Deliberative democracy and the digital public sphere: Asymmetrical fragmentation as a political not a technological problem

Abstract

Political communication and opinion formation have always been central topics in democratic theory. Today, all eyes are on the new digital landscape and the ways that it is affecting these central elements of democracy. The diagnosis in both the popular press and scholarly research is that the digital revolution has been anything but good for democracy: “Today conventional wisdom holds that technologies have brought the world addictive devices, an omnipresent surveillance panopticon, racist algorithms, and disinformation machines that exacerbate polarization, threatening to destroy democracies from within” (Bernholz et al., 2021, p. 3). Assessing the threat of the present information revolution is especially relevant for theories of deliberative democracy that place communication and deliberation at the center of the democratic system.

In this essay, I focus on Jürgen Habermas’ version of deliberative democracy and the assessment of the digitalization of the public sphere that follows from it (Habermas 2022c). This assessment identifies fragmentation and privatization as the most serious threats to a properly functioning public sphere. While I agree that fragmentation and privatization are threats to the democratic function of public sphere, I question whether digitalization is their primary cause and suggest that we should be focusing on political actors who intentionally pursue strategies that fragment and polarize the public sphere. Thus, the culprit here is not so much technology and acquisitive platforms as authoritarian political elite intent on dulling the power of the public sphere to hold political actors to account.

Deliberative democracy is a broad research paradigm. Very generally, it can be described as a “talk-centric” rather than “vote-centric” view of democracy (Chambers, 2003, p. 308) in which democracy is studied and evaluated “from the point of view of the quality of the processes through which individuals come to discuss, debate and mutually justify their respective stances before voting or taking other sorts of political action” (Scudder & White, 2023, p. 12). This central normative core has been developed, studied, and theorized at what might be called two levels of democracy. On one level, we see the development and indeed proliferation of citizen deliberative initiatives. These concrete exercises in deliberative democracy bring citizens together in face-to-face designed settings with good information, trained moderators, and procedural norms that promote participant equality in the deliberative and decision-making process. Here, deliberation is a practice structured within an institution. There are thousands of these initiatives across all democracies, and within non-democracies, with immense variation in design and function (Farrell & Curato, 2021). Their use and insertion into democratic systems is on the rise and, in many places, significantly addressing democratic deficits. In deliberative mini-publics and citizen assemblies, however, there are no social media, no fake news, no affective polarization, and many of the pathologies associated with post-truth are absent or mitigated. We can learn a great deal from these initiatives and translate some of it into the wild and open spaces of democracy writ large, but they do not add up to a democratic system, and it would be a mistake to think that deliberative democracy begins and ends with face-to-face citizens deliberate initiatives.

The second branch of deliberative democracy seeks to develop a fuller theory of democracy that is applicable at the macro level. This fuller theory begins with a principle of democratic legitimacy and connects that principle to a systemic analysis of democracy. There are variations in the articulation of the principle of legitimacy (Cohen, 1997; Dryzek, 2017; Habermas, 1996; Mansbridge et al., 2012). But they all tie it to processes of inclusive reason-giving, where all affected have an equal right and opportunity to participate in collective opinion formation that is translated into action (or will) through, for example, elections.

But where and how does “reason-giving” precisely go on in mass democracy? It is in translating the principle of deliberative legitimacy into the macro level that we see variation in theories. What is very rare to see (but what critics of deliberative democracy often claim to see) is the idea that this view of legitimacy translates into a utopian vision of democracy in which each and every citizen ought to engage in (or aspire to engage in) the epistemically demanding practice of deliberation about public matters (Habermas, 2022b). Applying the principle of legitimacy to the macro level involves looking at different parts of a full democratic system as playing different functions in instituting the ideal of legitimacy.

The version that I investigate here is a two-track model of democracy introduced by Habermas and developed in different ways by subsequent theorists of deliberative democracy (Chambers, 2017; Cohen & Fung, 2021; Lafont, 2020). The two players here are the informal public sphere and the formal institutions of government, especially legislatures. Other models—for example, Dryzek (2002), Bächtiger and Parkinson (2019), and Mansbridge et al. (2012)—all highlight different elements in the system and put less stress on the public sphere as the clearing house linking citizens to government.

