A Brief Review and Critique of Technocentric Approaches to Democratic Erosion

Much of the focus on disruptive communication has centered on the political effects of poor-quality online information following the 2016 Brexit referendum in the United Kingdom and the unexpected electoral victory of Donald Trump in the United States a few months later (Freelon and Wells Reference Freelon and Wells2020). Journalists and scholars quickly linked Trump’s victory to online disinformation and conspiracy theories. Many Americans, it was claimed, were radicalized by algorithmically amplified lies, mendacity, and hate on their social media feeds. “It seemed possible,” noted political scientist Yascha Mounk (Reference Mounk2018), “that the rise of digital technology, and the concomitant spread of essentially costless communication, have set up a direct clash between two of our most cherished values: freedom of speech and the stability of our political system.” Such claims mark a rapid reversal in earlier scholarly assessments that digital media were a kind of “liberation tech” responsible for a new wave of democracy sweeping the world (Arendt Reference Arendt2023; Diamond Reference Diamond2015; Diamond and Plattner Reference Diamond and Plattner2012; Howard and Hussain Reference Howard and Hussain2013; O’Neil Reference O’Neil1998; Rawnsley Reference Rawnsley2005; Seethaler and Melischek Reference Seethaler and Melischek2023; Voltmer Reference Voltmer2013; Voltmer and Rawnsley Reference Voltmer, Rawnsley, Haerpfer, Bernhagen, Welzel and Inglehart2019).

The growing volume of work on various forms of disruptive communication has been most notable in political communication scholarship, but similar trends have developed across the social sciences, as well as in medicine, public health, epidemiology, environmental science, and computer and data science (Navarro-Sierra, Magro-Vela, and Vinader-Segura Reference Navarro-Sierra, Magro-Vela and Vinader-Segura2024). The result has been a babelesque eruption of terms, methods, and lines of parallel yet siloed research. Included among the many overlapping and poorly defined concepts are fake news (Grimm Reference Grimm2020; Lazer et al. Reference Lazer, Baum, Benkler, Berinsky, Greenhill, Menczer and Metzger2018; Nielsen and Graves Reference Nielsen and Graves2017; Tandoc et al. Reference Tandoc, Ling, Westlund, Duffy, Goh and Lim2018; Vargo, Guo, and Amazeen Reference Vargo, Guo and Amazeen2018), computational propaganda (Murphy, Keane, and Power Reference Murphy, Keane and Power2020; Woolley and Howard Reference Woolley and Howard2016; Reference Woolley and Howard2019), rumors (Berinsky Reference Berinsky2023), misperception (Flynn, Nyhan, and Reifler Reference Flynn, Nyhan and Reifler2017), conspiratorial thinking (Sunstein and Vermeule Reference Sunstein and Vermeule2009, 205; Uscinski et al. Reference Uscinski, Enders, Diekman, Funchion, Klofstad, Kuebler and Murthi2022), disinformation (Lewandowsky and Van der Linden Reference Lewandowsky and van der Linden2021; Roozenbeek et al. Reference Roozenbeek, van der Linden, Goldberg, Rathje and Lewandowsky2022), malinformation (Wardle and Derakhshan Reference Wardle and Derakhshan2017), and the most popular catchall term misinformation (Southwell et al. Reference Southwell, Niederdeppe, Cappella, Gaysynsky, Kelley, Oh, Peterson and Chou2019; Vraga and Bode Reference Vraga and Bode2020; West and Bergstrom Reference West and Bergstrom2021).

While there have been several noteworthy attempts to clarify and untangle these terms (see Wardle and Derakhshan Reference Wardle and Derakhshan2017), distinctions remain blurred and unstable. For example, disinformation is often defined as the intentional creation and propagation of factually unsound content for the purpose of mobilizing disruptive political movements to undermine the established order (Bennett and Livingston Reference Bennett and Livingston2018). In contrast, misinformation is usually understood as the intake or propagation of factually unsound information without the intent to cause harm or disruption. While perhaps theoretically helpful, such a distinction requires researchers to somehow decide the subjective intentions of the conveyer of factually unsound information. The task is made even more challenging by the mixed motives often behind the expression of fact claims or beliefs. One might, for example, give full-throated voice to a dubious claim for the purpose of signaling group loyalty or antagonizing opponents without necessarily believing it.

