The Prospects & Limits of Deliberative Democracy

Home , Anne Norton, Arend Lijphart, Deliberative democracy, Jane Mansbridge, Majority judgment

Summer 2017

Ian Shapiro Simon Beste · · Nicole Curato Selen A. Ercan · William Bazeyo · Mark E. Warren Simon Niemeyer · · · Hélène Landemore · Roy William Mayega Roy Nathan Tumuhamye · · Anne Norton Anne Norton · Baogang He André Bächtiger John S. Dryzek John S. with Claus Offe with Claus Offe · Julius Ssentongo Bernard Manin Lynn Atuyambe Atuyambe Lynn Cass R. Sunstein Journal of the American Academy of Arts & Sciences Journal of the American Academy · Carolyn M. Hendriks Carolyn Deliberative Democracy Deliberative Arthur Lupia The Prospects & Limits of The Prospects Dædalus Cristina Lafont Cristina Lafont Alice Siu James S. Fishkin & Jane Mansbridge, guest editors Mansbridge, Fishkin & Jane James S.

Dædalus Summer 2017 The Prospects & Limits of Deliberative Democracy knowledge and issues of public importance. explores the frontiers of the frontiers explores Dædalus and diversity, Representing the intellectual community in its breadth Representing edited by Karl Eikenberryedited & Stephen Krasner Bruce Fearon, Stephen John Jones & D. with James Paul H. Wise Stedman, Stewart Martha Patrick, Crenshaw, Vanda Lischer, Kenyon Sarah & Michele Barry, Felbab-Brown, Hendrik Spruyt, Will Reno, Biddle, Stephen & MiguelAila M. Matanock García-Sánchez, and Barry Posen Eikenberryedited by Karl Krasner & Stephen Kalyvas, Stathis N. M. Fazal, Tanisha Fukuyama, Francis Call & Susanna Campbell, Steven Heydemann, Chuck Clare Lockhart, Risse & Thomas Sumit Ganguly, Sonja Grimm, & A. Börzel Eric Stollenwerk, Tanja Nancy Beyene, Lindborg & Josephn & Abdeta Mesfi Seyoum Hewitt, John Stedman, Richard Gowan & Stephen Doucet, and Jean-Marie Guéhenno Lyse Lomawaima, K. Tsianina edited by Ned Blackhawk, Deloria, Philip J. Bryan McKinley Jones Brayboy, Douglas Medin, and Mark Trahant Loren Ghiglione, Ending Civil Wars: Constraints & Possibilities & Possibilities Constraints Wars: Ending Civil Native Americans & Academia on the horizon: on the & Opportunities Disorder: Threats & Global Civil Wars @americanacad U.S. $15; www.amacad.org; $15; www.amacad.org; U.S.

Dædalus Journal of the American Academy of Arts & Sciences

“The Prospects & Limits of Deliberative Democracy” Volume 146, Number 3; Summer 2017

James S. Fishkin & Jane Mansbridge, Guest Editors Phyllis S. Bendell, Managing Editor and Director of Publications Peter Walton, Assistant Editor Heather Mawhiney, Senior Editorial Assistant

Committee on Studies and Publications John Mark Hansen and Jerrold Meinwald, Cochairs; Bonnie Bassler, Rosina Bierbaum, Marshall Carter, Gerald Early, Carol Gluck, Linda Greenhouse, John Hildebrand, Jerome Kagan, Philip Khoury, Arthur Kleinman, Sara Lawrence-Lightfoot, Rose McDermott, Jonathan F. Fanton (ex of½cio), Don M. Randel (ex of½cio), Diane P. Wood (ex of½cio)

Inside front cover: Pieter van der Borcht’s The Difficulty of Ruling over a Diverse Nation (1578), adapted to appear as The Difficulty of Deliberating in a Diverse Nation–Difficult, But Today Not Impossible. Contents

6 Introduction James S. Fishkin & Jane Mansbridge

Engaging the Issue

14 Referendum vs. Institutionalized Deliberation: What Democratic Theorists Can Learn from the 2016 Brexit Decision Claus Offe

28 Twelve Key Findings in Deliberative Democracy Research Nicole Curato, John S. Dryzek, Selen A. Ercan, Carolyn M. Hendriks & Simon Niemeyer

New Thinking

39 Political Deliberation & the Adversarial Principle Bernard Manin

51 Deliberative Democracy as Open, Not (Just) Representative Democracy Hélène Landemore

Critics

64 Inequality is Always in the Room: Language & Power in Deliberative Democracy Arthur Lupia & Anne Norton

77 Collusion in Restraint of Democracy: Against Political Deliberation Ian Shapiro

85 Can Democracy be Deliberative & Participatory? The Democratic Case for Political Uses of Mini-Publics Cristina Lafont Responses

106 Deliberative Citizens, (Non)Deliberative Politicians: A Rejoinder André Bächtiger & Simon Beste

119 Deliberation & the Challenge of Inequality Alice Siu

Applications

129 Deliberative Democracy in the Trenches Cass R. Sunstein

140 Applying Deliberative Democracy in Africa: Uganda’s First Deliberative Polls James S. Fishkin, Roy William Mayega, Lynn Atuyambe, Nathan Tumuhamye, Julius Ssentongo, Alice Siu & William Bazeyo

155 Authoritarian Deliberation in China Baogang He & Mark E. Warren Dædalus

Journal of the American Academy of Arts & Sciences

Nineteenth-century depiction of a Roman mosaic labyrinth, now lost, found in Villa di Diomede, Pompeii

Dædalus was founded in 1955 and established as a quarterly in 1958. The journal’s namesake was renowned in ancient Greece as an inventor, scientist, and unriddler of riddles. Its emblem, a maze seen from above, symbolizes the aspiration of its founders to “lift each of us above his cell in the labyrinth of learning in order that he may see the entire structure as if from above, where each separate part loses its comfortable separateness.” The American Academy of Arts & Sciences, like its journal, brings together distinguished individuals from every ½eld of human endeavor. It was char- tered in 1780 as a forum “to cultivate every art and science which may tend to advance the interest, honour, dignity, and happiness of a free, independent, and virtuous people.” Now in its third century, the Academy, with its more than ½ve thousand members, continues to provide intellectual leadership to meet the critical challenges facing our world. Dædalus Summer 2017 Subscription rates: Electronic only for non- Issued as Volume 146, Number 3 member individuals–$50; institutions–$137. Canadians add 5% gst. Print and electronic © 2017 by the American Academy for nonmember individuals–$55; institutions– of Arts & Sciences $153. Canadians add 5% gst. Outside the United Introduction States and Canada add $24 for postage and han- © 2017 by James S. Fishkin & Jane Mansbridge dling. Prices subject to change without notice. Referendum vs. Institutionalized Deliberation: Institutional subscriptions are on a volume-year What Democratic Theorists Can Learn from the basis. All other subscriptions begin with the 2016 Brexit Decision next available issue. © 2017 by Claus Offe Political Deliberation & the Adversarial Principle Single issues: $15 for individuals; $38 for insti- © 2017 by Bernard Manin tutions. Outside the United States and Canada, Collusion in Restraint of Democracy: add $6 per issue for postage and handling. Prices Against Political Deliberation subject to change without notice. © 2017 by Ian Shapiro Claims for missing issues will be honored free Can Democracy Be Deliberative & Participatory? of charge if made within three months of the The Democratic Case for Political Uses of Mini-Publics publication date of the issue. Claims may be © 2017 by Cristina Lafont submitted to [email protected]. Members of Deliberation & the Challenge of Inequality the American Academy please direct all ques- © 2017 by Alice Siu tions and claims to [email protected]. Applying Deliberative Democracy in Africa: Uganda’s First Deliberative Polls Advertising and mailing-list inquiries may be © 2017 by James S. Fishkin, Roy William addressed to Marketing Department, mit Press Mayega, Lynn Atuyambe, Nathan Tumuhamye, Journals, One Rogers Street, Cambridge ma Julius Ssentongo, Alice Siu & William Bazeyo 02142-1209. Phone: 617 253 2866. Fax: 617 253 1709. Email: [email protected]. Editorial of½ces: Dædalus, American Academy of Arts & Sciences, 136 Irving Street, Cambridge ma To request permission to photocopy or repro- 02138. Phone: 617 576 5085. Fax: 617 576 5088. duce content from Dædalus, please complete the Email: [email protected]. online request form at http://www.mitpress journals.org/page/permissionsForm.jsp, or con- Library of Congress Catalog No. 12-30299. tact the Permissions Manager at mit Press Jour Dædalus publishes by invitation only and assumes nals, One Rogers Street, Cambridge ma 02142- no responsibility for unsolicited manuscripts. 1209. Fax: 617 253 1709. Email: journals-rights@ The views expressed are those of the author(s) of mit.edu. each article, and not necessarily of the American Corporations and academic institutions with Academy of Arts & Sciences. valid photocopying and/or digital licenses with Dædalus (issn 0011-5266; e-issn 1548-6192) is the Copyright Clearance Center (ccc) may re published quarterly (winter, spring, summer, fall) produce content from Dædalus under the terms of by The mit Press, One Rogers Street, Cambridge their license. Please go to www.copyright.com; ma 02142-1209, for the American Academy of ccc, 222 Rosewood Drive, Danvers ma 01923. Arts & Sciences. An electronic full-text version Printed in the United States by The Sheridan of Dædalus is available from The mit Press. Press, 450 Fame Avenue, Hanover, 17331. Subscription and address changes should be ad pa dressed to mit Press Journals Customer Service, Newsstand distribution by Ingram Periodicals One Rogers Street, Cambridge ma 02142-1209. Inc., 18 Ingram Blvd., La Vergne tn 37086. Phone: 617 253 2889; U.S./Canada 800 207 8354. Fax: 617 577 1545. Email: [email protected]. Postmaster: Send address changes to Dædalus, One Rogers Street, Cambridge ma 02142-1209. The typeface is Cycles, designed by Sumner Periodicals postage paid at Boston ma and at Stone at the Stone Type Foundry of Guinda ca. additional mailing of½ces. Each size of Cycles has been separately designed in the tradition of metal types. Introduction

James S. Fishkin & Jane Mansbridge

Democracy is under siege. Approval ratings for democratic institutions in most countries around the world are at near-record lows. The number of recognized democratic countries in the world is no longer expanding after the so-called Third Wave of democratic transitions.1 Indeed, there is something of a “democratic recession.”2 Further, some apparently democratic countries with competitive elections are undermining elements of liberal democracy: the rights and liberties that ensure freedom of thought and expression, protection of the rule of law, and all the protections for the substructure of civil society that may be as important for making democracy work as the JAMES S. FISHKIN, a Fellow of the electoral process itself.3 The model of party compe- American Academy since 2014, is tition-based democracy–the principal model of de- Director of the Center for Delib- mocracy in the modern era–seems under threat. erative Democracy, the Janet M. That model also has competition. What might be Peck Chair in International Com- munication, Professor of Commu- called “meritocratic authoritarianism,” a model in nication, and Professor of Political which regimes with flawed democratic processes nev- Science (by courtesy) at Stanford ertheless provide good governance, is attracting at- University. tention and some support. Singapore is the only suc- JANE MANSBRIDGE, a Fellow of cessful extant example, although some suggest China the American Academy since 1994, as another nation moving in this direction. Singapore is the Charles F. Adams Professor is not a Western-style party- and competition-based of Political Leadership and Dem- democracy, but it is well-known for its competent civil ocratic Values at the Harvard Ken- servants schooled in making decisions on a cost-ben- nedy School. efit basis to solve public problems, with the goals set (*See endnotes for complete contributor by elite consultation with input from elections rath- biographies.) er than by party competition.

6 Public discontent makes further difficul- Over the last two decades, another ap- James S. ties for the competitive model. Democra- proach to democracy has become increas- Fishkin & Jane cies around the world struggle with the ap- ingly prominent. Based on greater deliber- Mansbridge parent gulf between political elites who are ation among the public and its represen- widely distrusted and mobilized citizens tatives, deliberative democracy has the who fuel populism with the energy of an- potential, at least in theory, to respond gry voices. Disillusioned citizens turning to today’s current challenges. If the many against elites have produced unexpected versions of a more deliberative democracy election results, including the Brexit deci- live up to their aspirations, they could help sion and the 2016 U.S. presidential election. revive democratic legitimacy, provide for The competitive elections and referenda more authentic public will formation, proof most current democracies depend on vide a middle ground between widely mis- mobilizing millions of voters within a con- trusted elites and the angry voices of pop- text of advertising, social media, and efforts ulism, and help fulfill some of our common to manipulate as well as inform public opin- normative expectations about democracy. ion. Competing teams want to win and, in Can this potential be realized? In what most cases, are interested in informing vot- ways and to what extent? Deliberative de- ers only when it is to their advantage. The mocracy has created a rich literature in both rationale for competitive democracy, most theory and practice. This issue of Dædalus influentially developed by the late econo- assesses both its prospects and limits. We mist Joseph Schumpeter, held that the same include advocates as well as critics. As de- techniques of advertising used in the com- liberative democrats, our aim is to stimu- mercial sphere to get people to buy prod- late public deliberation about deliberative ucts can be expected in the political sphere. democracy, weighing arguments for and On this view, we should not expect a “gen- against its application in different contexts uine” public will, but rather “a manufac- and for different purposes. tured will” that is just a by-product of po- How can deliberative democracy, if it litical competition.4 were to work as envisaged by its supporters, Yet the ideal of democracy as the rule of respond to the challenges just sketched? “the people” is deeply undermined when First, if the more-deliberative institutions the will of the people is in large part manu- that many advocate can be applied to real factured. The legitimacy of democracy de- decisions in actual ongoing democracies, pends on some real link between the public arguably they could have a positive effect on will and the public policies and office-hold- legitimacy and lead to better governance. ers who are selected. Although some have They could make a better connection be- criticized this “folk theory of democracy” tween the public’s real concerns and how as empirically naive, its very status as a folk they are governed. Second, these institu- theory reflects how widespread this nor- tions could help fill the gap between dis- mative expectation is.5 To the extent that trusted elites and angry populists. Elites leaders manufacture the public will, the are distrusted in part because they seem normative causal arrow goes in the wrong and often are unresponsive to the public’s direction. If current democracies cannot concerns, hopes, and values. Perhaps, the produce meaningful processes of public suspicion arises, the elites are really out will formation, the legitimacy claims of for themselves. On the other hand, pop- meritocratic autocracies or even more ulism stirs up angry, mostly nondelibera- fully autocratic systems become compar- tive voices that can be mobilized in plebes- atively stronger.6 citary campaigns, whether for Brexit or for

