1 Introduction

Donald Trump has done political science a great service. He has pushed the United States out of the parochialism of American politics into the realm of comparative politics. And vice versa, he has made comparative scholars veer towards the US. Since he won, first the Republican nomination and then the presidency, US democracy has faced a problem that was thought to be exclusive to democracies in developing countries: public concerns about its survival. If democracies are consolidated when they can be expected “to last well into the future” (Valenzuela 1992, p. 70), the ascent of Donald Trump to the US presidency has put an end to democratic consolidation. In US public debate, the stability of democracy is no longer taken for granted. All of a sudden, the country has entered the universe of fragile democracies and turned into a crucial case in the comparative study of democratic persistence.Footnote 1

Strangely, though, the way we are looking at democratic politics when we discuss Donald Trump and the ways we have been looking at democratic politics in the comparative study of political regimes are startlingly dissonant. Many of our democratic worries about Donald Trump seem to arise from factors that do not bother us in the comparative study of democracy: his unconstrained discourse, his violation of linguistic norms, and his programmatic amorality. As I contend, Donald Trump provokes democratic anxieties which we had ignored in our comparative research because he violates tacit cognitive, normative, and causal assumptions which we had taken for granted. By challenging these implicit assumptions, Donald Trump has pushed them into the open. In this sense, his normative transgressions have worked in a way akin to sociological “breaching experiments” which unveil tacit expectations by breaking them.

This essay is primarily an exercise in disciplinary self-reflection. It strives to make sense of the way we have been responding to Donald Trump’s unsettling political career. The “we” I keep referring to is somewhat fluid. Its expansive version includes us, practitioners of comparative politics; its more narrow version refers to us, students of political regime change. My central claim is one of analytical incongruence. As I contend, our perceptions of the threats Trump poses to US democracy do not match our main theories of democratic instability. To establish this claim, I dedicate a fair part of this essay to assessing the nature and severity of Donald Trump’s democratic norm transgressions, since they delineate our precise object of both research and anxiety.

I begin by sketching the global debate on “democratic backsliding” whose scope and urgency has dramatically increased after Donald Trump’s appearance on the political stage. I then lay out three features of contemporary research on democratic stability which are strangely inconsistent with our democratic anxieties about the 45th US president: its emphasis on structure, democratic “high crimes”, and the stabilizing force of self-interest. I go on to explain the basic logic of the “breaching experiments” introduced by ethnomethodologist Harold Garfinkel in the early 1960s. In the main section of the paper, I describe the fundamentals of Donald Trump’s practical breaching experiment by drawing a preliminary balance sheet of his compliance with fundamental democratic norms, which allows me to identify the demonstrative value of his breaches: they are forcing us to recognize the relevance of political actors, political language, and normative commitments.

2 Democratic anxieties

Over the past years, concerns about worldwide “democratic backsliding” have been spreading and deepening in academic, political, and diplomatic circles.Footnote 2 Even though the New York think tank Freedom House laments more than a decade of steady “decline in global freedom” (Freedom House 2018), until now instances of democratic progress and regression have largely offset each other (see e.g. Varieties of Democracy 2017). The worldwide sense of democratic crisis seems to derive less from a clear pattern of democratic retrocession than from the absence of a clear pattern. We have seen transitions from democracy to authoritarianism taking place in highly diverse circumstances, in relatively poor countries (such as Nicaragua and Bolivia) as well as in relatively rich ones (such as Thailand and Venezuela). And they have been driven by the manipulation of a wide range of societal cleavages, such as class conflict (as in Venezuela and Thailand), religious conflict (as in Egypt and Turkey), or national self-assertion (as in Russia and Hungary) (see e.g. Bermeo 2016; Diamond 2016; and Erdmann 2011). The disconcerting diversity of contexts and origins helps spreading the vague, alarming suspicion: “it can happen here” (Sunstein 2018).

In numerous instances, the “death” of new democracies has not been “rapid” but “slow”.Footnote 3 Military and executive coups continue to slay democratic regimes (see Tansey 2017). Yet, the modal transition from democratic rule today unfolds as a “slow and at times opaque” “process of successive authoritarian advances” (O’Donnell 1992, pp. 19, 33) in the hands of illiberal governments. Prime examples of contemporary leaders who have orchestrated gradual transitions from electoral democracy to electoral authoritarianism have been Russia’s Vladimir Putin, Venezuela’s Hugo Chávez, Turkey’s Recep Tayyip Erdoğan, and Hungary’s Viktor Orbán.Footnote 4

Before Donald Trump’s irruption into the US presidential election, concerns about the “global democratic recession” (Diamond 2015) had been largely confined to new democracies. Despite decades of debate about “the crisis of democracy” (see e.g. Merkel 2014), established democracies appeared to be essentially immune to their illiberal subversion. The US 2016 election has changed this overnight. It induced an instant radical shift of the democratic mood. Due to a surprising concatenation of contingent events, the unthinkable has become thinkable and the sense of democratic fragility has spread to “advanced” democracies. In a rare case of interdisciplinary convergence, politicians, journalists, historians, philosophers, psychologists, and political scientists alike have been warning that one of the world’s oldest democracies might be sliding into authoritarian rule.Footnote 5

3 Disciplinary blinkers

Comparative political science has been caught ill-prepared by democracy’s renewed fragility. After the turn of the millennium, we had shifted much of our attention to study authoritarian regimes.Footnote 6 Taking the stability of established democracies for granted (even while surveying their defects), we all but abandoned the study of consolidation in new democracies (see e.g. Cheibub 2014). Somewhat painfully, we now discover that the theoretical tools and empirical findings which we have accumulated in the comparative study of political regimes over decades are of little use for understanding the most common form of democratic regression today, that is, the stepwise subversion of democracy by illiberal governments (see also Waldner and Lust 2018). Arguably, their limited usefulness derives from their almost exclusive emphasis on probabilistic structural explanations, high-profile violations of democratic norms, and utilitarian sources of norm compliance.

3.1 The absence of actors

Our body of past research on democratic continuity teaches us little about democratic backsliding because its main explanatory thrust lies on structural factors, such as economic wealth, social inequality, political institutions, mass culture, and the international environment.Footnote 7 Enduring, structural causes, however, cannot account for short-term changes, such as processes of democratic subversion (see also Waldner and Lust 2018). Only political actors and their actions can.