Despite little success at reconciling these scattered concepts, several related hypotheses have quickly emerged to explain social media’s role in democratic backsliding. First is the recognition that the priorities of the social media behemoths are shaped not by a sense of collective civic responsibility or regard for democratic principles, but by insatiable appetites for financial growth and market domination. This includes developing algorithms to engineer user behavior by recommending and amplifying often-extreme content (Zuboff Reference Zuboff2019, 8). Second, such content is thought to radicalize individual users by pulling them into proverbial rabbit holes where their worst instincts and darkest prejudices are ignited and reinforced (Horwitz Reference Horwitz2021). Put differently, algorithmically amplified content is thought to reinforce the tendency to uncritically embrace new information aligned with prior beliefs and reject information that is at odds with them, all irrespective of factual accuracy. This line of thinking has drawn on earlier work in cognitive psychology (Baumeister and Newman Reference Baumeister and Newman1994; Boxell, Gentzkow, and Shapiro Reference Boxell, Gentzkow and Shapiro2017; Fiske and Taylor Reference Fiske and Taylor1991; Kruglanski and Webster Reference Kruglanski and Webster1996; Kunda Reference Kunda1987; Reference Kunda1990; Taber and Lodge Reference Taber and Lodge2006). The idea that social media content reinforces motivated reasoning is the foundation upon which much of the technocentric explanation of democratic backsliding rests. Third, because of these cognitive tendencies, citizens are thought to grow more polarized and socially sorted (Iyengar and Westwood Reference Iyengar and Westwood2015; Mason Reference Mason2018; Rawlings Reference Rawlings2022), from which a downward recursive spiral of democratic dysfunction emerges. Namely, polarization opens space for yet more disinformation, which further exacerbates polarization (Pennycook and Rand Reference Pennycook and Rand2021; Thaler Reference Thaler2021). Finally, layered over this democratically dysfunctional dynamic is the disruptive influence of foreign adversaries using social media to exploit existing social and cultural conflicts to their political advantage (Harold, Beauchamp-Mustafaga, and Hornung Reference Harold, Beauchamp-Mustafaga and Hornung2021; Jamieson Reference Jamieson2020; Sanovich Reference Sanovich, Woolley and Howard2018).

This simple set of ideas—built upon the fuzzy collection of underlying concepts noted above—describes the logic of much of the contemporary debate about democratic erosion and digital technology (Tucker et al. Reference Tucker, Guess, Barberá, Vaccari, Siegel, Sanovich, Stukal and Nyhan2018). We begin our argument by offering a critical assessment of this largely individual-level information processing approach to democratic erosion. We argue that institutions, history, and social and economic power structures are elided from view when democratic disruption is understood primarily as an individual-level media effect. Furthermore, the dominant research focus on individual effects misses the question of how scattered individuals become organized in ways that connect to movements, parties, elections, and other political processes. Later, we suggest another way of understanding the role of communication in democratic erosion by showing how political operatives and leaders help to create networked organizations that enhance the voter ranks of parties. The mobilization of extremist networks often comes at the expense of traditional party organization, agendas, and gatekeeping capacities, and may even result in the capture of parties by more radical factions. First, we offer a critical reflection on the state of the technocentric model when based on individual-level cognitive effects.

Despite the conceptual ambiguities and methodological challenges, what are increasingly called “mis/disinformation studies” have skyrocketed. In broad terms, research has centered on three themes: (1) how information of dubious quality is created, targeted, spread, and amplified on digital platforms; (2) how erroneous information affects an individual’s ability to build accurate and democratically functional representations of reality; and (3) how factually inaccurate representations of reality and other irreconcilable political stances can be prevented or corrected. To give the reader a sense of the scope of the technocentric explanation of backsliding, we offer noteworthy examples of work in each of these three research domains.

Algorithmic amplification of disinformation and misinformation, sometimes referred to as computational propaganda, has received considerable attention (Duan et al. Reference Duan, Li, Lukito, Yang, Chen, Shah and Yang2022; Hindman Reference Hindman2018; Shin et al. Reference Shin, Hameleers, Park, Kim, Trielli, Diakopoulos, Helberger, Lewis, Westlund and Baumann2022; Woolley and Howard Reference Woolley and Howard2016; Reference Woolley and Howard2019). It argues that automation of various types boosts the signal strength of poor-quality information, thus putting individual citizens at greater risk. In some instances, foreign states and domestic operators use trolls, bots, and sock puppets as super-spreaders of disinformation on social media platforms (Ahmad et al. Reference Ahmad, Webb, Desouza and Boorman2019; Bradshaw and Howard Reference Bradshaw and Howard2019; Jamieson Reference Jamieson2020; Keller et al. Reference Keller, Schoch, Stier and Yang2019; Linvill and Warren Reference Linvill and Warren2020; Weissmann et al. Reference Weissmann, Nilsson, Palmertz and Thunholm2021; Woolley and Guilbeault Reference Woolley and Guilbeault2017; Zannettou et al. Reference Zannettou, Caulfield, De Cristofaro, Sirivianos, Stringhini and Blackburn2019).

