1 Introduction

Human fervor is two-sided. On the bright side, it operates as the great driving-force for human progress, well-being and prosperity, gain of knowledge and pursuit of happiness (Goklany 2007; Watson 2008; Harari 2015; Norberg 2017; Pinker 2018). On the dark side, it motivates humans to tyranny and despotism, dogmatism and hatred, ruthlessness and cruelty (White 2011; Pinker 2013; Harari 2015, 2019; Snyder 2017; Brogaard 2020). Human fervor may bring out the best as much as the worst in us.

In this paper, I use fervor as a synonym for zeal. Speaking of fervor, I mean a passionate enthusiasm. Entrenched in human nature, it can be seen as a constant rather than a variable in analyzing human society. Thanks to our enthusiasm we have learned how to control fire and think about ourselves. Our ambitions have led us to invent agriculture and domesticate animals, but also to build imposing temples and gorgeous cathedrals, dig tunnels through mountains, engineer tremendous bridges and construct skyscraping buildings. We have invented the wheel, steam engine and airplanes, antibiotics, computer, internet and social media. Our fervor encourages us to excel as individuals in order to heal deadly diseases, lessen poverty and eradicate hunger in large parts of the world, but also to enhance the lives of billions of people, fly to the stars and come together with a touch of a finger. Actually, we are working to become deities ourselves, by acquiring for us “divine powers of creation and destruction” and upgrading “Homo sapiens into Homo deus” (Harari 2017, p. 69). At the same time, our fervor provides us with the means to efficiently surveil a whole populace, effortlessly oppress fellow citizens, kill each other with lethal precision, destroy our planet and annihilate the human species once and for all. Therefore, we would be well advised to find a way of governing ourselves that prevents us from the gloomy part of human passions. Otherwise, we are doomed.

Liberal democracy has been invented to be exactly such a temperate mode of governance. While promising the sovereignty of the people, it does limit the absolute power of the people at the same time (Zakaria 1997, p. 30, 2003, p. 101 f; Plattner 1998, p. 174, 2004, p. 107; Mény 2003; Kaina 2009, p. 169; Galston 2018, p. 9–11; Urbinati 2019, p. 10). Liberal democracy seeks to restrain human passions by implementing institutions of checks and balances and codifying indefeasible and inalienable rights to protect the few from the fickle majorities. Additionally, representative procedures are designed to ensure “accountability of rulers to the ruled” (Cheibub and Przeworski 1999, p. 222) and allow for peaceful change of power (Popper 2015, p. 223, 242). The interplay of those institutions, procedures and rights aims for tempering political zeal and containing human ambition in exercising political rule. In this paper, I accordingly refer to a procedural definition of democracy (as opposed to a substantial one—Møller and Skaaning 2013, p. 5). By liberal democracy I mean a kind of representative governance that combines three core attributes: electoral rights, civil liberties, and the rule of law (Møller and Skaaning 2010, p. 270 f; see also Urbinati 2019, p. 10). Put differently, “liberal democracy (…) refers to popular rule conditioned by representation, fundamental civil liberties, and the rule of law” (Møller and Skaaning 2013, p. 37).Footnote 1 To function this way, liberal democracy is heavily dependent on cultural underpinnings, such as the respect for compromise and pluralism, the willingness to cooperate with strangers, tolerance of ambiguity and public spirit.

Recently, however, this type of government is showing signs of dysfunctions (e.g. Krastev 2017; Levitsky and Ziblatt 2018, 2021; Przeworski 2019; Runciman 2019; Schäfer and Zürn 2021). Facing multiple challenges, liberal democracy is in trouble to keep on working as intended. Human fervor, so my argument goes, generates outcomes that have detrimental impact on liberal democracy. Since human fervor is not new, the question arises: Why is our fervor causing harmful effects on liberal democracy right now?Footnote 2 For a long time, liberal democracy has been proven its unparalleled ability to cope with new challenges successfully. Meanwhile, however, our passionate enthusiasm brings about a slew of results which question the capability of liberal democracy to adapt without abandoning itself. While some of these results challenge the proper functioning of liberal democracy, others come with the risk of destroying its social and cultural prerequisites.

My paper addresses two sources that foster the rise of dysfunctional liberal democracy. First, I will deal with the bounded capability of liberal democracy to cope with “untidy” problems. Second, I will discuss several symptoms for overstrained institutions. In doing so, I will also show why liberal democracy is facing the rise of “tragic” institutions. To begin with, I will offer some considerations on the crucial question of how we can recognize a dysfunctional democracy when we see one.

2 How to grasp the idea of dysfunctional democracy—Some preliminary considerations

The words dysfunctionality and dysfunctional are widely used in human life. Physicians, for example, speak of dysfunctional organs, psychologists of dysfunctional families, geneticists of dysfunctional protein, software engineers of dysfunctional software. But what does it mean when political scientists speak of dysfunctional democracy? How can we grasp a dysfunctional democracy, both theoretically and empirically? And do we really need another democracy “with adjectives” (Collier and Levitsky 1997) to analyze what is going on in contemporary democracies? These are big questions and as yet there are no elaborated answers to them. This is partly caused by familiar challenges for research on democracy, such as defining and measuring democracy, differentiating “good” and “bad” democracies, studying transition to and decline of democracy. Again, we are faced with the intricacies of an “essentially contested concept” (Gallie 1956, p. 183–187) which allow for different meanings. Minimalist definitions of democracy compete with ambitious ones, alleged “Western” views of democracy with supposed “Non-Western” understandings of it, liberal democracy with other types and sub-types of democracy. The proliferation of democracies with adjectives, be it for analytical purpose or for assessing the quality of given democracies, makes things even more complicated. Thus, how might we make sense of dysfunctional democracy? I suppose, we need a kind of travelling concept that is able to cover different views of democracy without risking conceptual stretching (Sartori 1970). So, where to begin?

