Representing and Being Represented in Turn’ – A Symposium on Hélène Landemore’s Open Democracy

SYMPOSIUM `


Jane Mansbridge
Harvard University, US jane_mansbridge@hks.harvard.edu We have never needed this book more than now. Democratic legitimacy is declining in most countries across the globe. Yet we need increasingly more government regulation to handle our growing interdependence. So activists, political theorists, and political scientists are applying brainpower to experience in order to devise ways of making that regulation-and the government behind it-more legitimate.
Open Democracy has a more radical agenda than simply shoring up today's representative governments. Landemore wants to 'reinvent popular rule for the twentyfirst century' (xviii) by placing ordinary citizens at the center, in contrast to the ' extraordinary' individuals elected to office (34,80). She reconceptualizes the principle of democratic representation by decoupling it from elections, making ample room in democratic theory for non-electoral forms of democratic representation. Responding to the alienation experienced by so many citizens in today's democracies, the book then explores analytically the most promising democratic innovations in the last few decades, explaining why these measures satisfy democratic needs in ways that electoral representation cannot do.
To frame the achievements of Open Democracy and both the criticisms and the appreciations in this symposium, this brief introduction will look first at the larger macro system in which particular innovations are embedded and then at the microstructures of the innovations themselves.
On the macro side, I use the word system not to imply systematic, with all the rigidity that term suggests, but rather to suggest a larger entity internally connected in such a way that each part influences other parts (although not all other parts) and those parts in turn influence others, to some degree predictably. For deliberative democracy, several colleagues and I have suggested that our normative understanding improves when we can think of specific deliberative forums as embedded in a larger deliberative system, so that we judge those forums, along with the typical behaviors they encourage or allow, in part by their effects on that larger system (Mansbridge et al. 2012). The macro democratic system includes many different arenas, including everyday talk in the deliberative system and informal representation in the representative system. It includes direct citizen agency through voting in referenda, initiatives, and recall, and through action in protests, mass gatherings, and social movements. It includes the many forms of influence and being influenced that run from rich donors' and corporations' massive effects on the political system to the passing comments of ordinary citizens. It includes structural power and resistance. It includes the many ways of constructing democratic selves and being constructed, from the reciprocal construction of representatives and the represented, through the advantaged classes' powerful social constructions, the organizational and individual constructions of identity in social moments, and the bottom-up social constructions of everyday life.
How might the innovations in Open Democracy fit into the larger macro democratic system in any given city, region, country, or the world? Will they simply improve the entire system, or might some democratic innovations drive out valuable existing features? Joshua Cohen asks in this symposium: Are these open democratic innovations an alternative or a supplement to liberal democratic electoral representative democracy? I read Landemore as answering 'both.' Replace when warranted; supplement when needed. Ethan Leib asks how the purely discursive accountability of a randomly chosen assembly can substitute adequately for the power to replace that citizens wield in electoral democracy. Peter Stone and Daniela Cammack also question the possible effects of Landemore's desire to reduce the importance of the one instance of citizen agency that potentially involves all citizens: that moment in the ballot booth when a citizen chooses (or at least changes the margins of public support for) a representative. Landemore does want to displace electoral democracy as the central democratic institution, but she advocates other direct forms of citizen action, such as voting in initiatives and referenda. I read her as insisting that open democratic institutions enhance agency overall.
On the micro side, we must look carefully through democratic eyes at the actual working of these innovations, just as we scrutinize the institutions of existing democracies. Which of the many plural, and sometimes conflicting, values in democracy do these innovations enhance? How do they advance-if at all-equal power, the equal opportunity to influence, the freedom of both participants and non-participants, the quality of deliberation, and the quality of the outcomes? Christopher Achen asks in this symposium about groupthink and self-selection among those chosen at random for citizens' assemblies. Cohen asks about the theory and practice of handling disagreement. Ethan Leib asks if the members of citizens' assemblies drawn by democratic lottery may be given too open and unstructured a field of inquiry, with too little stress on their role responsibility to act in the public interest. I read Landemore, whose earlier work has focused on such micro dynamics, as agreeing that these are critical questions as the field moves forward. These questions demand study, experiments in design to reduce negative effects, and wide dissemination of the lessons from successful experiments.
The symposium also raises larger philosophical questions, such as whether the participants in an assembly drawn by lot can 'represent' the citizens who had no role in selecting them. Drawing on the experience of ancient Athens, both Landemore and Cammack answer 'yes.' Athenian citizens viewed their democratic entities drawn by lot, such as the Council of 500 and the 500 or more citizen judges in court cases (also translated 'jurors'), as acting ' on behalf of' of the polity as a whole. So too today most citizens take as acting on behalf of the polity both town meetings attended by self-selection and representatives chosen by self-selected voters, although in most cases that attendance and voting takes place in a highly unequal selection environment. Representation is always partial, and the multiplication of forms of representation may help some forms compensate for the failures of others.
The quality of the questions this book prompts reflects the quality of the book itself. Open Democracy is a vivid, spirited, compelling defense of moving from the consent of citizens to their exercise of power. It shows in myriad, well-documented, and subtly analyzed ways how to open up the democratic process so that citizens can enter. The book ' aims to change minds, develop intuitions, and expand our imaginations' (xvii). I think it highly likely that it will succeed in that aim. In my view, it is only through such innovations and reconceptualizations that democracy-begun in the Greek city-states and evolved in the last six hundred years to meet the needs of nationscan meet the steadily increasing demands of the future.

Joshua Cohen
Apple University, Apple, Inc., US Joshua_cohen@apple.com Carl Schmitt (1926) said that deliberation belongs to the world of the parliament, not to the world of mass democracy. A deliberative democracy-marrying public reasoning and broad public opportunities for participation-is simply not possible. Not possible because of the practicalities of getting large numbers engaged in public discussion; cognitive and motivational limitations; and the sheer heterogeneity of interests, identities, and convictions.
