves a response to a question that has by and large already been defined at the outset.

The plebiscitary character of social media reactions and the crowd-like form of the contemporary public sphere may raise fears that, rather than a democracy, what we are being delivered in the social media era is an ochlocracy. “Ochlocracy” is a term used by Hannah Arendt to speak of the power of the crowd and the mob, which, using violence and intimidation, contributed to the rise of Nazism and Fascism (Reference ArendtArendt 2013). Speaking of reactive democracy as an ochlocracy would mean that it is an anti-democratic space wholly dominated by irrational and authoritarian instincts. As previously discussed, there are many worrying aspects of “reactive democracy,” and the examples of how it has been geared towards authoritarian ends are manifold. At the same time, it is also necessary to acknowledge that democracy has often involved similarly simplified and massified mechanisms of popular participation. Choosing among candidates is not very dissimilar from choosing among different emotional reactions, just as voting in a referendum is not too dissimilar in principle from the choice of “liking” or not liking. Mass democracy involves all sorts of simplifications in gathering and recording the preferences of the citizenry.