“What if the democracy we believe we serve no longer exists? And the Republic has become the very evil we have been fighting to destroy?” That’s what the beautiful Padmé, former queen and now senator in the galactic senate, says to young Anakin Skywalker in Star Wars: Revenge of the Sith (George Lucas. USA, 2005). Indeed, the political situation has become quite confusing in this film. War is raging all over space, Jedi knights are fighting clone warriors, and the Chancellor of the Galactic Republic wants nothing more than to break up that republic so that he can install himself as sole ruler. Toward the end of the film, the Chancellor proclaims the end of the Republic before the great assembly in the House of Representatives. In a world that has become confusing, he argues, only a fascist system can ensure order. So he immediately proclaims the “first Galactic Empire.” All the democratically elected representatives applaud enthusiastically. “So this is how liberty dies, with thunderous applause,” Padmé comments sarcastically.

The theme of a destroyed or dysfunctional democracy features in many sci-fi films, reflecting the secret fear many citizens have of the failure of our democratic system. These fears have particularly increased in our time. One idea to solve this problem is the so-called liquid democracy.

In a short, simply animated info videoFootnote 1 on YouTube, the idea is explained nicely in less than 4 min. “Direct democracy involves every citizen voting on every issue. Indirect democracy has designated representatives whose jobs it is to be aware of current events and use this information to make a well-informed decision on our behalf. [...] There are issues with both of these models. In a direct democracy, all citizens are not informed on all the issues and many citizens do not have time to debate and make these tough decisions on a day-to-day basis. Indirect democracies are accused of creating a disconnect between the citizens and the policies that they must abide by. [...] Liquid democracy is truly a blend of the two. You can choose to vote as an informed citizen or you can choose to delegate your vote. It is this fluid alternation that gives its name to liquid democracy. But is this even possible? Yes, it is. With revolutionary technologies. [....] This is a choice that every citizen should have. Don’t you agree?”

Indeed, there is widespread unease about the perceived insufficient participation of the citizens. Left-wing and right-wing populist movements have mobilized the masses with the idea that aloof elites rule the country while the opinions of ordinary citizens no longer carry any weight—sometimes quite successfully, as in the Brexit campaign or Donald Trump’s presidential campaign.

This discomfort has a long history and an important witness, namely the Enlightenment philosopher and pioneer of the French Revolution Jean-Jacques Rousseau. For Rousseau, it is of central importance that citizens consult together in assemblies in order to (ideally) cast a unanimous vote at the end. He calls the jointly worked out goals the volonté générale, the general will. Only in this way, Rousseau believes, can the original freedom of the individual be restored. Rousseau imagined this original freedom as that of a man of nature who lives his life on his own, without being oppressed by others or having to take others into consideration. In modern civilization, however, we are dependent on cooperation and community, and therefore the original freedom to determine for oneself the rules by which one lives can only be restored by unanimously adopting, in the community of citoyens, the citizens of a republic, rules that apply equally to all. Any association of private interests in the form of interest groups, lobbying or parties would, according to Rousseau, be pernicious because it would threaten the moral character of the republic, because the willingness to ascertain and follow the common will requires that one distance oneself from one’s private interests and see them merely as one of many determinants of the common interest.

The objections to this republican understanding of democracy, the Rousseauian utopia of the volonté générale, are obvious. How are millions of citizens to come together at an assembly? How can groups with common interests be prevented from joining together and forming factions, parties, lobby groups? Above all, how can it be ensured that every voice, every opinion is given equal consideration in determining the common will? In mass democracy, how can it be achieved that complex issues to which legislation has to respond can be transmitted to all those involved in the opinion-forming process? Under the traditional conditions of democracy, as they existed until recently, the Rousseauian idea of the republic, or more generally: the idea of direct, immediate democracy, of a political opinion-forming process that includes everyone, is not feasible.