Despite the criticism of the structural-functional approach, the conceptual framework by Almond et al. (2000) seems to be a useful point of departure here. I see two important merits of their framework. First, they have given us conceptual tools for comparative political analyses in unlike places, in particular structure and function (Almond et al. 2000: 35–42). Second, they have conceptualized both structures and functions as variables. As for structures, Almond et al. (2000, p. 39) name political parties, interest groups, legislatures, executives, bureaucracies and courts that “are found in almost all modern political systems”. Furthermore, they relate these structures to several functions (Almond et al. 2020, p. 40–42): System functions comprise political socialization, political recruitment and political communication; process functions cover interest articulation, interest aggregation, policy making as well as policy implementation and adjudication; policy functions include regulation of behavior, extraction of resources and the distribution of benefits and services within the population.

However, in an autocracy like China those structures work in basically different ways compared to a democracy like Sweden (Almond et al. 2000, p. 39, 42–46). The same institutions—such as political parties, parliament, the president, bureaucracy or mass media—obviously do not serve the same end. In fact, across political regimes, the functioning of structures may and frequently does profoundly differ in purpose. The particular intended purpose originates from the established idea of political rule.

Against this backdrop, the conceptual framework by Almond et al. (2000) provides a great advantage for making sense of dysfunctional democracy. We may use the notion of dysfunctional democracy for liberal democracy just as well as democracies with adjectives. The same structures make it possible. In this sense, dysfunctional democracy is a travelling concept because we can apply it to varying types of democracy. Though, in democracies, structures have to function similarly because they have to serve the same intended purpose. This purpose is derived from the idea of democratic political rule. The less the functioning of existing structures aligns with the purpose of democratic political rule, the greater the probability that we see something other than a democracy. Surely, democracy is just one idea of political rule that is usually selectively implemented (Fuchs et al. 2002, p. 431). Therefore, there is not just one type of democracy but many. Nevertheless, implementing the idea of democracy requires institutions which are conducive to the purpose of democratic political rule (Dahl 1998).

One may object that we still have to solve a severe conceptual problem: There is no single view of that idea of governance. Various interpretations of democratic political rule entail numerous definitions and competing concepts of democracy. That is true. But every idea of political rule has to rest on some core principles and a few pivotal values. Otherwise, it is hardly feasible to put it into practice. This also holds for democratic political rule. The idea of democracy cannot be stretched endlessly. When it means everything it means nothing. While the idea of democratic political rule remains diffuse we are not able to design appropriate institutions and procedures to make democracy work. What is more, we lack normative standards to justify—or criticize—this kind of governance (Kaina 2009, p. 157). Last but not least, we could also face an epistemological problem (Kaina 2009, p. 164): once the notion of democracy does not align to its normative basics anymore we simply leave the conceptual terrain of democracy. In doing so, we risk misunderstanding reality by employing a false or at least inappropriate concept.

Therefore, to conceptualize dysfunctional democracy, we need not bother with all possible views and definitions of democracy. We just have to step back and specify the basics. Following Jan-Werner Müller (2021a), I propose to confine democracy’s basics to three core principles: liberty, equality and uncertainty.Footnote 3 The functioning of structures of given democratic political systems has to realize these principles as best as possible. As long as structures work as intended, we speak of functional democracy. In contrast, a democracy becomes dysfunctional once its structures do not longer work as intended and bring about detrimental effects to the core principles of democratic political rule. We can even use this notion of dysfunctional democracy when we choose to define other than the aforementioned core principles of democratic political rule. At any rate, we need to disclose the basic principles of democratic political rule before we can explore dysfunctions of given democracies.

These considerations raise another important question. How can we realize whether parts of structures in existing democracies no longer function as intended? I suggest, we can discover such dysfunctions by outcomes. In case several outcomes no longer serve the purpose of realizing the basics of democratic political rule properly, we face signs of dysfunctional democracy. Put differently, outcomes reveal whether the functioning of structures in an existing democracy still delivers what it should (see also Runciman 2019, p. 4).

In contemporary democracies this cannot be taken for granted anymore. The recent literature on democracy’s crises (e.g. Dahrendorf 2003; Mair 2013; Marsh 2014; Urbinati 2016; Grayling 2017; Mounk 2018; Ginsburg and Huq 2018; Levitsky and Ziblatt 2018; Runciman 2019; Castells 2019; Przeworski 2019; Schäfer and Zürn 2021), democratic backsliding (Bermeo 2016) and de-consolidation (Foa and Mounk 2016) as well as the rise of authoritarian populism (Norris and Inglehart 2019), to name only a few buzzwords for democracies in trouble, highlights prevalent concerns about the proper functioning of democracy. In the next two sections, I will discuss two sources for emerging dysfunctions in contemporary liberal democracies: first, the bounded capability of liberal democracy to cope with “untidy” problems and, second, the issue of overstressed institutions.