146 (3) Summer 2017 7 Introduction elected office. In their contributions to this These are some of the challenges facing issue, both Claus Offe and Hélène Lande- those who might try to make deliberative more explore the crisis of legitimacy in democracy practical. representative government, including the The earliest work on deliberative democ- clash between status quo–oriented elites racy began by investigating legislatures.7 In and populism. Deliberative democratic this issue, Cass Sunstein, in contrast, looks methods open up the prospect of prescrip- at deliberation among policy-makers with- tions that are both representative of the in the executive branch. Bernard Manin entire population and based on sober, evi- looks outside government toward debates dence-based analysis of the merits of com- and public forums that can improve the de- peting arguments. Popular deliberative in- liberative quality of campaigns and discus- stitutions are grounded in the public’s val- sions among the public at large. ues and concerns, so the voice they magnify Much of the energy in deliberative de- is not the voice of the elites. But that voice mocracy efforts has focused on statisti- is usually also, after deliberation, more ev- cal microcosms or mini-publics, in which idence-based and reflective of the merits of citizens, usually recruited by random sam- the major policy arguments. Hence these pling, deliberate in organized settings. In institutions fill an important gap. some settings, relatively small groups of fif- How might popular deliberative democ- teen or so deliberate online with an elect- racy, if it were to work as envisaged by its ed representative.8 In other settings, the supporters, fulfill normative expectations groups can be given access to balanced of democracy, thought to be unrealistic information and briefing materials that by critics of the “folk theory”? The issue make the best case for and against various turns on the empirical possibility that the options. They can also be given access to public can actually deliberate. Can the peo- competing experts who answer their ques- ple weigh the trade-offs? Can they assess tions from different points of view. Then, competing arguments? Can they connect at the end of the deliberations in these or- their deliberations with their voting pref- ganized settings, there is some way of har- erences or other expressions of preference vesting their considered judgments. Sever- about what should be done? Is the problem al of the essays discuss Deliberative Polling, that the people are not competent, or that which brings together a random sample of they are not in the right institutional con- citizens for a weekend of deliberation and text to be effectively motivated to partici- gathers data, as in an opinion poll, from pate? These are empirical questions, and the random samples both upon recruit- the controversies about them are part of ment and then again at the end of the de- our dialogue. liberations. The method also permits qualitative data by recording the discussions, This issue includes varying definitions, both in moderated small groups and in ple- approaches, and contexts. The root notion nary sessions where questions generated in is that deliberation requires “weighing” the small groups are directed at experts rep- competing arguments for policies or candi- resenting different points of view. Other dates in a context of mutually civil and di- mini-publics, such as “citizens’ juries” verse discussion in which people can decide and “consensus conferences,” are usually on the merits of arguments with good in- smaller (a couple of dozen instead of two or formation. Is such a thing possible in an era three hundred people) and arrive at some- of fake news, social media, and public dis- thing like an agreed-upon statement or ver- cussions largely among the like-minded? dict as a recommendation to the public or

8 Dædalus, the Journal of the American Academy of Arts & Sciences to authorized policy-makers. Some ran- necting the people to policy-making. Nicole James S. domly selected mini-publics even make Curato, John Dryzek, Selen Ercan, Carolyn Fishkin 9 & Jane binding decisions. Hendriks, and Simon Niemeyer offer a sys- Mansbridge The basic rationale for the mini-public tematic overview of what they regard as the approach is that if the random sample that key findings of the deliberative democra- is gathered to deliberate is representative cy research around the globe. Their find- of the population, and if it deliberates un- ings are optimistic and differ from some of der good conditions, then its considered the critical perspectives presented later in judgments after deliberation should rep- the issue. resent what the larger population would The second group of essays might be la- think if somehow those citizens could en- beled “new thinking.” Bernard Manin pro- gage in similarly good conditions for con- poses that the core of deliberation is cap- sidering the issue. A great deal depends on tured by what he calls the “adversarial the mini-public actually being representa- principle,” according to which public dis- tive and on the account of good conditions cussions should be organized to allow a to which it is exposed. “confrontation of opposing positions.” Im- Whenever an application of delibera- plementing this idea is more complex than tive democracy depends on a randomly se- first appears and has a history going back to lected mini-public, that application raises Ancient Athenian institutions. Manin of- the issue of degree of empowerment. Can fers various suggestions, including some for or should such mini-publics supplant de- modern televised debates. Hélène Lande- mocracy by competitive elections? No con- more asks whether deliberative democracy tributor to this issue makes that argument. can be saved from the current crisis of rep- But in several cases, duly appointed admin- resentative democracy around the world. istrators have committed in advance to im- Her positive answer depends on an ambi- plementing the recommendations of such a tious sketch of an “open democracy,” in mini-public and, in some cases, those rec- which institutions would be inclusive and ommendations are binding. How much can power accessible to ordinary citizens, in- randomly selected groups be relied upon for cluding through representation in delibera- authoritative public decisions and in what tive bodies of randomly chosen citizens, cit- ways? Cristina Lafont argues against re- izens’ initiatives, and crowd-sourced law- lying solely on such groups for decisions, making and policy processes. but opens the door to discussions of a pos- The next two groups of essays alternate- sible albeit limited role for them. She use- ly present and respond to some of the main fully poses the problem from the perspec- criticisms of deliberative democracy. Ar- tive of the vast majority of citizens who will thur Lupia and Anne Norton argue in their not be in a mini-public: how do the deliber- elegant phrasing that “inequality is always ations connect with them if they have not in the room.” If the outcome of delibera- deliberated? tion is inevitably distorted by the more advantaged participants dominating the dis- The essays are organized roughly in five cussions, the results are not likely to repre- groups. To introduce the topic of deliber- sent the true views of the rest of the group. ative democracy, Claus Offe sketches the Rather, any such results would reproduce conflict between distrusted elites and the the inequalities and power relations among populism of Brexit and other plebiscitary the participants. Inequality among partic- processes, arguing that deliberation via ran- ipants is one of the major challenges to the dom sampling could help fill the void, con- larger idea of implementing deliberative

146 (3) Summer 2017 9 Introduction democracy–a challenge that must be pur- some parliamentary contexts, particularly sued with great seriousness. those characterized by “coalition settings, Responding to critics of deliberation, second chambers, secrecy, low party disci- Alice Siu reflects on the role of inequality pline, low issue polarization, and the strong using data from Deliberative Polls, both presence of moderate parties.” Their insti- online and face-to-face, finding far less dis- tutional prescription for parliament con- tortion than critics expect. She also offers trasts sharply with Shapiro’s. Regarding surprising findings on who takes the most public deliberation, they draw on Europolis, talking time, who has the greatest influence a European-wide Deliberative Poll with a on the outcomes, and who offers more “jus- sample of ordinary citizens, and provide ev- tified” arguments, supplying reasons for idence that the citizens were able to reason their positions. But this is an ongoing em- in ways comparable to those of the parlia- pirical question. No one has yet systematic- mentarians. ally studied the role of inequality under dif- In her essay, Cristina Lafont makes a ferent deliberative designs. More research case against giving any decisional status with controlled experiments could clarify to mini-publics. Although she grants that this issue further. deliberating mini-publics may make rea- Ian Shapiro robustly defends the model sonable decisions when the participants of competitive democracy as the alterna- have considered the options in good con- tive to deliberative democracy. He believes ditions, to grant them power over decisions that through party competition we can fos- on this basis would be to give “blind defer- ter an “argumentative ideal” that has ele- ence” to a “special version of elite concep- ments of deliberation, but does not suffer tions of democracy.” On the representative- from either the lack of realism of the delib- ness argument for granting them power, the erative model or the potential veto power public might think that the participants in of intense minorities that emerges when a mini-public “share our interests, val- consensus is the decision rule or goal. He ues, and policy objectives,” so their views champions an argumentative version of the will “coincide with what we would have Westminster two-party competition mod- thought if we had participated.” Yet most el in which each side must make its case. He larger mini-publics (including those that also criticizes the room for deliberation of- collect post-deliberative opinions in con- fered in multiparty proportional represen- fidential questionnaires) are not designed tation systems, in an argument that con- to produce consensus. In this respect, they trasts with the position offered by André differ from the model of deliberation most Bächtiger and Simon Beste in their contri- criticized by Shapiro. Hence there is almost bution to this issue. always, at least in the larger mini-publics, a Bächtiger and Beste contest the “standard majority view and a minority view revealed argument that politicians do not want to de- in the final confidential questionnaires or liberate and citizens are not able to.” They vote. Lafont argues that an individual voter draw on extensive empirical work with the who has not participated cannot be sure “Discourse Quality Index,” which exam- whether she would have been in the ma- ines the reasoning offered by deliberators jority or in the minority after deliberation. in legislatures, especially on the question of Why should she be bound by the majority whether they offer justifications for their as- view post-deliberation if she might have sertions. They find that, despite the current come out with the minority view? cynicism about representative democracy, No essay in this issue stands as an explic- room for genuine deliberation appears in it response to Lafont, as we fortuitously

10 Dædalus, the Journal of the American Academy of Arts & Sciences had for the first two critics. So we will try high-quality deliberation. In this case, how- James S. to respond here by asking: if the citizens in ever, high-level policy-makers, rather than Fishkin & Jane the broader public believe in democracy, the people themselves or their elected rep- Mansbridge then why might they not take as serious- resentatives, are doing the deliberating. ly the recommendations of deliberating James Fishkin, Roy William Mayega, Lynn majorities as they do the decisions of non- Atuyambe, Nathan Tumuhamye, Julius deliberating majorities? In a deliberating Ssentongo, Alice Siu, and William Bazeyo mini-public, the final reported views are examine the first Deliberative Polls in Af- what the people in microcosm concluded rica. Those skeptical of the capacity of ran- on the basis of in-depth deliberation. If a domly selected bodies to make intelligent decision is taken on the basis of the major- decisions have assumed that if such proce- ity after deliberation, there will certainly dures are viable at all, they must apply only be dissenters, as with any majority decision. or primarily in developed countries with Much depends on what we mean by the highly educated populations. Can these public taking the results seriously. Lafont methods be applied to populations with argues forcefully against any trust-based low literacy and very low educational lev- argument that might suggest “blind def- els? Can the people in such communities erence” to the majority in a randomly se- reason usefully about the trade-offs of ma- lected mini-public. Perhaps, however, duly jor policy choices affecting their commu- elected officials might delegate some re- nities? Can they do so in ways useful for sponsibility to such a group. How much policy? The difficult issues of disaster re- decisional status should the recommenda- lief and population pressure in rural Ugan- tions of a mini-public have? Should these da pose a test case for the question: who mini-publics be an official part of a decision can deliberate? In these first African De- process or only part of the dialogue in the liberative Polls, random sampling and de- public sphere? Are there contexts in which liberation allowed the people who must live they could bear the full weight of an institu- with development policies to be consulted, tional decision? The question of role poses with reasonable results, even in such diffi- a central challenge for deliberations based cult conditions. on mini-publics. In the final essay of the issue, Baogang He The final section focuses on applications. and Mark Warren look outside the purview The essays shed light on the questions: who of competitive democratic systems to ask deliberates, and in what context? As Cass whether the practice of deliberative democ- Sunstein notes, the term deliberative democracy may be feasible within authoritarian racy was coined in a study of how delibera- regimes, such as China. They ask: why have tion took place in the Senate, in ways that, some Chinese authorities embraced and to some degree, matched how the Consti- supported the form of a randomly selected tution’s framers thought the Senate ought mini-public for “grass roots experimen- to act.10 Deliberation is a crucial part of tation” for local government decisions? government in the executive and judicial Can deliberating mini-publics be properly branches. Sunstein distills his experience conducted for budget and other local deci- in government to offer a compelling picture sions in a society that lacks the civil liber- of deliberation taking place within the pol- ties and individual rights familiar in com- icy teams grappling with interagency issues petitive democracies? What are the effects and the production of good policy in the ex- and prospects of what they call “delibera- ecutive branch of the U.S. government. His tive authoritarianism?” Will such experi- account seems to satisfy all the criteria for mentation lead to further institutional de-

146 (3) Summer 2017 11 Introduction velopment in line with democratic values reform mass politics and public debate to or will it simply serve to legitimate current avoid not only fake news, but also the in- power relations and institutions, preclud- creasing pressures of narrow-casting in the ing long-term reform? commercial media, self-sorting into information bubbles on social media, and geo- This issue examines a wide range of de- graphic sorting by ideology as people move liberative democratic practices and appli- to more politically homogeneous commu- cations. It includes competitive democra- nities. It should leave the reader asking: cies, authoritarian regimes, and developed What challenges and critiques are most and developing countries. It opens up de- telling for deliberative democracy? How bates on how to improve deliberation in serious are the ways in which deliberation legislatures and other governmental bod- can go awry? Whatever conclusions our ies, and on what institutional roles and de- readers reach on these questions, this is- cision power randomly selected citizens sue depicts a vibrant area of democratic ex- might have after they have been able to dis- perimentation at a time when many have cuss issues in some depth under good con- lost confidence in the processes of electoral ditions. It asks how we might effectively representative democracy.

endnotes * Contributor Biographies: JAMES S. FISHKIN, a Fellow of the American Academy since 2014, is Director of the Center for Deliberative Democracy, the Janet M. Peck Chair in International Com- munication, Professor of Communication, and Professor of Political Science (by courtesy) at Stan- ford University. He is the author of When the People Speak: Deliberative Democracy and Public Consulta- tion (2009), Deliberation Day (with Bruce Ackerman, 2004), and The Voice of the People: Public Opinion and Democracy (1995). JANE MANSBRIDGE, a Fellow of the American Academy since 1994, is the Charles F. Adams Professor of Political Leadership and Democratic Values at the Harvard Kennedy School. She is the author of Why We Lost the ERA (1986) and Beyond Adversary Democracy (1983), and coeditor of Political Negotiation (with Cathie Jo Martin, 2015) and Deliberative Systems (with John Parkin- son, 2012). 1 Samuel P. Huntington, The Third Wave: Democratization in the Late Twentieth Century (Oklahoma City: University of Oklahoma Press, 1991). 2 Larry Diamond, “Facing Up to Democratic Recession,” Journal of Democracy 26 (1) (January 2015): 141–155. 3 Steven Levitsky and Lucan A. Way, Competitive Authoritarianism: Hybrid Regimes After the Cold War (Cambridge: Cambridge University Press, 2010). 4 Joseph A. Schumpeter, Capitalism, Socialism and Democracy (New York: Harper and Row, 1942), 263. 5 Christopher H. Achen and Larry M. Bartels, Democracy for Realists: Why Elections Do Not Produce Responsive Government (Princeton, N.J.: Princeton University Press, 2015). 6 For the legitimacy of the distortions from economic inequality, see Martin Gilens, Affluence and Influence: Economic Inequality and Political Power in America (Princeton, N.J.: Princeton University Press, 2014). 7 The work began with Joseph Bessette, The Mild Voice of Reason: Deliberative Democracy and American National Government (Chicago: University of Chicago Press, 1994), followed by Amy Gutmann and Dennis Thompson, Democracy and Disagreement (Cambridge, Mass.: Harvard University

12 Dædalus, the Journal of the American Academy of Arts & Sciences Press, 1996); Amy Gutmann and Dennis Thompson, Why Deliberative Democracy? (Princeton, James S. N.J.: Princeton University Press, 2004); Jürg Steiner, André Bächtiger, Markus Spörndli, and Fishkin Marco R. Steenbergen, Deliberative Politics in Action: Analyzing Parliamentary Discourse (Cambridge: & Jane Cambridge University Press, 2004); and further work inspired by this Swiss team and others. Mansbridge 8 See, for example, Michael A. Neblo, Kevin M. Esterling, Ryan P. Kennedy, David M. J. Lazer, and Anand E. Sokhey, “Who Wants to Deliberate–And Why?” American Political Science Review 104 (3) (August 2010): 1–18. 9 On citizens’ juries, see the work of Ned Crosby and Peter Dienel. On binding decisions and other features of a variety of randomly selected mini-publics, see Yves Sintomer, Petite histoire de l’expérimentation démocratique: Tirage au sort et politique d’Athènes à nos jours (Paris: La Découverte, 2011). 10 Bessette, The Mild Voice of Reason.