The potential of digital communication, and especially internet communication, as well as the use of complex software to control opinion-forming and decision-making, such as liquid democracy, however have made these objections to a republican form of democracy obsolete.Footnote 2 At first glance, digital platforms seem to be an excellent way to elicit such a common will and to strengthen democracy by expanding opportunities for participation. It is therefore understandable that the old republican ideals are currently experiencing a renaissance and are putting pressure on the cumbersome, multi-level, institutionalized procedures of political decision-making in parliamentary, representative, constitutional democracy. From the left, this pressure is being exerted in the form of initiatives, petition platforms, social media groups, i.e. in the form of a new civic engagement, from the right in the form of the emotionalization of political opinion-forming and the devaluation of scientific expertise. These characterizations, however, oversimplify the situation. There is also populism on the left of the political spectrum that takes no notice of scientific arguments and relies on emotionalization. And of course, there are also defenders of scientific rationality on the right of the political spectrum.

There is much to be said for using the new technological possibilities of digitalization and the internet to make democratic opinion-forming more inclusive and substantive.Footnote 3 More inclusive by involving all sections of the population, and more substantial, in that the easy availability of information through digitalization is used.Footnote 4 Unfortunately, the utopia of the digital republicFootnote 5 will never be realized in this way. Three theorems show why.

First, Condorcet’s “problem of cyclical majorities” (also known as the “Condorcet paradox”), then Kenneth Arrow’s “impossibility theorem,” and finally Allan Gibbard and Mark Satterthwaite’s “manipulation or strategic voting theorem.” Although all these theorems were proven some time ago (Arrow 1951; Gibbard/Satterthwaite 1973; Condorcet as early as 1785), they have not penetrated the general consciousness beyond specialist circles. This even applies to the relevant discipline, political science. This is probably mainly due to the fact that the theoretical consequences of these theorems for political practice have not been sufficiently clarified.

Condorcet’s paradox can be described as follows: A, B, and C stand for three factions in parliament. None of them can form a majority on their own. Only two factions together can form a majority. A vote takes place that obeys the simple, binary majority rule, i.e., an alternative is elected if and only if it gathers more votes than other alternatives. Let us assume that three alternatives X, Y, Z are to be voted on (e.g., candidates for the chancellorship or bills). If one now votes in alphabetical order, i.e., first X runs against Y, X wins in a coalition of A and B. Then there remains the vote between X and Z, this time Z wins in a coalition of B and C. Here, normally, the voting would be over. However, a test vote shows that Y would have prevailed over Z. This violates the so-called transitivity condition. This condition requires that whenever an alternative X is preferred to Y and at the same time an alternative Y is preferred to Z, X is also preferred to Z. The order of voting alone is decisive here. This means that an assembly leader can cause the alternative he prefers to be voted for simply by choosing a certain voting order. This phenomenon is called “susceptibility to manipulation” (please see Table 17.1).

Table 17.1 The Condorcet paradox
Full size table

Now, one might assume that such cyclical voting results are extremely rare. But in fact, the probability of cyclical voting results increases with the number of people involved in the decision and the number of alternatives. Quite a blow to the ideal of the democratic voting process! But it gets worse. Around 170 years after Condorcet, the US economist Kenneth Arrow proves that it is not possible to fulfil four indispensable conditions of collective and democratic rationality in a process of collective decision-making.

The first condition he calls “D” for non-dictatorship. This means that there must not be a single person who determines the outcome of collective decisions regardless of the preferences of others.

The second condition is what Arrow calls “P” for Pareto efficiency. This postulate, which goes back to the Italian scientist Vilfredo Pareto, requires that common preferences of all are realized, which in application to collective decisions means that the unanimity principle applies: if all those involved in the decision prefer X over Y, then this should be reflected in the outcome of the vote. For example, it cannot be the case that in a committee where all persons consider a candidate X to be better than a candidate Y, in the end, Y wins the vote.

