Introduction

Hélène Landemore’s Open Democracy: Reinventing Popular Rule for the Twenty-First Century (2020) is one of the most widely discussed books in recent democratic theory. The book is unique in both its radicalness and ambitiousness. It is uniquely radical because it proposes nothing less than abolishing the familiar institutions of representative democracy, that is, parliaments, parties, and elections. And it is uniquely ambitious because it argues for replacing those traditional institutions with a variety of more ‘open’ institutions for collective deliberation and decision-making, most notably randomly selected citizens’ assemblies. In short, the institutions we inherited from the nineteenth century should give way to more inclusive and dynamic forms of collective self-rule that are fit for the twenty-first century; or so Landemore argues.

Obviously, there is much to admire about Landemore’s book. It provides innumerable refreshing intellectual impulses, encouraging us to imagine the possibilities inherent in a thoroughgoing overhaul of the democratic institutions we have all grown accustomed to. Ultimately, however, I remain sceptical about Landemore’s proposals. This is not because of a disdain for ambitious theoretical projects, but because her argumentation exhibits several, in my view quite critical, blind spots. In this short piece, I discuss what I take these blind spots to be. Importantly, the remarks that follow do not amount to a wholesale rejection of Landemore’s ideas—but they put pressure on some of the more central premises of her argument for an open democracy.

Building the Theory: What Problem is Open Democracy Responding to?

Let us start by looking at the heart of the case for open democracy. This is predicated on the idea that our representative democracies are failing empirically because of flaws that are built into the classic ideal of representative democracy, as we inherited it from the eighteenth and nineteenth centuries.Footnote 1 The problem, in other words, is that we have institutionalised the wrong ideal of democracy. What makes it wrong is that it:

  1. a)

    introduces, via elections, ‘systematic discriminatory effects in terms of who has access to power, specifically agenda-setting power’ (Landemore 2020, p. 26); and it

  2. b)

    brings with it ‘a type of party politics that is itself not all that conducive to deliberation or its prerequisite virtues, such as open-mindedness’ (Landemore 2020, p. 26).

According to Landemore, the implication of (a) is that the ideal that ‘all citizens are supposed to have equal opportunity of access to power’ (Landemore 2020, p. 36), which is central to the very idea of democracy, can never be sufficiently instantiated. The implication of (b) is that deliberation does not take place to a sufficient extent: as she puts it, ‘we know from empirical evidence that parties and partisanship do not go well with deliberation’ (Landemore 2020, p. 39).

The All-Too-Brief Treatment of Political Parties

It does, of course, not follow from this analysis that we should endorse any particular alternative conception of democracy. More argument is needed to get there, and Landemore proceeds by unpacking the different problems that plague the classic ideal of representative democracy. I want to focus in particular on her treatment of political parties—central representative-democratic institutions that are charged with fostering non-deliberative attitudes and communication. This way of portraying parties strikes me as unduly dismissive, and so Landemore’s tentative conclusion that we might be better off with a ‘no-party democracy’ (Landemore 2020, pp. 146–149) appears rash.

One problem with Landemore’s treatment of parties is that her exclusive focus on deliberation leads her to overlook that parties have a number of other democratically relevant functions that are not merely or primarily deliberative, and that other institutions most likely cannot perform as effectively. For one thing, parties construct and signify the primary conflict lines in society, the main political alternatives that citizens can choose between. This is an enormously important task, not least because, in so doing, parties also show that conflicts are contingent and dynamic, rather than natural and static (this point has been most powerfully made by Claude Lefort; for a recent discussion of this and the link to parties, see Geenens 2020). For another thing, the mobilisation function of parties that requires the creation of partisan identities and mindsets is actually quite essential to get people to care about politics in the first place, for it ‘links the distant and formidable, often inscrutable institutions of the state with the passions, interests, prepossessions, and beliefs of ordinary citizens’ (Muirhead 2019, p. 85). One can accept this point without assuming that all existing parties perform this function equally well, or that their political goals are admirable.Footnote 2

A further counter-argument against Landemore’s treatment of parties is that it is doubtful that parties, conventionally conceived, are always and necessarily harmful to deliberation, as Landemore suggests. Parties are certainly harmful to deliberation if one adopts what I would call an ‘apolitical’ conception of deliberation, according to which deliberation is essentially a dispassionate, quasi-scholarly form of rational discussion (which most people, scholars included, are in any case unable to have about topics they genuinely care about). If, on the other hand, one conceptualises deliberation in a more inclusive and realistic fashion—which is how most deliberative democrats today think about deliberation (see, most recently, Habermas 2022)—then it is far from clear whether parties and partisanship undermine deliberation.Footnote 3