In response to Schmitt, minimalist democrats drop deliberation and preserve democracy by lowering expectations. Democracy is peaceful competition for office-which is, as January 6 forcefully reminded us, better than the alternative.
Habermas' two-track model has provided a promising alternative to minimalism. We get mass participation in dispersed, informal public discussion, which identifies troubles and puts themes on the political agenda. Then we have a translation of those topics and themes into law and administration through the formal public sphere of legislatures, agencies, and courts. But this translation takes place without direct engagement by citizens in acts of lawmaking.
Hélène Landemore's wonderful book offers a more ambitious marriage. Open Democracy brings mass democracy and deliberation together in an alternative to electoral democracy, crystallized in the idea of an open mini-public: '[1] an all-purpose, randomly selected body [2] open to the input of the larger public via citizen initiatives and rights of referral as well as [3] a permanent online crowdsourcing and deliberative platform, and [4] ultimately connected to a demos-wide referendum on central issues (including, ideally, via the same online platform used for crowdsourcing and deliberation, this time used for electronic voting)' (218). Therefore, you get non-electoral, democratic representation via random selection, and you get broad inclusion in agenda-setting, discussion of alternatives, and decision making-features that are all absent from the two-track model. I find the general line of argument creative and important. To build on it, I will raise three questions, which aim to clarify our understanding of what Landemore is proposing and suggest directions that the ideas might be developed.
First, I am not sure whether open democracy is an alternative to electoral democracy, with competitive elections gone from the picture, or a supplement to conventional representative institutions.
For example, in the discussion of participatory budgeting, Landemore mentions that practitioners talk about ' co-governance,' where the 'implied partners are the existing (electoral) representative institutions' (95) A second question is about the virtues of open democracy as against electoralist democracy. Electoralism introduces 'systematic discriminatory effects in terms of who has access to power, specifically agenda-setting power ' (26). The open alternative expands agenda-setting power.
Agenda-setting power comes in two types: informal and formal. Landemore focuses entirely on the formal type: the power to add a bill or an amendment, say, to the legislative agenda. But-as Habermas's view emphasizesmuch important political work is done by informal agenda setting. Consider Black Lives Matter, or Occupy, or MeToo, or countless other efforts to change the focus of publicpolitical discussion by saying 'Something is really wrong here that must be addressed.' Landemore, I assume, thinks something like this: 'Yes, informal agenda setting power is essential, and yes, it is different from putting items on a formal agenda. I am simply talking about the latter: making decision-making bodies more porous.' Fair enough. But just as we should have some hesitation about decision-making bodies turning into talk shops, so, too, we want to be cautious about narrowing the openended explorations of concerns by turning informal public discussion into practical problem-solving. During the Vietnam War, we were often asked what our plan was for ending the war. The right answer was that we did not need a plan in order to know that something was wrong. Any more than Occupy needed a tax bill.
Thus my second question: How do we preserve the independence of open-ended moral-ethical exploration in a political world that effaces the boundaries between public discussion and policy debate? Maybe-and I intend this second question as a cautionary note, not as an objection-limited access to formal agenda-setting power has some advantages.
My third question is about disagreement. Landemore says 'The question of whether homogeneity of the population is a prerequisite for this kind of experiment to work is, in my view, the real crux of the matter' (187). She is concerned about disagreement because the public reasoning characteristic of deliberation may make demanding assumptions of background agreement. Landemore offers three responses to these concerns.
First, Landemore offers the case of medieval Iceland in the summer of 999. She describes an ' astonishing' switch from paganism to Christianity that occurred after some impressive speeches in the Althing, Iceland's public assembly. Landemore offers this striking example in support of the conclusion that democratic deliberation, which produced the rapid religious conversion, may itself yield the greater cultural homogeneity required successful for public deliberation. Given the cultural heterogeneity of democracies, I think we should be wary about generalizing on this avowedly ' astonishing' case.
A second argument appeals to the Hong-Page ' diversity trumps ability' theorem, which explains how collective decisions are improved by diverse perspectives. The Hong-Page theorem, however, assumes a shared objective function. Thus it appears to assume away the very kinds of disagreement that Landemore is concerned with addressing.
A third, more promising response comes in a comparison of Switzerland and Belgium, which suggests that disagreement is more navigable under democratic conditions. Here we need more about the kinds of disagreement that are compatible with deliberative decision making (as distinct from the Deliberative Polling discussed in Luskin et al. 2014), and why the deliberative character of the decision making might make majoritarian decisions more acceptable to all.
So, the third question is: How does deliberativeness help democracies to accommodate and process the disagreements that they inevitably contain?

Daniela Cammack
Charles and Louise Travers Department of Political Science, University of California, Berkeley, US daniela.cammack@berkeley.edu Hélène Landemore's excellent new book rests on a startling premise: that modern representative (i.e., electoral) democracy is a form of oligarchy. The ancient Greeks would have agreed. Electoral democracy resembles the species of oligarchia practiced in ancient Larissa and Abydos, in which the dēmos (citizen assembly) elected those who sat on the governing council (Aristotle Pol. 1305b: 28-34).
Landemore wants us to think more imaginatively about non-electoral forms of representation. Here, too, ancient Greek democrats are her allies. Some may be surprised by Landemore's suggestion that ancient Greek democracies used political representation (73), but she is right, as is shown by the frequent appearance, in Athenian political prose, of the term hyper, ' on behalf of' or 'in the interests of,' with respect to political figures (Cammack 2021b). Ambassadors, who were elected by assemblygoers, acted ' on behalf of' the polis, that is, the entire citizen community. Orators in the assembly-who were self-selected-spoke and acted ' on behalf of' the dēmos. Councillors-who, in democratic Athens, volunteered into annual local lotteries to serve on a panel of five hundred for a year to set the agenda and attend to other matters for the assembly-spoke and acted ' on behalf of' the polis and dēmos. And judges (dikastai, sometimes translated 'jurors')-who volunteered into a daily lottery to sit on judicial panels several hundreds strong-gathered, deliberated, and decided ' on behalf of' the polis, dēmos, and plēthos, that is, the wider multitude.