146 (3) Summer 2017 13 Referendum vs. Institutionalized Deliberation: What Democratic Theorists Can Learn from the 2016 Brexit Decision

Claus Offe

Abstract: This essay proceeds in three steps. First, it will briefly outline the often invoked “crisis” of representative democracy and its major symptoms. Second, it will discuss a popular yet, as I shall argue, wor- ryingly misguided response to that crisis: namely, the switch to plebiscitarian methods of “direct” democracy, as advocated, for example, by rightist populist forces in many European Union member states. The United Kingdom’s Brexit referendum of June 2016 illuminates the weaknesses of this approach. Third, it will suggest a rough design for enriching representative electoral democracy with nonelectoral (but “aleatory,” or randomized) and nonmajoritarian (but deliberative and consultative) bodies and their pecu- liar methods of political will formation (as opposed to the expression of a popular will already formed).

One core question of political theory is how best to make collectively binding decisions: who should make those decisions, and by what rules and procedures? The modalities of decision-making are not just something to be determined at the founding, or “constitutional” moment, of a political community once and for all times by some pouvoir constituant (constituent power). The question of whether our rules and procedures are still “good enough” or whether they are in need of amendments and adjustments is an ongoing challenge in the background of any political pro- CLAUS OFFE, a Foreign Honorary cess, and certainly one that qualifies as democratic. Member of the American Academy Yet how should we decide how to decide? The dif- since 1995, is Professor Emeritus ficulty of any conceivable answer to this question of Political Sociology at the Hertie derives from its tricky recursive logic. The answer, School of Governance in Germa- in order to be recognized as valid and binding, must ny. He is author of Europe Entrapped (2015), Varieties of Transition: The itself be decided upon–but how and by whom? If East European and East German Ex- we were able to deduce the “right” mode of deci- perience (1997), and Modernity and sion-making from a robust theory of a divine order, the State: East, West (1996). as in an ideal-typical theocratic regime, the problem

14 would go away. Conversely, if we had a sci- are quintessential political systems “on the Claus entific theory about whose decision-mak- move,” driven by the legitimacy of rule and Offe ing competencies and methods would yield its effectiveness. optimal policy results and rational problem In the course of the last forty years of the- solutions (as was the claim of “scientific” oretical self-reflection and empirical ob- state socialism), the problem of deciding servation of the stability, modes of oper- how to decide would also evaporate and ation, and trajectories of change of liberal the one best way of running a country and representative democracies, many propo- its economy would reveal itself beyond any sitions have been advanced that converge doubt. Given the modern obsolescence of on the diagnosis of a “crisis,” or the creep- either of these certainties, we need to face ing deformation, of liberal representative the fact that neither constitutional meth- democracy. This multifaceted crisis exists ods of arriving at decisions nor the resulting in the absence of explicitly nondemocratic decisions themselves (that is, policies) are (totalitarian, theocratic, or otherwise au- capable of having unquestionable validity. thoritarian) countermodels and theoreti- At best, political procedures can be consis- cal doctrines of how political rule should be tent with widely shared normative premises conducted. To oversimplify: The vast ma- of fairness, and policy outcomes can be re- jority of contemporary mankind believes in grettable–or not. and endorses (some version of the above) democratic principles and promises.1 At the Any account of what we mean by liberal same time, large minorities and sometimes representative democracy will, rather un- majorities of inhabitants of existing liberal controversially, include the following fea- democracies are dissatisfied with, and feel tures: Liberal democracy is a political sys- left out by or alienated from, the democratic tem applying (at least, so far) only to nation- routines and practices they experience. We states and their subnational territorial com- may thus say that abstract liberal democ- ponents. The right to rule derives, directly racy is celebrating its near-global victory, or indirectly, from periodic and contested while concrete and existing democracies elections through which the composition of are widely looked at with discontent and legislative assemblies and governments is frustration over failures of both the legiti- determined. It is premised upon the dichot- macy and effectiveness of democratic rule. omy between rulers and ruled, or (elected) More specific, liberal democracies of the elites and (voting) nonelites. Citizens, re- Organisation for Economic Co-operation gardless of other resources they control, en- and Development have experienced symp- joy equal political rights and freedoms (vot- toms of stress and malfunctioning over the ing, communication, association) as a mat- last generation that have activated a glob- ter of constitutional guarantee. Rule of law al discourse of political theorists and prac- and division of powers constrain the use of titioners to suggest innovative remedies. state power and its monopolistic exercise, What are the deficiencies or illnesses to thus making its use at least minimally ac- which these remedies are targeted? To gen- countable. As an empirical generalization, eralize, symptoms of this dissatisfaction in- we can add that democracies are constant- clude the following. ly challenged and self-scrutinizing politi- 1) Apathy and other forms of nonpar- cal systems that face on-going controver- ticipation and political alienation are on sial demands for their own revision, devel- the rise and are undermining the increas- opment, and improvement. Democracies ingly nominal equality of political rights. are continuously being renegotiated. They The least advantaged strata of populations

146 (3) Summer 2017 15 What (by education, economic, and class status, minorities, and everything “foreign,” in- Democratic and also by age, gender, and minority sta- cluding, in the eu context, Brussels as the Theorists Can Learn tus) show the strongest features of (self-) location of its executive branch. The kind from the exclusion. As many people in these cate- of social protection populists offer derives 2016 Brexit Decision gories do not vote or participate through not from constituted state power to achieve membership in parties and other formal collective goals through policies, but from organizations, a vicious cycle is set in mo- territorial borders of nation-states. Popu- tion by which elites of such organizations list movements and parties are, in many find little strategic incentive to respond to cases, not instrumentally focused on poli- the interests and values of the marginalized cy, but expressively focused on the politics groups. At the upper end of the socioeco- of protest, obstruction, and the assertion nomic hierarchy, investors, financial insti- of some kind of identity against a distrust- tutions, employers, and a host of organized ed “establishment” and political class, as interests enjoy de facto privileges of shap- well as minorities and foreign or suprana- ing political agendas and constraining the tional powers. They also focus on “strong” resources that elected governments have leaders whose space of action must not be available for the conduct of policies. unduly constrained by liberal constitution- 2) Political parties and elites have suffered al and other inhibitions, thus giving rise to from a rapid loss of trust concerning both the oxymoronic phenomenon of illiber- their willingness and ability to respond to al democracy and more-or-less soft forms nonelites and to promote desired kinds of of electoral authoritarianism. Its preferred social and economic change. The “monito- form of legitimation (of both leaders and ry” tactics of commercial and social media, policies) is by reference to plebiscitarian with their “gotcha” incentives, further dis- acclamation and referenda, which allegedly credit elites. As major socioeconomic prob- are best suited to reveal the true, authentic, lems (such as low growth, precariousness unified, and uncorrupted will of the people of employment, widening inequality, so- –a will that, in reality, is often but a mere cial exclusion, and international conflicts) artifact of media and party campaigns con- have come to be seen as beyond the reach of fronting the “establishment,” foreign forces, any conceivable government, the perceived and minorities. political purchasing power of the ballot de- 4) The space left to maneuver for govern- clines. In many cases, the parameters set by ing elites, and hence the extent to which the political economy of capitalist democ- they can relate at all responsively to pop- racies have enforced a convergence of major ular interests and demands, is increasing- political parties that makes them virtually ly limited by the international political indistinguishable in terms of programs and economy (globalization) with its neolib- ideology. The result tends to be restricting eral imperatives of competitiveness, aus- competition to the appeal of leading per- terity, debt consolidation, and tax compe- sonalities. tition, giving rise to a condition now often 3) If political mobilization and contes- described as “postdemocracy.” Parameters tation occur at all, they do so, to a rapidly that determine peoples’ life chances and liv- growing extent, in rightist populist ways: by ing conditions–whether in their roles as appeals not to shared interests or some ver- workers, consumers, savers, or citizens re- sion of the common good, but to primordi- ceiving state-provided services and trans- al and ethnonational identities and “moral fers–are set by technocratic supranational majorities,” and in confrontational oppo- elites at places and levels that have largely sition to established elites, outside groups, escaped the reach of national policy-making

16 Dædalus, the Journal of the American Academy of Arts & Sciences and its democratic accountability, while criminal enemies of the state, or ethno- Claus nation-states suffer from a decline of their linguistic minorities. Although Switzer- Offe “governing capacity,” facing conditions in land has the oldest and most famous tradi- which they by themselves are unable to pro- tion of direct democratic legislation in Eu- vide for their citizens’ socioeconomic, civil, rope (usually preceded in that country by and military security and the integrity of extensive and reasonably balanced pub- their physical environment. lic debates on issues), these practices have spread in more limited forms to other coun- The battle cry of rightist populism is: tries in Europe, with hot spots in the right- “Let us, the people decide” and take con- ist populist regimes that have emerged in trol out of the hands of untrustworthy na- many of the post-Communist polities. In tional elites and illegitimate supranational Hungary, a national referendum on a man- forces. The arsenal of plebiscitarian meth- datory eu migrant quota was held (and lost ods (which, to be sure, are sometimes also by the government due to insufficient turn- advocated by some nonpopulist forces) out) in October of 2016. Yet probably the includes referenda on policy issues, citi- most consequential referendum held in Eu- zen initiatives to hold such referenda, and rope to date appeared in precisely the Euro- agenda initiatives to force legislatures to ad- pean country where parliamentary repre- dress certain policy issues. The use of sur- sentative democracy was born: the United vey research for identifying popular pref- Kingdom. erences and then elevating them to the sta- The Brexit referendum of June 23, 2016, tus of policy priorities on leaders’ platforms asked citizens to vote on whether the United can sometimes be seen as cases of social sci- Kingdom should leave the European Union ence–assisted populism. Thirty-six of the or remain a member state. Note that this forty-seven member states of the Council referendum was called for, but not initi- of Europe have by now adopted one or all ated by, a rightist populist political party. of these direct-democratic devices as part To the contrary, it was politically designed of their constitutional repertoire. In 2012, by David Cameron, a Conservative yet pro- the eu itself introduced the European Cit- European prime minister, who intended to izen Initiative as a device of supranational curb the growing political influence of the direct democracy. In recent years, these in- populist United Kingdom Independence struments of direct democracy have been Party (ukip), thus turning, he hoped, the applied to policies as varied as whether to means of populists against their ends. To permit or ban the construction of minarets, the surprise of most observers, that plan restrictions on migration, the public use of failed when a narrow majority of voters ac- a minority language, the acquisition of ag- tually voted Leave. Was it a wise decision to ricultural land by foreigners, same sex mar- let the question of Britain’s eu membership riage, the (retroactive) imposition of inher- be decided by referendum? In addressing itance taxes, and the introduction of a basic this question, I shall refrain from discussing income. For example, in the context of the the substantive political question of wheth- recent failed military coup in Turkey, Presi- er Brexit is a “good” move, confining myself dent Erdoğan has gestured at holding a ref- to the issue of whether the method used in erendum on reintroducing the death pen- making the decision was an adequate one. alty. The target groups of these referendum Here is a rough summary of the events. In campaigns may be Muslims, migrants, sex- the 2014 general elections to the European ual minorities, wealthy heirs, foreign real Parliament, ukip, the British anti-eu po- estate speculators, European institutions, litical party, won a relative majority of 27.5

146 (3) Summer 2017 17 What percent of the vote, with most of its votes was 51.9 percent Leave versus 48.1 percent Democratic taken from those defecting from the Con- Remain, with the citizenry sharply divid- Theorists Can Learn servative Party. Recognition of this grow- ed along class, age, and regional lines, but from the ing threat prompted incumbent Conserva- not equally sharply along party lines. Giv- 2016 Brexit Decision tive Prime Minister Cameron to commit en a turnout of 71.8 percent of all eligible himself in January 2013 to holding a refer- voters, roughly 37.3 percent of the elector- endum on the Brexit issue by the year 2017 ate will have caused (if it actually comes to if he were reelected in the national elec- that) Britain’s exit from the eu by a mar- tions of May 2015. His decision was a con- gin of just four percentage points.2 cession to the rightist populist demand to When making their decision on referen- let “the people” express its will directly, dum day, citizens were largely left with their rather than being represented by distrusted own individual means of will formation elites suspected of being corrupted by their (their beliefs and preferences) and with- own or other special or “foreign” interests. out much clear guidance from the political Populists are to be classified as “rightist” parties as to which of the alternatives, to- when framing the people in terms of nativ- gether with their entirely unknown impli- ist ethnic belonging versus some strange, cations, to choose. The two major parties foreign, and (as such) threatening enemy. were either openly divided (Conservatives) Cameron’s promise to hold a referendum or deeply ambivalent (Labour) about what was intended to serve the dual purpose of to recommend to their voters. Yet the only 1) increasing British bargaining power in party that was clear and committed on the ongoing negotiations with eu partners issue (ukip) had no chance of achieving the (who were seen as averse to further ukip parliamentary representation through ma- gains and the prospect of Brexit and hence joritarian British electoral law to follow its ready to grant concessions to the British option through. The division of pros and government on the key issues of Euro- cons was almost orthogonal to the major mobility and “ever closer” integration) and party cleavage. Similarly divided were the 2) immunizing the Conservative electoral media, with some of the tabloid press en- base against further defections of voters, as gaging in a vehement denunciation of the Eurosceptic Conservative voters were now eu, often with little regard for the truth of offered the option of expressing their Leave their claims.3 Moreover, both camps relied preference without having to switch to sup- heavily on fear as a negative economic mo- porting ukip. tivation: the Leavers feared losing control Both of these purposes were, to an ex- over the fates of “our” country to “Brus- tent, achieved, the second more fully than sels” (or of having to compete with foreign the first. The turn to plebiscitarian meth- migrant labor for jobs),4 and the Remain ods (which are foreign to the United King- camp feared the adverse economic conse- dom’s constitutional traditions) came at quences (jobs, trade, investment, exchange the price of undermining the authority rates) of Brexit. Appeals to the advantag- of Parliament, the members of which op- es, political attractions, prior commit- posed Brexit by a large majority. Having ments, hopes, and promises of remaining won the 2015 elections and being bound were rarely advanced, implying that there by his referendum promise, Cameron ini- were few. Left in a state of disorientation tiated the eu Referendum Act, which was and anxiety, and being informed by the passed by the House of Commons in De- media and polling organizations that the cember 2015. When the referendum was contest would be a tight one (suggesting eventually held on June 23, 2016, the result that every vote or abstention could make