The public sphere is a sphere of political communication that stands between civil society and the state. Communication is highly mediated and disaggregated and ranges along a vast multi-dimensional continuum from everyday talk to formal deliberation and includes a growing number and variety of digital platforms. When it works, the public sphere provides “a close-to-the ground, locally infirmed, dispersed arena for detecting problems, exploring them and bringing them to public view, suggesting solutions, and debating whether the problems are important and worth addressing” (Cohen & Fung 2021, p. 28). Democracies function properly when they respond to and act on problems, concerns, and issues that confront real people in civil society. Democracies further the ideal of deliberative legitimacy to the extent that the process through which problems, topics, proposals, and solutions are raised, honed, and make it onto legislative agendas meet two conditions: that the process includes all those affected as equal participants and that the process be governed by “the force of reason” (Habermas, 2022a, p. 150). Epistemic quality and equal participation are tightly linked in this picture of democracy (Chambers, 2017).

In Habermas’ version of the two-track model, there is some division of labor between inclusivity and epistemic quality. First, political communication is asymmetrical, with most citizens participating as audience. In other words, most citizens are readers, viewers, listeners, and consumers of messages that one hopes become the subject of internal deliberation, reflection, informal everyday talk, and consideration. Second, not all ideas, claims, positions, and demands raised in the anarchic public sphere can be justifiably translated into a legislative agenda, and, in any case, all of them need clarification, articulation, and translation in order to make it onto an agenda. Therefore, the system must “filter,” to use Habermas’ word, the claims and demands by putting them through a feedback loop of public scrutiny and then ever more rigorous processes of justification (Habermas, 2009, p. 159). As the political talk moves closer to the center, traditional rules of deliberation and arguing become more rigorous, and the conversation looks more and more like deliberation (at least ideally). The more epistemically rigorous function of arguing and deliberating over clear policy options takes place higher up the system ladder. But the content of deliberation—what gets deliberated—comes out of communicative processes that filter, clarify, and prioritize claims and narratives in the public sphere. When this is working well, the result is considered public opinion, that is, public opinions that have been shaped and constituted by an inclusive process of public debate and problem-solving.

The two-track theory of deliberative democracy is a highly stylized and ideal picture of the proper function of the public sphere. No real public sphere comes close, although we occasionally see glimpses of national debates that approximate some of the functional characteristics. Many conditions and prerequisites—from legal protection of speech and association to adequate levels of social equality—need to be in place to achieve even a minimum of functionality. But now I want to focus on media systems and the way that deliberative democratic theory is meeting the challenge of the digitalization of the public sphere.

This situation has a negative impact on both epistemic and normative functions of the public sphere, which is to say, on the circulation of reasons and the inclusiveness of debate.

Networked communication leads to the demise of gatekeepers, who were never perfect and often lent on the side of the status quo or interests of the owners of media, but who nevertheless ensured some minimum standards of epistemic quality and political responsibility (Cohen & Fung, 2021, p. 41; Habermas, 2022a, p. 160). The new media neither produce nor edit nor select; instead, they furnish platforms where anybody and everybody can connect to anybody and everybody and curate information using profit-oriented algorithms. There is a huge uptick in the amount of information accessible and a downtick in the quality of that information. This situation (and the fragmentation I discuss next) is then being exploited by bad actors to push political agendas via disinformation (Chambers & Kopstein, 2022). Disinformation, manipulation, and propaganda are endemic to all democratic public spheres. There is concern, however, that digitalization of communication, especially the outsized role of social media in the circulation of information, may exponentially increase the danger of falsehood. But part of the story here is not just that there are no gatekeepers, but also the creation of enclaves insulated from the types of debate and criticism that can expose falsehood. This leads to the second problem with the new digital landscape.