Besides, structural explanations are probabilistic. They estimate average effects of contextual parameters and regime properties on the likelihood of regime change or survival. These explanations are never deterministic but leave room for contingencies. Even if we know, for example, that “economic explanation alone goes a long way” (Przeworski 2003, p. 129) in explaining democratic survival, it does not go all the way. The possibility of outliers, exceptions, and surprises always exists. That is, the possibility that deviant actors may disrupt structural equilibria always exists. As political scientists, it appears, we tend to deal with such remote possibilities in the bipolar fashion in which we as human beings tend to deal with low-probability outcomes: we either ignore or magnify them (see Kahneman 2011, Ch. 29). Donald Trump made us switch from one pole to the other.

3.2 The narrowness of norms

The classic definition of democratic consolidation as a situation in which democracy has turned into “the only game in town” (Linz and Stepan 1996, p. 5) describes a situation of universal rule compliance in which “all the actors in the polity [have] become habituated to the fact that political conflict will be resolved according to the established norms” (Linz and Stepan 1996, p. 5). They know, accept, and obey the fundamental rules of the democratic game, and expect others to do so as well. Yet which are those rules? What does democratic compliance entail?

Until recently, the literature on democratic instability focused its attention on those public and dramatic instances of democratic rule violation which account for the vast majority of “quick deaths” of democracy: military coups, executive takeovers, and armed rebellions (see e.g. Przeworski 2005; Svolik 2015a, b, p. 730). At the same time, the broad literature on “intermediate regime categories” (from “defective democracies” and “hybrid regimes” to “electoral authoritarianism”) recognized that the forceful closure of representative institutions represents only one existential threat to democracy among many others.Footnote 8 Political actors can damage democracy in countless ways below the level of open coups and rebellions. Given its interest in authoritarian governance, much of the hybrid-regime literature focused on what we may call “democratic capital crimes”, that is, democratic norm breaches that are “severe and systematic” enough to kill democracy. In the electoral realm, for instance, it studied transgressions such as repression, censorship, party bans, voter intimidation, or electoral fraud, which threaten the democratic spirit of multiparty elections. More recent research on “democratic backsliding” tends to replicate this focus on incisive norm violations that oblige us to raise the regime question: Is country X still a democracy? Or did the illiberal incumbent already push it into the realm of electoral authoritarianism?Footnote 9 As I will argue, Donald Trump’s transgressive behavior lies outside our standard catalogue of democratic offenses. It obliges us to broaden it, or rather, to recognize that it includes transgressions we had not held to be essential.

3.3 The narrowness of motives

If democracy requires actors to comply with its basic norms (however defined), what makes them do so? Which are the motivational foundations of compliance? Does democracy require democrats? Does it require normative commitments to democratic principles? Or can it work without them and rely on no more than actors’ self-interested calculations of material benefit?

Even while recognizing the possibility of building “democracy without democrats” (see e.g. Salamé 1994), early studies of regime change placed actors’ normative commitments to liberal democracy at the very center of their explanatory enterprises. Their central analytic distinction ran between the friends and enemies of democracy, between “democratic actors” committed to democracy, on the one side, and “authoritarian actors” hostile to it, on the other. In essence, the classic, multi-volume project on democratic breakdown (edited by Linz and Stepan 1978) as well as the later collection on democratic transitions (edited by O’Donnell et al. 1986) explained political regime dynamics by the evolving power balance between these two categories of actors. For scholars of regime transitions, distinguishing between these polar camps (and the amorphous group in between) represented a central analytic challenge.Footnote 10

Normative commitments to democracy continue to be an important subject matter in the comparative study of public opinion.Footnote 11 Macro-level research, however, has largely abandoned the notion that political actors’ normative commitments play a significant causal role in political regime dynamics. Prevalent theories of democratic stability discard the relevance of normative motives and assume democratic stability to be primarily grounded in material self-interest. They hold democracy to be sustainable when it constitutes a “self-enforcing equilibrium” which is “in nobody’s interest to change” (O’Donnell 1994, p. 58). Assuming the universal primacy of self-interest, these theories conceive actor preferences for democracy or dictatorship as derivative of their underlying material interests (see e.g. Acemoglu and Robinson 2006; Boix 2003; Przeworski 2005; Svolik 2015a; Teorell 2010, p. 26).

The few “idealist” scholars who still insist on distinguishing between “democratic” and “authoritarian” actors on the basis of their “normative preferences” know that they are fighting an uphill battle. For instance, in their longitudinal study on Democracies and Dictatorships in Latin America, Scott Mainwaring and Aníbal Pérez-Liñán recognize that even the modest “idea that a value commitment to democracy could be a major asset for the durability of democracy is far from consensual” (Mainwaring and Pérez-Liñán 2013, L. 1233, emphasis added). Our responses to Donald Trump, however, seem to reveal a surprising consensus: we do not trust rational actors who lack “a value commitment to democracy”.

3.4 The puzzlement

Given this state of the literature, our democratic anxieties about Donald Trump look puzzling. Despite the disciplinary emphasis we put on structures, basic norms, and self-interest, many of us, comparative scholars of political regimes, have responded with disbelief, shock, and alarm, both to the public behavior and the public persona of Donald Trump. Our systemic worries seem to ignore the weight of structures; they do not seem to arise from the types of transgressive behavior which we commonly associate with the demise of democracy; and they revolve around a politician who appears to pursue no more than his personal interest, rather than a transformative authoritarian program. How comes? Which are the sources of our anxieties? What is it in Donald Trump that awakes our democratic fears?

As I claim, our alarmed reactions within the political science community have been informative. They tell us that Donald Trump has been violating strong normative as well as causal assumptions about the foundations of liberal democracy which we had been embracing in a tacit manner, in contradiction with our explicit theories of democratic equilibria. In this sense, the 45th US president has subjected us to a political “breaching experiment” whose normative breaches have been exposing our hidden disciplinary assumptions about the normative bases of democracy.

