Introduction

Cristina Lafont’s powerful critique of deliberative minipublics strikes at the central strategy that has energized efforts to actually apply deliberative democracy to real public problems.¹ Every effort to make deliberative democracy practical needs to take account of her critiques. I will consider her argument in five main steps: (1) The argument that such minipublics should not have any real ‘decisional-power’ (15) and 2) that it is not democratically acceptable to rely on them as a ‘second best’ because they only engage a small portion of the population (138–139). Nothing but a ‘first best’ strategy will do in her view. 3) The challenge of achieving a first best solution, which I have previously outlined in terms of the ‘trilemma’ of democratic reform (Lafont 2019: 106). (4) Why I believe her solution, a ‘participatory conception of deliberative democracy’, does not actually offer a solution to the trilemma because it is insufficiently participatory (Lafont 2019: Part III). (5) My own approach to dealing with the trilemma, as outlined in my book Democracy When the People Are Thinking, is, I will argue, both more participatory and more deliberative. I lay out an actual solution. While it is elaborate and expensive (who said it would be easy?) there is no theoretical or practical impediment to realizing it, except for collective political will, except for a shared decision to move forward.

Power for the Minipublic?

Her argument against ‘decisional-power’ for the minipublic has, at its core, a concern for the rights of democratic participation for those who have not been randomly selected to deliberate. Why should they support decisional power for the microcosm when its views may differ from their own actual views? Granted that if it is a good sample, the microcosm should represent the public’s views before deliberation in aggregate. But she quotes me, correctly, as saying:

the thoughtful and informed views created in the experiment are not widely shared because the bulk of the public is still, in all likelihood, disengaged and inattentive…Deliberative Polling overcomes those conditions, at least for a time, for a microcosm, but leaves the rest of the population largely untouched. (Lafont 2019: 118, italics in original)

Her conclusion: ‘This is precisely the problem!’ In my view: ‘This is precisely the solution.’ In her view, proponents of decisional power for deliberative minipublics are asking for ‘blind deference’ to conclusions at least some citizens actually disagree with. If the transformation of opinion is widespread, it may be that the final considered judgments may be different from the actual conclusions many citizens, who have not deliberated, actually have.

But Lafont and I agree that the ones who have not deliberated are likely to be ‘disengaged and inattentive.’ They have no particular reason to get involved in thinking about the complexities of policy issues posing difficult trade-offs. If government listens to them, via conventional polls, they will be empowering what is likely to be an aggregation of the public’s mere impressions of sound bites and headlines. It is not too much of an overstatement to say that policymakers face a stark choice—listen to the people via conventional polls or do not listen to them at all. In other words, listen to the people when they are not thinking (or well informed)—or just ignore their views. Neither seems very satisfactory as a realization of democratic aspirations. A well-designed deliberative minipublic offers the potential for a way out of this dilemma. It is that ‘way out’ that Lafont is criticizing in her critique of my position.

Consider a simple example where there was a transformation of opinion in the microcosm. In 2011, the Japanese government was considering privatizing the pension system. After all, Japan is rapidly aging and the ratio of workers paying into the pay-as-you-go government pension system is changing—there are more and more retired to be supported by fewer and fewer active workers. A privatization of the pension system, roughly along the lines of the Chilean reform, has an appealing logic: If each worker has his or her own pension fund, and invests it, then the government does not have to worry about the changing ratio of elderly to retired. There is a fund for each retired person.

In initial polls close to 70% of the public supported this ‘funded’ system. In the Deliberative Poll, 69% supported it before deliberation but only 29% supported it after deliberation. Instead, there was increased support for a rise in the consumption tax if the money were used to shore up the finances of the current pay-as-you-go pension system. The people of Japan, in microcosm, did not wish to shoulder the risk of private investment for their individual retirements. Once they fully understood the proposal, they did not, by and large, support it. As it happened the Deliberative Poll had major media partners and the government was well aware of these results. In the end, the logic of the solution in the Deliberative Poll was the logic followed by the government.

Lafont’s argument is that even though the microcosm changes, you do not know if you would be one of the people who would change, or one of the people whose views would stay the same after deliberation. If you were initially in the 69%, why should you support the privatization represented by the changing views of the microcosm? In her words, why should you ‘blindly follow’ someone you may not agree with if you are not part of the deliberations?