The lion’s share of “mis/disinformation” research has been focused on the second and third research questions above: How can individuals build more accurate representations of reality? And how can inaccurate understandings be prevented or corrected? Discussions of public communication in liberal democracies often assume that the capacity to resolve (or at least tolerate) differences hinges on the capacity of individuals to update prior beliefs in the direction of greater openness to considering evidence and reason (Bennett and Kneuer Reference Bennett and Kneuer2024). Much of the research literature on public information quality has focused on testing techniques for discouraging confirmation bias or motivated reasoning. Inoculation strategies, media literacy initiatives, fact checking, and correction strategies are central to this effort (Bayes and Druckman Reference Bayes and Druckman2021; Benegal and Scruggs Reference Benegal and Scruggs2018; Dwoskin Reference Dwoskin2021; Ecker et al. Reference Ecker, Lewandowsky, Cook, Schmid, Fazio, Brashier, Kendeou, Vraga and Amazeen2022; Jerit and Zhao Reference Jerit and Zhao2020; Man-Pui et al. Reference Man-Pui, Jones, Jamieson and Albarracín2017; Tandoc et al. Reference Tandoc, Ling, Westlund, Duffy, Goh and Lim2018; Walter and Tukachinsky Reference Walter and Tukachinsky2020; Walter et al. Reference Walter, Brooks, Saucier and Suresh2020). For example, studies have tested the effectiveness of “prebunking” factually unsound claims (Cook, Lewandowsky, and Ecker Reference Cook, Lewandowsky and Ullrich2017; Guess et al. Reference Guess, Lerner, Lyons, Montgomery, Nyhan, Reifler and Sircar2020; Roozenbeek et al. Reference Roozenbeek, van der Linden, Goldberg, Rathje and Lewandowsky2022). This is also referred to as inoculation theory, as the basic supposition is rooted in the metaphorical reference to a vaccine, an injection of a modulated strain of a pathogen to trigger a defense mechanism.

Considerable research attention has also been given to assessing the effectiveness of correcting factually unsound beliefs. The results of these approaches, however, have not always been encouraging. Correction effects, for example, are not only short-lived (Porter and Wood Reference Porter and Wood2019), but they have even been found to trigger a “backfire effect”: experimental attempts to correct factually unsound beliefs have led to a deepening of the subject’s convictions (Nyhan and Reifler Reference Nyhan and Reifler2010; Nyhan, Reifler, and Ubel Reference Nyhan, Reifler and Ubel2013). Unsurprisingly, this finding was met with considerable alarm. If efforts to correct unsound beliefs lead to their entrenchment, what hope is there for democratic dialogue, debate, and compromise? Concerns were abated only after subsequent research failed to replicate a backfire effect (Porter and Wood Reference Porter and Wood2019; Wood and Porter Reference Wood and Porter2019). As it turns out, the perceived backfire effect was probably the result of a poorly worded correction treatment. But this episode points to another, and perhaps even more serious, concern.

Replication failures in social psychology and cognitive science, as we saw with the backfire effect, turn out to be uncomfortably common (Maxwell, Lau, and Howard Reference Maxwell, Lau and Howard2015). According to findings published in Science, most of a sample of a hundred published findings in social and cognitive psychology failed to replicate (Open Science Collaboration 2015). Other studies have come up with similar findings (Camerer et al. Reference Camerer, Dreber, Holzmeister, Ho, Huber, Johannesson and Kirchler2018). Despite these concerns, individual-level cognitive science–inspired studies have flourished in the past decade. This might be explained by the generous flow of resources in support of misinformation studies. According to one estimate, between 2016 and 2021, over $70 million went to a handful of universities for the purpose of starting up new research centers and institutes that would, among other things, find ways to correct the factually unsound beliefs stemming from online sources (Livingston and Miller, Reference Livingston and Millerforthcoming).Footnote ¹ Funding at this scale has also attracted the attention of data and computer scientists, who often came to the topic without a grounding in social science theory. Even political communication research done by political scientists often reads like a subfield of computer and data science.

In this environment, questions of social and economic power, or the efforts of parties and leaders to advance illiberal agendas, typically recede into the background. Moreover, when treated as an irrational outcome of faulty information processing, “polarization” may be seen as an inherently unwelcome and aberrant condition in a democracy that can somehow be corrected with better communication and civic education. Yet if one acknowledges systematic racism, gross wealth and income inequalities, and the chronic stresses they create, irreconcilable socioeconomic and cultural differences seem less a consequence of faulty cognitive processes than understandable emotional responses to lived conditions (Kreiss Reference Kreiss, Persily and Tucker2021; Kreiss and Reddi Reference Kreiss and Reddi2021; Marwick, Clancy, and Furl Reference Marwick, Clancy and Furl2022).

We do not mean to suggest that the research literature we have just reviewed is without merit. It has certainly revealed important individual-level effects often associated with digital communication technologies. What it fails to do, however, is engage societal, institutional, and organizational-level processes and effects. Instead, politics is reduced to messages and brain functions, with little consideration of how individuals belonging to different political subcultures, movements, or parties may actually use “alternative facts” and communication logics to connect with one another and to challenge opponents. Indeed, the results of the 2024 US presidential election demonstrate that fact checking and correction have little effect on tempering even the most preposterous claims. When challenged on the claim that Haitian immigrants were eating the pets in Springfield, Ohio, Trump’s running mate J. D. Vance admitted it was not true, yet it accomplished its goal of focusing media attention on their number one issue. And so, both Trump and Vance continued to invent claims of horrific societal breakdown across America, while all manner of media continued to spread those images, spun differently to reflect the competing communication preferences of political influencers and audiences.