3 The problem of “untidy” problems

Many problems of contemporary liberal democracies are not new. Some of these troubles are even inherent features of democratic political rule (e.g. Schmidt 2008, 2010, p. 499–504). Other challenges are novel and we do not yet look through their impacts on the functioning and effectiveness of democratic governance. The capability of liberal democracies to cope with these challenges is not only demanding but also indeterminable. Today, we do not even know for sure, whether autocracies perform far better than liberal democracies when it comes to reacting successfully to these challenges. A great part of humanity’s current problems stems from human fervor and make many of them most confusing. Both the quality and the quantity of such tricky problems brings about perilous ramifications for democratic governance. On the one hand, those problems produce ill effects on the ability of contemporary liberal democracies to deal with novel challenges in a productive manner. On the other hand, such complicated problems abet the rise of dysfunctional democracy. The emergence of “untidy” problems originates from several sources of humans’ fervor. In this section, I will point out some of these sources.

Thanks to our enthusiasm as human beings, problems increasingly cross boundaries. Many problems pass territorial borders—above all those of sovereign nation-states. A very different kind of territorial limitation will be exceeded when problems transcend the threshold between physical and digital world (Gershenfeld 2017, p. 436). Other problems cross temporal boundaries since they produce legacies for unborn generations of human beings. And some problems break speed limits—such as rapid self-replications of misinformation, rumors and guesswork, but also escalation cascades which get out of control. The crucial point is that border-crossings intensify the problem at hand because of a shift of scales. A change of scales heightens risks, increases insecurity and is prone to produce system failures because it may disrupt the system’s tolerance of disturbance. Last but not least, there is a special kind of scale shifting as soon as problems cross a final frontier. Final borderlines designate a point of no return. Crossing such frontiers is irreversible since we cannot go back. Humanity can no longer resolve problems that passed a point of no return. We can only live with the consequences and try to confine the fallout. Problems such as global climate change, earth heating, sea contamination or replacing human by artificial intelligence are heading for such final frontiers.

Our fervor has been improving our technological skills to optimize, enhance and even transcend our humanness. Maybe, we will soon approximate another point of no return: the final frontier between humanism and transhumanism (von Becker 2015). We use technological means to repair, improve and upgrade physical, mental and social skills of human beings (Woyke 2010, p. 22). Cybernetic implants or neuronal interfaces between human brains and machines are no longer a sci-fi fantasy (Dolan 2017, p. 171)—they have coming into our reach. Neuroscience has been making breathtaking progress: “Optogenetics is a new field of biotechnology that let us transform brain activity into light and light into brain activity” (Keysers 2017, p. 401). Those techniques enable us to cause a state of mind at will. We could easily switch certain feelings on or off—affection, gladness and compassion, but also anxiety, rage and hatred. As a result, we would be able to manipulate the human mind, implanted with feelings one has never had. By means of CRSPR genome editing as well as decoding and reprogramming of DNA by bioinformatics we could finally defeat terrible diseases (Highfield 2017; Tooby 2017; Topol 2017), but also create perfectly designed humans (Dolan 2017; Pagel 2017). How can liberal democracy cope with the implications? Maybe, we will provoke new variations of apartheid. In this respect, transhumanism could turn out to be one of the world’s most dangerous ideas (Fukuyama 2004; Sorgner 2016). In the future, political equality may not only suffer from great socio-economic gaps alone, but also from disparities between humans and super humans. More than fifteen years ago already, Francis Fukuyama (2004, p. 42) warned in a paper for Foreign Policy: “If we start transforming ourselves into something superior, what rights will these enhanced creatures claim, and what rights will they possess when compared to those left behind?”. Our increasing technological skills entail both important political and pressing ethical problems regarding the proper functioning of liberal democracy. In particular, we have to find answers to the question of how technological progress makes an impact on liberty and equality.

Another challenge is our amour fou with data. An excessive “data love” (Simanowski 2014) in today’s societies spawns mountains of data (see also O’Neil 2016). We permanently expand our comfort zones and, by doing so, concurrently offer novel threats for our liberty to have a seat on our sofa. We even voluntarily pay for our “nanny techs” (Ramge 2018, p. 90) who impose their standards on us and patronize us with their “good” advices. Our liberty is no longer exposed to governments alone. It is also endangered by anonymous data farmers who boast of knowing more about me than I do.

One more pressing problem originating from humans’ increasing technological skills is ever-refined communication technology. Never before was it so easy to join other people almost everywhere on earth—and, at the same time, to isolate from each other, found new tribes, stoke tribal hostilities and raise battlements around “Extremistan” (Taleb 2011, p. 109). Never before was it so easy to spread knowledge—and, at the same time, to deluge the world with foolishness, triviality and triteness. Never before, was it so easy to find like-minded persons, maintain contact with them and overcome social isolation. But also never before was it so easy to establish anti-social enclaves of “uninhibited aggression and malignity” (Pörksen 2018, p. 19). Communication areas of such sort turn civilizing communication filters off (Pörksen 2018, p. 75) and corrode the reliability of social and civil norms. Public deliberation, democratic competition and peaceful change of power cannot properly function any longer when related social norms dissolve.

Yet, many problems are not only confusing because they cross borders, but also because they are intractable. Policy research also speaks of wicked problems, a notion that was introduced to the social sciences by Horst Rittel and Melvin Webber nearly five decades ago. According to Rittel and Webber (1973), there are several properties of wicked problems which make them very different from tame problems (see also Blackman et al. 2006, p. 70). For example, wicked problems are ill formulated. It is difficult to say what precisely the problem is. Wicked problems allow only for temporarily solutions which can be “good” or “bad”, “better or worse”, “satisfying” or “good enough” (Rittel and Webber 1973, p. 163). Furthermore, “every wicked problem is essentially unique” (Rittel and Webber 1973, p. 164), “solving one won’t help you with the others” (Rosen 2012, p. 203). Each wicked problem “can be considered to be a symptom of another problem” (Rittel and Webber 1973, p. 165) and “is interconnected to a lot of other problems; pulling them apart is almost impossible” (Rosen 2012, p. 203). Last but not least, wicked problems are “never definitely resolved. We just run out of patience, or time, or money. It’s not possible to understand the problem first, then solve it; rather, attempts to solve it reveal further dimensions of the problem” (Rosen 2012, p. 203). Phenomena such as world hunger, malnutrition and global warming are typical wicked problems (e.g. Durant and Legge 2006, p. 310). By now, some researches characterize global climate change even as a super wicked problem (Levin et al. 2012). Other examples for wicked problems are mass migration, the post-work society appearing on the horizon of Artificial Intelligence (AI) and digitalization or the “weaponization of everyday things” (Davies 2019, p. 18).