In fact, employing more inclusive conceptions of deliberation, the important theoretical work by Bonotti (2018), Muirhead (2014), Rosenblum (2008) and White and Ypi (2016) has shown that there are good reasons to think of parties as potential enablers and supporting structures of deliberation. This is not only because parties, as noted, highlight what alternative political positions there exist in society, and thus give structure to public debate. Their internal organisation can also provide unique possibilities for political deliberation. This has already been intimated by one of the doyens of deliberative democratic theory, Joshua Cohen (1989), and Wolkenstein (2019) has shown that there is ample empirical evidence in support of this claim, arguing moreover that deliberative spaces within parties may fruitfully be tapped as resources for reinvigorating the party–citizen link and, by extension, representative democracy (also see Ebeling and Wolkenstein 2018).

It is worth adding to this that the strategies for party reform that have been proposed in recent theoretical scholarship are animated by normative concerns that chime with Landemore’s own ambition to increase political institutions’ ‘accessibility or, in other words, openness to the ordinary person’ (Landemore 2020, p. 2). Scholars such as Biale and Ottonelli (2019), Invernizzi-Accetti and Wolkenstein (2017), White and Ypi (2016, chap. 10) and Wolkenstein (2019) all argue for making party organisations more porous to the inputs of ordinary party members (and possibly to an even wider public of unaffiliated citizens), while emphasising the importance of ‘open and visible mechanisms of internal deliberation’ (Biale and Ottonelli 2019, p. 517). To be sure, none of these authors proposes to de-institutionalise parties and turn them into the sort of ‘transitory, fluid, and self-reconfiguring’ associations that Landemore (2020, p. 147) at one point presents as an attractive alternative to parties.Footnote 4 But arguably they are animated by making political parties more ‘open’ and deliberative so as to reconnect them with society. Landemore does not even mention this in passing, though clearly at least some of these novel proposals for reforming parties may well be compatible with her own suggestions—rather than leading to the stark conclusion that a ‘no-party democracy’ might be the way forward (Landemore 2020, pp. 146–149).

Empirical Variations in Democratic performance—and Citizens’ Attitudes

The first part of the case for ‘open democracy’ (which, again, is meant to prepare the ground for Landemore’s own normative proposals) raises yet another quite fundamental question: If representative democracy is failing because of shortcomings that are built into the classic ideal of representative democracy, why are there such enormous differences in how well real-world representative democracies function—and, correspondingly, differences in how satisfied citizens are with ‘their’ democracies? It is uncontroversial that some representative democracies (think, e.g., of some of the Nordic countries) function better in terms of equal representation, policy outputs, etc. than others (e.g., the US). But if the problem is the very ideal of representative democracy—if, that is, this conception of democracy is indeed not ‘fit for purpose’ in the twenty-first century—should not democracies be failing universally?

It remains somewhat unclear to me how Landemore would respond to this question, not least because the book never looks more closely at the different ways in which representative democracies can fail (or, indeed, succeed) empirically, shifting the conversation instead to the level of abstract ideals. One possible response would be to bite the bullet and claim that the crisis of representative democracy is indeed universal. This would be most consistent with Landemore’s proposition that representative democracies are failing because representative democracy is, at bottom, an inappropriate democratic ideal. But this response calls for an explanation as to (i) what exactly is dysfunctional about those representative democracies that seem to work comparatively well (e.g., according to leading democracy indices such as V-Dem), and, even more challengingly, (ii) what those citizens who are relatively satisfied with their democracies (to cite just some examples, 72% of Swedish citizens and 68% of Dutch citizens are satisfied with democracy, see Pew Research Center 2020, p. 23) misperceive or fail to understand about the political system they live under. It is difficult to see how this could be done in a convincing fashion.