18 Dædalus, the Journal of the American Academy of Arts & Sciences a big difference), voters were left to rely on In this implicit debate of identity versus Claus their gut feelings, rather than an informed interest, the elderly and the less-educated Offe judgment, on the merits of the two alter- considered eu membership both a cultur- natives.5 The dichotomy of a referendum al and economic threat and hence gravitat- further induced the voters to ignore the nu- ed toward the Leave option, while the best- merous intermediate solutions that might educated, younger (below age forty-five) have been worked out through bargaining voters welcomed diversity within Britain following the formal declaration of Brexit. because they could “compete with ease in One of the damages the reliance on the ple- an internationalised labour market.”8 biscitarian method can do stems from its How has the Brexit referendum per- one-sided fixation on voting at the expense formed in realizing the democratic prin- of the two other modes of democratic po- ciple of equality of political rights to make litical communication: arguing and bar- one’s voice heard? Good democrats know gaining.6 Plebiscitarian procedures thus that those affected by the law must have a impoverish the tool box of democratic pol- voice in making the law. Yet voting rights itics by eliminating the space for postvot- in the Brexit case became effective only by ing reasoning and compromise-finding in passing three filters: First, in the United the institutional framework of representa- Kingdom, you must be a citizen, not just a tive democracy. They privilege the fast, im- resident, to be eligible for voter registration pulsive snapshot reaction generated by pas- in national elections/referenda. Millions of sions and visceral instincts over the more mainland eu citizens residing in the Unit- time-consuming balancing of interests and ed Kingdom were thus not allowed to regis- the typically lengthier process of persuasion ter and vote. That would be immaterial had through argument. As a consequence, con- the referendum been on a “purely British” sistency is not required: voters can simulta- issue. But here the category of people most neously opt for lower taxes and greater ex- directly affected by Brexit are exactly those penditures, or for cheaper gas and stricter migrant workers from member states re- environmental standards. siding in the United Kingdom. After Brex- Not only were the two major parties split it, these migrant workers are likely to be de- in their preferences between Remain and prived of some or all of their socioeconomic Leave, but voters were also “cross-pres- rights as eu citizens.9 sured” at the individual level. Many voters Second, you must register in order to be were motivated by the issues of immigra- admitted to the voting booth. “Many peo- tion and “sovereignty,” with the support for ple chose not to register to vote because the Leave alternative fueled by an identity- they feared the debt collection agencies based opposition to having to adopt “for- that are allowed access to the electoral reg- eign-made” eu laws (“let’s take back con- ister.”10 As many as seven million eligible trol of our country”). Yet, at the same time, adults were not registered to vote in the many of the same voters “regarded the eco- United Kingdom in 2016, perhaps in part nomic impact of leaving the eu negatively. due to that deterrence effect. . . . No less than 40 [percent] reckoned that Third, you must vote. Thirteen million re- Britain would be worse off economically if gistered voters did not turn out. They were it left the eu. . . . The two central issues of the disproportionally young, renters, mem- campaign were seemingly pulling voters in bers of ethnic minorities, and recent mov- opposite directions.”7 Fears for the econo- ers. Older people voted in greater propor- my, based on socioeconomic interest, pro- tion. They generally voted for Leave, while vided a reason for voting in favor of Remain. among those aged eighteen to twenty-four,

146 (3) Summer 2017 19 What 73 percent voted (if they voted) for Remain. sions of the largely unanticipated referen- Democratic But the youngest age groups also had the dum outcome (for Britain and for the geo- Theorists Can Learn largest share of abstainers. Again, a paradox political role of the eu and its prospects for from the shows up in that those affected by the out- further disintegration), over four million 2016 Brexit Decision come for the longest time span (the young) voters signed a petition in the days after the had the lowest impact on that outcome, and referendum that called for holding a second those least affected the greatest impact. referendum, thus indicating a widespread So much for the democratic egalitarian- sense of regret, as well as alarm, over the ism of voting in referenda. In regular elec- outcome. Yet such a repetition would seem- tions, contending political parties provide ingly have required another Referendum some guidance to voters and tend to make Act as its legal basis. It would have opened an effort to mobilize in demographically the horrifying perspective of an endless balanced ways. Now another problem of chain of further referenda on the outcomes referenda is that there is no way to make of prior referenda: vote until the outcome sure that the answer voters give is actually seems right! If the first is seen by voters as their answer to the specific question they ill-considered and in need of self-correc- are asked: in this case the question of eu tion, why should the second fare better?13 membership or not. Chances are that the How can the decision to let the relative answer the Leavers gave was the answer to majority of those participating in the ref- an entirely different question, such as: “Do erendum decide on a complex, highly con- you want to seize the opportunity to send sequential yet, at the same time, most un- a hostile message and cause trouble to the predictable national issue be justified as the hated political establishment–be it the na- “right” procedural decision–rather than as tional or the one in Brussels?”11 If this is the the (eventually failed) opportunistic calcu- question being actually answered (and an- lus of a leading politician to maintain his swering “yes” is less inhibited because of power over his party and the country? In a widespread belief that the Remain camp other words: what is this outcome’s pro- would win anyway), there is no reason for cedural source of validity and normative voters to stick to their answer for even a sin- bindingness? The procedural design of the gle day after the vote. When surveyed im- Referendum Act was ill-considered. It failed mediately after the referendum, “7 [per- to make use of the several safety valves cent] of those who voted Leave feel like they that can be applied in referenda in order to did not make the right choice,” while no strengthen the normative bindingness, or less than 29 percent considered their vote legitimacy, of the outcome: that is, its pros- instrumentally futile as the two goals of the pects of being durably and universally rec- Leave campaign could not, in fact, both be ognized as reasonable and hence valid, rath- accomplished in the upcoming Brexit nego- er than as a regrettable collective misstep. tiations with the eu: namely, the interest- For one thing, a quorum, or minimally re- related goal to stay in the single market and quired turnout of voters, could have been enjoy its economic advantages and the iden- stipulated, such as a 75 percent require- tity-related goal to limit freedom of movement. The stipulation of such a threshold, ment of eu citizens and to “take back con- however, might have provided the oppor- trol.”12 Concerns of interest and those of tunity for the Remain side to sabotage the identity seem to have pulled voters in dif- referendum by launching an abstain cam- ferent directions. paign. Another possibility might be a super- Given the vast and highly uncertain majority requirement, such as a 60 percent short-term as well as long-term repercus- threshold for the winner.14 Adopting such a

20 Dædalus, the Journal of the American Academy of Arts & Sciences supermajority rule would avoid deciding a whereas the voting public can only blame Claus matter of this magnitude by a slim and pos- itself (that is, nobody in particular, since the Offe sibly even accidental and unstable majority. vote is secret and nobody can be held ac- A third safety measure could have been the countable) in case the results of a referen- use of federal constraints. Given that the dum turn out to be widely seen as mistaken. United Kingdom is a multinational polit- A further provision that was, in fact, de- ical entity, one or more of its constituent ployed in the Brexit referendum was the pro- nations–Northern Ireland, Wales, and in cedural stipulation that the government is particular Scotland (where the Remain vote not strictly bound to implement the result, achieved a substantial majority)–could but can treat it as merely advisory. As sover- have been procedurally protected from de- eignty resides in Parliament, it is, arguably, feat by a (narrow) overall national major- that representative body that must eventu- ity by granting Scotland autonomy rights ally decide whether or not to endorse and concerning the issue of eu membership. implement, through its law-making, the In fact, the referendum result has strength- referendum decision. In theory, the only ened Scotland’s claim for national autono- thing that even the most sovereign body my, thus putting into political jeopardy the cannot do is abdicate its own law-making very unity of the United Kingdom. Finally, a powers and transfer them to another body, test vote (as sometimes taken in party groups such as the multitude of citizens voting in a of legislative bodies) could have been pro- referendum. It seems to follow that a prime vided, the result of which would have in- minister cannot self-bindingly promise formed voters about dispositions of their voters that he or she will follow their ex- fellow citizens and encourage them to re- pressed preferences as if they constituted vise or assert their own dispositions accord- an act of legislation. Absent a parliamen- ingly in the second (and only valid) round. tary or at least executive ratification of the Applying some or all of these provisions (presumed) popular will as expressed in a could have been justified by the fact that referendum, such a referendum cannot be the Brexit referendum was a one-shot and binding. For example, the invocation of Ar- highly consequential decision, which will ticle 50 of the Treaty on the European Union create consequences that are certain to be (teu)–the article that prescribes the first felt in the long term. In contrast, the “nor- step of the procedures of actually exiting mal” democratic procedure of holding con- the Union–must be an act of Parliament tested elections is defined by its periodic- or at least, if “royal prerogative” were to ap- ity, meaning that governing authority is ply (which is bitterly contested), a decision granted pro tempore and that losers of an of the prime minister, who in turn might election will have another chance in four be seen as in need of winning the legiti- or five years’ time, with both competing macy of her or his decision through an en- parties and members of the general public dorsement through regular elections (rath- given a learning opportunity to revise plat- er than a nonelectoral accession to office, forms and preferences during the interval. as in the case of Prime Minister Theresa An election constitutes both a government May). These manifold ambiguities and and an opposition of losers, while a refer- disputes illustrate the extent to which the endum constitutes a fait accompli that can “will of the people” is a largely elusive sub- no longer be challenged.15 If after an ordi- stance contingent on the procedures by nary legislative election, policies are consid- which it is being assessed. Holding a ref- ered to have gone wrong, there is someone erendum has not been, in the instance of to blame (and punish) in the next election, Brexit, a way to settle a question, but an in-

146 (3) Summer 2017 21 What advertent move to open a constitutional weeks, while the most prominent Leave pro- Democratic Pandora’s box. The attempt to fight pop- tagonist, Boris Johnson, moved up to the Theorists Can Learn ulism by adopting its own plebiscitarian position of Britain’s Foreign Secretary. The from the weapon has not only misfired, but has had new prime minister’s signature tautology– 2016 Brexit Decision a destructive impact upon the principle of “Brexit means Brexit,” being void of any in- representative government. formation about what Brexit means–rati- To be sure, a parliamentary validation of fies the unconditional surrender of repre- the referendum decision might well be the sentative to plebiscitarian will formation. It result of principled argument and prop- also gives carte blanche to rulers to define the er deliberation, weighing the merits of meaning ex post. As constitutional scholars the “advice” the voting public has offered Richard Gordon and Rowena Moffatt have against alternative policies. Yet the sover- stated with unfathomable yet inconclu- eignty of Parliament, in the sense of hav- sive juridical wisdom: “In practice, the . . . ing the last and decisive word, has largely referendum outcome will bind the govern- been rendered nominal by the referendum ment. In theory it is advisory but in reality and the legislature’s prior decision to hold its result will be decisive for what happens that referendum. By adopting the eu Refer- next.”17 At the time of the submission of endum Act, thereby (seemingly) passing its this essay in December 2016, the answer to legislative responsibilities to the “people,” this question is by no means settled by the the Parliament has virtually destroyed its referendum, but remains a pending case be- recognition as a body to be credited with fore the highest court of the country. the capacity to form policy on the basis of informed, considered, and balanced argu- Given all these premises, dilemmas, and ment. It has eschewed its responsibility to consequences, the Brexit referendum must do so, thereby confirming, in a way, the car- be considered a clear and unambiguous les- icature populists paint of members of the son on what democracies ought not to do. “political class.” If Parliament abdicates its Holding referenda with a 50 percent ma- law-making authority on as weighty an is- jority on important substantive policy issue as eu membership, what should pre- sues with substantial yet unknown long- vent it from doing so on other issues in the term results is a misguided remedy to the future?16 ills of liberal democracy. Referenda encour- Having unleashed the plebiscitarian forc- age the accountability-free expression of es voicing fear of foreign control and for- poorly considered mass preferences and eign migrants, neither the political par- de-emphasize requirements of consisten- ties nor the members of Parliament could cy, compromise-building, and the reflec- henceforth afford to advocate any solution on consequences. By inviting citizens tions to future uk-eu relations that could to leap into the dark, they create irrevoca- be denounced as defying the referendum’s ble facts and preclude learning. They often “advice.” Politicians cannot be expected betray minimal standards of rational policy to commit electoral suicide by refusing to formation, traces of which are institution- follow the “will of the people,” the expres- alized in even the most corrupted practices sion of which they themselves had allowed of parliamentary debate, party competition, for, even if only as part of a power game. and mass media reporting. They anonymize These problems (and not an electoral or par- the locus of accountability. If these critical liamentary defeat) made the committed Re- generalizations are only partly right, the mainer David Cameron disappear from the urgent question is: can we think of better scene of uk national politics in a matter of and smarter–more reliably “regret-avoid-

22 Dædalus, the Journal of the American Academy of Arts & Sciences ing”–modes of making highly salient deci- can preference formation be improved so Claus sions? Otherwise, we may regret decisions as to make the citizen preferences that will Offe that fail to take sufficient account of the fu- later be translated into policies by govern- ture, other people, and the facts.18 How can ing elites more regret-proof? we minimize these forms of rational regret The first of these two major reorienta- while maintaining the basic tenets of liberal tions of democratic innovation involves democratic theory: namely, equality of civil complementing the universe of the adult and political rights, freedom of opinion, and permanent legal residents of the territory the division of state powers? The remain- of a state (or municipal entity or province), der of this essay will consist of a short and who are the ultimate source of popular sov- schematic account of what should be done ereignty, with a small body (“mini-public” instead on the basis of deliberative demo- or “deliberative panel”) of persons that is cratic theory. (as accurately as possible) statistically representative of the whole. Constituting ac- Apart from large literatures on new social tive citizenship by lot is an ancient idea, movements, civil society, and social capi- dating back to the times of Athenian de- tal, a major conceptual and theoretical in- mocracy (and found, to some degree, in novation in democratic theory over the last Renaissance Italian city republics), that fell generation has been the idea of deliberative into discredit in the course of the French democracy. Compared with conventional and American revolutions with the crypto- approaches in democratic political theory, aristocratic notion that the people can be deliberative theory performs a dual shift represented only through elected bodies of emphasis. In one shift that has become and leaders.20 Lotteries as a procedure of re- increasingly popular among theorists, and cruiting people for public roles are typically even to some degree in practice, delibera- regarded as risky because they rely on high- tion is brought to the public through a par- ly optimistic assumptions concerning both tial move from territorial representation to the readiness and the competence of those aleatory,19 or randomized, representation– chosen by lot to perform the needed public an analog to jury selection through sorti- roles. Yet both the readiness and compe- tion in the common law countries. This use tence objections can be dealt with through of randomly selected citizens also serves to appropriate institutional precautions. partially dissolve the conventional dichot- The readiness of randomly selected can- omy of ruling elites representing voting yet didates to assume the tasks assigned to them ruled nonelites. Few suggest replacing cur- by lot can be enhanced through a compen- rent political institutions with such bodies; sation that follows a rule of thumb such as they are intended to complement existing “no loss, no gain,” with a cap of, say, 150 per- institutions to help correct their known de- cent of the median income, depending on ficiencies. the complexity of the issue under consid- The second shift moves from an ideal eration. To enhance that readiness, the du- of maximizing the citizens’ expression of ration of the time in “office” might also be political preferences (in participatory de- limited to a maximum of six months, for ex- mocracy, as many people as possible should ample. Nevertheless, civic duty to partici- have a chance to voice their preferences on pate in deliberative mini-publics will proba- as many issues as possible and as directly as bly remain hard to enforce, and participants possible) to maximizing the citizens’ capac- who see themselves as being coerced will ity to form preferences and judgments on likely not properly perform. Techniques of public affairs they will not later regret. How stratified sampling may offer a solution in