Border-crossing and wicked problems generate enormous hurdles for decision-makers’ ability to act. “Untidy” problems increase the complexity of systems. Therefore, democratic governance is facing a knowledge problem (Kaina 2002, p. 82). It is increasingly difficult for politicians to recognize, orientate oneself and chart a course for action. For this reason, “untidy” problems provoke a structural deficit of expertise in political realm. Sometimes, it is difficult for governing elites in contemporary liberal democracies to get access to reliable information. Regularly, however, the bigger problem is to pick relevant information from towering mountains of information garbage and waste of data. As a giant surge of information is hardly controllable any more, it becomes far more difficult for all of us to focus our attention. The more politicians have trouble to manage their attention, the more grievous is the impact on their capacity to act.

What is more, in liberal democracy citizens must show much patience for problem-solving. More and more citizens of contemporary liberal democracies, however, may less willing to do so (Dahrendorf 2003, p. 10 f). Social scientists have been discussing global consumer capitalism as one reason for citizens’ shrinking patience with politics (Bauman 2007, 2013; Flinders 2013; Needham 2003; Morozov 2013, p. 200–201). According to Catherine Needham (2003, p. 33) and Matthew Flinders (2013, p. 67, 83–84), “consumer democracy” cannot work since it erodes vital social prerequisites for democratic life: “Consumerism is a model that prioritizes the individual over the community, encourages passivity, downgrades public spaces, weakens accountability, and privatises citizenship” (Needham 2003, p. 33). The fundamental danger to liberal democracy is that consumerism turns citizens into selfish and resentful fellows “whose expectations of government can never be met” (Needham 2003, p. 33). As the sky is the only limit, all kinds of volatile wishes and political demands are free to go wild (Bauman 2013, p. 76). Accordingly, the “surplus of expectations” (Kielmansegg 2016, p. 13) among citizens is incessantly raising and gradually transmute to an unbridgeable “‘expectations gap’ between what is promised/expected and what can be delivered by politicians and democratic states” (Flinders 2013, p. 57). Furthermore, citizen consumers do not want to be bothered with concerns for the public good (Flinders 2013, p. 83; see also Needham 2003, p. 6). What follows, is a damage of the cultural pillars of democratic governance (e.g. Flinders 2013, p. 66 f; Needham 2003, p. 33; Morozov 2013, p. 200).

Yet, “untidy” problems do not only challenge the governments and legislatures of today’s liberal democracies but also unnerve many people. How can ordinary citizens make sense of what is going on when even experts struggle to do so? While the world around us is increasingly inscrutable and confusing, numerous people are feeling to lose control over their life and be exposed to nameless powers. They defend themselves against this enervating unease by using several strategies.

One strategy is “politics of crowd” (Davies 2019) in order to compensate individual impotence with “the power of numbers” (Bauman 2013, p. 179). Another strategy aims to the shelter of community (Bauman 2013, p. 179, 2017, p. 53). Accordingly, a “new wave of tribalism” (Bauman 2017, p. 52) does not only build on human desire for belonging but also promise individuals certainty, security and safety in an incomprehensible world (Bauman 2017, p. 79). Some people flee to “Fantasyland” (Andersen 2018) where they invent ludicrous conspiracy theories (Butter 2018; Hendricks and Vestergaard 2018). Others seek to regain control with a kind of egocentric dogmatism. Those people only believe their own truth, and they doubt everything they cannot personally verify (Castells 2019, p. 21). In this regard, they are alike the Pirahã people, a tribe of natives in the Amazon region. For the Pirahã, there is nothing real what they cannot understand by personal experience (Everett 2010). The regression of modern humans to archaic, pre-scientific mindsets brings about at least two casualties. The first one is people’s mind, the second one rational discourse. Both of them have also a hard time whenever emotions trump arguments. Individual attitudes solely based on impressions and feelings become decoupled from actualities (Noelle-Neumann 1987, p. 27 f). As a result, it can be claimed everything about the world and nothing can be acknowledged as a safe matter of fact. What follows, is the end of debate since there are only statements brought into position to other statements. Thereby, interpersonal communication is coming to be entrenched, and dialog is running to a dead end.

Fear is one of the most powerful human emotions (Castells 2019, p. 25). As we know so far, it is deep-rooted in human evolution since our brains were implanted with loss aversion (Kahneman 2012, p. 347). The intricacy of many problems today’s humanity is facing delivers unending food for our fears. But also fake news, strategic disinformation or the viral spread of conspiracy theories are ceaseless fuel for our anxieties (Jaster and Lanius 2019). We can hardly make this undone, because highly developed technologies facilitate an efficient, cheap and ubiquitous infrastructure for polluting our brains (Urner 2019). One consequence is a prevalent belief that “the world is going to hell” (Pinker 2018, p. 40). Another one is politics of fear that seeks to exploit people’s desire for protection (Castells 2019, p. 25).