The first point to make is that it need not be ‘blindly.’ If the project is done well, there will be coherent and identifiable reasons for the change on a major issue. In the Japanese case, it was obviously the risk of taking responsibility for your own retirement, rather than having something guaranteed by the government. The support was cut in half when people considered this, with the strong majority becoming a bare minority. If you are willing to follow the results of a democratic consultation, which one should you follow? Do you follow the first poll, which was largely an impression of sound bites and headlines reflecting little thought and little information, or the second, which was the result of a long weekend of intense discussion and the provision of balanced information and civil, small group discussions? Lafont might answer that you should follow neither. You just believe whatever you believe, as an individual, and that is the end of the matter. But the government and the society have to make a collective decision. So, the question is what is the appropriate public input to that decision? If there is a commitment to collective and democratic decision, then there is the question of which majority should be followed. To shift the example, if a jury were to take a vote before the trial and then one after (a bizarre design for obvious reasons) would you convict or acquit based on the first vote or the second? In the Japanese case, should the government follow the many polls that showed that when the public has not thought about the issue they wanted the ‘funded system’ or should the government have followed the Deliberative Poll that shows a gigantic shift of opinion once people understood and thought about the implications of one pension system or the other? Both are random samples in which most people have not participated. But only the second reflects the public’s considered judgments. The key premise in the move from individual to collective decision is some notion of democracy including a commitment to majority rule of thoughtful opinion. In the case of most policy issues do we want a majority when the people are not thinking or a majority when they are?

The Trilemma and Second Best

The second main argument I want to consider is Lafont’s critique of my work that I rely on the ‘trilemma’ of democratic reform to rationalize settling for ‘second best.’ The trilemma is my name for the apparently forced choice between political equality, deliberation, and mass participation. It is a kind of dilemma with three corners. She is correct that I have long argued that commitment to any two of these principles will routinely undermine the third (Fishkin 2009). And, hence, she says, I simply settle for the combination of deliberation and political equality (achieved through random sampling) for a deliberative microcosm and give up on the idea of having all three.

However, in Democracy When the People Are Thinking (Fishkin 2018), I explicitly introduce the trilemma as a challenge that can be overcome by deliberative systems that will facilitate a stage of mass deliberation (Fishkin 2018: 7–9). I argue: ‘Ultimately, we need more than deliberative microcosms of the people. We need a deliberative macrocosm—a deliberative society. How can this be done?’ (Fishkin 2018: 7). I then refer the reader to the last sections of the book, particularly the discussion of deliberative systems (Fishkin 2018: Part IV). There, I re-introduce the idea of Deliberation Day, a proposal that Bruce Ackerman and I put forward in 2004 (Ackerman, & Fishkin 2004). The idea was to take a national holiday and pay everyone a stipend to engage in deliberations on the model of the Deliberative Poll in their local communities. Imagine hundreds of thousands of gatherings of 500 each, broken into discussions of a dozen or so, with plenary sessions where they gather together to ask questions from the small groups fielded by competing experts and politicians from the relevant parties. The proposal was treated as utopian because of the cost. However, the cost can be greatly cut by strategies to conduct the national Deliberation Day discussions online with vast numbers of small groups. We already do Deliberative Polls online, with video-based discussions in the small groups and plenary sessions with competing experts. So why not consider the expansion to a mass scale by using technology? In fact, we are collaborating with a team at Stanford led by Professor Ashish Goel in Management Science and Engineering on developing an automated moderator that could be used to scale the deliberative process to large numbers of diverse participants recruited via social media. (Fishkin et al. 2019). These are moderated small group discussions just like those in the Deliberative Poll. They end in a period of reflection and decision about the most important questions worth asking in plenary sessions with panels of competing experts. We have successfully applied it inside Deliberative Polls in various venues, including most recently in the City of Tokyo on solar energy issues. We are applying the automated moderator now to ever larger trials and plan realistically to start scaling it.