This and other evidence suggest that ancient Greek democrats were perfectly comfortable with the idea that a relatively small number of ordinary citizens might act on behalf of the community. But besides these mini-publics, the ancient Greeks used another form of representative political institution, which also deserves open democrats' attention: the open mass meeting (ekklēsia).
Was the ancient Greek ekklēsia a representative political institution? Yes. Ancient authors routinely described the dēmos (the agent constituted at the meeting: Cammack 2019) as gathering, deliberating, and deciding ' on behalf of' the polis and plēthos, using the same term (hyper) as they used to describe the activities of ambassadors, orators, councillors, and judges. That fits what we know about the numbers who attended such meetings. In Athens, during most of the classical period, a typical assembly comprised around a fifth of the citizen body. In other democratic communities, the proportion was smaller: say 1,000-3,000 out of citizen bodies up to 20,000 (Robinson 2012: 232-3).
Ancient Greek councils, courts and mass assemblies are all examples of 'synecdochical' political representation, in which a part stands for a whole. 1 Within that sphere, Landemore privileges what she calls ' open mini-publics' (13, 219)-large, all-purpose, randomly selected assemblies of between 150 and a thousand or so people gathered together for an extended period of time-such as the Athenian council and (perhaps) courts. She favors what we may call the dialogical model of deliberation: she wants all the members of the deliberating group to have the opportunity to participate in exchanging with one another reasons for their positions (139). 2 That, obviously, is not possible in an assembly of thousands. Nonetheless, I think an ancient Greek democrat would have seen at least three reasons for using open mass meetings along with open mini-publics-reasons that open democrats should endorse.
The importance of agency. Open democrats ought to require at least one political process in which citizens' participation depends only on their own agency, thus maximizing the political impact of ordinary citizens (which I take to be the point of democracy) and helping to justify rule by those citizens. Only if we know that everyone who wished to take part in making a decision could have done so, can we infer that those who did not participate were comfortable with others making a decision on their behalf. Landemore calls this stance tacit majoritarianism (109-110), and modern elections, for all their faults, have this feature. 3 Mass meetings held in spaces too small to fit the entire citizen body do not fully meet this criterion. Nevertheless, open mass meetings come much closer to making citizen agency dispositive than randomly selected mini-publics, which, by design, do not allow each citizen the agency to determine participation.
The importance of decision-making. Early on, Landemore argues that Rousseau was mistaken: popular sovereignty is not just about taking part in final decisions, but also about controlling the agenda and deliberating on the dialogical model (56-60). Control of the agenda was certainly important to democratic Athenians, and the people retained that control through the assembly's supremacy over the council (see Landauer 2021). But the ancient Greeks put most weight on being a decision maker-the term bouleuesthai, 'to deliberate,' implied coming to a decision (Cammack 2020). Unlike Landemore, I am not persuaded that we owe one another reasons for our positions. Giving others a good faith account of why we hold the positions we do is a gift that honours the recipient; it is a sign of trust, especially if the reasons in question are personal; it cannot be required, even putatively, without diminishing the dignity of both speaker and listener. But reason giving is not necessary for respectful and constructive democratic politics. What we owe one another are not reasons for our positions but equal opportunities to take part in decision making and to make public, if the speaker wishes, any potentially salient considerations beforehand. At the Athenians' open meetings, everyone could vote and anyone could speak. Few did speak, but that is not always a demerit. People may often feel that someone else has already aired their position effectively (as Landemore notes, 109), or they may have other honorable reasons not to come forward in public. Either way, open democracy's purview need not be restricted to numerically small gatherings. Instead, open democrats should consider supporting mass meetings, mass debates, and mass votes wherever possible.
The power of numbers. There is something powerful about seeing a mass of people: it's a reminder of the power of collective agency, including its physical power to prevail. The ancient Greeks called the power to prevail-especially physically-kratos and enshrined it in the term dēmokratia. That term reminds us of the main advantage that a dēmos has over the smaller socio-economic, political, and even rhetorical elites that often try to dominate it. Today, the physical power of a mass of people is typically visible only in revolutions and protests, when crowds gather together. But the ancient Greeks gave mass gatherings a place in everyday politics, and open democrats should consider doing the same. There is no harm-in my opinion, quite the contrary-in reminding political leaders and office holders that the crowd has resources, including physical ones, that its leaders cannot match.
Ancient Greek democracy was indeed representative, albeit in an unfamiliar mode-the 'synecdochical' modeand it offers us significant imaginative resources. Open democracy, as Landemore sketches it, draws on one kind of synecdochical representation-the 'lottocratic,' as seen in the open mini-public. But it neglects another, what I'll call simply the kratic, alluding to the one real bit of political leverage ordinary citizens have at their disposal: their kratos, that is, their overwhelming power to prevail when they come together collectively.

Peter Stone
Trinity College Dublin, the University of Dublin, IE pstone@tcd.ie Hélène Landemore's new book offers a new 'institutional paradigm of popular rule,' one that is 'primarily nonelectoral yet (more) democratically representative than any existing form' (p. xvii). This institutional paradigm gives pride of place to both lottocratic representation, via randomly selected mini-publics, and self-selected representation, via new forums for volunteer participation (80-81). Most of the buzz regarding this book, I would say, is being generated by the interesting and innovative institutional proposals Open Democracy makes, and by her examination of contemporary democratic experiments that Landemore views as baby steps towards her new institutional paradigm. But this is only half the story told by Landemore's new book. Open Democracy also offers ' an alternative normative conception of popular rule, one true to the democratic values of inclusiveness and equality, and one we can use to imagine and design more participatory, responsive, and effective institutions' (p. xviii).