4 Breaching experiments

In the 1960s, ethnomethodologist Harold Garfinkel directed his students to perform so-called “breaching experiments”. These sociological experiments were designed to uncover the implicit cognitive or normative presuppositions that undergird ordinary social interactions. In classic examples, experimenters would pretend not to understand common-sensical expressions and ask for clarification. Or they would assume inappropriate social roles (such as acting like boarding students in their own homes) or attribute inappropriate social roles to others (such as treating costumers as salespersons) (see Garfinkel 1984, 2001). Unsurprisingly, such unexplained breaches of “the routine grounds of everyday activities” (Garfinkel 1984, p. 35) take subjects by surprise. The “nastiness of [their] surprise”, Garfinkel posited, “should vary directly” with the strength of the assumptions they hold about “normal” dynamics of social interaction (2001, p. 381).

As a matter of fact, the “unsuspecting subject[s]” (Garfinkel 2001, p. 385) of these experiments tended to respond with high emotional intensity to the breaches of trust they were confronted with. Garfinkel reported expressions of “astonishment, bewilderment, shock, anxiety, embarrassment, and anger” (p. 386). The experimental disruption of normality (under pretensions of normality) left some subjects, not just “bewildered” (p. 382) and “indignant” (p. 383), but badly “shaken” (p. 385). Their “suffering was dramatic and unrelieved” (p. 391). “No pains were spared” (p. 392).Footnote 12

Breaching experiments continue to enjoy some popularity in contemporary sociology classes (see e.g. Braswell 2014) that direct their students to violate common rules of communication or interaction in order to uncover “hitherto undiscovered assumptions about commonsense reality” (McNall and Johnson 1975, p. 50). Their “breaching experiments” usually consist in doing either mildly funny or mildly irritating things “out of the normal”, such as making music or noise in the quiet reading room of a public library, to record the consequent expressions of puzzlement or acts of resistance by the target audience. They aim at making explicit what we can leave implicit in our everyday interactions. As Garfinkel explained, they “must be thought of as demonstrations rather than as experiments” (2001, p. 381).

Though Donald Trump has been excelling in the Garfinkelian art form of “making trouble but not so much as to be jailed” (McNall and Johnson 1975, p. 50) we may safely assume that his multifaceted breaches of democratic expectations have not been designed to teach us lessons in democratic theory. If we wish to appreciate the didactic, demonstrative value they possess nonetheless, we need first to understand their precise nature. The US president is said to have broken “just about every basic rule of democratic politics” (Mounk 2018, p. 118). But what have been the exact contours of his transgressive acts and dispositions? Which exactly has been the political breaching experiment he has been performing on us?

Answering this question demands taking some analytic distance from ongoing debates on Donald Trump’s breaches of democratic norms. In terms of their substantive focus, these debates have been quite disorderly, frequently mixing alleged violations of democratic and other types of norms, such as moral propriety or financial integrity (e.g. Frum 2018; and Klaas 2017). They often choose their objects of critique in inductive, associative ways. The president says and does things and his critics cry foul because they recognize some resemblance to actions undertaken by modern dictators (“the autocrat’s playbook”).Footnote 13 In evaluative terms, much of these debates have been conducted in a dichotomous, not to say Manichean, manner. Almost any transgression of democratic norms, regardless of their severity, has served to put Donald Trump into the category of the “elected autocrat” (Levitsky and Ziblatt 2018).

Here I propose a more “deductive”, more systematic and comprehensive approach. Excluding all considerations that go beyond constitutive liberal-democratic norms, I review the US president’s compliance with five sets of basic norms which the comparative literature holds to be largely uncontroversial (except for the last one): the support of democratic ideas and institutions, the renunciation of violence, the respect of electoral integrity, the respect of constitutional constraints, and linguistic accountability. With the help of illustrative comparisons from contemporary experiences of democratic subversion, I will strive to reach rough assessments of the severity of his normative breaches in these realms. Given the unending whirlwind of public controversies over his democratic transgressions, I will limit myself to his first two years in office. Besides, I will exclude all international considerations: the damage which external authoritarian interference invited, aided, or tolerated by Trump might have done to US democracy as well as the damage which the US president might have been inflicting on the democratic cause in other countries.

5 The experiment

As empirical scholars of comparative democracy, we sometimes must remind ourselves of the obvious: Democracy, the rule by the people, is a normative concept. Its core principles of equality and liberty stand against those that have ruled human societies during most of our collective existence: hierarchy and oppression. The set of modern representative institutions we describe as democratic strive to put those principles to practice: constitutional government, individual rights and liberties, the rule of law, competitive elections, representative legislatures, civil society, and mass media. Democracy requires those institutions. Without them, it cannot be said to exist. Yet, democracy also requires behavior which is consistent with its principles and institutions. It requires active compliance. Otherwise, its principles are empty rhetoric, its institutions empty shells. Which, then, are the basic behavioral demands that democracy puts on political actors?Footnote 14

5.1 Democratic support

The softest demand the democratic game places on its players is to support the game itself in public speech and writing.Footnote 15 In its most narrow version, this implies the ideological demand to support democracy and reject its systemic alternatives. Yet, due to the formal or informal sanctions which most contemporary democracies put on anti-democratic speech, anti-democratic actors will seldom declare open hostility to democracy. Today, in the age of democracy, where everyone is supposed to profess at least lip service to liberal democratic values, even dictators have learnt to speak the language of democracy. Generic public declarations of democratic commitment have thus become worthless. They fail to distinguish democrats from anti-democrats.

In consequence, actors who are “structurally suspicious”, be it because of their violent past (such as former authoritarian rulers and former guerrilla fighters) or their illiberal ideologies (such as communist parties and religious movements), face persistent suspicions that their public discourse and behavior is deceptive. In an analogous way to the subjects of oppression under authoritarianism, they are suspected of engaging in strategic “preference falsification” (Kuran 1995; Havel 1985). Almost invariably, critical debates about their democratic convictions revolve around distinctions like appearance vs. reality, mask vs. face, front stage vs. backstage (see e.g. Driessen 2012; Richards 2001).

Accordingly, we often demand that democratically suspicious actors do more than to profess general support for the fundamental rules, institutions, and principles of liberal democracy. We ask them to support the concrete rules and institutions that happen to govern their real existing democracy (see e.g. Booth and Seligson 2009; Higley and Gunther 1992). If they fail to do so, they do not reveal themselves as unequivocal enemies of democracy. But they do reinforce existing suspicions. They kindle the question whether they might eventually be willing to abandon, breach, or attack democratic institutions.