Why not just use human moderators? The Deliberation Day idea envisions very large numbers of small groups, certainly hundreds of thousands eventually. For such a scenario, automation would greatly assist the scaling. But the merits of the automated moderator, continually improved, must be tested empirically in controlled experiments. I go into detail about this in order to show the seriousness of our interest in scaling deliberation outside the confines of the random sample. To the extent we are successful in this aim, the area of disagreement with Lafont will largely disappear, at least on the first claim. Suppose virtually everyone is deliberating so no one will be in the position of having to blindly follow deliberations in which they did not take part. My point is that the Deliberation Day idea is not merely utopian. It is my attempt to answer the problem of creating a ‘first best’ alternative that is both deliberative and participatory. Even if it has not yet been achieved, there is no particular technological or empirical constraint that would prevent it from being realized. It is just a question of collective political will.

My position with respect to Lafont’s critique is that the deliberative systems approach culminating in Deliberation Day should be put to the same test as Lafont’s proposal for a solution that is both deliberative and participatory. Despite the many admirable aspects of Lafont’s last section, I can find no comparable blueprint for how she would achieve a system that is both deliberative and participatory. It is the participatory side that is under-specified in my view.

Adding Deliberation Day produces a change in focus. The design is obviously distinct from a Deliberative Poll because the scale is not limited to those in a stratified random sample. Hence there is a loss of focus on the hypothetical claim that justifies the connection of the Deliberative Poll to policy making: this is what the people would think about an issue under good conditions for thinking about it. When the samples are no longer representative, then a key premise of that argument is lost. Generalizing from unrepresentative samples will not tell us much about what the whole population would think. However, something else is gained—the beneficial experience of deliberation for the participants as well as the potential to transform opinion in the broader society. If nearly everyone deliberated, there would be no need for the Deliberative Poll. We would have a Deliberative Macrocosm (rather than microcosm), in effect, a deliberative society. We might even imagine exit polling from the mass deliberation could then be used to represent the aggregate views. More importantly, almost everyone would participate and very few would be left out. Very few if any would be in the position that Lafont objects to, having to follow the opinions of those who deliberate while one is left out. Indeed, assuming all registered voters are invited, then if someone is left out, that is by their own choice. Of course, there are always complexities—those who are ill or handicapped or who need childcare or assistance. In the best realizations of the Deliberation Day idea special efforts would be made to assist such registered voters and the whole process would be as inclusive as possible. We are well used to these methods fostering inclusion in the Deliberative Poll. There is no reason not to make similar effects with Deliberation Day.

Lafont is correct that while I endorse Deliberation Day in combination with other deliberative processes as my ‘first best’ solution, I do offer Deliberative Polling as a second best that can give voice to the public’s considered judgments and facilitate public opinion and public will formation. After all, the book argues that the crucial missing ingredient in our current competitive democracies is some site for the formation of the public will. Everyone is trying to distort and even strangle the will of the people while at the same time invoking it once an election is concluded. Unless there is a coherent connection between the public will and what is actually done, the whole point—and the key legitimating function—of democratic practices will be lost. Admittedly, second-best solutions are partial solutions, but they are urgently needed during the current democratic recession. This scenario is second best because it fails to make adequate provision for mass participation. However, like the legal cases Lafont relies on, the Deliberative Polling results are likely to be widely covered in the media. In that sense they would involve the broader public. But the actual acts of participation by millions would certainly not take place. The key in my view is to cross the line between what Bernard Manin (1997) calls ‘audience democracy’ for spectators and actual active deliberation.

Let’s sketch the trilemma in more detail to show why it is such a challenge. I believe the strategy I outline in Democracy When the People Are Thinking offers a possible solution because it relies on the proposal for Deliberation Day, which would actually engage the bulk of the population in mass-deliberative participation. Hence, I believe Lafont misses the target when she criticizes me for not solving the trilemma and being just satisfied with second best as a result. She relies almost entirely on Fishkin 2009 rather than Fishkin 2018. In the former I sketch the problem. In the latter I attempt to outline a solution. We will return to that below.

More generally, the same criticism can be levied at her approach to a ‘participatory conception of deliberative democracy’ as her approach is not, at least in my view, really participatory. It relies on the theoretical opportunity to bring a legal claim, which the vast bulk of the population will not participate in. ‘Participatory’ in the view defended here and specified above, means action by the ‘bulk of the population’, not a few people bringing legal claims on their behalf. Voting in a national election is an example of a practice that is participatory at the mass scale, even though it is not usually deliberative. New institutions are required to effectively combine mass participation and deliberation. The ‘deliberative systems’ section of my book outlines one attempt to specify such a solution, primarily because it culminates in Deliberation Day.