As for the functioning of liberal democracy, these societal trends raise some important questions. First, can public opinion in liberal democracies withstand masses of toxic information without damage (Santos 2019, p. 123)? Second, how can we still find compromises when more and more people barricade themselves into opinion silos and heavily guarded sanctuaries for their thoughts, emotions and world views? Third, what is happening to the built-in correction mechanisms of democratic governance which have assured liberal democracies an impressive ability to adapt so far? Can we still learn from mistakes and failures when subjective realities and individual verities compete against each other so that everything is possible now but nothing sure anymore? How can we recognize mistakes at all when we do not have shared criteria for truth any longer? Fourth, what do pervasive attacks on truth imply for liberty and freedom? Only those who know the truth are able to make a free choice: for a medical treatment, against a job offer, for a political party (Moini 2020, p. 37). Fifth, can liberal democracy survive many people’s growing need for reducing intricacy, simplifying the world, making reality comprehensible and tidy, getting emotional wellbeing and feeling cognitive safety?

Finally, there is another strategy to fight the world’s complexity. One may long for unambiguousness (Bauer 2018) to retrieve a lost feeling of control. Thomas Bauer (2018, p. 87) believes that “strategies of disambiguation” are caused by a precarious tolerance to ambiguity. Humans are equipped with an instinctive distaste for ambitious, ambivalent and contradictory situations. We dislike circumstances which resist a clearly arranged either-or-dichotomy: “yes” and “no”, for example, “right” or “wrong”, “good” or “evil”. However, the world is not a fairytale. Growing up, we must learn that reality is ambivalent and contradictory—whether we like it or not. The better we bear with this, the greater our tolerance to ambiguity—and vice versa. People featuring low tolerance to ambiguity find incomprehensible, vague, undecidable, ambivalent, contradictory, shortly, “untidy” situations unbearable. They simply reject reality and, just like the children’s book figure Peter Pan (Barrie 2015), refuse to grow up. In 1929, Aldous Huxley (1994a, p. 66) called people’s protest against reality childish; and only three years later he (1994b, p. 99–100) sharply criticized the “cult of infantilism” which was arising at the beginning of the twentieth century. Thus, the emergence of Peter Pan-society, some observers have recently diagnosed (Pfaller 2018; Bolz 13,14,a, b; Kissler 2020; see also Gottschalk 2018), is far from being an invention of our time. The crucial question is: Does it matter for the proper functioning of liberal democracy? In the 1980s, Dan Kiley (1984) characterized the Peter Pan Syndrome (PPS) from a social-psychological perspective. Accordingly, individuals suffering from PPS have reached adult age, but never grown up. Among other attributes, they are characterized by irresponsibility, anxiety and narcissism. A Peter Pan-society would be a society of self-centered individuals who do not care for their fellow citizens, compromise and cooperation. There is some empirical evidence that eroding social norms, in particular among younger generations in the US, correlate with disdain for democracy (Howe 2017).

Indifference, however, is just one variant of refusing ambiguity; another one is fundamentalist zeal (Bauer 2018, p. 30, 83). Referring to Roberto Simanowski (2014), I see a third attempt to tame ambiguity. I want to call it ambiguity denial by machine programming. This strategy of refusing ambiguity will be risky inasmuch as the if-then logic of algorithms may increasingly determine societal processes so that reason will be reduced to formal logic (Simanowski 2014, p. 84). Debates become superfluous when we give our decisions—including our moral ones—up to the machine algorithm which was programmed afore. If-then-relations do not allow for “but” and “nonetheless”, for ambivalence, irony and skepticism; they are indifferent towards context and the delegation of decisions to a principle (Simanowski 2014, p. 84). By means of machine algorithms we are able to automate decisions without the need to decide ourselves (Simanowski 2014, p. 101). In doing so, we may drastically reduce complexity by using binary-coded decision rules such as 0/1 or yes/no. Those possibilities promise us enormous relief—and make us probably susceptible for this promise. However, we have to pay a price for this strategy of disambiguation: the dichotomization of society, the replacement of thinking by steps of calculation, the infantilization of decisions, the loss of collective memory (Simanowski 2014, p. 85, 100, 114, 132). For liberal democracy this might be too high a price. Liberal democracy is a type of governance which seeks to ensure liberty and equality by acknowledging pluralism, diversity and complexity. There is much to suggest that liberal democracies will show signs of dysfunctions when tolerance to ambiguity will further shrink within society. In line with this thought, Evgeny Morozov (2013, p. xiii) argued: “Imperfection, ambiguity, opacity, disorder, and the opportunity to err, to sin, to do the wrong thing: all of these are constitutive of human freedom, and any concentrated attempt to root them out will root out that freedom as well.”

However, liberal democracies also need to solve problems successfully. Without system effectiveness (Dahl 1994), that is, when they continue failing to improve the lives of their citizens (Przeworski 2019, p. xi), they can neither properly function nor survive. In this context, one important question arises: Do the institutionalized rules of decision-making in liberal democracies still function good enough to serve the greatest happiness of the greatest number of people? We cannot take this for granted. For some time, the core institutions of liberal democracy have shown signs of overburdening, and hitherto reliable rules of collective decision-making reveal indications of wearing.