Thus, on the one hand, Landemore promises a set of institutions that will realize popular rule better than ever before. On the other hand, she offers a new definition of what 'popular rule' means. She does not simply propose institutional innovations that will succeed according to our existing democratic yardstick; she wishes to remake the yardstick by which we assess democratic innovations like these. The new yardstick is important, because what Landemore calls 'popular rule' is significantly different, I would argue, from existing understandings of the term. What Landemore's open democracy promises to deliver may be attractive, but I am skeptical that the term 'popular rule' should be used to describe it. 4 According to Landemore, the American Founders ' explicitly presented as a superior feature of their intended republic the fact that it was not meant to rest on demoskratos, or people's power, but instead on the power of elected elites' (4). In support of this claim, she quotes James Madison's boast that the U.S. political system was characterized by 'the total exclusion of the people in its collective capacity from any share in' government (4). Landemore's characterization of the American Founders is fair. But note that for Madison, denying the people the chance to act collectively meant denying them popular rule in any meaningful sense. In Madison's system, the people might almost be said to play a role like that of Queen Elizabeth II today: she has a fancy title, but her contribution to the political system is purely ceremonial and symbolic.
The U.S. political system has changed much since the days of Madison. The American people collectively select members of the U.S. Senate and House of Representatives, and indirectly select the President. This may be a pale imitation of the level of power enjoyed by the Athenian demos, but it is still meaningful. Landemore's open democracy, however, offers even less space for the people in its collective capacity to act. Her system minimizes, if not eliminates, election. Referenda play no major role in it either. In Open Democracy, Landemore out-Madisons Madison.
If open democracy reserves no important space for the people to act in its collective capacity, can it be said to embody popular rule? Landemore can answer in the affirmative only by offering an ' alternative normative conception of popular rule,' one that in no way depends upon the people acting as a body. For Landemore, popular rule is something each individual member of the demos, not the demos as a whole, enjoys. She sees 'much of the value of popular rule' coming from ' empowering all members of the demos equally, and in particular giving them all an equal right of access to the deliberation shaping the laws and policies that govern us all' (8). Politics, for Landemore, must be something in which each of us can take part, not simply politicians, bureaucrats, or other elites. It is this fact that makes the experiments she catalogues-from the unsuccessful attempt to rewrite Iceland's constitution (Chapter 7) to France's Great National Debate (184-187)-so attractive to her.
Landemore describes her efforts to specify a normative conception of popular rule as an exercise in ' concept clarification…allowing us to have a better understanding of what democracy means and what regime forms can legitimately count as such ' (19). I believe what she offers is less concept clarification and more concept revision. She claims that the ideal of popular sovereignty is 'both sublime and somewhat vague' (6). Fair enough, but this does not mean that the ideal is completely without content. I am unfamiliar with any prior understanding of popular sovereignty or popular rule which left no place for the people, in its collective capacity, to act.
Landemore may be right to contend that popular rule requires nothing of the people in its collective capacity, no ability of the people as a body to play a meaningful role in politics. But if she is right, then our entire understanding of popular rule, of popular sovereignty-indeed, of democracy itself-will need to be radically revised. I do not believe Landemore is always clear in Open Democracy on the radicalism of her revised conception of popular rule, and I would urge her in the future to elaborate and defend further this radical revision.

Christopher H. Achen
Center for the Study of Democratic Politics, Princeton University, US achen@princeton.edu I am honored to offer my thoughts on Hélène Landemore's Open Democracy. It is an inspirational vision of human control over our governments. It is also a well written, forceful argument that pushes back against a variety of critics-all in all, a pleasure to read. No one interested in contemporary democratic theory should miss it.
Landemore's argument is not a purely abstract exercise in democratic philosophy. It engages at some length with actual political experiments she favors, particularly citizens' assemblies in Canada, Iceland, and Ireland, along with other attempts at open democracy. Hence, I will offer brief comments here on the actual performance of citizen assemblies, leaving detailed evidence, references, and related topics to Achen (2021).
Where citizens' assemblies are concerned, perhaps the most striking difference between deliberative democracy theorists and political scientists lies in their factual judgments. To empirical scholars, the pathologies of deliberative group decision-making are well known, dating to the studies of 'group dynamics' by Cartwright and Zander (1953) and continuing through many dozens of subsequent important studies too numerous to list here. Jury decision making is also well studied in both experimental and observational settings, with depressing results. Even highly educated, professionally trained groups are subject to powerful irrationalities in their deliberations, including 'groupthink' (Janis 1982).
One might expect the literature on deliberative democracy to be deeply engaged with the best-established scientific findings about group deliberation, both its successes and its chronic aberrations. Yet very little discussion of that kind appears anywhere. 5 This raises the concern that citizens' assemblies may not work well in actual politics.
Landemore argues that deliberation has already proven successful in the British Columbia, Iceland, and Ireland citizens' assemblies. Theorists have generally adopted the rosy view of these experiments laid out by their proponents and participants, while political scientists have typically kept a close hand on their wallets. Landemore (152) describes the Irish case in particular as 'the jewel in the crown of deliberative democracy,' referring to citizens' assemblies' successful recommendation of liberalized gay marriage and abortion laws. Among the Irish themselves, however, the evaluation has been mixed at best, with references to 'the myth of the citizens' assembly' and 'behind the hype.' The two Irish moral issues were easily understood and long deliberated. In the case of abortion, prior referendums had codified abortion restrictions thirtyfive years earlier and then somewhat liberalized the law a decade later. After 2010, the horrifying death of a young woman, combined with ongoing Catholic Church scandals, substantially weakened the anti-abortion forces, so that by the time of the new referendums, the polls showed that a fully convinced majority for change already existed on both topics. Irish politicians then used the citizens' assemblies in ' a deeply cynical exercise' and ' an act of political cowardice' for a largely successful effort to get the assemblies to take the blame for crossing the Roman Catholic Church. In the end, the assemblies and the citizenry voted the same way-a success of a sortbut the fix was in. The congruence occurred for reasons unrelated to those advanced by deliberative theorists.