To fully understand why it might be hard for Landemore to offer good explanations for (i) and (ii), consider that she would first have to argue that those democracies that still perform well do so despite being instantiations of a fallacious ideal of democracy that should be rejected and overcome. Note that the challenge is not to show that well-performing representative democracies may be aberrations in an age of broader democratic decline, but to demonstrate that well-performing representative democracies perform well for reasons that are more or less unrelated to their institutional setup (i.e., being centred on parliaments, parties and elections). This would be a highly implausible argument to make, and one that underlines the limits of Landemore’s principle-driven apriorism. Second, Landemore would also have to argue that any citizen who is satisfied with how her national democratic system performs is simply wrong about her national democratic system. Setting to one side the question of whether or not this might be a rather undemocratic argument to make, it clearly jars with Landemore’s epistemic credo of ‘diversity trumps ability’ (see Landemore 2012), in that it implies that, when it comes to the evaluation of democratic institutions, the theorist’s verdict concerning the correct democratic ideal(s) trumps the verdict of (potentially very many) citizens. This strikes me as self-defeating.

‘Open Democracy’ as an Alternative Democratic Ideal

I turn now to Landemore’s own democratic ideal, namely, ‘open democracy’. This is not a clearly-defined model of democracy, but a broader, principled approach to thinking about democracy that is predicated on a commitment to more inclusive and unconventional forms of political participation, beyond elections and traditional partisan activism. Throughout her book, Landemore tends to describe ‘open-democratic’ institutional designs in fairly broad and general terms, so while one gets a good sense of what kinds of institutions she considers central—most importantly, elections by lot are introduced as the key institutional innovation for facilitating more inclusive participation—it sometimes remains unclear how exactly an open democracy might look in practice. Nor does she define a ‘cut off point’ at which a traditional representative democracy with elections, parties and parliaments that is enhanced with, say, empowered randomly selected assemblies, would cease to be a traditional representative democracy and become an open democracy proper.

Unfortunately, the Icelandic case that is continuously presented as something like a ‘pure’ case of an open-democratic design (see Landemore 2020, chap. 7) does little to illuminate these issues, since it constitutes a very particular case of ‘higher law-making’ after a momentous national crisis. Because of these particularities, the case is difficult to treat as a blueprint for how ‘ordinary’ law-making should be conducted in a more ‘open’ fashion: drafting a constitution is, after all, very different from the day-to-day business of devising policies and regulations.

The Icelandic Example: Higher vs. Ordinary Law-Making

This gives rise to the following problem: While Landemore’s discussion of the Icelandic constitution-writing process shows that applying the principles of open democracy could make higher law-making tasks such as drafting a constitution more inclusive, she does not show, in my view, that the much more complex and specialised tasks of ordinary law-making could successfully be made more ‘open’ in the same way. And I would argue that Landemore also cannot show that on the basis of the evidence she marshals, simply because it is difficult to see what exactly the Icelandic example could teach us about the potential of making ordinary law-making more ‘open’ through crowdsourcing, online participation, randomly selected citizens’ assemblies, etc.

Consider first that the task that citizens were given in the Icelandic constitution writing process—the task of writing a constitution—is demandingly complex, yet it is also a task that involves imagining a better society and laying out basic rights as well as specifying their interpretation and the institutions that should enable their exercise. As such, it requires the formulation of general rules and broadly applicable normative provisions, which is something that most citizens will likely be able to do if given time to reflect. Ordinary law-making, in contrast, is usually about solving very specific policy problems that more often than not are difficult to understand for any non-specialist (think of the myriad complex question to do with climate change and sustainability) and/or exceedingly boring (think of the regulation of fishing nets or bathtubs). Ordinary law-making is indeed quite technical. This is why it is usually left to policy-makers and experts in economics, sociology, law, political science and the natural sciences.Footnote 5

Reflecting for a moment on these more general differences between higher and ordinary law-making suffices to see that Landemore’s arguably persuasive demonstration that higher law-making can successfully be made more ‘open’ does not neatly translate into a convincing argument that open procedures of deliberation and decision-making like those that were adopted in the Icelandic constitution-writing process could fruitfully be used in day-to-day politics. The tasks citizens would have to master are simply very different. Besides, I believe that there is also a real risk that ordinary law-making in a system of open democracy would end up looking much the same as it did under ‘old-style’ representative party democracy: because policy-making requires specialised expertise that would take any non-expert citizens a very long time to acquire, open democracy might unintentionally give policy-makers and experts disproportionate power in the process of deliberation and decision–making—at least if policy-makers and experts are meant to assist citizens along the way, as Landemore (2020, p. 192) suggests. To my mind, this would seem to defeat the purpose of an open democracy.