4 From overstrained institutions to “tragic” institutions

Since the American Revolution about 250 years ago, liberal democracy has extraordinarily succeeded. By linking the ideas of democracy and representation, it became feasible to implement democracy in large-scale territories (Dahl 1998). For this purpose, humans invented certain institutions which help to solve recurrent problems of democratic governance in territorial states with large populations (Dahl 1989, 1998). Among the most important institutions of such sort are free, fair and regular elections, citizens’ freedom of expression and speech, freedom of assembly and freedom of information. Others are independent courts and a free press. For a long time, these institutions have served modern liberal democracies very well. Otherwise, the success of liberal democracy would not have been possible. Meanwhile, however, researchers on democracy see indications for dysfunctions of some core institutions of liberal democracy (e.g. Dahrendorf 2003; Mair 2013; Krastev 2014, 2019; Levitsky and Ziblatt 2018, 2021; Castells 2019; Runciman 2019; Przeworski 2019; Berman and Kudnani 2021; Müller 80,81,a, b; Schäfer and Zürn 2021).

On its own, dysfunctions in open systems are neither surprising nor necessarily concerning. In fact, shortcomings in the way institutions in liberal democracies work have been discussed and analyzed since this form of government was put into practice. Furthermore, dysfunctions are not inevitably hazardous. They may bring about prolific reforms and innovative adaptations of the institutional system. Compared to other kinds of political rule, liberal democracy has been successful so far precisely because of its remarkable ability to learn and adapt to challenges. Recently, however, some core institutions of liberal democracy are suffering from overburdening; and it might be, that they are not able to recover. In this case, we do not deal with institutional change but institutional decay. Several symptoms may typify overstressed institutions, and I will deal with two of them. The first symptom points to a paradox, the second one to a tragedy.

Beginning with the first symptom, we are possibly facing a baffling contradiction: Institutions do not work while they function as intended. In the words of David Runciman (2019, p. 4): “All [institutions] can continue to function as they ought while failing to deliver what they should”. If that is true, we may take this paradox as a sign of self-referential feedback loops in a closed system. Closed systems may be able to cope with problems they know already, but they are regularly unable to react to novel, still unknown problems in a productive way (Easton 1965). The more we find empirical evidence for this paradox, the more liberal democracy is at risk to lose its particular superiority to autocracies stemming from its ability to correct itself (see also Krastev 2019, p. 31).

Naturally, for a reliable diagnosis we have to look at the empirical underpinnings of our finding. In this paper, I cannot do so in detail. Rather, I am going to exemplify this paradoxical symptom of institutional overstress. The notion of institutions I use conforms to the common denominator of several variants of neo-institutionalism (Peters 2005). Accordingly, I will speak of institutions as (formal and informal) rules “that shape human interaction” (North 2002, p. 3) by both enabling and constraining individual behavior (Kaiser 2001, p. 255). Those rules tell us how we should act and, at the same time, how we should not, which mode of behavior is appropriate, and which one is not (Göhler 2011, p. 192).

Supposing we find empirical evidence for the aforementioned paradox—what exactly could that mean? It could mean: the rules still work as they ought by allowing some actions and reliably precluding other ones; yet, the outcome of the behavior becomes decoupled from the rule’s purpose—namely, to solve a problem in a certain way. Put differently, rules which do not serve its purpose or any purpose at all may still function as they ought, but they fail to deliver what they should (Runciman 2019, p. 4). As a result, there are rules which still constrain and allow behavior in some way, but it does not matter anymore. Those rules are inconsequential since the observable outcome may also come about without them. Regarding the problem that has to be resolved, those rules do no longer make a difference, and it does not matter anymore whether or not the rule will be applied. The self-referentiality of such institutions systematically undermines their legitimacy insofar as they become meaningless.

In today’s liberal democracies, political decision-making is occasionally showing signs of such dysfunctions. Usually, in liberal democracies, political decisions are frequently made by applying a kind of restricted majority rule (Sartori 1997, p. 33). In contemporary liberal democracies, however, we observe a strange phenomenon. The application of the majority rule generates “false” outcomes: although the majority rule is applied, political decisions are made by enthusiastic minorities instead of majorities. A typical and well known example refers to the effects of majority voting systems, in particular, in form of the first-past-the-post principle (FPTP or FFP). Strictly speaking, governments in democracies with majority voting systems are usually minority governments, even though the notion of “majority voting” suggests the opposite. A.C. Grayling (2017, p. 139) remarked, for example, that British governments are “standardly based on about 35% of the vote (and therefore an even smaller percentage of the total electorate)”. Manuel Castells (2019, p. 71) noted a similar outcome in the French legislative elections of 2017: The election result that gave Emmanuel Macron’s party La République en Marche “an overwhelming absolute majority comprised just 15.7 per cent of the electorate”. Another example is the Brexit referendum held in 2016. 52% of participants voted to leave the European Union. Yet, this narrow majority represents only 37% of the electorate and 26% of the population of the UK (Grayling 2017, p. 164, 194). In this case as well, the outcome of applying the majority rule became disconnected from the rule’s purpose. The aftermath of the Brexit referendum illustrates even more drastically how the outcome of rule application became detached from the purpose of the rule. In May 2019, the repercussions of the Brexit referendum forced Theresa May, then British Prime Minister and leader of the governing Conservative Party, to resign. Two months later, her successor as party leader and head of government was appointed by the members of the Conservative Party. Again, May’s successor was selected by applying the majority rule. However, only the members of the Conservative Party were enfranchised—about 160,000 persons. In other words, 0.35% of the UK electorate assigned Boris Johnson to be the present Head of British government (Schulte von Drach 2019). This portion is near to that of Ireland where 0.2% of the population was enfranchised—in 1830, though (Fukuyama 2015, p. 415).