Communication in the public sphere is not directed toward consensus or producing a unified public will; it produces multiple and often conflicting opinions about a shared problem or salient concern. To the extent that the digital public sphere makes it impossible to have shared topics of conversation, it will be unable to perform its function as a clearing house for public opinion and will formation (Chambers & Kopstein, 2022).

There is a paradox here. In one sense, the new digital public sphere is excessively inclusive, but it fails to create the conditions for that inclusivity to perform its democratic function, which is to have many voices contributing to the shaping of political agendas. Instead, users are increasingly siloed in enclaves talking to each other. But there is a further dimension to this situation that particularly worries Habermas. This has to do with the privatization of communication in the digital public sphere. Networked communication creates the opportunity for us all to be authors, not just readers. But as authors, we are more like people writing personal letters that are massively and widely circulated than self-conscious public figures contributing to a public debate. “Different standards apply to the composition of printed matter addressed to an anonymous reading public than to private correspondence” (Habermas, 2022a, p. 165). The problem has to do with the self-consciousness of the author as performing a public role within a democratic public space. Do I see myself, for example, as a journalist adhering to a professional code and contributing to a public debate? Or as an individual sharing my personal opinions with “friends”? The digital public sphere becomes the place for sharing our private reflections about public matters and mostly with likeminded consumers. Here, Habermas worries that citizens will no longer see the public sphere as the (potential) place of inclusive public debate over shared problems. This, he suggests, would be a radical and democratically unfortunate transformation of the public sphere.

Habermas’ diagnosis of the dangers of digitalization is grim. Epistemically, the worry is threefold: first, citizens will lose access to trustworthy sources of facts and information upon which to build considered opinions; second, citizens will become skeptical and distrustful of all sources of facts and information, including trustworthy ones; and, finally, citizens will believe that fellow citizens have lost access to trustworthy sources of fact and information and so lose trust in democracy. But the second normative dimension is more serious. Here, the overarching worry is about fragmentation, privatization, and the migration of political communication into closed enclaves. Under these conditions, inclusiveness loses its democratic force as a means for all those affected by a decision to take part in the making of the decision. Political communication becomes millions upon billions of personal letters sent to friends or to the whole world in a virtual and boundless messaging universe.

How accurate is this diagnosis? Most people entering the debate about how the digital revolution impacts democracy admit that it is early days in this revolution and so often use the future conditional tense. Like with the climate crisis, there is a lot of “this will happen if we do not act.” Habermas is no different. He is sounding an alarm about trajectories. These are real threats, but the jury is still out whether we can get a handle on them and act in sensible ways to mitigate the worst. He suggests that regulation and education will play an important part in stemming fragmentation. I agree. But I want to offer a slightly different diagnosis of the problem of fragmentation as well as the problem of privatization. On the question of fragmentation, I want to suggest that the problem is not produced by the new technology but rather by some political uses of the new technology. And secondly, the transformation from reader to author is also a highly asymmetrical phenomenon where the vast majority of users are in fact consumers and not creators of content.

Echo chambers and filter bubbles are not as widespread as some have feared (Chambers & Kopstein, 2022; Guess et al., 2018). In this respect, there is some good and some bad news. The good news is that the pathologies of untrustworthy and siloed information are asymmetrically distributed across and within public spheres. Across public spheres, national context matters. The United States, for example, has seen more information pollution and fragmentation during elections than most European public spheres, although all liberal democracies have some. This suggests that it is the American political context (aka polarization) and not digital technologies themselves that is at the bottom of some of the pathologies.

There is also asymmetry within a national public sphere. Data gathered between 2016 and 2018 showed a portion of the American population caught in a right-wing media ecosystem that exhibited “all the characteristics of an echo chamber that radicalizes its inhabitants, destabilizes their ability to tell truth from fiction, and undermines their confidence in institutions” (Benkler et al., 2018, p. 383). This phenomenon did not extend beyond politically constructed right wing eco-systems, however. Politically determined asymmetry also characterizes trust in mainstream media sources. Information produced by legacy media, although shared and circulated via social media, is still filtered by gatekeepers and governed by traditional norms of truth-telling, fact-checking, professional journalism, and minimum levels of civility (Benkler et al., 2018). Opinion polls that show mistrust in legacy media is highly skewed toward conservative and right-wing information consumers (Swift 2016).