Following our review of institutionalist approaches in the next section, we suggest a different way of thinking about the importance of digital technologies: as tools for connecting and organizing scattered and often extreme individuals and fringe groups and integrating them into movement, party, and election ecosystems. This picks up one promising thread from the technological analyses described above: the importance of digital platforms in organizing social and political networks (Klinger, Kreiss, and Mutsvairo Reference Klinger, Kreiss and Mutsvairo2023; Klinger et al. Reference Klinger, Bennett, Knüpfer, Martini and Zhang2023). We expand current accounts of platforms and power to include a framework for understanding different types of digitally organized networks, along with an explanation of how networked organizations are managed by parties and movements. This moves the level of technological analysis from individuals to organizations and enables the integration of technological and institutional approaches. The result offers a better explanation of how existing conservative parties can make illiberal transitions so quickly, as well as explain how many new parties on the radical/illiberal Right have developed such impressive voting publics. As we argue below, institutionalist approaches have yet to account for these transformative effects of digitally networked organization in accounts of democratic backsliding or erosion.

A Brief Review and Critique of Institutionalist Approaches to Democratic Erosion

Most democracy scholars have concluded that contemporary threats to liberal democracy do not come from a sudden military coup d’état or rapid institutional changes, but rather by way of a slow erosion of democratic norms, practices, and institutions (Bermeo Reference Bermeo2016; Kneuer Reference Kneuer2021). It is beyond the scope of this paper to offer an extensive review of this long-established and copious literature (for comprehensive reviews, see Grillo et al. Reference Grillo, Luo, Nalepa and Prato2023; Lust and Waldner Reference Lust and Waldner2015; Waldner and Lust Reference Waldner and Lust2018). However, it is important to be clear that our reference to “institutionalist” approaches includes a broad range of social, economic, and political organizational factors, with a focus in this analysis on conservative parties and how they manage (or fail to manage) their voter publics. We show how hybrid media systems and the politically managed construction of digital surrogates add explanatory power to institutionalist thinking about democratic erosion.

In explaining democratic backsliding, many institutionalist scholars have emphasized the effects of contingent decisions made by political elites (Capoccia Reference Capoccia2005; Linz Reference Linz, Linz and Stepan1978). Related work argues that democratic stability is a function of the normative commitment to democracy by political leaders (Mainwaring and Pérez-Liñán Reference Mainwaring and Pérez-Liñán2014). Other scholars contend that executives who are unconstrained by countervailing institutions or power centers are more likely to initiate backsliding (Fish Reference Fish, Anderson, Fish, Hanson and Roeder2001; Kneuer Reference Kneuer2021; Van der Linden et al. Reference van der Linden, Leiserowitz, Rosenthal and Maibach2017). There is also a long tradition of connecting qualities of civil society to democratic stability (Almond and Verba Reference Almond and Verba1965; Verba, Brady, and Schlozman Reference Verba, Brady and Schlozman1995). Others have examined democratic backsliding through a cultural lens. Inglehart and Welzel (Reference Inglehart and Welzel2005) and Norris and Ingelhart (Reference Norris and Inglehart2019), for example, argue that stable democracy requires the transition from values defined by the pursuit of basic survival needs to more secular and self-actualizing values.

While various aspects of leadership, civil society, power, institutions, and political culture are important to understanding democratic stability or erosion, we prefer to start with the economic dilemmas facing conservative parties, particularly the problem of voter aversion to policies that may produce greater inequalities in wealth and income (Acemoglu and Robinson Reference Acemoglu and Robinson2006; Boix Reference Boix2003). As Waldner and Lust (Reference Waldner and Lust2018, 102) put it, “As income inequality rises, democracy’s costs for the wealthy increase, lowering the probability of democratic transitions.” But it would be too simplistic, they argue, for one to rely solely on a crass class analysis. A second, more “heterogeneous axis” involves identity-based divisions that fall “along religious, linguistic, racial, or other descent-based attributes” (103). Class politics interact with these other sources of identity. They also note that “it is not accurate to claim that social divisions are first formed and then influence political processes and structures; political structures and processes also influence group identity formation. Political entrepreneurs, for example, might deliberately facilitate certain forms of group formation and impede others” (103, emphases added).

Ziblatt (Reference Ziblatt2017) develops a detailed analysis of the ways in which conservative parties may become “political entrepreneurs” by disguising often-unpopular and class-biased economic programs with more visceral social packaging to expand their voter bases. He begins with a common dilemma faced by conservative parties, especially during times of growing material inequality: their most natural political alignment is with economic elites, yet the wealthy are few in number, and often not wildly popular. This creates what Ziblatt calls the conservative dilemma. In a time of growing inequality and broad voting rights, to remain competitive, conservative parties must find a way to build a cross-class coalition. Put differently, to resolve the dilemma, conservative parties must focus elections on nonmaterial cultural and social issues, or what Ziblatt calls cross-cutting cleavage issues. Issues of this sort can mobilize a cross-class coalition that gives the party a fighting chance at the ballot box.