The election of Donald Trump in 2016 is a further example for dysfunctional majority rules. Due to a peculiarity of the US voting system, the businessman, reality TV show star and nominee of the Republican Party became the 45th President of the United States of America. His competitor Hillary Clinton, candidate of the Democratic Party, won by a large margin of 2.9 million votes, “more than any defeated candidate in US history” (Runciman 2019, p. 14). Yet, while Clinton won the so called popular vote by more than 65.8 million votes, the majority of the Electoral College voted with 304 of 538 members for Trump and secured him the US Presidency (Federal Election Commission 2017, p. 5 f). Only 6% of US citizens “chose Donald Trump as the Republican nominee, and 28% of eligible voters made him president” (Müller 2021b, p. 187). Once more, the application of majority rule entailed a minority decision (see also Levitsky and Ziblatt 2021). Against the backdrop of an ever-increasing ideological polarization in the US society, it would hardly come as a surprise when such dysfunctions deepen political disruption and undermine the willingness of US citizens to accept the presidential election system.

Referring to the electoral victory of Donald Trump in 2016, there is another indication of how the outcome of action became detached from the rule’s purpose. The US founding fathers distrusted ambitious individuals striving for political power, and they distrusted the skills of “the people” to make wise decisions alike. For this reason, the implementation of the Electoral College fitted in the architecture of institutional checks and balances. Originally, the Electoral College was devised as a gatekeeper which contrast the electorate’s caprice with the reasonable, deliberated decision of a few well-informed and capable personalities “to stop unsuitable individuals reaching the Oval Office” (Grayling 2017, p. 112; see also Levitsky and Ziblatt 2018, p. 39). While the Electoral College worked in 2016 as it ought, it did not delivered what it should. It failed as a gatekeeper because of two shortcomings of the original design: first, the Electoral College plays no role in determining who seeks the presidency by popular vote; and, second, the US constitution never mentions political parties and therefore ignores the possibility that Electors became party loyalists (Levitsky and Ziblatt 2018, p. 40).

These examples indicate a symptom of overstressing when it comes to a key institution of liberal democracy. Still, the rule works inasmuch as it effectively influences behavior. Yet, the rule brings about outcomes which miss or event thwart the rule’s intended purpose. The rule does not fail in being an incentive for political actors to behave in a certain way. Rather, the usual modus operandi of the rule cannot prevent that the rule’s purpose become detached from the outcome.

The second symptom of overstressed institutions I want to sketch, is somewhat different from the first one. In this case, institutions are overstrained because the preconditions for its functioning break down. Thus, a rule cannot be saved, even though it is seen to be a proper one. Those institutions are “tragic” institutions since they are legitimate but useless. Technically, they are still suitable for solving a problem in certain way, but they are no longer able to do so since the functional preconditions have been destroyed. “Tragic” institutions do not simply malfunction for being ineffective or generating inappropriate behavior. “Tragic” institutions fail because the requirements for their applicability do not exist any longer.

This danger might lurk in the institutional heart of today’s liberal democracies. Democratic elections do not only assure a peaceful change of power, they are also to legitimize persons who are entitled to make generally binding decisions. What is more, voters decide who will keep the right to govern a further term of office and who will be punished by voting out. In order to function this way, at least two preconditions are needed. First, voters have to be able to address accountability of political decisions. Second, accountability has to be meaningful for citizens’ voting decision. Much is already said about the first precondition. Meanwhile, the second precondition is probably coming under pressure as well. Two observations may undermine this assumption.

First, the complexity of “untidy” problems entails a temporal displacement of political outcomes. The impact of many political decisions is increasingly shifting to the future. Often, the effects of bad political decisions become visible not before those who made them are retired. Thus, voters cannot hold politicians who have withdrawn from politics some time ago accountable by voting them out. Failed infrastructure projects such as the Berlin-Brandenburg airport are striking examples for this sort of suspended political accountability. But also the outcomes of a protectionist trade policy and unpredictable foreign policy of the Trump administration will probably become fully apparent for the US electorate not until several new administrations will have been in office. The same holds true for the repercussions of the Brexit referendum. As for voting decisions, accountability becomes meaningless that way. As a result, we might increasingly observe erratic election results (Krastev 2014, p. 24). Furthermore, citizens in liberal democracy may increasingly become frustrated with democracy stimulating rage and angry mass protests. Public eruptions of disgust from politics and politicians may transform liberal democracy into “democracy of feeling” (Davies 2019, p. 3) when the “politics of crowd” is increasingly seen as an alternative to government by elections (Krastev 2014, p. 31).

Second, accountability may also lose its significance for voting decisions by technological progress. Already today, the proliferation of machine algorithms is risky for liberal democracy since such tools are misused to manipulate public opinion. However, who can be hold accountable for decisions made by machines instead of humans? In several US states, for example, “robo-graders” are used already to score the essays of university students (Küchemann 2018). Also in the US, judges who made decisions on the early release of prisoners got support by AI (Ramge 2018, p. 26). However, they parted from doing so because the AI had learned human prejudices and made decisions biased by ethnic discrimination (Ramge 2018, p. 26). Nevertheless, what should stop overburdened legislators to trust the recommendations of algorithms when it comes to vote about bills and amendments? Yet, AI-systems will not be found on voting lists. Humans not machines will be elected to parliament and government office. Once in office, politicians need not feel responsible for their decisions and the results of their actions when the algorithm and not the human being creates and enforces the rules (Simanowski 2014, p. 51 f). Accordingly, the democratic rule of holding elected politicians accountable will collapse once machine algorithms destroyed the preconditions for applying this rule. Soon enough, it could also become relevant in just another way.