Focusing on echo chambers and fragmentation as a technological problem of information curation designed to maximize profit suggests a false symmetry in information pathologies. All social media users receive curated information, but not all social media users are equally victims of misinformation, or susceptible to narratives undermining legacy media sources. Citizens who find themselves in pernicious echo chambers shielding them from the possibility of correcting false beliefs are often the target of misinformation strategies and campaigns designed by political elites and not necessarily the victims of systemic features of social media platform algorithms (Starbird et al., 2019).

Furthermore, empirical evidence suggests that in the aggregate people's digital information ecosystems are more porous than we have thought (Burns, 2019). There is so much information out there that one needs to make a conscious effort if one wants to limit one's information to items that fit with one's preexisting views. On the one hand, homogeneity of political information consumption correlates with political knowledge and activism, implying that isolation is an information consumption choice and not a systemic effect (Guess et al., 2018). On the other hand, homogeneity also shows up more on the right and intensifies as the consumer's views are more extreme (Sindermann et al., 2021). This too suggests that it is more than algorithmic curation at work. Of course, the existence and resilience of these echo chambers cannot be good news for democracy, but it does suggest (and empirical data supports this) that the diagnosis of a general centrifugal trend produced by the technology itself needs to be revised.

There are also some unexpected ways that new digital technologies are creating the opportunity for a centripetal dynamic in public debate. Many factors have contributed to the unified European response to the crisis in Ukraine. Claire Berlinski argues that the latest advances in Google translate, which saw huge breakthroughs beginning in 2016, are one of those factors (Berlinski, 2023). Google translate has created a European public sphere where journalists, pundits, and citizens communicate instantaneously across multiple language barriers. Millions of Europeans follow Zelenskyy whose tweets are automatically translated into the followers native language, and social media are creating links between ordinary Ukrainians and Europeans where they are exchanging information in real time. Google translate has an enormous potential for the future of supra and intra national debate.

On the transformation of the public from readers to authors, I also think there is a political and asymmetrical dimension to this phenomenon. While it is true that, as Cohen and Fung say, one of the most radical transformations of the public sphere is from a one-to-many model of communication to a many-to-many model. And this in turn means the demise of gatekeepers and the rise of what Habermas has identified as the privatization of communication. All this is true and is embodied in the idea and ideal of “influencers.” Here is the dream that anybody with a smart phone can gain millions of followers and potentially shape public opinion. But what is still true is that although perhaps anybody can rise to be an influencer, not everybody can. The fact remains that the vast majority of users are consumers and not creators of content. Trump's Twitter account is a good example of the paradox here. On the one hand, at the height of his Twitter power, he was just shy of 89 million followers, more than 99.9 % of whom, however, never tweeted themselves. Many of those who did tweet were journalists covering Trump's tweets. This is a massive public audience and many of his tweets generated a lot of talk and chatter.

On the other hand, as Habermas has so astutely observed regarding the problem of privatization, Trump's tweets often expressed personal vendettas and grudges, or random thoughts thrown out into the public sphere to see where they would land. Thus, although public, Trump's tweets were private messages and failed a minimal publicity test of being contributions to public debate. Twitter made this possible, but Twitter did not produce the phenomenon. Trump and many similar authoritarian leaning populists have no interest in the public sphere as a space of pluralist debate. Indeed, they have an interest in wrecking its democratic function (Chambers Kopstein, 2022).