Parties promote these issues with the assistance of “outside interest groups and civil society organizations” that he refers to as party surrogate organizations (Ziblatt Reference Ziblatt2017, 174). Hacker and Pierson (Reference Hacker and Pierson2020, 23) offer a detailed description of contemporary surrogate organizations: “What parties are not always equipped to do is generate intensity sufficient to motivate potential voters and convince them to put their economic concerns to the side. For these purposes, other kinds of organizations—single-issue groups, cultural institutions such as churches, and certain kinds of media outlets, for instance—are often more effective. These organizations can focus on building strong emotional bonds with citizens and tapping shared identities.”

Though surrogate organizations can help to resolve the conservative party dilemma of how to expand voter appeal, they also present potential risks to party integrity. Surrogates can take extreme positions on highly emotive issues, and in the process pull the party along with it. Increasingly radicalized parties and their allied surrogate organizations can make unreasonable demands on the state or pull a party in illiberal or authoritarian directions. This concern is in keeping with long-standing concerns about the role of civil society organizations in a democracy (Almond and Verba Reference Almond and Verba1965; O’Donnell Reference O’Donnell1988; Verba, Brady, and Schlozman Reference Verba, Brady and Schlozman1995; for an overview, see Bermeo Reference Bermeo2003, 10).

Ziblatt brings these concerns to his treatment of “surrogate organizations,” particularly in cases where the interests of surrogates are not well aligned with party priorities. For example, in a desire for better ratings, an aligned media organization might appeal to key audience demographics by taking extremist positions on divisive issues, or by promoting fringe politicians that spread destabilizing extremist rhetoric on immigration, minority gender identities, non-Christian religious sects, or “white replacement” theories. The sworn testimony provided by Fox News personalities, and Rupert Murdoch himself, in the Dominion Voting defamation case following the 2020 US election underscored that Fox News can be understood as a Republican Party surrogate organization. However, Fox shifted away from the party mainstream as it became more interested in preserving its market share among radical-right viewers. According to their private message exchanges, Fox prime-time hosts continued to promote Donald Trump’s election fraud narrative despite their strongly stated private misgivings and internal tensions with the news division (Ellison and Gardner Reference Ellison and Gardner2023). That position helped to consolidate the influence of the Make America Great Again (MAGA) movement in the Republican Party, which in turn expanded the flow of disruptive communication in mainstream media. Hacker and Pierson (Reference Hacker and Pierson2020, 24) summarized this dynamic in the following way: “In a worst-case scenario, the party falls into a spiral of weakening control over the most extreme elements of its coalition. Ultimately, conventional politicians who are cross pressured by competing demands may be outflanked, supplanted by demagogues who are happy to work with such elements and know how to do so. Reliance on surrogates can thus lead a party down the path to extremism.”

What we add to this account of the conservative dilemma is the idea that formally organized surrogates such as Fox News or conservative Christian groups are increasingly joined by organized grassroots online extremist factions. These digitally networked organizations may compete with traditional surrogate groups for control of the party and its political agenda. And when digitally networked surrogates are successful (e.g., by electing representatives), they can disrupt surrounding democratic processes, such as party gatekeeping, election integrity, journalistic freedom, or judicial independence.

In addition to more conventional Republican Party surrogate organizations like the National Rifle Association, various Christian and family-values organizations, or the Koch donor network, dozens of influential factions of the MAGA movement can now be regarded as “digital surrogate organizations” of the Republican Party. Those digital surrogates include QAnon, the Boogaloo Bois, the Proud Boys, various white nationalist/supremacist organizations, wellness and anti-vaccine networks, local armed poll-watching groups, and hundreds of online media sites. Networking these radical factions not only brings more extreme voters and candidates into the party, but also creates a pull for more conventional party surrogates to become more extreme, as in the case of election denial conspiracies on Fox News, or the shifting center of gravity among Christian nationalist surrogates toward the creation of a religious state. As a result, party boundaries became far more “porous”—that is to say, the party is unable to maintain clear boundaries between it and its surrogates. Though William F. Buckley might have once distanced the conservative movement and the Republican Party from Robert Welch and the John Birch Society (Felzenberg Reference Felzenberg2017), it is more difficult to stop QAnon and other online conspiracy theory organizations from encroaching on today’s Republican Party, particularly when prominent politicians from Trump to members of Congress begin echoing their positions (Trickey Reference Trickey2021).

The next section of this paper describes how digital media networks form and operate at the organizational level and link to movements, parties, and elections. This model of networked surrogate organization offers a useful bridge between technocentric and institutionalist explanations of democratic backsliding. Our framework operates at a level of analysis better suited to link media platforms and user networks to institutions. As shown earlier in our critique of the technocratic literature, the cognitive individual level of analysis largely examines democracy one citizen at a time, while failing to show how those mis/disinformed individuals become organized in ways that attach their emotional truths to political movements, parties, and leaders.