Such motivational suspicions have haunted debates on populism. Populist actors exploit the structural distance that separates citizens from their representatives in liberal democracy. They play a game of calculated ambiguity. While they present themselves as champions of democracy, they declare established democratic elites and institutions to be the source of all evil that befalls ordinary citizens. Given their ambivalent high-wire act between abstract support and concrete condemnation of democracy, the democratic commitments of populist actors have been subject to persistent controversy. Scholarly assessments span the whole range of possibilities. Populists have been analyzed as illiberal threats to democracy (Müller 2017; Weyland 2013), semi-democratic actors (Schedler 1996), democrats with exclusionary ethnic conceptions of the demos (Mudde 2010), democrats with alternative conceptions of democracy (Mounk 2018), and democrats with selective definitions of liberal democracy (Slater 2013; Rovira Kaltwasser 2012).

Since the early days of his presidential nomination campaign, Donald Trump has been identified as “the perfect populist” (Lind 2016). Surely, during his electoral campaign, he cherished to invoke “the people”, “the forgotten ones”, and portrayed existing democratic institutions as sites of treason and corruption. Yet, in office, he has clearly situated himself on one side of interparty competition. The central political conflict he mobilizes is not the populist cleavage (people vs. elite) but the partisan cleavage (Republicans vs. Democrats). Aside from playing the tunes of interparty polarization, he embraces a combination of ethnic and economic nationalism (older immigrants vs. more recent ones, ethnic Americans vs. all other nations) and has adjusted his self-description accordingly.Footnote 16

Still, even though Donald Trump may not qualify as a populist, he does partake in the typical discursive ambiguities of populism. While various critics have labeled him a “fascist” or “aspirational fascist” (Beinhart 2018; Connolly 2017; and Tucker 2016), he fully embraces the existing democratic order. He professes nothing but pride in the history of American democracy. He does not promise a fascist Führerstaat or the formal transformation of the US into a hereditary oligarchy. In contrast to other democratically ambiguous actors, he is not suspected either of dissimulating his authoritarian convictions. He has been accused of falsifying objective facts, but not his personal preferences. Even his critics see him as an open book. Due to his apparent lack of an ideological core and his “radical honesty” (Bruni 2018) about his own evolving thoughts and emotions, he has not been suspected of pursuing a hidden anti-democratic agenda.

Yet, while refraining from attacking democracy in the abstract, Donald Trump frequently criticizes specific democratic actors and institutions. Their appreciation seems to be strictly contingent on their conformity with his personal and political ambitions. If they serve his purposes, fine. If not, he showers them with contempt. This can be seen in his recurring rhetorical attacks against the media, judicial actors, and the electoral process. Much has been written about his hostile attitude towards critical news media, which are a central pillar of liberal democracy according to all democratic theory and “the enemy of the American people” according to Trump.Footnote 17 Notoriously, the 45th US president has also been harshly dismissive of judicial actors he dislikes (be it courts, judges, prosecutors, the fbi, or the justice department) (see e.g. Baker 2017; Levitsky and Ziblatt 2018, pp. 178–179). In the closing weeks of the campaign, he started denouncing the presidential election as being “absolutely [] rigged”, “one big fix”, “one big, ugly lie” (Martin and Burns 2016), without offering even a hint of empirical evidence.

When political actors fail to pay honor to democratic institutions, the consequences are seldom direct. Their verbal attacks may induce others to reject and weaken these institutions, or else, to defend and strengthen them. In and by themselves, they do not impair institutional performance. Despite his rhetorical attacks on the press, Donald Trump has not been restricting freedom of opinion in practice. Despite his attacks on judicial actors, he has been heeding the limits they have set on his policies (see e.g. Posner 2018). In the first instance, the discursive denial of institutional respect is what it is: a discursive breach of democratic obligations. Its initial relevance is primarily diagnostic. It reveals the speaker’s weak attachment to democratic values.

5.2 The renunciation of violence

The primary behavioral imperative of liberal democracy is its absolute ban on political violence. Violence is “the greatest enemy of democracy as we know it. [It] is anathema to its spirit and substance” (Keane 2004, p. 1). For the demos to speak and to be heard, arms must remain silent. Democracy prohibits the use of violence as an instrument of accessing state power (coups and rebellions), competing for state power (electoral and ethnic violence), exercising state power (political repression and the selective enforcement of law), influencing state policies (terrorism and coercive corruption),Footnote 18 and settling political conflicts among citizens (hitmen, paramilitary forces, lynching mobs).

Democracy, however, demands more from political actors than refraining from exercising violence in person. Democrats need to do everything in their power to prevent others from doing so. Political actors have a wide spectrum of possible strategies at their disposition through which they can induce others to commit political violence without bloodying their own hands. They can command, delegate, license, invite, encourage, condone, conceal, deny, belittle, and gloss over acts of political violence. Democrats must avoid any of this, unambiguously. They must show zero sympathy with the perpetrators of violence, regardless of who they are, and full solidarity with its victims, regardless of who they are. Their responsibility extends to their formal and informal agents as well as to their actual and potential allies. It also includes the obligation of recognizing their opponents as legitimate participants in the political process, rather than treating them as enemies who are legitimate targets of political violence.

Spanish political scientist Juan Linz, who saw the Spanish civil war break out at the age of ten, had been deeply troubled about ambiguities in the face of violence. “What”, he asked in his seminal 1978 reflections on elite attitudes towards democracy, “would be an effective ‘litmus test’ of loyalty to a democratic regime? An obvious possibility is public commitment to legal means for gaining power, and rejection of the use of force. Ambiguities in such public commitments are certainly prima facie evidence of semiloyalty” (Linz 1978, p. 29). To exemplify, he cited the “willingness to encourage, tolerate, cover up, treat leniently, excuse, or justify the actions of other participants that go beyond the limits of peaceful, legitimate pattern of politics in a democracy” (Linz 1978, p. 32).

The democratic ban on political violence thus imposes a broad set of negative and positive obligations on political actors. How does the 45th president of the US fare in this regard? The picture is mixed. In contrast to the global avant-garde of illiberal presidents, his personal involvement in domestic repressive violence has been close to zero. He has not conducted a wave of mass arrests against journalists, civil activists, politicians, and public officials, as Recep Tayyip Erdoğan has done after the 2016 coup attempt (see e.g. Kingsley 2017). He has not unleashed a campaign of extrajudicial killings against sellers and consumers of psychoactive drugs, as Rodrigo Duterte has done in the Philippines (see e.g. Berehulak 2016). Still, in his rhetoric, he has shown some of the ambiguities Juan Linz identified as symptoms of democratic semi-loyalty.