146 (3) Summer 2017 23 What case the characteristics of the sample devi- els would perform a purely consultative Democratic ate far (by gender, age, socioeconomic, ed- function,21 helping citizens form prefer- Theorists Can Learn ucational, and minority status) from those ences that they would then express in elec- from the of the constituency as a whole. The logisti- tions and possibly referenda. And citi- 2016 Brexit Decision cal problems of organizing face-to-face de- zens must be provided access to those rec- liberation sessions on national legislation ommendations through the reporting of in geographically large countries might be print media, brochures, and (public) elec- alleviated by first selecting (possibly, again, tronic media. The role of deliberative bod- by lot) two municipal units from which the ies should be strictly advisory, addressing samples are to be drawn. Although in com- both elites and voters. That role should posing that sample a measure of self-selec- also be limited to the specific issue of pub- tion cannot be avoided, the statistical repre- lic policy about which a deliberative panel sentativeness of members of the mini-pub- is commissioned to elaborate a recommen- lics thus selected should be much superior dation. The lay policy-makers who jointly to that of the composition of ordinary legis- author such a recommendation may con- lative bodies. The relatively small size of de- clude with a consensual recommendation liberative panels (probably fewer than one or with majority and minority positions. In hundred candidates) must be big enough the latter case, a second order consensus on to allow for representativeness on all rele- what stood in the way of a consensual rec- vant variables, yet small enough to allow ommendation should be provided. The two for serious and inclusive face-to-face argu- panels may also disagree in their consensu- ing under the supervision of a trained facil- al recommendations. If the recommenda- itator. The virtue of lottery representation tion is both consensual within panels and would consist not only in providing a polit- identical between the two locations, this is ical role to ordinary citizens, but in deny- likely to translate into the highest degree ing such a role to political parties and orga- of persuasiveness and impact on elector- nized interests. Unlike the parties and inter- al outcomes. This impact is due to the en- est groups, randomly selected citizens are lightened vicarious judgment that “people unlikely to have the interest or the capac- like us” have formed on the issue at hand. ity to entrench themselves in their public The more consensual the recommendation role of deliberators. within and between panels, the stronger its Even thornier than the issue of readi- influence ought to be and probably will be ness to participate is the issue of compe- on the decisions that voters and elected rep- tence. Members of issue-specific deliber- resentatives will make. ative panels need to acquire a measure of The premise from which theorists of de- understanding and expertise, as do mem- liberative democracy by sortition start is bers of legislative bodies, in order to ar- the assumption that citizens do not simply rive at minimally reasoned conclusions. have political preferences and attitudes, in- Such expertise can be provided by an ade- cluding preferences and aversions to par- quate number and diversity of recognized ticular policies. Rather, they continuously experts made available to members of a form these preferences in a process of on- mini-public as providers of information. going confirmation, revision, and learn- Concerns about deficiencies in the knowl- ing. Most of the time and on most issues, edge and experience of members of delib- most peoples’ preferences are incomplete, erative mini-publics are further reduced by inconsistent, insufficiently informed, con- the fact that no political decision-making tingent, fluid, and subject to relations of power is vested in them. Deliberative pan- trust, as when we adopt the point of view

24 Dædalus, the Journal of the American Academy of Arts & Sciences of others because we happen to feel con- to some “most different” design) selected Claus fident about the adequacy of their judg- subterritorial entities (counties or cities). A Offe ment. The capacity of forming thoroughly statute would regulate the size of the panel, considered judgment can today no longer the sampling method, the mode of oper- be vested in individual representatives (as ation (including a budget for expert assis- Burke claimed), but must emerge from the tance and compensation payments), the discursive confrontation of diverse mem- role of facilitators and moderators, and bers of an organized body. The key demo- the scheduling of meetings. Their work of cratic act of voting is about the expression (at most) six months would result in poli- of preferences, whereas the activity taking cy recommendations (consensual or other- place in randomized deliberative panels wise) in the form of an executive summary, (as well as, mostly implicitly, in many oth- together with the reasoning from which er theaters, such as peer groups, schools, re- the recommendations derive. The identity ligious communities, media, the arts, con- of members would ideally be kept anony- sumption, and not least the political pro- mous through the time of deliberations so cess itself ) is that of the formation and as to shield the deliberators from outside in- (de)consolidation of those preferences fluence. Neither governments nor citizens through learning. The presence of deliber- would be pressured to follow those poli- ative panels–and the public perception of cy recommendations. Governments (and, the conclusions they arrive at–allows ordi- perhaps, political parties) might, however, nary citizens to get an idea about what hap- be formally required to publish an official pens when “people like us” spend time and statement specifying the reasons why they energy on refining their preferences, find- did not follow the advice, in cases in which ing out for themselves and others what they they decide not to do so. hold to be the right position on particular It is impossible to know whether the out- policy issues. The role of the citizen delib- come of the Brexit decision would have erators will be strictly limited to that of an been different if it had been processed advisory agency assisting citizens (includ- through an institutional arrangement of ing elected and appointed officials) in the will formation such as the one just out- process of their will formation. lined. Whatever the answer, British voters For such reflexive preference learning to and elites would at least have been more take place at the level of mass constituen- certain that they made the right decision cies, deliberative panels need to be institu- than they can possibly be after the experi- tionalized: that is, made part of the rules ence of the Brexit referendum. regulating the process of legislation. To illustrate, one conceivable institutional design would be the following. A deliberative panel would come into being at the initiative of at least 20 percent of the members of the state or federal legislature. These members would also define the policy issue on which the panel is commissioned to deliberate. The panel would deliberate one year prior to a decision to be taken by the legislature or executive branch on the policy. Such panels would always come in pairs, with both being active in two (according

146 (3) Summer 2017 25 What endnotes Democratic 1 Theorists This generalization does not apply to the Chinese case of industrial capitalism presided over Can Learn by a “Communist” party. But that model is neither intended for export nor appealing to elites from the or masses of Western societies. 2016 Brexit 2 Decision The day after the referendum, Philip Stephens, chief political commentator of the Financial Times, commented in undisguised horror: “Who would have thought pragmatic, moderate, incrementalist Britain would tear down the political temple? This week’s referendum result was a revolt against the status quo with consequences, national and international, as profound as anything seen in postwar Europe.” Philip Stephens, “How a Cautious Nation Came to Tear Down the Political Temple,” Financial Times, June 24, 2016, https://www.ft.com/content/ b90a7278-3a02-11e6-9a05-82a9b15a8ee7. 3 The ironic label “post-truth” has been attached by several commentators to populist movement practices. This label is not only deserved by the generous use populist campaigns have made of outright lies, but also, as in the Trump campaign, by their anti-intellectual aversion to expertise and educated intelligence. 4 Princeton economist Ashoka Mody has put it well: “Cameron misjudged . . . by making an economic case for remaining in the European Union rather than attempting a serious political argument for Europe–one based on shared values.” Ashoka Mody, “Don’t Panic: Britain’s Economy Can Survive Just Fine Outside the European Union,” Independent, July 4, 2016, http:// www.independent.co.uk/news/business/analysis-and-features/dont-panic-britains-economy -can-survive-just-fine-outside-the-european-union-a7118736.html. 5 There is more than a grain of truth in Edmund Burke’s famous claim: “Your representative owes you . . . his judgment; and he betrays, instead of serving you, if he sacrifices it to your opinion.” In the case of Brexit, it was the plain cowardice of representatives facing a populist challenger that caused this sacrifice. Edmund Burke, “Speech to the Electors of Bristol,” November 3, 1774. 6 Jon Elster, “Arguing and Bargaining in Two Constituent Assemblies,” Journal of Constitutional Law 2 (2) (March 2000): 345–421. 7 John Curtice, “Brexit: Behind the Referendum,” Political Insight 7 (2) (2016), http://pli.sagepub .com/content/7/2/4.full.pdf+html. 8 Ibid. Wolfgang Streeck has argued that “the losers under neoliberal internationalism [globalization] place their hopes on their nation states.” Wolfgang Streeck, “Where Are We Now? Responses to the Referendum,” London Review of Books 38 (4) (July 14, 2016), http://www.lrb .co.uk/v38/n14/on-brexit/where-are-we-now#streeck. This causal explanation of the outcome commands a great deal of empirical plausibility. But equally great is the temptation to exclaim, with the words of the great Austrian poet Ernst Jandl, “What an error!” 9 A reciprocal loss of socioeconomic status rights applies to British citizens who were economically active in eu member states and, after Brexit, are now relegated to the status of third country nationals. 10 Benjamin D. Hennig and Danny Dorling, “In Focus: The eu Referendum,” Political Insight 7 (2) (2016), http://pli.sagepub.com/content/7/2/20.full. 11 This was widely seen by commentators to have happened in a Dutch referendum held on the highly technical as well as politically rather marginal issue of a Dutch Approval Act on a Eu- ropean Union–Ukraine Association Agreement, which was held in The Netherlands on April 6, 2016. The outcome was a turnout of 32.8 percent, with 61 percent voting against the Act. In an interview after the referendum, the members of the rightist nationalist Citizens’ Commit- tee eu that had successfully campaigned for holding it admitted not caring about Ukraine, but rather were just against the eu political system. See Wilmer Heck, “Oekraïne kan ons niets schelen,” ncr, March 31, 2016, https://www.nrc.nl/nieuws/2016/03/31/oekraine-kan -ons-niets-schelen-1606419-a969298.

26 Dædalus, the Journal of the American Academy of Arts & Sciences 12 See James Crouch, “Voters React to Post-Referendum World,” Opinium, July 1, 2016, http:// Claus opinium.co.uk/voters-react-to-post-referendum-world/. The thought that voting for Brexit Offe means “taking back control” is plainly delusionary, at least in the short and medium term. Article 50, which is still binding for the United Kingdom, stipulates that “the Union shall ne- gotiate and conclude an agreement with [the uk].” In the interest of the Union to prevent the Brexit decision from becoming a template that other member states might follow, the eu is likely to opt for the harshest possible terms in negotiating Britain’s exit arrangement, thus “taking control” over the economic fates of the United Kingdom to an unprecedented extent. The Lisbon Treaty, Article 50 of the Treaty of European Union, December 13, 2007. 13 There is, however, a strong argument for having a second referendum at a later point. As the proponents of Brexit had no plan (and could not have one) concerning the many and very different versions of what is going to happen next in re-embedding Britain into the international political economy, the eventual outcome of negotiations with the eu must also be subject to a (dis)approval by voters. As Simon Wren-Lewis has cogently argued: “I cannot see the logic in saying people should have a direct say in whether to leave the eu, but no direct say on what to leave for.” Simon Wren-Lewis, “Why We Must Have a Second Brexit Referendum,” Social Europe, August 29, 2016, https://www.socialeurope.eu/2016/08/must-second-brexit-referendum/. The emphasis is mine. 14 As one commentator has noted: “It is highly unusual [in mature democracies] that, particularly on issues of great constitutional significance, a simple majority of those who happened to vote on a particular day should be regarded as binding.” Brendan Donnelly, “After Brexit: The Light at the End of the Tunnel is Several Oncoming Trains,” Social Europe, July 18, 2016, https://www.socialeurope.eu/2016/07/light-end-tunnel-several-oncoming-trains/. 15 “The 48 [percent] of voters . . . who wanted to remain in the European Union now suddenly find themselves substantially unrepresented in the British Parliament.” Ibid. 16 This question touches on the thorny issue, not to be dealt with in the present essay, of what kind of policy issues are “safe” to be processed by plebiscitarian methods. Referenda on con- stitutions can arguably enhance the self-binding effect and thus the constitution’s validity and longevity. The adoption of legal rules, the consequences of which are easily understood, predictable, and largely uncontroversial (such as local referenda on opening hours of stores), would also seem unproblematic. The same can surely not be said of the plebiscitarian adoption, now common in several eu countries, of rules that discriminate against moral, ethnic, migratory, sexual, religious, or criminal minorities (as in the current initiative of the Turk- ish president to hold a referendum on the reintroduction of the death penalty). 17 Richard Gordon and Rowena Moffatt, Brexit: The Immediate Legal Consequences (London: The Constitution Society, 2016), 7. 18 Future-regardingness, other-regardingness, and fact-regardingness can serve, taken together, as a standard of political rationality. See Claus Offe, “Crisis and Innovation of Liberal De- mocracy: Can Deliberation Be Institutionalised?” in Citizens in Europe: Essays on Democracy, Con- stitutionalism and European Integration, ed. Claus Offe and Ulrich K. Preuss (Colchester, United Kingdom: ecpr Press, 2016), 73–98. 19 See Hubertus Buchstein, “Elective and Aleatory Parliamentarism,” in Parliamentarism and Dem- ocratic Theory, ed. Kari Palonen and José María Rosales (Opladen, Germany: Budrich, 2015), 255–278; and David van Reybrouck, Against Elections: The Case for Democracy (London: The Bod- ley Head, 2016). Some also use the word “sortition” and others “the lot” to describe this feature of representation through random selection. 20 Rousseau was still convinced that a democracy must be built on a mix of territorial and aleatory representation. See chapter three of book four of Jean Jacques Rousseau, Du contrat social. 21 See Patrizia Nanz and Claus Leggewie, Die Konsultative: Mehr Demokratie durch Bürgerbeteiligung (Berlin: Wagenbach, 2016). See also Baogang He and Mark E. Warren, “Authoritarian De- liberation in China,” Dædalus 146 (3) (2017) for the spectrum between advisory/consultative deliberative panels and empowered ones.

146 (3) Summer 2017 27 Twelve Key Findings in Deliberative Democracy Research

Nicole Curato, John S. Dryzek, Selen A. Ercan, Carolyn M. Hendriks & Simon Niemeyer

Abstract: This essay reflects on the development of the field of deliberative democracy by discussing twelve key findings that capture a number of resolved issues in normative theory, conceptual clarification, and associated empirical results. We argue that these findings deserve to be more widely recognized and viewed as a foundation for future practice and research. We draw on our own research and that of others in the field.

Deliberative democracy is a normative project grounded in political theory. And political theorists make a living in large part by disagreeing with and criticizing each other. In fact, it is possible to evaluate the success of a political theory by the number NICOLE CURATO is Australian Re- of critics it attracts, and the vitality of its intramural search Council Discovery Early Ca- disputes. By this measure, deliberative democracy is reer Research Fellow at the Uni- versity of Canberra. very successful indeed. Yet if the normative project is to progress and be applied effectively in practice, JOHN S. DRYZEK is Australian Re- it needs to lay some issues to rest. search Council Laureate Fellow Deliberative democracy is not just the area of con- and Centenary Professor at the University of Canberra. tention that its standing as a normative political theory would suggest. It is also home to a large volume of SELEN A. ERCAN is Senior Re- empirical social science research that, at its best, pro- search Fellow at the University of ceeds in dialogue with the normative theory. Indeed, Canberra. the field is exemplary in this combination of politi- CAROLYN M. HENDRIKS is Asso- cal theory and empirical research. Deliberative ideas ciate Professor at the Australian have also attracted the attention of citizens, activists, National University. reform organizations, and decision-makers around SIMON NIEMEYER is Australian the world. The practical uptake of deliberative ideas Research Council Future Fellow at in political innovation provides a rich source of les- the University of Canberra. sons from experience that can be added to theoriz- (*See endnotes for complete contributor ing and social science. This combination has prov- biographies.) en extremely fruitful. Rather than proving or falsi-