The first Irish assembly also successfully recommended abolition of the unloved blasphemy section of the constitution, which had gone unenforced since the 1850s. Otherwise, on the 50 other proposals by the Irish assemblies, along with those of deliberative assemblies in British Columbia, Ontario, Prince Edward Island, and Iceland, the number of successful adoptions into law thus far is exactly zero.
Politicians have not been the only barrier. When publicity was adequate and turnout was normal, voters have typically rejected assembly proposals, often by outsized margins. Political theorists like to cite the near success of the first British Columbia referendum, but they should stop doing so. It was conducted in what one close academic observer called a public 'silence.' Two subsequent referendums on the same topic, with more publicity, better citizen knowledge, and more public deliberation-closer to the conditions deliberation theorists insist on-resulted in one-sided beatings for the citizen assembly proposal. In all these cases, what has gone so wrong?
A large part of the problem is that the assemblies are highly unrepresentative, leading to proposals often seen as undesirable by the voters and other governmental actors. Landemore (59) writes about Iceland, 'The selection process was thus technically near random sampling, subject to some self-selection, combined with stratified sampling.' However, she (158-162) notes that 80 percent of those randomly drawn for the assembly refused to participate. In Canada and the Netherlands, the corresponding refusal percentages were well over 90 percent.
As always, this kind of selection induced a bias in the resulting assemblies, with the well-educated grossly overrepresented-more than double their proportionate share in British Columbia, Ontario, and the Netherlands, for example. The presence of the less-educated, often poor or minorities, were systematically reduced. This is the same kind of sampling that has led to serious polling errors in recent American presidential elections. Most people do not want to participate in surveys or citizens' assemblies. Those who do, whatever their charms, are a really bad sample.
Unless one plans to use the police to compel the unwilling to attend (or to take from other social needs in hard-pressed public budgets the very substantial amounts required to recruit the reluctant), unrepresentative citizens' assemblies are inevitable. They will usually be disproportionately populated by the educated upper middle class, the kinds of people who like to do that kind of thing. That group has long been prominent in reform politics, beginning with the American Progressives a century ago and continuing to the present day. They bring their middle-class good intentions and moralistic style to their square-offs against conventional politicians, whom they often despise as corrupt insiders but whom ordinary people usually prefer at the ballot box. Landemore largely steers around that culture clash, but it runs all through the current literature. Citizens' assemblies sing with an educated upper-middle-class accent. 6 Then too, as recent populist agitation in Western societies suggests, citizens' assemblies may not always consist of the bien pensants. Ordinarily, 90 percent of those randomly chosen decline to participate. But in hard times, assemblies may be vulnerable to aroused citizens who are randomly selected and then surge into participation to protest their economic losses or to protect their racial privileges, to control abuses or to punish scapegoats, and to get rid of oligarchy or to get rid of democracy. Thus in all cases, we will need Aristotle's balance-other governmental structures that protect us from ourselves at our worst moments, when we are least fair-minded, least self-aware, and least willing to listen to reason.
Hence in my view, ' open democracy' on its own will not work. Listening to any particular group of people is always a good thing, but only within a broader deliberative system (Parkinson & Mansbridge 2012). Thus, we have every reason to try additional open democracy experiments-though slowly and cautiously at first until detailed empirical research tells us why they have loomed so large in political theory and in activist enthusiasm, but have produced so little real change. Perhaps in time, we may work out the kinks and want to give sortition and organized deliberation a formal, ongoing role in democratic government. If so, Landemore's brave, idealistic vision will have helped us see a better future. For that hope, we are in her debt.

Ethan J Leib
John D Calamari Distinguished Professor of Law, Fordham Law School eleib1@law.fordham.edu Hélène Landemore's Open Democracy invites us to think more about the analogy between citizen representatives who serve in mini-publics and those who serve as members of a criminal jury. Probing that analogy further illuminatesbut also might limit-the project of making democracy more 'open,' in her terms. Here are three lessons from the jury. First, Landemore (2020: xvi, 153-54) urges what seems to be a rhetorical question: If we can have a reasonably productive set of deliberative citizen representatives making constitutions, as her case study of Iceland is poised to show, why can't we also have them do ordinary lawmaking? Part of what the jury model may teach us, however, is that citizen representatives are better and worse at different sorts of tasks. Heavily structured and cabined deliberation about facts (or single issues) may work better than open-ended policy deliberation at the ordinary lawmaking level. Even if a second case study in the book (the French Convention on Climate Change) shows that we can use citizen representatives in not-quiteconstitutional moments, we may not be able easily to extrapolate from one structured context to one much less ordered. Ultimately, context-specificity about how and when open democracy and jury democracy work seems to still be, as it were, an open question.
Second, to get juries to do their jobs well, we must try to acculturate and inculcate members in their role responsibilities. Designing mechanisms of enforcement for sustaining role fidelity can have dramatic and important effects on the deliberation citizen representatives undertake. At one point in the book Landemore reports what kind of a role the 150 French participants in the Convention case study said they took on (119). Thirtyfive percent said they spoke in their own name only. She reads 60 percent to have taken on some kind of representative role (although only 22 of 150 claimed to speak for the 'public,' whereas 25 spoke in the name of people 'like themselves' and 24 in the name of a ' cause'). There is something a little passive about the vision of open democracy here in which citizen representatives construct their own orientations to their roles. A more proactive approach might be better. As with the jury, we should show mini-public participants inspiring videos, remind them that they are instrumentalities of the state in their decision making, and, perhaps more controversially, demand that they take oaths (of office?) upon installation. In citizens' assemblies like these, many citizen representatives become something like trustees, our fiduciaries, for our governance, and need to be reminded early and often of their role obligations, when we can define them adequately. If a juror in a jury room claimed to be voting her own interests-or the interests of her race-we might tell her: you got this project wrong. Oaths, in short, can be powerful in orienting role.