Above all, Donald Trump has been permissive towards the violence of ideologically proximate third parties. On numerous occasions, he has been seen as encouraging or condoning illegal violence. At campaign rallies, he invited his followers to “beat the crap” out of protesters and promised to cover their legal defense bills. Once, he obliquely suggested arm owners might recur to legitimate self-defense in case Hillary Clinton would restrict access to fire arms (indeed, in a context of deep partisan division over gun control, his advocacy of “gun rights” ensures that his political supporters control an immense and disproportionate arsenal of assault weapons—which they may never use for political purposes but can at any time).Footnote 19 After his election victory, he hesitated to condemn attacks against Jewish institutions or deadly violence by neofascists in Charlottesville, Virginia. He deems the notion of holding police officers accountable for the use of lethal violence unpatriotic and has invited immigration officers to do their job without undue respect for human dignity (“please, don’t be too nice”).Footnote 20 Together with his trademark rhetorical aggressiveness against adversaries, Donald Trump’s rhetorical ambiguities towards violent allies and agents are bound to breed consequences in the so-called real world.Footnote 21

5.3 Electoral integrity

Since the end of the Cold War, the most common form of dictatorship are not military and single-party regimes anymore, but electoral autocracies which hold multiparty elections, yet subject them to severe and widespread manipulation (see e.g. Levitsky and Way 2010; Schedler 2013). Accordingly, as noted above, the most common form of democratic termination are not military coups anymore, but gradual, government-led transitions to electoral authoritarianism (see e.g. Bermeo 2016; Svolik 2017). Governments have a variegated “menu of manipulation” at their disposal which allows them to preserve the formal façade of competitive elections, while hollowing out their democratic substance (Schedler 2013, Ch. 3). In the transition to electoral authoritarianism he commandeered, Venezuela’s Hugo Chávez, for example, deployed the entire arsenal. He rewrote the constitution to his liking; captured judges and electoral authorities; ruled by decree, through repressive laws, and the transgression of laws; colonized the state bureaucracy and civil society; stifled private media and created his own propaganda machine; harassed, banned, and prosecuted opposition parties, sacked their elected officials, or stripped them of formal authority (see e.g. Corrales 2015, 2011, Ch. 2).

What has Donald Trump done so far to threaten the integrity of the electoral process? During his campaign and after, he repeatedly denounced “the system” and “the election” to be “rigged” (see e.g. Martin and Burns 2016; Wines and Haberman 2018). He is right, of course.Footnote 22 For a long time now, US democracy has suffered deep, structural damage from its radical openness to the influence of money (see e.g. Winters 2011) and the partisan manipulation of electoral rules (district boundaries and voter identification) (see e.g. Anderson 2018; McGann et al. 2016; Klaas 2017, Ch. 5; Schaffer 2008, Ch. 2). Donald Trump, though, has not been the creator of these structural flaws, only their beneficiary.

Has he further harmed the democratic integrity of American elections? Not that much, actually. Having lost the popular vote to Hillary Clinton, he kept complaining, as he had done during the campaign, about “serious voter fraud” (Shear and Haberman 2016), a rhetorical phenomenon that has eluded empirical documentation. He continued to play on the topic through his short-lived Presidential Advisory Commission on Election Integrity (see Wines and Haberman 2018) and took it up again after the 2018 midterm elections (see Thrush and Peters 2018). Yet, as in other spheres, the main harm done was discursive. Rather than undermining the integrity of elections, he has been undermining the credibility of elections.Footnote 23

5.4 Constitutional constraints

In the 1990s, in early debates on the institutional foundations of democratic stability, Latin American countries were thought to inhabit the worst of all worlds. Combining presidential systems of separation of powers and fragmented multiparty systems, they seemed condemned to engender irresolvable confrontation between minoritarian presidents and legislative majorities, inviting the military to step in and play the arbiter (see e.g. Linz 1990; Mainwaring 1993). Subsequent empirical studies have qualified or shed doubt on “the perils of presidentialism” (see e.g. Negretto 2006; and Cheibub 2007), just as subsequent empirical events have increased confidence in the conflict-solving capacities of presidential democracies. Since the early 1990s, Latin American countries have solved recurrent “presidential crises”, not through military intervention, but through the constitutional ousting of presidents who had lost the public trust (see e.g. Hochstetler 2006; Pérez-Liñán 2007).

In the light of preceding debates, it is not without irony that current dangers to democracy do not seem to arise from minority presidents who attack hostile legislatures, but from presidents who command legislative majorities and use them to redefine the rules of the political game. Juan Linz identified the tendency to produce “unified” rather than “divided government” as one of the core “virtues of parliamentarism” (Linz 1990). Today, executive control of the legislature constitutes the Archimedean point for hegemonic projects of “executive aggrandizement” (Bermeo 2016).

Putin, Chávez, Erdoğan, Orbán, and company have all followed the same basic script. Act one: they gain control over rule-making after winning legislative majorities. Act two: they extend their controlling powers to the final settlement of conflict by subjecting supreme courts to partisan control. Act three: once they are in control of the legal system (the content, implementation, and adjudication of law) they are (legally) free to colonize the entire state bureaucracy (including the management of elections) and to domesticate the media and civil society.