Understanding Digitally Networked Organizations

Rather than centering our analysis on individual-level cognitive effects, we focus on the ways in which managed and curated communication on digital platforms can alter the organizational landscape of politics (Chadwick Reference Chadwick2013; Cooren et al. Reference Cooren, Kuhn, Cornelissen and Clark2011; DeLanda Reference DeLanda2006; Kreiss Reference Kreiss, Persily and Tucker2021; Putnam and Nicotera Reference Putnam and Nicotera2009; Putnam, Nicotera, and McPhee Reference Putnam, Nicotera, McPhee, Putnam and Nicotera2009). Our core assertion is that networked communication can, given the right conditions, constitute new organizational forms (Bennett and Segerberg Reference Bennett and Segerberg2013; Mirbabaie et al. Reference Mirbabaie, Brünker, Wischnewski and Meinert2021; Suk et al. Reference Suk, Abhishek, Zhang, Ahn, Correa, Garlough and Shah2021; Vaast et al. Reference Vaast, Safadi, Lapointe and Negoita2017). Related to this, we also argue that these digitally constituted organizations on the Right are likely to deepen existing threats to democracy posed by more traditional civil society groups and partisan media organizations, or what we, following Ziblatt, have called party surrogate organizations.

The organizational challenge is that many extremist groups are typically scattered and poorly organized, and thus present more challenging problems to mobilize than traditional organizations such as churches. Organizations that exist partially or even entirely online are made possible by a variety of platform algorithms and “affordances”—capabilities inherent in the design of a technology, such as hashtags, subreddits, Facebook groups, or recommendation engines that create patterned relationships among people (DeLanda Reference DeLanda2006). But that is only part of the story. Understanding how algorithms and affordances help people to become organized for public action requires broadening the understanding of communication by recognizing that “communication is often much more than a means of exchanging information and forming impressions, or an instrument for sending updates and instructions to followers” (Bennett and Segerberg Reference Bennett and Segerberg2013, 8). In some cases, highly active users, professional influencers, and political operatives link groups and resources across multiple platforms, with the result that these routinized communication patterns “become dynamic organizations in their own right” (8). The communication-as-organization perspective has been applied more broadly in the area of organization studies (Cooren et al. Reference Cooren, Kuhn, Cornelissen and Clark2011). Our application involves how the relative ease of communicating with formerly marginalized fringe elements attracts opportunistic politicians, media sites, and political organizations with the aim of mobilizing them into voting blocks.

To capture these organizational dynamics, Bennett and Segerberg (Reference Bennett and Segerberg2013, 46) array different organizational forms along a continuum that is anchored at one end by conventional hierarchical or “brokering organizations [that] carry the burden of facilitating cooperation and bridging differences when possible.” The other end of the continuum is characterized by crowd-enabled networked organizations that emerge with little formal hierarchy by using technological affordances (e.g., links, hashtags, reposts, archiving) to do some things that characterize more formal organizations, such as distribute resources, set priorities, and coordinate action. These relatively horizontal crowd networks have appeared in various forms, from mass protests such as the “Arab Spring” and Occupy Wall Street to more sustained issue networks such as #MeToo, #FridaysForFuture, or #BlackLivesMatter. Technology-enabled networked organizations are often “subject to notable reconfiguring as subnetworks shift their activities and the crowd responds to external events” (46). In other words, they are nimble, fluid, and constantly evolving.

Found between these two poles are many now familiar hybrid forms of organizationally enabled connective action in which more conventional political organizations assist in networking scattered groups, but may not formally lead or direct them, and often do not even acknowledge their coordinating roles. These hybrid organizational networks are increasingly important to democratic civil societies, from loose coalitions of economic justice or climate change groups on the Left to extremist militia and conspiracy networks on the Right. For example, as discussed in the next section, the national network of “Stop the Steal” groups that led to the insurrection following the 2020 US election included many of the familiar MAGA digital surrogates that were enabled and repurposed by resources from Trump-aligned media sites such as Breitbart or Gateway Pundit, and political operatives like Roger Stone. The result was the creation of broad public support for, and in some cases participation in, the 2021 coup attempt—inspired, in part, by a digital and social media tsunami that included more that twenty million views of #StopTheSteal videos on YouTube alone (Atlantic Council DFRLab Reference DFRLab2021). The next section explores how Bennett and Segerberg’s “communication as organization” model offers a constructive update to core political science models of democratic backsliding that emphasize the role of more conventional civil society and pressure groups.

The Political Dynamics of Illiberal “Digital Surrogate Organizations”

Much of the existing literature on what Bennett and Segerberg (Reference Bennett and Segerberg2013) call “connective action” has focused on progressive or left-wing social movements. It is important to note that the kind of “asymmetrical polarization” (Benkler, Faris, and Roberts Reference Benkler, Faris and Roberts2018) that distinguishes a fragmented Far Left from a more coherently networked Far Right in many democracies reflects the diversity of political issues and identities that have become characteristics of left-wing cultures based on direct action, participatory democracy, and aversion to more formal coalition and party organization (Bennett, Segerberg, and Knüpfer Reference Bennett, Segerberg and Knüpfer2018). There have, of course, been impressive examples of massive mobilizations on the Left based largely on forms of technology-enabled organization. For example, the Occupy Wall Street protest networks in 2011–12 became importantly organized through processes of sharing memes that affirmed common identities (e.g., “We Are the 99 Percent”), with networked crowds circulating and curating various connective resources from protest tactics to advice on what to do when arrested. Those patterns of coordinated communication produced familiar organizational capacities, such as making and sharing agendas, planning and acting, and responding to external threats and opportunities (see Bennett, Segerberg, and Walker Reference Bennett, Segerberg and Walker2014). However, unlike the Far Right, those importantly networked protest organizations tended to dissolve back into their constituent parts due to lack of interest in forming more permanent coalitions, sharing common ideologies or narratives, or supporting common political parties, as explained in the next section.