In tweeting the way he did, Trump was surely following a play book outlined by one of his early advisors, Steve Bannon, who remarked in a 2018 interview, “The Democrats don't matter. The real opposition is the media. And the way to deal with them is to flood the zone with shit” (Lewis, 2018). The intention is to generate a level of cacophony so that it is impossible to sort out truth from untruth, much less deliberate about the best path forward. Failure to adopt the appropriate standards of public speech that should accompany contributions to public problem solving is part of a political strategy aimed precisely at undermining the democratic function of the public sphere. In some weird sense, Bannon's statement shows that he has come to the same conclusion as many deliberative democrats: public debate and opinion and will formation are where all the democratic action is. Elections follow opinion and opinion is endogenous to the communicative context. Rather than putting forward an alternative ideological message, this strategy is about making the whole process of opinion and will formation impossible. This strategy is certainly facilitated by new technology, but it is not a necessary outcome of that technology.

The solution to this problem involves regulation, education, but also political activism. Deplatforming Trump from Twitter and Facebook was certainly an important step but voting him out of office had the biggest impact on the circulation of his non-public political communication as it is not clear the platforms would have kicked him off if he had won the election. My point is not that fragmentation and privatization are not serious, potentially devastating pathologies of the public sphere. My point is that we should not ignore the role of political actors who intentionally seek to produce fragmentation and privatization precisely because they know how devastating such developments would be for democracy.

Habermas argues that the centrifugal dynamics of digital communication coupled with the personalization of messages within this new sphere will have a deleterious effect on the democratic function of the public sphere as the host of public debate. “The fragmentation of the public sphere prevents citizens who participate in the process of political opinion-and will-formation from focusing their attention on the same issue—including relevant problems that stand in need of regulation.” This danger is compounded by a “lack of public awareness that this is even a problem” (2022b, p. 26). I want to question this last statement and suggest that the very wide and inclusive debate we are having about social media and its role in democracy is evidence that we have yet to become irreversibly fragmented.

The post-2016 public disclosures about fake news, disinformation campaigns, Russian bots, micro targeting, voter suppression, the amplification of crazy conspiracies theories, the role of Cambridge Analytica, curating extremism, echo chambers, and the complete failure of most platforms even to know what was going on, let alone step in and moderate the free for all, was a shocking eye opener for the public (Chambers Kopstein, 2022). Even the least informed and engaged citizens would have been hard pressed not to have noticed the alarm bells ringing across the media landscape, including the platforms themselves claiming they would put things right and straighten out the mess. The worries, close to panic, about the digital transformation of the public sphere have fueled an astonishing and unprecedented explosion of institutes, centers, programs, think tanks, NGOs, academic studies, conferences, colloquia, and research devoted to assessing and studying the new digital landscape across all domains. This symposium is of course part of that knowledge mobilization. Money and resources have poured in from public and private sources to support work and research in this area. The flood of research and study has been paralleled by an equally constant stream of discussion and reflection in the popular press as well as endless public commissions and congressional committees disturbed by post-truth and the black box of algorithmic curation. Much of the scrutiny in the popular press and governmental investigations has taken aim at rapacious platforms and the Big Tech moguls who run them as the primary villains in this narrative. Despite its massive market share, Facebook is the platform everyone loves to hate and revelations about shenanigans, double standards, and ethics violations sells copy. Big tech exposés and whistle blowing are now a journalistic genre in their own right. Furthermore, survey after survey shows citizens across the ideological spectrum and indeed across the globe are concerned with the ways that social media has undermined trust in information (Rainee et al., 2019; Watson, 2019). This all adds up to a moment of opinion and will formation where we are focusing our attention on a shared problem in need of regulation.

Parliaments and legislatures in all liberal democracies have the regulation of the digital public sphere on their agendas. It is impossible to tell how effective that regulation will be in protecting and strengthening the democratic functions of public debate. We are at the very beginning of this transformation. The focus is on privacy and content moderation and not so much on fragmentation and privatization. Perhaps there are few regulatory tools to halt fragmentation. But if fragmentation is an asymmetrical phenomenon in which sections of the population (and not everybody) seek to cut themselves off from the circulation of reliable information and shared epistemic starting points, which is to say, if fragmentation is to some extent the product of consumer choices pushed by political propaganda, then exposing and publicly criticizing the political strategies of opinion manipulation used on social media and messaging apps is as important as opening up the black box of algorithmic curation.