28 fying key hypotheses, deliberative practice Deliberation is essential to democracy. Social Nicole Curato, has sharpened the focus of the normative choice theory appears to demonstrate that John S. Dryzek, Selen A. Ercan, project, showing how it can be applied in democratic politics must be plagued by ar- Carolyn M. many different contexts. bitrariness and instability in collective de- Hendriks & We believe that conceptual analysis, logic, cision. Notably, for political scientist Wil- Simon Niemeyer empirical study, normative theorizing, and liam Riker, clever politicians can manipu- the refinement of deliberative practice have late agendas and the order in which votes set at least some controversies to rest, and are taken to ensure their preferred option we provide the following set of twelve key wins.8 But if their opponents are also clever, findings that can be used as the basis for fur- they can do the same. And in that case, there ther developments. can be no stable will of the people that can possibly be revealed by voting (in, say, a leg- Deliberative democracy is realistic. Skeptics islature). So, how can meaning and stability have questioned the practical viability of de- be restored to democracy? There are essen- liberative democracy: its ideals have been tially two mechanisms, once dictatorship is criticized as utopian and its forums have ruled out. The first is what rational choice been dismissed as mere experiments, with theorist Kenneth Shepsle calls “structure no hope of being institutionalized effec- induced equilibrium,” under which formal tively.1 rules and informal understandings restrict But skeptics have been proved wrong by strategizing, including the ability to manip- the many and diverse deliberative innova- ulate agendas and the order in which votes tions that have been implemented in a va- are taken.9 The second is deliberation. riety of political systems.2 Both state and Political theorist David Miller and, lat- nonstate institutions demand more deliber- er, John Dryzek and political philosopher ative forms of citizen engagement. Policy- Christian List have demonstrated formal- makers and politicians convene citizens’ fo- ly that deliberation can, among other re- rums to elicit informed views on particular sponses: 1) induce agreement to restrict the issues.3 Studies find that deliberating citi- ability of actors to introduce new options zens can and do influence policies, though that destabilize the decision process and impacts vary and can be indirect.4 Delib- 2) structure the preferences of participants erative forums are also being implement- such that they become “single-peaked” ed in parliamentary and electoral contexts.5 along one dimension, thus reducing the Outside the state, citizen forums are funded prevalence of manipulable cycles across and implemented variously by civil society alternatives (in which option A beats B in a organizations, think tanks, corporations, majority vote, B beats C, and C beats A).10 and international organizations to advance Empirical research confirms this effect.11 a particular cause, foster public debate, or This result explains why all democratic promote democratic reform.6 settings, in practice, feature some combina- The recent turn toward deliberative sys- tion of communication, which can be more tems demonstrates that deliberative demo- or less deliberative, and formal and infor- cratic ideals can be pursued on a large scale mal rules. The more deliberative the com- in ways that link particular forums and munication, the better democracy works. more informal practices, such as commu- Democracy must be deliberative. nication in old and new media.7 Delibera- Deliberation is more than discussion. Delib- tive democracy is not utopian; it is already erative democracy is talk-centric. But talk implemented within, outside, and across alone can be pathological, producing wild- governmental institutions worldwide. ly mixed results from an ideal deliberative

146 (3) Summer 2017 29 Twelve Key perspective.12 Resolution here requires dis- tique has been raised by political theorist Findings in tinguishing carefully between deliberation Chantal Mouffe, who criticizes delibera- Deliberative Democracy and discussion. tive democrats for missing the crucial role Research Empirical observation reveals that de- that passion plays in politics and for em- liberation is more complex than original- phasizing the rationalism of liberal dem- ly theorized, involving both dispositional ocratic political thought.20 and procedural components. The purely Deliberative democrats have responded procedural rationalist model of delibera- by foregrounding the varied articulations of tion is normatively problematic because it reason-giving and consensus requirements is empirically questionable.13 Distinguish- of deliberation. Most have acknowledged ing between deliberation and discussion in- political philosopher Iris Young’s concep- troduces an emotional dimension in which tion of “communicative democracy” and dispositional factors, such as open-minded- have conditionally embraced greeting, rhet- ness, are important.14 oric, humor, testimonies, storytelling, and The overall content of this disposition other sorts of communication.21 Even the has more recently been referred to as the originally somewhat rationalistic criteria “deliberative stance,” which political the- of the widely used Discourse Quality Index orists David Owen and Graham Smith have have evolved to include storytelling as one defined as “a relation to others as equals en- indicator, recognizing the importance of gaged in mutual exchange of reasons orient- personal narratives in political claim-mak- ed as if to reaching a shared practical judge- ing.22 Recent developments in deliberative ment.”15 Achieving a deliberative stance in theory have begun to recognize the plurali- citizen deliberation involves careful facili- ty of speech cultures. The turn to delibera- tation and attention to “emotional interac- tive systems has emphasized multiple sites tion.”16 Its achievement in group settings of communication, each of which can host can be a pleasurable experience and consis- various forms of speech that can enrich the tent with ideals of human cognition.17 Scal- inclusive character of a deliberative system. ing these effects up to the wider deliberative The increasing attention paid to delibera- system requires careful attention to institu- tive cultures is also part of this trajectory, tional settings.18 in which systems of meanings and norms Deliberative democracy involves multiple sorts in diverse cultural contexts are unpacked of communication. Some democrats have to understand the different ways political charged deliberative democracy with be- agents take part in deliberative politics.23 ing overly rationalistic. For political scien- Deliberation is for all. The charge of elitism tist Lynn Sanders, deliberation works un- was one of the earliest criticisms of delib- democratically for it excludes “those who erative democratic theory: that only privi- are less likely to present their arguments in leged, educated citizens have access to the ways that we recognize as characteristical- language and procedures of deliberation. ly deliberative.”19 Sanders refers to wom- However, empirical research has estab- en, racial minorities, and the poor, whose lished the inclusive, rather than elitist, char- speech cultures depart from “rationalist” acter of deliberative democracy. forms of discourse that privilege dispas- Findings in deliberative experiments sionate argumentation, logical coherence, suggest that deliberation can temper rath- and evidence-based claims as practiced in er than reinforce elite power. Political sci- the most exclusive kinds of scholarly de- entists James Druckman and Kjersten Nel- bates, parliamentary procedures, and judi- son have shown how citizen conversations cial argumentation. A similar kind of cri- can vitiate the influence of elite framing.24

30 Dædalus, the Journal of the American Academy of Arts & Sciences Simon Niemeyer has shown how deliber- participants that are less partisan, using Nicole Curato, ative mini-publics, such as citizens’ juries independent facilitators, or ensuring de- John S. Dryzek, Selen A. Ercan, (composed of a relatively small number of liberations are public. Carolyn M. lay citizens), can see through “symbolic pol- Empowering or generative forms of Hendriks & itics” and elite manipulation of public dis- power are central to the communicative Simon Niemeyer course through spin doctoring.25 Real- force of deliberative governance.31 Author- world deliberative processes provide con- itative power is also necessary for delib- siderable evidence on deliberation’s poten- erative democracy, which requires lead- tial to build capacities of traditionally mar- ers who are receptive to the concerns of ginalized groups. Economist Vijayendra affected publics and have the legitimate Rao and sociologist Paromita Sanyal’s work authority to consider and act on the pub- on gram sabhas in South India is a landmark lic’s preferences and concerns.32 Actors in study, demonstrating village-level deliber- and around deliberative processes can also ations’ capacity to mobilize civic agency strategize to advance agendas and address among the poor, counteracting resource inequalities.33 scarcity and social stratification.26 Brazil’s Deliberative democrats recognize that National Public Policy Conferences–one coercive power pervades social relations, of the biggest nationally successful exercis- but understand that certain kinds of power es in public deliberation–illustrate how or- are needed to maintain order in a deliber- dinary citizens influence public policy once ative process, to address inequalities, and they acquire the opportunity to take part in to implement decisions.34 consequential deliberation.27 Productive deliberation is plural, not consen- These examples illustrate deliberative de- sual. A seeming commitment to the pursuit mocracy’s record in curtailing, rather than of consensus–that is, agreement on both perpetuating, elite domination by creating a course of action and the reasons for it– space for ordinary political actors to create, once provided a target for critics of delib- contest, and reflect upon ideas, options, and erative democracy, who stressed its other- discourses. worldly character and silencing of dissident Deliberative democracy has a nuanced view voices.35 However, contrary to these argu- of power. Early critics of deliberative de- ments, deliberative democrats have rare- mocracy worried about its political na- ly endorsed consensus as an aspiration for iveté, particularly its neglect of power and real-world decision-making (as opposed to strategy.28 However, deliberative democ- one theoretical reference point). racy is not naive about power, but rather Decision-making in deliberative de- has a nuanced approach to it. mocracy can involve voting, negotiation, In the deliberative ideal, coercive forms or workable agreements that entail agree- of power, defined as the threat of sanction ment on a course of action, but not on the or use of force against another’s interests, reasons for it. All of these benefit from de- are absent because they distort communi- liberation, which can involve clarification cation.29 But deliberative practice reveals of the sources of disagreement, and under- that coercive power is ubiquitous: it per- standing the reasons of others. Rather than vades the very process of argumentation consensus, deliberation should recognize and communication, affects the remit and pluralism and strive for metaconsensus, organization of deliberative procedures, which involves mutual recognition of the and shapes the broader policy context.30 legitimacy of the different values, prefer- Procedural designs can, however, limit ences, judgments, and discourses held by coercive power by, for example, selecting other participants.36

146 (3) Summer 2017 31 Twelve Key At first sight, this acceptance of plural- practice par excellence, but rather are used Findings in ism and metaconsensus might seem to as a tool to democratize other facets of po- Deliberative Democracy contradict the findings of political scientist litical life and deepen the quality of political Research Jürg Steiner and colleagues that the more participation. consensual a system of government, the Second, the political projects of partici- better the quality of deliberation that oc- patory and deliberative democracy are in- curs in its legislature. Consensual democ- timately linked. Pateman’s aspirations for a racies–notably the Nordic countries, The “participatory society,” in which various as- Netherlands, Germany, and Switzerland– pects of our social and political lives are de- are also arguably the world’s most success- mocratized, are not distinct from delibera- ful states on a variety of indicators, sug- tive democrats’ vision of a society in which gesting a strong correlation between delib- all citizens affected by a decision have ca- eration and public policy success, though pacities and opportunities to deliberate in correlation here does not necessarily imply the public sphere.39 This has been articulat- causality. However, the concept of consen- ed by “macro” deliberative theorists, whose sual liberal democratic states (as opposed focus is to improve the quality of political to adversarial) does not imply consensus in participation in the public sphere, whether the strong sense we identified. Consensu- online or offline, mediated or face-to-face, al states are still pluralistic, but their plu- such that citizens can affect political pro- ralism is channeled into workable agree- cesses on issues they care about. ments, not adversarial point-scoring. Deliberative transformation takes time. De- Participation and deliberation go together. A liberation by definition requires amena- sharp distinction between participation bility to preference transformation, but and deliberation is drawn by political theo- such transformation may not be a good rist Carole Pateman, who argues that delib- measure of the quality of deliberation.40 erative democrats have shown “little inter- While large changes in preferences can est in the last thirty years of participatory occur early in deliberative processes, this promotion” and instead focus on mini-pub- change can reflect anticipation of absorb- lics or “new deliberative bodies.”37 ing information and group deliberation as This distinction misfires. First, while it much as the effect of deliberation proper.41 is true that a large number of deliberative The goal of deliberation is for citizens scholars research mini-publics, these stud- to determine reflectively not only prefer- ies are motivated by the desire to better un- ences, but also the reasons that support derstand how lessons learned from small- them.42 As we have already noted, at the scale deliberative forums can be scaled up group level, this involves the formation of to mass democracies and enhance the qual- a kind of metaconsensus featuring mutu- ity of political participation. So, for exam- al recognition of the manner in which be- ple, John Dryzek and ecological economist liefs and values map onto preferences.43 Alex Lo have shown how particular rhetor- This process takes time and deliberation ical moves can increase the quality of rea- does not necessarily follow a smooth path. soning in a mini-public, which has direct Initial changes to preferences can even be implications for how climate change should partially reversed. The initial opening up be communicated in the public sphere (fur- of minds (as part of taking a deliberative ther examples will be provided in our dis- stance) and uptake of information rep- cussions of time, group polarization, and resents a dramatic threshold in the transi- divided societies).38 Mini-publics, in other tion toward deliberation proper, producing words, are not valorized as democratic changes that represent catharsis as much as

32 Dædalus, the Journal of the American Academy of Arts & Sciences deliberation. It is subsequent reflection that Their experiment involved citizens delib- Nicole Curato, produces deliberative preferences, only af- erating immigration in Finland, and after John S. Dryzek, 44 Selen A. Ercan, ter the stance is achieved. Consequently, deliberation, a group that was moderate- Carolyn M. reported results from very short delibera- ly to extremely hostile to immigrants shift- Hendriks & tive processes may only reflect the path to- ed toward a generally more tolerant opin- Simon Niemeyer ward, rather than the result of, deliberation. ion. After unstructured discussion, a simi- True deliberative transformation takes lon- lar group was, on average, more extreme. ger than that. Deliberation does, then, provide solutions Deliberation is the solution to group polariza- to group polarization, most obviously when tion. Cass Sunstein has claimed that a “law it moves beyond unstructured discussion. of group polarization” causes “deliberative Deliberative democracy applies to deeply di- trouble.”45 For if a group is made up of peo- vided societies. Deeply divided societies char- ple whose opinions range from moderate acterized by mutually exclusive religious, to extreme on an issue, after deliberation, national, racial, or ethnic identity claims the group’s average position will be closer challenge any kind of democratic politics, to the extreme. Thus, deliberation leads to including deliberative politics, which some unhealthy political polarization. There are skeptics believe belongs only in more order- three reasons why deliberative democracy ly and less fraught settings. Popular politi- does not succumb to this. cal solutions for deeply divided societies in- First, polarization depends crucially on stead involve power-sharing negotiated by group homogeneity, in which initial opin- elites from different blocs, leaving no space ions vary from moderate to extreme in a for public deliberation (indeed, communi- single direction, such as the degree of de- cation of any sort) across the divide.49 nial of climate science or the degree of sup- There is, however, growing empirical ev- port for public education. For anyone de- idence showing that deliberative practic- signing a deliberative forum, the solution es can flourish in deeply divided societies is simple: make sure there are participants to good effect, be it in association with, or from different sides on an issue. James at some distance from, power-sharing ar- Fishkin says this is exactly how his delib- rangements. Evidence comes from formats erative opinion polls resist polarization: a ranging from mixed-identity discussion random selection of participants ensures groups located in civil society to more struc- a variety of initial views.46 tured citizen forums with participants from Second, what Sunstein describes as polar- different sides.50 Mini-public experiments ization could, in many cases, be described as on deeply divided societies, for example, clarity. This is especially important for op- generate crucial lessons on how conversa- pressed groups struggling to find a voice.47 tions in the public sphere can be organized Talk with like-minded others can give peo- in such a way that they aid in forging mu- ple, individually and collectively, the confi- tual respect and understanding across dis- dence subsequently to enter the larger pub- cursive enclaves. As political scientist Rob- lic sphere; enclave deliberation can have ert Luskin and colleagues have noted, once positive effects in the deliberative system. assembled, conflicting groups in divided so- Third, political scientist Kimmo Grön- cieties can “have enough in common to per- lund and colleagues have demonstrated mit meaningful and constructive deliber- that polarization only applies under un- ation.”51 Such deliberation can promote structured conversation;48 polarization is recognition, mutual understanding, social not found when groups are run on standard learning about the other side, and even sol- deliberative principles with a facilitator. idarity across deep differences.52