The Fate of American Democracy and the Future of the Republican Party

The Republican Party has been transformed by the networked organization of diverse MAGA groups linking grassroots citizens with media hubs, political organizations, and elected officials—all joined in sharing convenient myths and disinformation. As a result, a collection of once-isolated extremist factions has taken its place alongside traditional surrogate organizations to impact the identity and electoral future of the Republican Party. As Sidney Tarrow (Reference Tarrow2021) has observed, when extremism develops into organized movements, those movements can take over parties, and in the case of the MAGA movement, push a conservative party into illiberal territory. This observation is aligned with the logic of Ziblatt’s conservative dilemma model.

The unfortunate outcomes of the conservative dilemma in the United States can be thought of as failures of institutional gatekeeping. As Levitsky and Ziblatt (Reference Levitsky and Ziblatt2018, 36) argue, “[P]rotection against would-be authoritarians has not been [found in] Americans’ firm commitment to democracy but, rather, [in] the gatekeepers—our political parties.” The scale of the Republican Party gatekeeping failure beggars the oft-heard explanation in political science circles that the democratization of candidate selection processes after the civil unrest of the 1960s somehow later opened the door to Trump. That conventional wisdom is ironic in the sense that it blames too much democracy for the failures of Republicans. More importantly, it does not account for the long list of corrosive institutional actions and the related spread of disinformation over decades by Republican elites. Though party leaders and some of their backers, such as surrogates in the “Kochtopus” network (Mayer Reference Mayer2016), may not have endorsed Trump in 2016, their own legacy of disinformation and eroded institutional trust left them in no position to stop him or his insurgency once it took hold of the party.

This does not mean that the new Republican Party is one happy family, or that its course, much less that of the nation’s, is clear. Nearly everywhere one looks in the collection of Republican Party surrogates, one sees instability and boundary problems. For example, where are armed militia groups like the Proud Boys, the Boogaloo Bois, or the Patriot Prayer group taking party boundaries? What should we make of the evidence that local police departments and the US military have been infiltrated by far-right white supremacist militias (Burris Reference Burris2020; Kennedy Reference Kennedy2020; Levine Reference Levine2020; Montgomery Reference Montgomery2017)? As party boundaries are changing, the old strategies of “legalized voter suppression” are being supplemented by calls for violence, including threats and attacks on elected Republican officials and election workers who do not cleave to the new, more radical order.

Indeed, the embrace of violence by party operatives is itself a key warning sign of democracy’s threatened collapse. As Nancy Bermeo (Reference Bermeo2003, 234) put it: “Even profound polarization … is never, in itself, a sufficient condition for regime collapse. Democracies will only collapse if actors deliberately disassemble them and the key actors in this disassembling process are political elites” (emphases added). Beyond those who condone violence in defense of their version of white nationalist democracy, many elected representatives have intensified efforts to dismantle voting rights and replace secure election provisions with partisan control. This combination of accelerated institutional corruption and the failure of many Republican representatives to condemn and punish threats and violence against state institutions and opponents signals what Bermeo (238) refers to as a failure of the party’s “distancing capacity” to separate itself and its members from acts of violence and lawlessness. Such distancing involves condemning and prosecuting all those who engage in violence, even when they present themselves as current or potential party allies. The act of distancing serves as an antidote to the contagion of polarization. If parties can muster this sort of organizational control, they show themselves to be solutions to the problem of disorder and not contributing factors.

Trump and his enablers in the Republican Party have not just failed to distance themselves from violence; they have at times actively promoted it. During many political rallies, Trump called on his supporters to act violently against protesters and journalists (Cineas Reference Cineas2021; Sullivan Reference Sullivan2019). In May 2020, one report found 54 cases of violence linked to Trump’s rhetoric that were captured in court documents and police statements (Levine Reference Levine2020), including his invitation to those attending his January 6 rally in the Ellipse Park near the White House to go to the Capitol and “fight like hell.” A study of Trump’s rhetoric between 2015 and 2024 showed a dramatic increase in violent language aimed at immigrants, political opponents, journalists, and others (Savin and Treisman Reference Savin and Treisman2024). Of course, we will never know the incidence of civil strife that might have occurred had Trump lost the 2024 election.