I construct the trilemma from three principles: political equality, deliberation, and participation. My argument is that our conventional political institutions fail at realizing all three. Pick any two and employ our familiar democratic practices to realize them and you can reasonably expect a failure to fulfill the third. The attraction of fulfilling all three is that the combination of two, especially political equality and deliberation, is insufficient without mass involvement. The two—political equality and deliberation—would provide a basis for considered judgments resulting from the deliberations reflecting all persons equally. But without mass participation, these judgments have not engaged the actual participation of all the citizens who are expected to live with those policies. Their views are represented via political equality (as embodied in well conducted random sampling) but that is not the same as actual participation.

This first option, combining political equality and deliberation, is the one that would be fulfilled with a deliberative microcosm chosen by random sampling. This is the option Lafont criticizes me for, as it leads to some people being obliged to follow the choices of others. Random sampling, by its very design, leaves out most of the population. In theory random sampling is not the only method of combining political equality and deliberation. If one could somehow get nearly everyone to deliberate then their equal participation could be combined with deliberation. In fact, that is the point of Deliberation Day, a concept we will return to. But for most purposes, the most practical way to realize the combination of political equality and deliberation is via a deliberative microcosm chosen by lot or random sampling. It embodies political equality because it gives everyone an equal chance of being selected. And it can be designed to realize good conditions for deliberation. But since most of the population is left out, it neglects mass participation.

Consider a second option, combining deliberation and mass participation. Most citizens most of the time will not spend the time and effort to deliberate in depth on the issues. Those that will take up the opportunity can be reasonably expected to be unrepresentative of the rest of the population. Imagine public meetings convened by well-meaning civic groups. Deliberation takes time and effort (and civic infrastructure to organize the opportunities). The condition of citizens in mass society renders them easily subject to the incentives for ‘rational ignorance’ on most public issues (Downs 1957). If I have one vote or one voice in millions why should I spend a lot of time and effort on some complex issue for which my individual vote, or my individual opinion, will not make any appreciable difference? Of course, some will take up the opportunity to respond to public issues. But self-selected, large-scale participation in a deliberative process will inevitably violate political equality by being unrepresentative.

The remaining option in the trilemma, combining participation and political equality, undermines deliberation. As noted, individual citizens in the large-scale nation state have little rational incentive to become informed or to spend any significant time and effort deliberating about public issues. Their individual votes or opinions will make no appreciable difference. While it is possible to have serious public discussions on important issues, discussions for which the public can provide an audience, it is far more difficult and unusual to get the bulk of the mass public actually deliberating. Perhaps Bruce Ackerman (1991) is right that there can be ‘constitutional moments’ when the public is so engaged in depth on constitutional issues that they get effective mass consent on the public’s shared understanding of its constitution. But he has only made the case for the American Founding, Reconstruction, the New Deal and most recently the civil rights movement (Ackerman 2014). The extraordinary nature of those periods illustrates that they do not occur by institutional design. And the norm is what he calls ‘normal politics’ when such mass-deliberative mobilization is mostly a distant memory.

If the engaged deliberations of the mass public are most realistically envisioned as occurring only in a specific period (a constitutional moment, which can last for more than a moment) then most people, most of the time, will be in the position of living with a constitution (or a revised constitutional interpretation) that they had no part in deliberating about. Rather they are outsiders to the process, the captive of what may be long-dead generations. In that sense they are not so different from the citizens outside the minipublic convened in a current generation. In both cases, we need not think of them as ‘blindly following’ the dictates of what others have decided. In both cases the reasons can be brought alive so that we can have a living constitution and a living deliberative polity.

If the reasoning of governing bodies like the legislature is not brought to life for the public, then we are all in the position of ‘blindly following’ the commands of government, even if decisions are made by duly constituted authorities. J. S. Mill offered an ambitious picture of how a legislature should connect its reasoning to that of the public. His ideas can extend very directly to the problem of microcosmic deliberation in a mini-public and its relations to those who are not deliberating. Mill thought that a legislature should act as what he called a ‘Congress of Opinions’:

Where every person in the country may count upon finding somebody who speaks his mind as well or better than he could speak it for himself—not to friends and partisans exclusively, but in the face of opponents, to be tested by adverse controversy; where those whose opinion is over-ruled feel satisfied that it is heard, and set aside not by a mere act of will, but for what are thought superior reasons and commend themselves as such to the representatives of the majority of the nation; where every opinion in the country can muster its strength and be cured of any illusion concerning the number or power of its adherents. (Mill 1861, 68)

The Congress of Opinions has a distribution of opinion that is like that in the country as a whole (‘where every opinion in the country can muster its strength and be cured of any illusion concerning the number or power of its adherents’). Each person can find that his perspective is advocated ‘as well or better than he could speak it for himself’ and then it is ‘tested by adverse controversy,’ by continuing dialogue in which opinions expressed are answered and, presumably, those are answered in turn in a continuing dialogue. And, finally, when conclusions are reached, those ‘whose opinion is over-ruled feel satisfied that it is heard, and set aside not by a mere act of will, but for what are thought superior reasons.’ We get a picture of a deliberative body where people are informed by the arguments of others, where there is some measure of substantive balance in the exchange of arguments, where the diversity of views is comparable to that of the society as a whole, and where the representatives are participating conscientiously and weighing the arguments on the merits.

So, the deliberative minipublic, like Mill’s vision for an ideal legislature, does not have to condemn its citizens to blindly following its conclusions. It can connect with the citizens on either side of the main disagreements and bring those positions to life in the debate.

Nevertheless, the argument is that our familiar democratic practices cannot normally be expected to realize all three values simultaneously. It may happen on rare historic occasions, such as the crises that stimulate Ackerman’s constitutional moments, but the condition of the citizen in mass society does not normally favor widespread deliberative participation.

A further word about audience democracy. If done properly, the deliberative microcosm can achieve political equality and deliberation, but it does not fulfill our remaining value—mass participation. Mass participation is a proxy for a kind of universal consent. It makes us active deliberators, not just spectators. With the deliberative microcosm, the numbers involved are miniscule fractions of the population, sometimes selected by random sampling, sometimes by quota sampling, and sometimes from self-selected groups. But all the people who are not selected do not participate, except possibly vicariously through television or the Internet or through press reports. We have some modest evidence that viewing a microcosmic deliberation, such as a Deliberative Poll, on television can have a small effect on people’s views about themselves and their political efficacy and sense of civic engagement.² But such media effects do little actually to encourage people to deliberate themselves or become more well informed.

By its very nature, microcosmic deliberation is for the few, not the many. To engage our value of mass participation alongside political equality and deliberation, we would need to engage the many. However, as we saw in the discussion of mass democracy, once we engage the many—the millions of voters in a large-scale mass society—we run into problems of rational ignorance and the lack of incentives for those who participate to also become well informed. Once again, we face an apparent forced choice in which we cannot achieve all three values simultaneously.

If we combine participation and equality, we count everyone’s views equally and we have an expression of actual mass consent through that participation. But it is not generally informed or thoughtful consent. Why should it be? It is usually the acquiescence of an inattentive and possibly manipulated public. On the other hand, if we combine equality and deliberation, we count a representation of the public’s considered judgments, but the connection to the mass of voters is only through their being equally considered via random sampling. There is no realization of actual mass participation. In the case of the remaining option, deliberation and participation, we can reliably expect ‘participatory distortion’ or a lack of equal counting. It is mostly certain groups who are especially interested who will participate. Distorted participation will almost surely lead to distorted deliberation because some voices and perspectives will be left out while others are over-represented. Under most foreseeable practical conditions, one can go round and round this trilemma and never get all three principles satisfied. We need new institutional designs to get all three. And if we do not achieve all three, we will have a democracy that is distorted by the lack of political equality, the lack of mass participation, or the lack of deliberation.

The point of my discussion in Democracy When the People Are Thinking was that the Deliberation Day idea can be made the capstone of a plan to meaningfully realize all three values simultaneously, allowing for mass participatory deliberation in a context of political equality. However, there are serious questions whether Lafont has a comparable plan. Her section on ‘participatory deliberation’ relies not on an institutional design for widespread participation in deliberation but on the theoretical right of all citizens to bring a lawsuit. The theoretical availability of the exercise of this right does not make for mass participation. Most citizens in most countries will never undertake the effort and expense of a lawsuit or other legal claim about their rights. The percentage of the mass public who participate will be miniscule, perhaps comparable to the percentage of those who participate in the minipublics she criticizes for leaving too many people out. Lafont might counter that this right applies to everyone. And I agree. But this form of inclusion does not involve any mass participation. One could even argue that the minipublic constituted by random sampling, if done well, is also inclusive. Everyone has an equal chance of being chosen—a chance that is realized by only a few just as the effort to bring a lawsuit or make some other formal legal claim is exercised by very few. In both cases one could argue that those who are left out could be stuck with following the results, perhaps blindly, perhaps with an understanding of the reasons. In that sense I believe her critique of minipublics applies as much to her own solution.