146 (3) Summer 2017 33 Twelve Key Deliberative processes have been applied communicative dynamics of deliberative Findings in in divided societies such as South Africa, systems.57 Deliberative Democracy Turkey, Bosnia, Belgium, and Northern Deliberative democracy scholars deploy Research Ireland. Given the depth of the disagree- multiple research methods to shed light on ment among conflicting groups, delibera- diverse aspects of public deliberation in tive practices do not seek or yield consen- practice. Those who insist on using conven- sus (understood as universal agreement tional social science methods must recog- both on a course of action and the reasons nize that their results should be interpreted for it), but they play a crucial role in terms in light of this broader array of methods and of “working agreements” across the parties the breadth of understanding so enabled. to a conflict. Under the right conditions, deliberation in divided societies can help to We have surveyed what we believe to be bridge the deep conflicts across religious, a number of key resolved issues in the the- national, racial, and ethnic lines. ory, study, and practice of deliberative de- Deliberative research productively deploys di- mocracy. In a number of cases, we have reverse methods. Standard social science meth- plied to critics skeptical of the desirability, ods, such as surveys and psychological ex- possibility, and applicability of delibera- periments, are often used to study deliberative democracy. Our intent is not, however, tion. However, they do not do full justice to to silence critics. Rather, we hope that their the ability of deliberators to develop their efforts can be more tightly focused on the own understanding of contexts, which can real vulnerabilities of the project, rather extend to the kinds of social science instru- than its imagined or discarded features. ments that are appropriate and to questions However, we suspect that, in practice, our that should be asked. Standard methods summary of key findings will be more use- have a hard time capturing these dynamic ful to those seeking to advance or study the aspects of deliberative opinion formation, project, rather than those trying to refute it. and they tell us nothing about the broader For these scholars and practitioners, identi- political or social context in which public fying the resolved issues will leave them free deliberation occurs.53 to concentrate on unresolved issues. Innovative quantitative methods have been developed to remedy these short- comings:54 they can involve analyzing the content of deliberations to assess deliberative practice against normative standards, to measure the quality of deliberation, and to evaluate the intersubjective consistency of deliberators across preferences and values.55 Qualitative and interpretive methods have also generated empirical insights into public deliberation, particularly through in-depth case studies. Methods such as in-depth interviews and observation have been used to examine the views and behavior of political actors in and around deliberative forums.56 Frame and narrative analysis have been used to map discourses and analyze the

34 Dædalus, the Journal of the American Academy of Arts & Sciences endnotes Nicole Curato, * John S. Dryzek, Contributor Biographies: NICOLE CURATO is Australian Research Council Discovery Early Selen A. Ercan, Career Research Fellow at the Institute for Governance and Policy Analysis at the University Carolyn M. of Canberra, Australia. She has published articles in such journals as Policy Sciences, Policy Studies, Hendriks & and European Political Science Review. Simon Niemeyer JOHN S. DRYZEK is Australian Research Council Laureate Fellow and Centenary Professor in the Centre for Deliberative Democracy and Global Governance at the Institute for Governance and Policy Analysis at the University of Canberra, Australia. His recent books include Democratiz- ing Global Climate Governance (with Hayley Stevenson, 2014), Climate-Challenged Society (with Rich- ard B. Norgaard and David Schlosberg, 2013), and The Politics of the Earth: Environmental Discourses (third edition, 2013). SELEN A. ERCAN is Senior Research Fellow at the Centre for Deliberative Democracy and Global Governance at the Institute for Governance and Policy Analysis at the University of Canberra, Australia. She has published articles in such journals as International Political Science Review, Policy and Politics, Australian Journal of Political Science, and Critical Policy Studies. CAROLYN M. HENDRIKS is Associate Professor at the Crawford School of Public Policy at the Australian National University. She is the author of The Politics of Public Deliberation: Citizen En- gagement and Interest Advocacy (2011) and has published in such journals as Politics & Society, Po- litical Studies, and European Journal of Political Research. SIMON NIEMEYER is Australian Research Council Future Fellow at the Institute for Gover- nance and Policy Analysis at the University of Canberra, Australia. He has published articles in such journals as American Political Science Review, Politics & Society, Ethical Perspectives, and Australian Journal of Politics and History. 1 John Mueller, Capitalism, Democracy, and Ralph’s Pretty Good Grocery (Princeton, N.J.: Princeton University Press, 1999); and Daniel A. Bell, “Democratic Deliberation: The Problem of Im- plementation,” in Deliberative Politics: Essays on Democracy and Disagreement, ed. Stephen Macedo (Oxford: Oxford University Press, 1999), 71–87. 2 Stephen A. Coleman, Anna Przybylska, and Yves Sintomer, eds., Deliberation and Democracy: Innovative Processes and Institutions (New York: Peter Lang, 2015). 3 Tina Nabatchi, John Gastil, G. Michael Weiksner, and Matt Leighninger, eds., Democracy in Motion: Evaluating the Practice and Impact of Deliberative Civic Engagement (Oxford: Oxford Univer- sity Press, 2012). 4 Gregory Barrett, Miriam Wyman, and Vera Schatten, “Assessing Policy Impacts of Delibera- tive Civic Engagement,” in ibid., 181–203. 5 Carolyn M. Hendriks, “Coupling Citizens and Elites in Deliberative Systems: The Role of In- stitutional Design,” European Journal of Political Research 55 (1) (2016): 43–60. 6 See Nabatchi et al., Democracy in Motion. 7 John Parkinson and Jane Mansbridge, eds., Deliberative Systems: Deliberative Democracy at the Large Scale (Cambridge: Cambridge University Press, 2012). 8 William H. Riker, Liberalism Against Populism: A Confrontation Between the Theory of Democracy and the Theory of Social Choice (San Francisco: W. H. Freeman and Company, 1982). 9 Kenneth A. Shepsle, “Institutional Agreements and Equilibrium in Multidimensional Voting Models,” American Journal of Political Science 23 (1) (1979): 27–59. 10 David Miller, “Deliberative Democracy and Social Choice,” Political Studies 40 (1) (1992): 54– 67; and John S. Dryzek and Christian List, “Social Choice Theory and Deliberative Democracy: A Reconciliation,” British Journal of Political Science 33 (1) (2003): 1–23. 11 Christian List, Robert C. Luskin, James S. Fishkin, and Iain McLean, “Deliberation, Single-Peak- edness, and the Possibility of Meaningful Democracy,” Journal of Politics 75 (1) (2013): 80–95.

146 (3) Summer 2017 35 Twelve Key 12 Tali Mendelberg, “The Deliberative Citizen: Theory and Evidence,” in Political Decision Making, Findings in Deliberation and Participation: Research in Micropolitics, vol. 6, ed. Michael X. Delli Carpini, Leonie Deliberative Huddy, and Robert Y. Shapiro (Greenwich, Conn.: jai Press, 2002), 151–193. Democracy Research 13 Iris Marion Young, “Justice, Inclusion, and Deliberative Democracy,” in Deliberative Politics, ed. Macedo, 151–158. 14 George Loewenstein, Ted O’Donoughue, and Sudeep Bhatia, “Modelling the Interplay Be- tween Affect and Deliberation,” Decision 2 (2) (2015): 55–81; and Jason Barabas, “How De- liberation Affects Policy Opinions,” American Political Science Review 98 (4) (2004): 687–701. 15 David Owen and Graham Smith, “Survey Article: Deliberation, Democracy, and the Systemic Turn,” Journal of Political Philosophy 23 (2) (2015): 228. 16 Jane Mansbridge, Janette Hartz-Karp, Matthew Amengual, and John Gastil, “Norms of De- liberation: An Inductive Study,” Journal of Public Deliberation 2 (1) (2006). 17 Hugo Mercier and Hélène E. Landemore, “Reasoning is for Arguing: Understanding the Suc- cesses and Failures of Deliberation,” Political Psychology 33 (2) (2012): 243–258. 18 Simon J. Niemeyer, “Scaling Up Deliberation to Mass Publics: Harnessing Mini-Publics in a Deliberative System,” in Deliberative Mini-Publics: Practices, Promises, Pitfalls, ed. Kimmo Grön- lund, André Bächtiger, and Maija Setälä (Colchester, United Kingdom: ecpr Press, 2014). 19 Lynn M. Sanders, “Against Deliberation,” Political Theory 25 (3) (1997): 349. 20 Chantal Mouffe, “Politics and Passions: The Stakes of Democracy,” Ethical Perspectives 7 (2/3) (2000): 146–150. 21 Iris Marion Young, “Difference as a Resource for Democratic Communication,” in Delibera- tive Democracy: Essays on Reason and Politics, ed. James F. Bohman and William Rehg (Cambridge, Mass.: The mit Press, 1997), 383–406. 22 Jürg Steiner, André Bächtiger, Markus Spörndli, and Marco Steenbergen, Deliberative Politics in Action: Analyzing Parliamentary Discourse (Cambridge: Cambridge University Press, 2004). 23 Jensen Sass and John S. Dryzek, “Deliberative Cultures,” Political Theory 42 (1) (2014): 3–25. 24 James N. Druckman and Kjersten R. Nelson, “Framing and Deliberation: How Citizens’ Con- versations Limit Elite Influence,” American Journal of Political Science 47 (4) (2003): 729–745. 25 Simon J. Niemeyer, “The Emancipatory Effect of Deliberation: Empirical Lessons from Mini-Publics,” Politics & Society 39 (1) (2011): 103–140. 26 Vijayendra Rao and Paromita Sanyal, “Dignity Through Discourse: Poverty and the Culture of Deliberation in Indian Village Democracies,” The Annals of the American Academy of Political and Social Science 629 (1) (2010): 146–172. 27 Thamy Pogrebinschi and David Samuels, “The Impact of Participatory Democracy: Evidence from Brazil’s National Public Policy Conferences,” Comparative Politics 46 (2014): 313–332. 28 Ian Shapiro, “Enough of Deliberation: Politics is About Interests and Power,” in Deliberative Politics, ed. Macedo, 28–38. 29 Jane Mansbridge, James Bohman, Simone Chambers, et al., “The Place of Self-Interest and the Role of Power in Deliberative Democracy,” Journal of Political Philosophy 18 (1) (2010): 64–100. 30 Carolyn M. Hendriks, “Deliberative Governance in the Context of Power,” Policy and Society 28 (3) (2009): 173–184. 31 Ibid. 32 Jonathan W. Kuyper, “Deliberative Democracy and the Neglected Dimension of Leadership,” Journal of Public Deliberation 8 (1) (2012). 33 Jennifer Dodge, “Environmental Justice and Deliberative Democracy: How Social Change Organizations Respond to Power in the Deliberative System,” Policy and Society 28 (3) (2009): 225–239.

36 Dædalus, the Journal of the American Academy of Arts & Sciences 34 See Mansbridge et al., “The Place of Self-Interest and the Role of Power in Deliberative De- Nicole Curato, mocracy.” John S. Dryzek, 35 Selen A. Ercan, Iris Marion Young, “Communication and the Other: Beyond Deliberative Democracy,” in Carolyn M. Democracy and Difference: Contesting the Boundaries of the Political, ed. Seyla Benhabib (Princeton, Hendriks & N.J.: Princeton University Press, 1996), 120–135; and Aletta J. Norval, Aversive Democracy: Inher- Simon itance and Originality in the Democratic Tradition (Cambridge: Cambridge University Press, 2007). Niemeyer 36 John S. Dryzek and Simon J. Niemeyer, “Reconciling Pluralism and Consensus as Political Ideals,” American Journal of Political Science 50 (3) (2006): 634–649. 37 Carole Pateman, “Participatory Democracy Revisited,” Perspectives on Politics 10 (1) (2012): 8. 38 John S. Dryzek and Alex Y. Lo, “Reason and Rhetoric in Climate Communication,” Environ- mental Politics 24 (1) (2015): 1–16. 39 Pateman, “Participatory Democracy Revisited,” 10. 40 Lucio Baccaro, André Bächtiger, and Marion Deville, “Small Differences that Matter: The Impact of Discussion Modalities on Deliberative Outcomes,” British Journal of Political Science 46 (3) (2016). 41 Robert E. Goodin and Simon J. Niemeyer, “When Does Deliberation Begin? Internal Reflection versus Public Discussion in Deliberative Democracy,” Political Studies 51 (4) (2003): 627–649. 42 Bernard Manin, “On Legitimacy and Political Deliberation,” Political Theory 15 (3) (1987): 338–368. 43 See also Simon Niemeyer and John S. Dryzek, “The Ends of Deliberation: Metaconsensus and Intersubjective Rationality as Deliberative Ideals,” Swiss Political Science Review 13 (4) (2007): 497–526. 44 Simon J. Niemeyer, “When Does Deliberation Really Begin?” working paper series (Canberra, Australia: Centre for Deliberative Democracy and Global Governance, 2016). 45 Cass R. Sunstein, “Deliberative Trouble: Why Groups Go to Extremes,” Yale Law Journal 110 (1) (2000): 71–119. 46 James Fishkin, When the People Speak: Deliberative Democracy and Public Consultation (Oxford: Ox- ford University Press, 2009), 131–132. 47 Christopher F. Karpowitz, Raphael Chad, and Allen S. Hammond, “Deliberative Democra- cy and Inequality: Two Cheers for Enclave Deliberation among the Disempowered,” Politics & Society 37 (4) (2009): 576–615. 48 Kimmo Grönlund, Herne Kaisa, and Maija Setälä, “Does Enclave Deliberation Polarize Opin- ions?” Political Behavior 37 (4) (2015): 995–1020. 49 Arend Lijphart, Democracy in Plural Societies: A Comparative Exploration (New Haven, Conn.: Yale University Press, 1977). 50 Ian O’Flynn, “Divided Societies and Deliberative Democracy,” British Journal of Political Science 37 (4) (2007): 731–751. 51 Robert C. Luskin, Ian O’Flynn, James S. Fishkin, and David Russell, “Deliberating across Deep Divides,” Political Studies 62 (1) (2014): 117. 52 Bora Kanra, Islam, Democracy, and Dialogue in Turkey: Deliberating in Divided Societies (Aldershot, United Kingdom: Ashgate, 2009); and George Vasilev, Solidarity across Divides: Promoting the Moral Point of View (Edinburgh: Edinburgh University Press, 2015). 53 John S. Dryzek, “Handle with Care: The Deadly Hermeneutics of Deliberative Instrumenta- tion,” Acta Politica 40 (2) (2005): 197–211; and Selen A. Ercan, Carolyn M. Hendriks, and John Boswell, “Studying Public Deliberation After the Systemic Turn: The Crucial Role for Inter- pretive Research,” Politics & Policy 45 (2) (2017): 195–218. 54 Laura W. Black, Stephanie Burkhalter, John Gastil, and Jennifer Stromer-Galley, “Methods for Analyzing and Measuring Group Deliberation,” in The Sourcebook of Political Communication

146 (3) Summer 2017 37 Twelve Key Research: Methods, Measures, and Analytical Techniques, ed. R. Lance Holbert (New York: Rout- Findings in ledge, 2009), 323–345. Deliberative 55 Democracy See, for example, Katharina Holzinger, “Kommunikationsmodi und Handlungstypen in den Research Internationalen Beziehungen. Anmerkungen zu einigen irrefuhrenden Dichotomien,” Zeitschrift Für Internationale Beziehungen 8 (2) (2001): 243–286; Steiner et al., Deliberative Politics in Action; and Niemeyer, “The Emancipatory Effect of Deliberation.” 56 Carolyn M. Hendriks, “Praxis Stories: Experiencing Interpretive Policy Research,” Critical Policy Analysis 1 (3) (2007): 278–300. 57 John Boswell, Carolyn M. Hendriks, and Selen A. Ercan, “Message Received? Examining Transmission in Deliberative Systems,” Critical Policy Studies 10 (3) (2016): 263–283; and Ercan et al., “Studying Public Deliberation After the Systemic Turn.”