Finally, the jury model provides a way to think about embedded accountability (cf. Leib & Ponet 2012: 269). Landemore struggles in the book to come up with a theory of citizen representative accountability-and draws on the jury to argue that because the criminal jury is unaccountable in important ways, we can accept some level of unaccountability within mini-publics (104). But the jury offers a richer way of thinking about this issue. Beyond discursive accountability-the accountability effectuated by the need to answer to others' arguments within the deliberative process-there is also structural accountability, which arises through embeddedness in judicial institutions. Landemore might resist the ' closed' nature of jury room and dislike the supervision by elites, but it is hard to deny that both the structure of the deliberation and its overseers can get us successfully channeled deliberative energy, offering us a form of accountability that can be scaled.
Take Landemore's discussion of the attempted abolition of minarets in Switzerland through direct democracy (202). She quickly tells us we don't have to worry about these illiberal results because the Supreme Court there was having none of it. Yet rather than conclude that deliberation has to be embedded in some form of legal and political system that provides oversight (as the analogy to the jury would reveal), she tells us it was really the people themselves who came to see they shouldn't get rid of minarets. It might be significant to acknowledge that elite input here was pretty central to harnessing the wisdom of the collective. The jury model may have been right all along: some judge somewhere (or some institutional check somewhere) can make sure things don't go off the rails when democracy gets too open.
Admittedly, this final lesson from the jury sits in tension with the thrust of Open Democracy. Landemore's project has a clear allergy to forms of ' dualism' that keep 'the power of the people' too much at the margins. Landemore is sure that even with what one might call in Habermasian terms 'sluiceway reform'-the sluiceways for Habermas (1996: 356) are the mechanisms by which deliberation in civil society makes its way to the center of the more formal political sphere-'fluidification wouldn't really challenge the dualism' (38 n. 20). Yet it will surely be beneficial in the years to come to think more about how sluiceway reform could go some way to deliberatively democratizing our politics: we probably need to bring elites (including elected and appointed officials) and citizens into dialogue with one another, not just, as Landemore puts it, to keep ' experts on tap ' (192). Without more attention to sluiceway reform (cf. Ponet & Leib 2011: 1249, more mini-publics might ban minarets. And that isn't the kind of ' open democracy' anyone should want.

Reply to Critics
Hélène Landemore Yale University helene.landemore@yale.edu Joshua Cohen asks one of the most important questions: Do I want open democracy as an alternative or as a supplement to conventional representative institutions?
At the theoretical level I do mean open democracy as an alternative, not a supplement, to representative democracy, for the simple reason that elections are optional in my model in a way that they are not in the historical paradigm of representative democracy (see Manin 1997). Open democracy could share with representative democracy a basic constitutional framework entrenching certain fundamental liberal rights (in which open democracy can be directly inserted or which it could evolve on the basis on its fundamental five institutional principles). But I see them as distinct models of democracy.
That said, I also recognize that actual reforms will probably have to start from where we are and accommodate existing institutions, such that a hybridization between representative democracy and open democracy is more likely than a pure open democracy. Still, I think it's important to keep these models analytically separate.
Additionally, the reason why I can imagine a nonelectoral democracy is because my conception of the main locus of sovereignty is the legislative power. I would be fine with an elected executive, as long as it is clearly subordinated to Parliament. So perhaps parties and the electoral machinery could be preserved to select officials for executive and perhaps also leadership functions. When it comes to the legislative function, however, which has the power of making the law, I prefer random or stratified random sampling.
Finally, a pure open democracy is plausible only in contexts where periodic elections are not already entrenched, which means that for most Western countries we are going to have to evolve a partly electoral version of open democracy. In my view though there should be serious transfers of power from the elected to the lottocratic bodies. If we simply create additional lottocratic bodies without curbing or redefining the power of existing elected bodies, we risk ending up with conflicts of legitimacy and system paralysis.
One might still want to know: Where would a pure open democracy be possible? Perhaps in developing countries or former soviet union countries that are so disillusioned with the kleptocratic, clientelist, corrupting effects of electoral models imported from (or imposed on them by) the West that they may want to try something new entirely, provided they can transition while preserving or establishing the rule of law as a prerequisite. Or in an election-less country that could democratize by leap-frogging the Western electoral stage and going straight for an open democracy model. Or perhaps in cryptocurrency-based online communities, which are in search of governance models and whose commitment to decentralization and transparency seem aligned with the spirit of open democracy.
Cohen also asks what we can reasonably conclude about the feasibility of open democracy from empirical cases that are always partly electoral. First, notice that the example of Classical Athens used throughout the book is not electoral at all when it comes to political offices. I could also now appeal to the examples of proto-democracies that David Stasavage documents as pre-dating or emerging independently from the Greek example in various parts of the world (Stasavage 2020), none of which were based on elections. But it is indeed true that there are no modern, large-scale example of non-electoral democracies and the central cases of Iceland and France are no exceptions to the rule. Note, however, that what I tried to do is extract from these empirical examples the non-electoral components, which there are independent theoretical and empirical reasons to think would be viable on their own, without the electoral legacies. In fact, in both Iceland and France, the electoral components can plausibly be blamed for the partial failures of the documented participatory processes. The electoral elements (parties, parliaments) were in each case obstacles as much as preconditions for open democracy moments. Daniela Cammack and Peter Stone raise the similar objection that open democracy does not make enough room for the people as assembled masses. In fact, Peter Stone's reading is that I do not make any room at all for the people as a whole in my model, in a way that makes me ' out-Madison' Madison. I'm afraid I cannot accept this latter version of the objection, which distorts my views and turns them into those of a downright demophobe. Open democracy's first principle is participation rights. Prioritizing participation rights is a way to open the possibility of mass movements, mass meetings, voting in referenda, and indeed initiating such popular vote processes-all things Madison would have absolutely opposed. But I grant Cammack's more moderate criticism that I spend comparatively little time theorizing the role of the masses in open democracy and that, as Peter Stone also correctly remarks, the book ends up being a lot more about the concept and practices of democratic representation than popular rule per se.