The third condition is what Arrow calls “I” for irrelevance. It requires that the collective preference for, say, X over Y does not change simply because an alternative Z is added. Of course, the added alternative may be better than X, but the ranking between X and Y should not change by the mere addition of another alternative. If I prefer to go on holiday to Italy instead of France, then this preference should not change simply because a holiday in the USA has become possible as an additional alternative. Why should I prefer a holiday in France to a holiday in Italy simply because I could now also spend a holiday in the USA?

Arrow calls the last condition “U” for unrestricted domain. The better term is “preference sovereignty”: all participants in the vote can feed in their preferences, there are in a sense no prohibitions or restrictions on having certain preferences.

It seems obvious that these four conditions are far too weak to characterize democratic decision-making procedures. One would want, for example, some form of assurance that majorities matter, that anonymity and neutrality are assured, perhaps minority protection and individual rights. The shocking thing, however, is that Arrow has shown that these four minimum conditions for collective decision-making procedures cannot be fulfilled simultaneously, i.e. that there is no rule of collective decision-making that fulfils these four conditions at the same time.

Now, one could draw the general conclusion that democracy is impossible. However, that would be premature. A closer look at parliamentary, representative, constitutional democracy shows that its procedures of political decision-making are designed in such a way that they usually circumvent the problems shown in Arrow’s theorem. For example, cyclical voting results, as in the Condorcet paradox presented above, can only occur if there are at least three alternatives. Fortunately, the practice in parliamentary democracy of basing the respective government on a parliamentary majority formed by one or more parliamentary groups bound together by a coalition agreement precludes the possibility of several proposals, each of which has a chance of being endorsed by a majority.

Finally, the findings of Gibbard and Satterthwaite should be mentioned. They were able to show that there is no single process of collective decision-making that is not susceptible to strategic voting and manipulation.

A decision-making process is susceptible to strategic voting if at least one person involved is more likely to realize their preferences if they feed other preferences into the decision-making process than they actually have.

The devastating result of Gibbard and Satterthwaite’s theorem is that all processes of collective decision-making are susceptible to both strategy and manipulation. Fortunately, there is a safeguard here too in parliamentary, representative, constitutional democracies, and this lies in the role played by the publicly presented argument and the transparency of politicians’ decision-making behavior.

If a politician puts forward certain arguments in favor of a project, then it would at least require justification if she votes against it in the end. One could also say that the obligation to give reasons and the public formation of political opinion limit the scope for strategic and manipulative behavior in democracy. In this sense, the institutional order of parliamentary democracy can be interpreted as an attempt to circumvent the paradoxes and dilemmas of collective rationality and make political decisions possible. If one were to move to a digital republic, this feature of a “deliberative democracy” would disappear. The control of who voted when, how, and with which arguments in favor of which project would be impossible given the sheer number of participants, and political responsibility would diffuse in an amorphous mass of thousands and thousands of participants who merely vote “yes” or “no” by mouse click.

So, as beautiful and simple as Liquid Democracy presents itself in the YouTube video, it must fail in this form.

In fact, large-scale attempts at liquid democracy have so far proved unfeasible—mostly due to lack of participation.Footnote 6 To read this merely as an expression of saturation, disinterest, or convenience would be wrong. It is more likely that the resistance to the transition to a digital republic is fed by the deeper insight that it inevitably entails a loss of rationality and that—as we have seen—in the worst-case collective self-blockades in the form of cyclical preferences lead to serious chaos and instability.

And yet, objecting to the idea of a digital republic, demonstrating that it is not feasible for systematic reasons and that moving toward it is not desirable, is not incompatible with advocating massive use of the new technological possibilities. There is no contradiction. The goal of digital humanism is to strengthen power of judgment and decision-making and thus individual and collective autonomy. To achieve this, digital information and decision-making technologies are to be used as a supplement to parliamentary, representative democracies based on the rule of law—but they are merely a support, not a substitute.

In this sense, the enrichment of public space through the involvement of as many citizens as possible would not replace representative liberal democracy but strengthen it. The opportunities for this are more favorable today than ever before.