How does Donald Trump fit the picture? In his first two years in office, he enjoyed the relative comforts of “unified government”. Even though Republican majorities in both chambers were rather slim, and the internal discipline of the party allowed for certain lapses, he used his legislative majorities to work towards effective partisan control of the judiciary.Footnote 24 The contentious nominations of Neil Gorsuch and Brett Kavanaugh to the Supreme Court locked in a conservative majority at the peak of the judicial system. Democrats viewed them as part of a “Republican strategy to capture our judicial branch of government” (Senator Richard J. Durbin) (Stevenson 2017). The aggressive, sustained, and successful campaign by the president and his party to appoint ideologically proximate judges to federal appeals courts has been validating their fears (see e.g. Savage 2017; Hulse 2019). Of course, the degree of formal control which ruling parties and sitting presidents exercise over judicial appointments has always been a deep anomaly of US democracy. Their power to shape the ideological profile of the judicial system would be perceived with great apprehension in any other democracy.Footnote 25

5.5 Linguistic accountability

Liberal democracy is not a measurement machine, a mechanism for registering the spontaneous, pre-reflexive preferences of citizens. It is a system for defining, debating, and deciding collective problems and conflicts within a community of free and equal citizens. Abjuring violence, abhorring violence, its core medium is language. Against the power of stones, swords, and guns, it poses the power of words (see e.g. Touraine 1988). It is only through our shared language that, “we, the people”, know who we are, what we have in common, what divides us, what afflicts us, where we want to go. Liberal democracy requires elections, but the institutional core of popular sovereignty is not the electoral arena, but the public space that gives meaning to the electoral arena (see Habermas 1998). Its democratic quality hinges centrally on the integrity of public language, its responsible and accountable usage.

As users of language, we make claims—about the world, about others, about ourselves. As responsible language users we take these claims seriously. We accept the commitments they carry, and we accept being called to account for these commitments. Linguistic accountability involves the duty to clarify and explicate our claims (their meaning) in case of doubt, and to answer critical questions about their rationality: their truthfulness (when we make factual claims about objective realities), their appropriateness (when we articulate normative claims about social relations), their sincerity (when we put forward subjective claims about our inner world), and their consistency (with other things we say and do) (see e.g. Brandom 2000; Habermas 1984).

Linguistic irresponsibility comes in many faces. Donald Trump is the carrier of an extreme form. Numerous observers have noted traits of linguistic poverty, even pathology, in his public discourse, such as his simple syntax and limited vocabulary, his delight in transgressing rules of courtesy and decency, his fondness of personal insult, his habitual substitution of adjectives for arguments, the coarse nature of his rhetoric, his impatience with caution and precision, his corresponding love for hyperbole and over-generalization, and last but not least, his general contempt for language (“just words”). Donald Trump has been called “a serial norm breaker” (Levitsky and Ziblatt 2018, pp. 146, 195). He surely is a serial breaker of linguistic norms.Footnote 26

Within this context, the primary public concern about Donald Trump’s political language, first as a candidate and then as president, has been his shallow commitment to truthfulness. While Bill Clinton faced impeachment charges over a single lie, Donald Trump has been generating an ever-growing list of lies, half-truths, and distortions of reality, creating “alternative facts” and “alternative lies” alike.Footnote 27 Critics and adversaries alike have been describing him as a “compulsive” or “pathological liar” (Willingham 2016) and he has co-inspired a whole new genre of literature on democratic politics in “the post-truth era” (e.g. d’Ancona 2017).

Donald Trump not only generates an unending stream of arbitrary claims which are not (or only loosely) constrained by empirical realities or the rules of language. He fully commits himself to these claims with public displays of emotional certainty (often through the use of superlatives) and dismisses all pretensions of holding him accountable for them (often through repetition). His conception of truth is private and declarative: he tells the truth because he says so, his critics are liars, all fake, full point. He leaves no legitimate space for public argumentation. Herein lies the double core of his linguistic irresponsibility: in his demonstrable contempt for truth and his demonstrative rejection of linguistic accountability.

Linguistic responsibility may seem to belong to the luxury department of democratic quality. As a matter of fact, students of deliberative democracy have been analyzing it as such: as essential to the quality, but not to the very existence, of democracy (see e.g. Gutmann and Thompson 2004). Yet, Donald Trump made us realize that linguistic responsibility is just as essential to democratic conflict resolution as it is, for instance, to judicial conflict settlement. In judicial proceedings, judges are called upon to apply rules to facts as they are established before court. If they get the facts wrong, they get everything wrong. Without truth, no justice. Modern judicial systems accordingly treat false testimony not as a stylistic mishap but a criminal offense. Democracies protect the truthfulness of public debate through informal norms and practices of accountability rather than penal law. Still, their existential vulnerability to linguistic abuse is similar. Irresponsible and unaccountable speech dissolves the twin democratic assumption of a shared reality and a shared language. By turning politics into a cacophony of unintelligible voices, it destroys the public sphere, the virtual meeting point of the demos. Irresponsible speech is antidemocratic speech—not in substance, but in form. “If words fall into disrepair, what will substitute? They are all we have” (Judt 2010, L 1523).

6 The revelation

Harold Garfinkel introduced his sociological breaching experiments as “a means of eliciting social order through the disruption of taken-for-granted realities” (Gregory 1982, p. 49). Their core purpose lies in the “‘rediscovery’ of common sense” (Garfinkel 1984, p. 75). By violating hidden assumptions of ordinary social interaction, these transgressive experiments intend to render them visible. The anxiety produced by the disappointment of tacit expectations is meant to serve an epistemic purpose: making them explicit.Footnote 28 Now, which are our hidden assumptions about the democratic order that Donald Trump’s political breaching experiment have been unveiling? As I wish to suggest, our anxious responses to his transgressions have been revealing implicit assumptions we hold about the power of individual actors, behavioral norms that constrain public speech, and normative motives that constrain political actors.

6.1 The weight of actors

Even while Donald Trump’s ascent to power has provoked intense debates about the threats he might pose to democracy, the United States of America are, of course, an entirely improbable case of democratic breakdown. The country scores high on almost all the structural and institutional characteristics which the comparative literature has identified as foundations of democratic stability, such as societal modernity and economic affluence, a broad middle class and a vibrant civil society, a pro-democratic citizenry, a strong judicial system, high regulatory and extractive state capacities, the protection of wealth from redistributive pressures, multiple experiences of alternation in power, high party-systemic institutionalization, strong vertical and horizontal divisions of power, and geographical proximity to the West.

Considering that our established theories of democratic consolidation lead us to expect US democracy “to last forever” (Przeworski and Limongi 1997, p. 165), what do fears about its subversion by Donald Trump tell us? Are they signs of scientific ignorance? Or expressions of partisan hysteria (see e.g. Moyn and Priestland 2017)? I believe they are neither. In fact, I hold them to be entirely rational and compatible with our scientific findings.