How do I then view the place of direct democracy and mass meetings in open democracy? I believe both that mass referenda are as close as we can get to having the whole people assembled and that moments of mass authorization via mass referenda are required for the legitimacy of the whole system. Indeed, if we care about equality and inclusion in the here and now (and not just over time, as allowed by rotation of lottocratic bodies), open democracy does require legitimizing moments of direct democracy like referenda. Ideally such referenda would be organized so as to be deliberative and highly participatory. It should also be possible for citizens themselves to initiate them directly (by collecting a sufficient number of signatures).
What I'm less sure about is how frequent those moments should be. My instinct is to leave it for the most part to the people themselves to decide on the frequency of such meetings by making participatory rights available and easy to use but not turning actual participation into a legal obligation. That said, having referenda triggered by law on a certain domain of questions or with a minimal frequency might be a good thing.
Additionally, and independently of my views on referenda per se, which are an aggregation of secret individual votes that need not be time-consuming to cast (especially if such voting could be conducted electronically), I do have some qualms about the open mass meetings that Cammack seems to favor as a way to make collective decisions.
My first reservation comes from the fact that I assume an economy relatively similar to that in which our institutions operate rather than the slave economy of ancient Greece, namely an economy in which people need to work to make a living and time is thus scarce for many people. If I had gotten a chance to fully develop the thought I introduce in the concluding chapter about substantive equality and the need to democratize the economy or create a universal basic income or some other solution, perhaps mass meetings would become a more plausible option.
My second reservation toward open meetings is that we need to be mindful of citizens' time, especially as we scale democracy to various levels and expand it to other domains and organizations. Mass meetings, especially if they purport to be deliberative, are time-consuming and might well end up taking, indeed, too many evenings.
Finally, a related reason to worry about open meetings, as distinct from referenda, is that what Cammack calls synecdochal representation is highly biased in practice. That was already the case in Ancient Greece (Aristotle remarks that farmers rarely attended the People's Assembly) and the examples we have today, from Participatory Budgeting to crowdsourcing experiments to social movements to the Great National Debate in France similarly show that people in such meetings, while a part of the whole, are not demographically like the whole at all, though probably more diverse than elected officials.
Indeed in giving place of pride to lottocratic representation over self-selected representation, I thought I was applying a lesson from Athens that I learned from Cammack herself, namely that the ancient Greeks came to recognize the vulnerability of mass meetings to capture by gifted orators and decided to move in the fourth century to a more lottocratic form of democracy, transferring some of the powers of the demos in the Assembly to the nomothetai. If the Greeks themselves thought that lottocratic synechochical representation was superior to self-selected synecdochial representation, I think we should pay attention.
Let me now turn to Chris Achen's claim that my inductions are too optimistic given the available evidence. Contrary to Achen, the state of the literature seems to me far from dire. First, on the effects of deliberation per se, recent research based on surveys and case studies provides evidence for an increase in participants' knowledge and sophistication during deliberative experiments (Abelson et al. 2003;Carman et al. 2015;Knobloch, Gastil, Reedy, & Cramer Walsh 2013;Luskin, Fishkin, & Jowell 2002). An RCT experiment by Carman et al. (2015) shows that deliberation increases participants' knowledge of evidence and shifts their attitudes regarding the role of evidence in policy making. Sophisticated quantitative studies by Westwood 2015, Gerber et al. 2018, and Esterling et al. 2021 vindicate the claim that deliberation, when properly conducted, is conducive to persuasion rather than polarization and, specifically, that this outcome is driven by deliberative mechanisms (rational justifications). Other work indicates that deliberation aligns fundamental preferences and produces other benefits (Farrar et al. 2010;Ingham & Levin 2018;Kim, Fishkin, & Luskin 2018;Knobloch, Gastil, Feller, Richards, et al. 2014;Setala et al. 2010). Finally, a comparison between the plenary sessions of an Irish Citizens' Assembly and an Irish parliamentary committee on the topic of abortion shows greater deliberative quality among citizens than parliamentarians (Suiter et al. 2021). All of this establishes that many of the intuitions of early deliberative democrats were right all along, contra much of the established doxa in psychology and some corners of political science and law that Achen relies on.
On the impact of citizens' assemblies (CAs) and other mini-publics, it is just not true that it has been null across the board. The two famous Irish CAs (respectively on marriage equality and abortion, as well as a host of other issues) have led to constitutional changes. Deliberative polls were behind the major reversal in Texas's energy policy, turning it from a pure oil and gas state into a leader in green energies (Galbraith & Price 2013). In Ostbelgien a Citizens' Council shaped the agenda and debates of the local parliament for over a year. Even in France, at least 25% percent of the CCC's recommendations ended up turned into law and led to the most ambitious French climate bill to date. So it cannot be said that the CCC did not move the needle. Many more examples can be found in the over 600 cases of deliberative minipublics documented by the OECD since the 1980s (OECD 2020, updated August 2021). The OECD study shows that in 81 percent of the 55 examples for which data is available, public authorities accepted at least half of the recommendations that citizens developed in these processes.