Toward a Comparative Framework for Communication and Democratic Erosion

There are, of course, important cross-national differences in forms and degrees of democratic erosion, from how different electoral systems constrain parties on the radical Right, to the reach of mainstream and partisan media systems, among other factors. At the same time, there are many indicators that illiberalism is on the rise in many nations, with similar discourses and institutional erosion patterns (Bennett and Kneuer Reference Bennett and Kneuer2024; Papada et al. Reference Papada, Altman, Angiolillo, Gastaldi, Köhler, Lundstedt and Natsika2023; Štětka and Mihelj Reference Štětka and Mihelj2024). It is also clear that there are broadly similar integrating themes on the illiberal Right in different nations, including white Christian nationalism (or, more broadly, ethnic religious nationalism), anti-immigration, traditional family values (including patriarchy and anti-LGBTQ politics), and the emergence of everyday political “theories” such as the “great replacement” and “deep state” conspiracies. Similar to our analysis, emerging research shows that these patterns are due to a combination of institutional and communication factors. On the institutional side, there is growing evidence of transnational networking across movements, parties, think tanks, and issue organizations (Ayoub and Stoeckl Reference Ayoub and Stoeckl2024). On the communication side, researchers have revealed how memes and messages flow over those transnational media networks (often fed by leaders and organizations on the institutional side), with the result that similar conspiracies and campaigns appear in different nations (Heft et al. Reference Heft, Knüpfer, Reinhardt and Mayerhöffer2021).

In addition to transnational networking, many nations have experienced similar domestic structural conditions that strain common citizen identifications and shared democratic norms: growing economic inequality, labor market disruptions due to economic globalization, growing cultural diversity due to immigration and expansion of various minority rights movements, and the failed promises of domestic neoliberal economics (Holmes and Krastev Reference Holmes and Krastev2020). These structural strains produce social divisions, new citizen rights claims, and perceived threats to the status of formerly dominant groups such as white men (Parker Reference Parker2021). Thus, politicians in different nations have spread similar political narratives of coded white male victimhood as part of larger myths of lost national greatness at the hands of immigrants, minorities, and globalist economic elites. For example, Hronešová and Kreiss (Reference Hronešová and Kreiss2024) show how Viktor Orbán and Donald Trump crafted similar narratives of victimhood in very different cultural and historical contexts. The case for common patterns of mobilization on the Far Right is further strengthened by research on European Union election campaigns showing how similar campaigns and themes occurred in different nations, leading to growing right-wing representation in the EU parliament (Pfetsch, Benert, and Heft Reference Pfetsch, Benert and Heft2023).

Beyond the core narratives of nativist victimhood and restoring lost national greatness, a host of other common illiberal themes reflect a combination of institutionalist (e.g., party transformation) and communication factors (e.g., networked surrogate organizations): ending immigration, demanding the cultural assimilation of citizens with immigrant backgrounds, and backlashes against gender politics, among others. For example, Off (Reference Off2022) draws on comparative cultural backlash theories to show how gender values have become predictive of Swedish radical-right voting. Reinhardt (Reference Reinhardt2023) analyzes the commonalities of anti-feminist discourses in the organization of radical-right politics. And Reinhardt, Heft, and Pavan (Reference Reinhardt, Heft and Pavan2023) explain how social media campaigns in Germany, Sweden, and Italy connect attacks on gender equality to radical-right parties. In short, many democracies are warring over gender rights, reflecting transnational campaigns by various “family values” and conservative Christian organizations attacking gender diversity and promoting heteronormativity (Ayoub and Stoeckl Reference Ayoub and Stoeckl2024). As a result, patriarchy has moved up the list of core right-wing values in many nations, accompanied by attacks on LGBTQ and women’s rights groups.

Processes of connective networked organization are also central to an analysis of how the Alternative for Germany (AfD) party managed to broaden the appeal of its most extreme factions despite state surveillance and legal rulings on the dangers presented by those factions (Klinger et al. Reference Klinger, Bennett, Knüpfer, Martini and Zhang2023). Another study of the growth of the German Far Right by Knüpfer and Klinger (Reference Knüpfer, Klinger, Livingston and Millerforthcoming) applies a variation of our framework to show how narrative networks connect citizens through media sites, party operatives, and political organizations using memes that travel across fringe networks—a process they refer to as “connective radicalization.”

Beyond the growing number of comparative cases of how networked organization expands the base and radicalizes right-wing parties, there are also promising comparative frameworks tracing the development of illiberal communication spheres that block resolution of political conflicts and mobilize popular support for the political capture of public institutions. Štětka and Mihelj (Reference Štětka and Mihelj2024) show how the gradual takeover of media systems turns news and cultural journalism into propaganda systems for state capture, mapping various national transitions from liberal to illiberal regimes. A somewhat different model of illiberal communication processes shows how the core norms of liberal democratic communication (e.g., tolerance, civility, inclusion, reasoned debate, independent journalism) are systematically opposed by illiberal communication spheres that violate basic values of reasoned debate and the moderation of public conflicts by broadly representative institutions and independent journalism (Bennett and Kneuer Reference Bennett and Kneuer2024). For example, they contrast the conventional liberal democratic norm that underrepresented “counter-publics” may seek greater political voice with the rise of illiberal “transgressive publics” seeking to exclude those others. In short, whether exploring common patterns of issue and identity discourse, the formation or transformation of parties, or the development of public communication systems, our model of connective networked organization has applications beyond the United States.