Political science is a craft of probabilistic inferences. It strives to discern empirical regularities, estimate average effects, and predict probable outcomes. Politics, by contrast, ineluctably contains possibilistic streaks. While it needs to calculate probable outcomes, it cannot ignore possible ones. It cannot ignore the fact that improbable events may happen after all. In relation to desirable events, politics is the art of the possible (and who could better testify to that point than Donald Trump himself). Its imagined capacity to make possible the impossible, or more modestly, to make the improbable come true, is a recurrent source of hope. In relation to undesirable events, politics is the art of the impossible. In the face of grave threats, it needs to take “all necessary precautions” to preempt their occurrence.

In the US (as well as elsewhere), fears about democratic subversion are not probabilistic, but possibilistic. The haunting question about authoritarianism in the US has always been (as in Lewis 2014 or Sunstein 2018): Can it happen here? Rather than: Will it happen here? Such concerns are expressions, not of irrationality, but of rational caution in the face of possible catastrophes. The odds that a democratic head of government will be able to destroy a complex, affluent, long-lived democracy are small. Still, they are arguably larger than zero. Democratic citizens are supposed to be vigilant and risk-averse towards potential threats to their rights and liberties. Rather than relaxing in the face of unlikely catastrophes, they must strive to discern and to preempt them. As our Trumpian anxieties reveal, we, probabilistic political scientists, do the same.

In our role of democratic citizens, we too embrace what cognitive psychologists call the “possibility effect”, which is, our human tendency to either ignore or overweight improbable outcomes (see Kahneman 2011, Ch. 29). For long, in so-called advanced democracies, we thought that we could dismiss any residual risks to their equilibria. Now, they make us panic. And plausibly so. In the face of existential threats whose (low) probabilities are not fixed by “nature” but contingent on our own behavior, their “overweighting” is not an expression of “bounded rationality” (Kahneman 2011, Ch. 29) but political prudence. Adopting such a “possibilistic” viewpoint, is not unscientific, although it does rest on a form of reasoning we are not accustomed to: the logic of “possibilism” (Hirschman 1986; Lepenies 2008).

6.2 The weight of words

How much damage have Donald Trump’s democratic transgressions inflicted on US democracy so far? Although academic experts tend to see “a sharp decline” in the quality of US democracy under Donald Trump (Carey et al. 2019, p. 701), none of his normative transgressions amounted to a major catastrophe. He has not suspended civil liberties, called in the military, organized anti-Mexican riots, or set up prison camps for Muslims. Despite some scratches on its surface, and some beneath, US democracy is still standing, vibrant with pluralism and activism. With all its structural defects, it has not been consumed by the fury of Donald Trump’s national narcissism. Many observers have interpreted this as a testimony to the enduring strength of democratic institutions: “The system worked” (Kristof 2018).

Given the paramount importance of impartial state institutions for democracy (see e.g. Lapuente and Rothstein 2014), the president’s judicial appointment policies risk deleterious long-term effects on democracy. The Republican Party’s success in enlarging the conservative majority on the Supreme Court and its sweeping campaign to shift the composition of federal appeal courts involve a creeping colonization of the judicial system. Although they have breached longstanding informal rules, these policies have remained within the bounds of the law, that is, within the bounds of a legal system that permits its own partisan capture. The expected political consequences are transparent: a more conservative bench is more likely to shield the president and his party from judicial accountability.

Otherwise, though, the 45th US president’s most severe democratic norm violations during his first two years in office have been discursive. This applies to his contempt towards democratic actors and institutions that threaten to thwart his ambitions; his recurrent ambiguities towards acts of political violence by ideologically proximate actors; his partisan support for unfair electoral boundaries and exclusionary voter registration laws, and his prodigious linguistic irresponsibility. Yet, as philosophers of language know, speech acts are acts, too. Arguably, they are the most important category of political acts in democratic politics. From the viewpoint of comparative regime studies, though, our sensation of alarm in the face of Donald Trump’s discursive norm transgressions looks puzzling, considering the relative neglect of such transgressions in our discipline.

In the comparative literature on political regimes, we have been studying a broad set of democratic norm violations, often on the basis of longitudinal cross-national datasets. Some of our datasets capture dramatic events and processes, like military coups, civil wars, acts of terrorism, riots, and electoral violence. Others provide annual assessments on vital dimensions of democratic quality, such as state repression, media freedom, electoral integrity, the rule of law, judicial independence, and corruption.Footnote 29 None, however, captures what has been igniting our worst fears in “the dangerous case of Donald Trump” (Lee 2017): his public discourse. The realm of public speech has been below the radar of comparative regime studies.

In general, in comparative politics, we have been reluctant to take public speech seriously. We have tended to regard it as a both illegitimate and unreliable type of empirical evidence. Given the symbolic nature of language, we often set speech in sharp contrast with “observable behavior” and thus exclude it from the realm of legitimate evidence. In addition, since speakers may be insincere and words deceptive, we habitually conclude that we should ignore what people say (their political discourse) and look exclusively at what they do (their political practice) (see e.g. King et al. 1994, p. 112; Przeworski 2010, L 278; and Rathburn 2008).

Both grounds for excluding political discourse from empirical analysis do not withstand scrutiny. (a) The notion that public discourse is “non-behavioral” overlooks that it belongs to the most important category of human action: speech acts (see e.g. Austin 1962; Searle 1969); and the notion that its symbolic nature renders it dependent on interpretation and thus “non-observable” overlooks the intrinsically interpretative nature of all social observation.Footnote 30 (b) The potential of human speech for deception and dissimulation is undeniable. Yet it has its counterpart in our human capacity of discerning inconsistency and insincerity. Donald Trump nicely illustrates the point: none of his critics have been accepting his discourse naively. Quite to the contrary, every word he utters in public and every sentence he tweets are subject to intense scrutiny.

Of course, “qualitative” or “interpretative” traditions of political research have no methodological reservations towards the study of language. Quite to the contrary, they place political language at the very center of their analysis.Footnote 31 And so do studies of political communication and electoral campaigns (for an overview, see Korte and Regge 2016). Furthermore, in recent years, due to both the electronic availability of huge amounts of text and technological progress in automated text analysis, more and more political scientists from the “quantitative” or “positive” camp, too, have been embracing “text as data” (see e.g. Grimmer and Stewart 2013). Our comparative research of regime dynamics, however, had been “speechless” until very recently. What political actors said in public had no explanatory value within our common analytical perspectives. We treated it as Trump would have: “It’s just words”.Footnote 32 It was only through the radical discursive disinhibition of figures like Hugo Chávez and Donald Trump that we have been “bringing language back in” to our democratic threat analyses, for instance, in emergent comparative research on populist communication (e.g. Aalberg et al. 2017; Hawkins and Kaltwasser 2017) and political polarization (e.g. McCoy et al. 2018).