Achen's reading of the evidence is puzzling. Regarding the Irish case for example, he chooses to interpret it as ' a largely successful effort to get the assemblies to take the blame for crossing the Roman Catholic Church.' A more charitable and plausible reading is to see it as a successful effort to give women's a fundamental right over their health and bodies using a participatory tool without which politicians would have been stuck for a couple more decades, precisely because their self-interest in an electoral system made them lack the courage to cross the Roman Catholic Church. The politicians may indeed have used the CA to do something they wanted to do anyway, but since they couldn't do it without the CA, how is that not proof enough that the exercise was necessary and impactful?
Second, Achen seems to attribute to advocates of deliberative democracy the belief that mini-public deliberations should produce radical outcomes that the larger population will rally to. In other words, he envisions for citizens' assemblies a role of visionary leadership. But that is not the role most deliberative democrats, including myself, theorize for CAs. What we have in mind instead is basic democratic representation. What I say, for example, is that randomly selected citizens assemblies will produce ideas and proposals that are both more aligned with the preferences of the larger population and draw on a more diverse pool of views and information than those of elected assemblies. As a result, their proposals are likely to be better and more likely to be accepted by the public than those of elected assemblies. The standard of success for deliberative democrats is not whether CAs can generate outstanding new policies that no one had thought of before. It's whether CAs can produce laws and policies that are better and more legitimate (in a sociological sense at a minimum and in a normative sense as well at best) than those produced by elected politicians. In the Irish, Icelandic, and French cases that standard was met (even as in the French and Icelandic case Parliament implemented respectively few or none of the proposals).
In the Irish case, however, Achen insists, 'the fix was in' because both the population and the politicians had evolved on the issue of abortion. One has to be quite cynical to look at a citizens' assembly giving majorities what they want after decades of politicians being unwilling or unable to do so and conclude that the congruence is 'a fix.' I also disagree that the experience proves that deliberation did not work as intended in that case. The congruence between the assembly and the rest of the country occurred precisely for the reasons I just rehearsed: a random sample is more likely to reflect the preferences, information, and perspective of the larger public. What deliberation among a mini-public makes uniquely possible, compared to the informal exchanges among the larger public is, in addition to the greater quality of reasoning, a deliverable: actual policy recommendations about whether and how the law should change, which a loose, disconnected public cannot formulate themselves (it is true, however, that not all minipublics are designed to produce such deliverables). Note that these recommendations can sometimes be extraordinary and visionary but the point is that they need not be.
Finally, Achen dismisses the mini-publics as 'highly unrepresentative' and therefore leading to proposals often seen as undesirable by the voters and other governmental actors-in essence the exact criticism I and others make of elected parliaments.
This claim is empirically false. CAs are indeed not perfectly representative, in part because, as Achen correctly points out, the take-up rate is still low. But they are not 'highly unrepresentative' by any means, and especially not in comparison with elected chambers. Quota sampling and incentivization via honoria, paid daycare, and other measures allow the sample to capture many dimensions of the target population, including gender, geographical origins, education levels, ethnicity, economic backgrounds, and sometimes partisan leanings and policy views. It is therefore also false to say that the assemblies speak with an upper-class accent in their composition. We can, however, discuss how to increase acceptance rates (some suggest making participation in citizens' assemblies mandatory, as in jury duty).
We can also discuss ways to improve the internal dynamics of deliberation within these assemblies, which may well in some cases reflect the greater influence of the wealthier and more educated members. Regarding that later criticism, however, I would note that 1) typically facilitators try to mitigate the influence of the wealthy and more educated; 2) the most empirically sophisticated studies available, of Deliberative Polls, show no such greater influence (Siu 2017); and 3) the set-up of citizens' assemblies in general is more empowering of traditionally oppressed minorities than any comparable environment.
It is fair to criticize deliberative assemblies for their current limitations. Still, it is hard to understand the urge summarily to dismiss CAs on account of some undeniably problematic but partly fixable and not outrageous distortions of representativity, when there are comparatively much larger, indeed mind-boggling, and essentially unfixable distortions of representativity in elected assemblies, which all over the world massively oversample the wealthy, urban, educated, male parts of the population as well as locally dominant races and ethnicities.
Interestingly, whereas Achen thinks I'm too optimistic about the empirical evidence, Ethan Leib, in a way, thinks I'm not making enough of it, especially not the evidence relative to criminal juries. I agree with Leib and I'm mostly grateful for his suggestions so I will simply end here by taking up his invitation to think more deeply about what we can learn about the norms and practices under which jurors operate to improve the design and deliberations of open mini-publics in a renovated democracy.

Notes
1 To be more precise, a part saliently undistinguishable from other parts of the community stands for a whole. The relevant contrast is ''metaphorical' political representation, meaning that someone saliently distinct from other members of the whole stands for the wholefor example, because they have been elected, or because they possess certain special qualities such as unusual rhetorical or organisational powers. I borrow the language of synecdoche and metaphor from Ankersmit 2019, 231--253; see further Cammack 2021b. 2 Alternative models of deliberation are ''internal' and '' audience': see Cammack 2020; Cammack 2021a. 3 The ancient equivalent was ostracism, in which any citizen who wished might drop off a potsherd (ostrakon) bearing the name of a citizen he would like to be exiled. 4 Landemore appears to use terms like 'popular rule,' 'popular sovereignty,' 'people's power, ' and 'popular power' interchangeably (e.g., 1, 6, 9, 53, 220). The concerns I raise here apply whichever term is adopted. 5 Mansbridge (1980) set an early example of looking closely at actual participatory systems, but her lead has only rarely been followed. 6 For example, the detailed abortion recommendations of the Irish citizen assembly were dropped; they were too much the opinions of the liberal educated classes and thus unlikely to pass. The successful referendum simply asked the voters to allow parliament to legislate on the topic.