My central question about the last part of the book is whether or not Lafont’s ‘deliberative participatory’ proposal is really participatory? If so, it would seem necessary to engage a scenario by which mass participation is realized. But she does not offer any scenario, any institutional design for doing so. Most people will be barely aware of lawsuits implicating their fundamental rights. And even if they were, what would they do about them? They might discuss them. They might demonstrate. But without an account of how legal contestation can bring about mass action, this theoretical move is massively incomplete. Same-sex marriage is a successful case. But during this administration many fundamental rights of many key groups have been the subject of lawsuits and the judicial decisions do not reliably lead to their defense. Even before this administration, the same Supreme Court that Lafont cites struck down the key provisions of the Voting Rights Act of 1965 on the grounds that in its view, voting rights for blacks no longer posed a sufficiently widespread problem. The court decisions Lafont hopes will provide redress for our most fundamental rights violations may sometimes be the source of those violations. We cannot focus just on the most successful case (same-sex marriage) and ignore all the other issues that have led to unsuccessful lawsuits directed at this administration (or any other).

I am in fundamental agreement with Lafont that we need more than participation in minipublics. I express this in the book by saying we need a deliberative and participatory society (what I call in the book a ‘deliberative macrocosm’). I also think that we probably agree on some applications of Deliberative Polling. It all depends on what she means by ‘decisional-power.’ If a deliberation makes recommendations and they are adopted by duly elected authorities, that reflects, in my view, some degree of decisional-power. In the Mongolian case and more recently in the Iceland case, a Deliberative Poll made recommendations about changes in the constitution. In Mongolia some of those changes have now been adopted. In Iceland we do not know yet, but they may be. And if they are, is that a violation of Lafont’s prohibition of the minipublics having any decisional power?³

In earlier periods my work in deliberative democracy was criticized because the projects were originally just television programs and had no discernible connection to decision-making. But we quickly moved to search for entry points where our existing democratic institutions could take account of the representative and informed views of the public. This was our effort to respond to the critique that deliberation was ‘mere talk’ with no effect. If we follow Lafont’s strictures about decisional power, we would find ourselves turning the clock back to the ‘mere talk’ stage. I believe that the work around the world with Deliberative Polls and Citizens Assemblies has achieved policy impact that embody a viable path to democratic reform. Perhaps Lafont will want to clarify her critique of ‘decisional-power’ for these gatherings as many cases fall in a gray area: advice is offered by the minipublic; it is not binding but it is also taken up by other components in the deliberative system. In that way it adds to the legitimacy and responsiveness of policy making when it is followed.

Lafont compares her deliberative participatory conception to voting. But voting is not just theoretically available to everyone (assuming we avoid voter suppression). It is the most common form of actual mass participation in most modern democracies. If Deliberation Day were participated in by most voters that would also be a form of actual mass deliberative participation. But if most people do not exercise the rights Lafont identifies for legal contestation are they not in the position of non-participants just like the citizens who are not selected in the Deliberative Poll? Are they not forced to accept, perhaps blindly, the results of legal contestations that they took no part in? Lafont seems to respond by saying that further participation in mass deliberation would be ideal but not required. Does that not seem to be a form of second best? Second best is the scenario when some people participate; first best would be when nearly everyone does. But then we need a viable proposal for how to get most people deliberating, even in theory. Lafont can say that she is a philosopher not a political scientist, but this whole field is an interdisciplinary nexus. We must work together to figure out how to resolve these issues. Lafont’s critique is so wide ranging and powerful that I fear it applies as much to her own solution as to the work of the rest of us. I look forward to further iterations of this dialogue where we can attempt to resolve these issues together. In the meantime, deliberative democrats, read this book!

Competing Interests

The author has no competing interests to declare.