Donald Trump is surely neither the first politician nor the only one who manipulated the truth (Löwenthal 2017; Arendt 2017; Jaster and Lanius 2019), and fake news have not been invented in the twenty-first century (Hendricks and Vestergard 2018, p. 32). However, Trump made his systematic and unabashed lies a trademark. He was successful in lying because of three favorable factors. First, his permanent lies did not make him an outcast. Second, he took advantage from many people wishing to live in a world unmolested by facts (Arendt 2017, p. 33 f; see also Hedges 2009; Andersen 2018). Third, he had the greatest benefit of liars on his side—to fake things in a way that they appeal to his audience (Arendt 2017, p. 75).

One can see this unsettling devaluation of truth in public debate from an optimistic angle of view such as Steven Pinker does. He points to “the rise of a new ethic of fact-checking” (2018, p. 375) to counteract a swelling flood of lies, half-truth, misinformation and deception. At least two factors, however, attenuate Pinker’s optimism. First, a study by three MIT researchers (Vosoughi et al. 2018) found evidence that false news travel faster than true ones on Twitter. Second, it may be shocking that the president of the world’s oldest democracy misused his office to spread lies all over the world. What is far more disturbing, though, is that he was allowed to do so in a liberal democracy. Obviously, it did not matter for his voters. In fact, in the presidential election in 2020, Donald Trump received about 11 million votes more than in 2016. A phenomenon known as “blue lies” may explain this gain of votes for Trump—as well as the popularity of Boris Johnson despite his loose relationship with truth (Dittert 2021). Blue lies are lies in the name and for the benefit of an in-group (Pinker 2018, p. 358; Hendricks and Vestergaard 2018, p. 133). Accordingly, ideological polarization which makes political preferences an integral part of social identities does not only erode social and political trust but also the prerequisites for political accountability.

The same holds for information that has become weaponized. Herbert Lin, Stanford expert on cyber policy and security believes that cyber-enabled information warfare is an existential threat to the future of humanity in its own and a threat multiplier of climate change and nuclear weapons (Lin 2019, p. 3). And while we still worry about this prospect, deep fakes overtrump fakes by now (Ball 2019; Chesney and Citron 2019). Deep fakes are sophisticated manipulated video clips or audio recordings. Using high-developed AI-technology or voice cloning software, one is able to let real people saying words and doing things they never said or did. We do not need much fantasy to imagine what damage deep fakes may do in social and political life. Above all, we may give everybody an easy and irrefutable opportunity to deny the facts lying before our eyes (Ball 2019). Politicians will use this chance as well. Albeit they are confronted with video or audio recordings that may destroy their reputation and damage their career, they may gainsay the evidence. If everything might be fake, everything might be real (Ball 2019). So, why not believe what suits? Yet, how can we hold elected politicians accountable when we cannot hold on reality anymore?

5 Concluding remarks

Research on democracy provides a plethora of empirical evidence that liberal democracy is in deep trouble. My paper aims at showing that much of this trouble is caused by our intelligence, creativity, enthusiasm and passion—shortly, by human fervor. When it comes to the concept of dysfunctional democracy, however, it is still an open question whether it may enrich research on democracy. On the one hand, exploring dysfunctions in given democracies may help us to keep some perspective: Not every disruption in contemporary liberal democracies has to be a crisis, a symptom of backsliding, a sign of de-consolidation or a proof of regression. On the other hand, the notion of dysfunction may just pour old wine in new skins. In order to use the concept of dysfunctional democracy, we therefore need to clarify the theoretical relationship between dysfunctions and other symptoms of trouble in given democracies. In doing so, we have to find answers to important questions, for example: Do dysfunctions generate crises of some sort, or is the causal link just the other way around? Are processes of backsliding, de-consolidation and regression an effect of dysfunctions or just other descriptions for the same thing? Do dysfunctions worsen crises and force a decision about whether a given democracy is recovering or finally breaking down? Is dysfunctional democracy a harbinger of democratic decline? And if a democracy does not properly function any more, do we still have to do with a democracy at all? Maybe, dysfunctional democracy is just an oxymoron. In a nutshell, what we need is a theory on dysfunctional democracy.

A first step would be to clarify the concept of dysfunctional democracy. We might do so by contrasting it to a functional democracy. Adam Przeworski (2019, p. 8) recently argued that a democracy “works well when whatever the conflicts that arise in society are channeled into and processed through the institutional framework—most importantly elections, but also collective bargaining systems, courts, and public bureaucracies—without preventing anyone from gaining access to these institutions just because of the substance of their demands. To put it succinctly, democracy works when political conflicts are processed in liberty and civil peace”. The proposal I made in Sect. 2 of this paper is basically in line with his argument, but also exceeds it for three reasons. First, referring to Runciman (2019), we should not solely look at the usual modus operandi of the institutional framework because it may function as it ought without delivering what it should. Second, following Müller (2021a), I propose to define the core principles of democracy by adding equality and uncertainty to liberty. Third, Przeworski limits the functioning of democracy to political conflicts, in particular to processing them within the institutional framework. Referring to Almond et al. (2000), I refer to a variety of functions—namely, system functions, process functions and policy functions. There is certainly a risk in broadening the conceptual scope of a democracy’s functions: the more functions we define the more dysfunctions we may discover (likewise: Przeworski 2019, p. 4). Yet, if our list of functions is too short we may underestimate the degree of dysfunctions in given democracies. It is up to future research on democracy to find a response to those theoretical challenges and robust answers to the question of whether liberal democracy can survive human fervor.