6.3 The weight of normative motives

Outside the community of his followers, Donald Trump has been provoking anxiety and irritation by the things he says and does, but perhaps even more so, by the person he reveals himself to be through the things he says and does. For us, political scientists, such emotional reactions are puzzling. Why should we be anxious, why get irritated, by the personal characteristics an individual decision-maker displays? In our theories of politics, we commonly assume actors to be homogenous. We expect them to be driven by similar motives, view the world in similar ways, and behave in similar manner under similar conditions. By assuming that they are all equal in their basic desires and beliefs we can explain their decisions as products of objective circumstances, rather than personal traits (see e.g. Brennan 1996).

In the academic world of comparative politics, the analytic assumption of actor homogeneity is commonsensical. In the so-called real world of politics, by contrast, it is almost unintelligible. Observers as well as practitioners of politics tend to assume that few things matter more to political outcomes than the identity of individual decision-makers. To them, political actors are not fungible. To the contrary, they appear as unique carriers of political world views, moral commitments, intellectual capacities, emotional intelligence, and so forth (see also Schedler 2007, pp. 59–61). Only antipolitical populists and their followers who condemn “the political class” tout court as a cartel of self-serving “traitors of the people” share the idea of homogenous actors, at least at the level of established political elites.

Under the assumption of homogenous actors, the entire debate on Donald Trump would make no sense at all, as it is premised on the opposite idea of radical actor heterogeneity. Neither his followers nor his detractors treat Donald Trump as an actor “like anybody else”. His irruption into politics has brought back the notion that qualitative differences among political actors may serve as prime explanatory factors of political action.

Donald Trump’s distinctive personal traits have been subject to intense debate. With certain obsessiveness, we have been talking about cognitive or emotional questions like presidential attention spans, impulse control, and tolerance to criticism. With psychiatrists weighting into the debate, the 45th president of the United States has been challenging our social scientific standard assumption of actor rationality (in the elementary sense of mental sanity) (see e.g. Cruz and Buser 2016; Lee 2017). With respect to his relationship to democracy, though, our core concerns have been normative. Donald Trump’s ongoing “breaching experiment” has pushed us to rediscover an explanatory variable which we had largely abandoned in the comparative study of political regimes, namely, variations in political actors’ normative commitments to liberal-democratic values.

What does Donald Trump’s record of democratic norm transgressions reveal about the strength of his democratic commitments? In terms of Juan Linz’ classic tripartite distinction, it seems plausible to locate him in the intermediate category of a “semi-loyal” actor who would not work to destroy democracy (as “disloyal” actors would) nor rise to defend it either (as “loyal” actors would). He seems to lack a normative commitment to liberal-democratic norms as well as an ideological commitment to authoritarian alternatives. As a matter of fact, he seems to lack any discernible normative commitments. His explicit political maxim lies in the maximization of personal, partisan, and national self-interest. Amorality is his very program. Neither “loyal” friend nor “disloyal” enemy of democracy, he is, above all, a rational man. He is the personification of our homogenous, utility-maximizing, rational actor.

So, here we have a puber economicus, a faithful disciple of our standard model of economic rationality, occupying the Oval Office … and instead of watching events unfold with analytical detachment, we are alarmed! While countervailing forces have prevented him from transgressing his presidential powers, he has been deploying them in unrestrained, uncompromising fashion. Analytically, his lack of self-restraint, his open rejection of democratic fair play and civility, his ruthless pursuit of partisan advantage, seem unremarkable. They are simple displays of rational self-interest. Yet, comparative scholars have seen them as dangerous breaches of informal norms, as “a supercharged version of constitutional hardball” that fractures the very foundations of democratic coexistence.Footnote 33

Classic sociological breaching experiments slice the rational fabric of everyday interactions in order to render it visible. In these experiments, the irruption of irrationality unveils tacit, taken-for-granted structures of rationality, such as common knowledge or shared language. In the case of Donald Trump, by contrast, it is the irruption of rationality (in its narrow, instrumental sense) that unveils tacit, taken-for-granted assumptions about the causal relevance of normative commitments. His rational breaching experiment reveals hidden, normative sources of democratic equilibria. As it demonstrates, we do not fear for the future of democracy only when its ideological enemies conquer the halls of power. We also fear for its future when rational, utility-maximizing actors do so. Donald Trump’s radical self-centeredness has led us to discover a new form of democratic anxiety, a new form of horror vacui: the dread of moral emptiness.

7 Conclusion

Through his cheerful breaches of democratic norms, Donald Trump has made comparative political science veer towards the United States. At the same time, by triggering democratic anxieties within the comparative community over behavior and attitudes we tend to ignore in our research, he has unveiled huge swaths of terra incognita at the very center of our discipline. By subjecting us to a practical “breaching experiment”, he has drawn our attention to both causal and normative assumptions about democratic stability which we had kept out of our conscious disciplinary field of vision. In a certain manner, Donald Trump has pulled the rug under our theories of democratic equilibria. He has made us realize that in practice we do not trust three core assumptions that sustain our theories. We do not trust the structural foundations of democracy which we tend to focus on. We do not trust either the protective force we tend to ascribe to rational self-interest. Nor do we trust the methodological privilege we tend to grant to behavior over speech. Donald Trump, the macho economicus par excellence, has revealed the fragility of our confidence in probabilistic, structure-induced, interest-based democratic equilibria. And Donald Trump, the serial breaker of discursive norms, has unveiled the fragility of our indifference towards public speech. The disciplinary implications are almost revolutionary. They suggest a triple paradigmatic shift. If we wish to comprehend the disruptive potential of illiberal actors, we need a possibilistic turn in comparative politics. If we wish to comprehend the role of language in democratic equilibria, we need a linguistic turn in comparative politics. And if we wish to comprehend the role of normative commitments in democratic equilibria, we need a normative turn in comparative politics.