Implementation—‘From Here to There’

Will all that has been said above in mind, I was particularly interested in the questions that are addressed towards the end of the book, under the heading of ‘from there to there’ (Landemore 2020, pp. 206–208). This section is devoted to the quite important issue of how realistic the prospect of establishing an open democracy is, understood in terms of feasibility. Landemore opens by arguing that ‘the feasibility and attractiveness of an open, possibly non-electoral democracy will vary depending on the point of departure or the kind of system we are trying to democratise’ (Landemore 2020, p. 206). What should we make of this?

Notice first that feasibility and attractiveness are treated here as separate categories, as if the likelihood of realising an ‘open democracy’ were independent of the extent to which citizens and their political representatives consider open democracy an attractive ideal. If this really is what Landemore wants to say, I wonder how she envisages the implementation of her proposals. Is open democracy most feasible under non-democratic circumstances, where an enlightened elite can implement a new democratic system in a top-down fashion? One hint that this is indeed how Landemore thinks of a smooth transition from the status quo to an open democracy is her startling claim that a ‘democracy of mini-publics’ might be ‘much easier to evolve’ in a ‘one-party dictatorship’ such as China (Landemore 2020, p. 207). It barely needs further explanation that this is not an especially democratic approach to improving democracy.

The Least Promising Conditions for Realising an Open Democracy?

It is perhaps even more interesting that Landemore argues that ‘an existing electoral democracy … is probably the least plausible testbed for a non-electoral model of open democracy, considering the transition costs and the unknown implications of replacing familiar elected assemblies with, say, mini-publics’ (Landemore 2020, p. 207, my emphasis). This is no small concession. What Landemore seems to be saying is that the sort of ‘open’ democratic design she argues for (at least its purest form, with randomly selected empowered citizens’ assemblies, the virtues of which are praised throughout the book) is most likely infeasible in the established representative democracies that she believes to be based on a fundamentally misguided ideal and in need of change. If, as I would argue, a normative proposal whose realisation is highly unlikely due to enormous transition costs can hardly be considered desirable (on this, see Gilabert and Lawford-Smith 2012, esp. p. 817), then it is very tempting to conclude that open democracy is not an especially desirable ideal.

Here are two further challenges facing the transition from an established representative party democracy to a Landemorean ‘open democracy’ that are not addressed. First, a transition of this sort will in most countries require far-reaching constitutional changes, probably even a fully fledged ‘re-founding’ of the state. After all, the core representative institutions that Landemore considers thus problematic—elections, parliaments and parties—are in most established democracies defined as central components of the constitutional order. Any radical departure from this will require some sort of ‘constitutional moment’, that is, an event or process in which fundamental changes to the status quo become possible, and can be made in a democratically legitimate fashion. Such moments occur only rarely, however. Deep, existential crises can trigger them, as was the case in Iceland. Planning and orchestrating constitutional moments, on the other hand, is enormously difficult, for such moments require a symbiotic agreement between a large number of citizens and their political representatives concerning the urgent necessity to re-invent the state or even the nation. Under conditions of strong political polarisation (where the proposals put forward by one side are always looked to with scepticism and distrust by the other side), or when citizens simply consider the present constitution an essential, collective identity-shaping source of valid norms (see Müller 2007 on ‘constitutional patriotism’), generating such a broad-based agreement will be close to impossible—even if citizens are dissatisfied with their political system.

Second, an entirely different concern has to do with the extent to which a transition from representative to open democracy (assuming for a moment that such a transition would indeed be feasible) would reliably solve the problems that Landemore associates with traditional representative democracy, notably the assumed lack of good quality deliberation. For the introduction of an open-democratic system would hardly ‘set the clock to zero’ and more or less immediately change the way in which citizens think about politics, cleavages and conflicts. ‘Sedimented’ views that have been shaped by decades of partisan struggle can hardly be overcome overnight and make space for the sort of rational deliberation about common ends that Landemore would like to see. More plausibly, collective deliberation and decision-making in an open democracy would be heavily affected by the conflicts and grievances that already played a central role in the political system that preceded it. Think, for example, of the US, where, as Disch (2021, pp. 14–15) observes in her recent book, political representatives have inculcated in citizens the notion that polarisation is deeply rooted and virtually impossible to overcome. One need scarcely add that it is also doubtful that an open democracy could offset the polarising effect of a partisan media and citizens’ necessarily selective exposure to information when they are not actively participating in deliberation or decision-making. Which returns us to the larger question of whether open democracy is a desirable normative ideal to begin with.