38 Dædalus, the Journal of the American Academy of Arts & Sciences Political Deliberation & the Adversarial Principle

Bernard Manin

Abstract: Retrieving an insight dating back to antiquity, this essay argues that the confrontation of opposing views and arguments is desirable in political deliberation. But freedom of speech and diversity among deliberators do not suffice to secure that outcome. Therefore we should actively facilitate and encourage the presentation of contrary opinions during deliberation. Such confrontation is our best means of improving the quality of collective decisions. It also counteracts the pernicious fragmentation of the public sphere. It facilitates the comprehension of choices. Lastly, arguing for and against a given decision treats the minority with respect. This essay proposes practical ways of promoting adversarial deliberation, in particular the organization of debates disconnected from electoral competition.

In a liberal democracy, the existence of conflicting opinions about the policies the country should adopt is a fact of life. In a democracy committed to the deliberative ideal, it is also a desirable situation. If such oppositions did not appear, the supporters of deliberation would have to encourage them. In contemporary democracies, there is one institution that requires the presence of opposing points of view: the judicial trial. Here, what we can call the “adversarial principle” obliges the judge to render a decision only after having heard both parties. The maxim audiatur et altera pars (“let the other side be BERNARD MANIN is Professor at heard”), gradually forged by the jurists of the Middle the École des Hautes Études en Ages, is today considered a fundamental legal prin- Sciences Sociales, Paris, and Pro- ciple. It is at the heart of our conception of a fair ju- fessor of Politics at New York Uni- dicial procedure. versity. His publications include But a trial and a political deliberation are two dif- “On Legitimacy and Political De- ferent processes. It is not clear that the rules that are liberation” (in Political Theory, vol. 15, no. 3, 1987), The Principles of Rep- appropriate for the settlement of legal disputes are resentative Government (1996), and also appropriate for collective reflection about the La démocratie délibérative (forth- decisions to take as a polity. In a trial, the judge is, coming). by definition, confronted with two points of view

39 Political that are opposed to each other. But in a po- is contentious, it will probably divide the Deliberation litical deliberation, the points of view be- body of deliberators. In these conditions, & the Adversarial ing advanced are not necessarily in oppo- the idea of an obligation of impartiality be- Principle sition. Faced with a collective problem, we comes problematic. To whom are the delib- often find not only supporters and adver- erators supposed to prove themselves im- saries of a given solution, but also support- partial, and for what reason? If, in politi- ers of solutions that, while different, are cal deliberation, the value of hearing both not always mutually exclusive. For the in- sides derives from reasons other than the junction to hear the other side of things to demand for impartiality, we need to dis- apply in these conditions, we must inten- cover those reasons. tionally organize deliberation so that it fos- Thus, the idea of regulating political delib- ters a confrontation of opposing positions. eration according to the adversarial princi- One could, for example, consider by turns ple raises several questions. It is this idea that the arguments for and against each prop- I would like to defend in this essay. I will at- osition advanced, or limit a debate only to tempt to show why it is desirable that polit- propositions that are mutually exclusive. If, ical deliberation be organized as a confron- in political deliberation, the opposition of tation between opposing points of view in different points of view ought to be inten- which the participants conform to the prin- tionally organized (as this cannot be simply ciple that the other side must also be heard. presupposed), we must show why it is desirable to deliberate in this adversarial, for- The idea of organizing collective delib- and-against manner. Such a demonstration eration as an adversarial debate, a debate is not necessary in the case of a trial. for and against a position, is not new. It has Another feature separates a political de- its origins, or at least one of its early illus- liberation from a trial. In a trial, the two trations, in classical antiquity. Recent con- parties in conflict bring their disagreement ceptions of deliberative democracy and the before a third person, the judge, so that she practices they have inspired (such as De- may decide the matter in a manner that is liberative Polls, citizens’ juries, and con- just and in accord with the law. The judge is sensus conferences) have led us to for- not a party to the conflict that she is called get an older idea of political deliberation, upon to adjudicate (no one can be a judge formulated by Greek and Roman histori- in her own cause). Moreover, she is not her- ans and theorists of rhetoric, from Hero- self affected by the decision that she impos- dotus to Quintilian via Thucydides, Aris- es on the parties to the suit. We might say totle, and Cicero. Today’s conceptions of that the trial presents a triadic structure in deliberative democracy put the emphasis which an exterior arbiter decides upon a on discussion, making it essential that the conflict between two parties. To fulfill this members of the deliberating group discuss function, the judge must be impartial. That among themselves, engage in dialogue, and she must equally hear both parties seems exchange arguments with one another. the logical consequence of this demand for The opposition of points of view, if men- impartiality. Political deliberation, on the tioned, occupies a secondary place. It is other hand, does not present such a struc- viewed either as a precondition to delib- ture. When citizens or representatives de- eration proper or simply as a natural conse- liberate about decisions to take, they do quence of pluralistic societies in which the not decide on affairs that are external to expression of opinions is free. Yet in the an- them; rather, they deal with their own af- cient idea of deliberation, the opposition fairs. If the question under consideration of points of view occupied a central place.

40 Dædalus, the Journal of the American Academy of Arts & Sciences To simplify, in the ancient conception, ora- pressed,” he has Artabanus explain, “it is Bernard tors advocating opposed policies each pre- not possible to choose the one which it is Manin sented arguments in favor of their position best to adopt.”5 Note that this argument is and against their opponent’s. These argu- purely epistemic: to hear contrary opinions ments were presented before an assembly is necessary for discovering the right an- that subsequently decided on the policy. swer. No consideration of fairness enters. It seems reasonable to assume that mem- The importance of the adversarial prin- bers of the assembly also discussed the ar- ciple in the ancients appears not only in guments among themselves. But the oppo- theoretical writings; we also find its reflec- sition of points of view–not mutual dis- tion in institutional practices. To wit, af- cussion–constituted the motor and chief ter the oligarchical revolutions and the res- element of deliberation. toration of democracy at the start of the It is in this way, for example, that Thucyd- fourth century bce, the Athenians adopted ides presents the major scenes of delibera- two institutions that mandated a for-and- tion in The Peloponnesian War: the debate at against debate: the graphè paranomon and Sparta about the decision to enter the war, the nomothetai.6 The graphè paranomon au- the Athenian debate over the punishment thorized a decision already voted on by the to inflict on the inhabitants of Mytilene be- assembly to be brought before the courts tween Cleon and Diodotus, and the debate (in which the judges were ordinary citizens in the Athenian assembly over the Sicilian selected by lot) on the grounds that the de- expedition where Nicias and Alcibiades cision was contrary to the law or simply confronted one another.1 In these scenes, harmful to the Athenian people. The plain- the orators who advance opposing points of tiff and the citizen who had proposed the view do not discuss among themselves and contested decree would then plead their do not seek to persuade their opposition. cases before the courts. The decree was an- Rather, they seek to convince the assembly nulled if the verdict went in favor of the to whom they address their speeches. Thu- accusation. Any of the assembly’s deci- cydides presents these scenes as delibera- sions could thereby be submitted to a sort tions of the assembly. At various points in of second reading before the courts. How- his account of the debate over Mytilene, he ever, this second reading, which possessed refers to it as “deliberation.”2 In such delib- greater authority than the first, needed to eration, the driving element is the hearing of include an adversarial debate, while the opposed persuasive speeches. first examination by the assembly might Similarly, on several occasions in Politics, have proceeded without oppositions. Aristotle indicates that the task of the as- The institution of the nomothetai illus- sembly is to deliberate on common affairs.3 trates even more clearly the benefits ex- In Rhetoric, we find a more precise descrip- pected from the use of the adversarial meth- tion of the assembly’s deliberative activity: od in the political realm. Here, the action orators arguing for and against the decisions did not unfold before the courts, and there being contemplated.4 Here, as well, the or- was neither plaintiff nor defendant. The ators speak, offering opposed opinions and adversarial form, nevertheless, was main- arguments, but the citizens deliberate. tained. Whenever it seemed desirable to ab- In a scene of deliberation among the Per- rogate and replace certain laws, proposals sians recounted in The Histories, Herodotus for new laws were put forth and announced reflects on the benefits expected from the in public venues in advance. These propos- method of opposed speeches. “If opinions als were then defended before the nomothetai contrary to one another have not been ex- by their initiators. At the same time, five

146 (3) Summer 2017 41 Political citizens were elected for the purpose of de- technical terms, or in terms of values–if the Deliberation fending the existing laws whose abrogation proposals for action have been submitted to & the Adversarial had been proposed. criticism beforehand. Criticism permits the Principle The institutions of the nomothetai and the elimination, or at least the reduction, of pro- graphè paranomon were adopted as means posals involving false factual beliefs, logical for protecting the newly restored democ- errors, or objectionable moral choices. We racy from the impassioned and hasty de- do not need to repeat here the arguments cisions from which the city had suffered, establishing the epistemic merits of criti- especially during the Peloponnesian War. cism. No one today would deny that criti- To protect against this outcome, Athenian cism is one of the best means at our dispos- democrats turned to institutions that con- al to test the quality, technical and moral, of ferred the most authoritative decisions on practical proposals. groups of citizens who would necessarily Nevertheless, the conclusion that we gen- have heard the arguments for and against erally draw from these arguments is that it the measures under consideration. is enough simply to establish the freedom to express criticism to produce its benefits. Just because the ancients conceived of This is without doubt how Mill reasoned. political deliberation as a confrontation We find an even more striking expression of between opposing views is not, in itself, a this position in the famous free-speech dis- reason for adopting their models. All that sent of Supreme Court Justice Oliver Wen- history can do is open us to perspectives we dell Holmes. Men, he wrote, will eventual- perhaps would not have otherwise thought ly realize “that the ultimate good desired is of. We must now ask ourselves why it might better reached by free trade in ideas–that be desirable to organize political delibera- the best test of truth is the power of the tion according to the principle of hearing thought to get itself accepted in the com- both sides. I see four principal reasons for petition of the market.”7 Yet the conclu- doing so: sion that the free exchange of ideas is a sufficient guarantee is not justified. Freedom 1) Improving the quality of collective of speech on its own does not ensure that decisions the right to criticize will be exercised. Fur- 2) Counteracting the fragmentation of thermore, as we shall see, the fact that crit- the public sphere icisms are put forward does not guarantee 3) Facilitating the comprehension of that they will receive proper consideration. choices Several mechanisms can prevent the 4) Treating the minority with respect. ability to criticize freely from leading to its exercise. I group these together here under Let us begin with improving the quali- the label of conformism. Social psychology ty of collective decisions. A long tradition has long told us (with elaboration in Rous- of thought, including, in particular, the re- seau) that people want to be liked. Being flections of John Stuart Mill and Karl Pop- esteemed and approved of by others also per, has highlighted what we can call the provides a gratifying image of one’s self. It epistemic virtues of criticism. Several argu- follows that when people perceive that, in ments have shown that to submit an idea to a given social environment, others’ opin- criticism constitutes one of the best means ions lean in a certain direction, they tend to of testing its validity. This holds for prac- bend their own expressions in the same di- tical ideas. A decision is more likely to be rection to gain the approval of their peers. of a high quality–whether in factual and They avoid expressing dissenting or critical

42 Dædalus, the Journal of the American Academy of Arts & Sciences views, reinforcing the mechanism of the bly have the liberty to oppose the measure Bernard “spiral of silence.”8 Even in the context of and criticize one another. Manin discussion, we observe the disposition to It appears, then, that if one wants to ob- conform to what is perceived as the norm tain from political deliberation the favor- within the group. This phenomenon is at able epistemic effects of criticism, the ex- the heart of what has been named, since pression of opposing opinions must be en- the works of psychologist Serge Moscovici, couraged, not merely permitted. the polarizing effect of group discussions.9 But there is another reason to ensure Thus, even if criticism is formally free, a that the participants in a deliberation are powerful social force works to marginal- actually confronted with opposing points ize, or even to stifle, its expression. of view; it concerns the reception of argu- Contrary to what liberal theorists often ments, rather than their production. Stud- affirm, social or cultural diversity within ies in social and cognitive psychology show a group does not suffice to assure a con- that, confronted with new information or frontation of opinions critically opposed evidence, people have a systematic propen- to one another. Suppose, for example, a de- sity to see in it a confirmation of their pre- liberating body whose members, while be- vious beliefs. In a now-classic experiment, ing diverse with regard to social position, psychologists presented the same ensemble education, and beliefs, also share a fear of of documents and studies concerning the some danger. Let us imagine, furthermore, death penalty and its effects to two groups that this assembly discusses a measure that of subjects selected on the basis of their an- would contribute to the reduction of this tecedent opinions: one group composed of danger; for example, strengthening the subjects favorable to the death penalty, the powers of the police. In this context, it is other subjects who are rather hostile to it. unlikely that collective deliberation will After being confronted with these docu- bring forth many arguments against the ments, the group that was favorable to the adoption of this measure, however much death penalty became more favorable to social or cultural diversity there may be it, and the group hostile to it became still in the group. Rather, the discussion will more hostile.10 This phenomenon is par- produce an accumulation of reasons in ticularly marked when the documents pre- favor of increased police power, with var- sented to the subjects were ambiguous and ious members finding, from within their called for interpretation. The propensity to own particular perspectives, diverse rea- find support for one’s antecedent beliefs is sons for adopting this course of action known as “confirmation bias.”11 that others, differently situated, may not Research has also shown that group dis- have seen on their own. Yet even if increas- cussion reinforces the effects of confirma- ing the prerogatives of the police did, in tion bias. Groups interpret information fact, contribute to the realization of the with more bias than do individuals; and desired end, the measure might also pres- they privilege information that supports ent undesirable effects or features in other their antecedent beliefs to an even greater ways. Collective deliberation should pre- degree than do individuals.12 Two mecha- cisely bring to light these potential nega- nisms explain this tendency. First, as not- tive effects and weigh them in the balance ed earlier, group settings accentuate ten- against the benefits of the measure. But in dencies that predominate among individ- our case, the assembly will systematically uals. If privileging information supportive underestimate these possible negative fea- of prior beliefs is already the dominant tures even though members of the assem- tendency among individuals, it is not sur-

146 (3) Summer 2017 43 Political prising that this tendency should be am- that the gathering of diverse points of view Deliberation plified in group discussion. But a second, will spontaneously produce a clash of argu- & the Adversarial and more surprising, mechanism is also ments pro and contra, nor that it will bring Principle at work. It seems that groups tend to dis- about a balanced consideration of views. cuss principally the information that was Mill was wrong to assume that, in a society already known to all the members before or an assembly composed of diverse mem- the beginning of the discussion. Within bers, opposing opinions would already be groups, it turns out, discussion turns essen- there, waiting to be set against one another tially on shared knowledge. Members of once they were allowed to be uttered. Mill the group are reluctant to discuss those bits wrote: “The most intolerant of churches, of information that are known only to one the Roman Catholic Church, even at the or a few other members prior to the dis- canonization of a saint, admits, and listens pa- cussion.13 Shared information appears to tiently to, a ‘devil’s advocate.’”16 He failed to have more weight in the eyes of the group see that the presence of a devil’s advocate members, and has a greater chance of be- was required, not merely admitted. And ing mentioned during the discussion, and through the requirement, the Church se- thus remembered later.14 Finally, informa- cured that objections to the canonization of tion supporting the position preferred by a given person were aired and considered, the greatest number of group members has even if no individual would otherwise have a greater likelihood of becoming the object spontaneously offered them.17 of discussion than information supporting the contrary position.15 The confrontation of opposing opinions Collective discussion thus tends to pro- also has merit beyond eliciting unshared duce a disproportionate volume of informa perspectives. It unifies the field in which tion and arguments in favor of the already- opinions are formed and expressed, coun- dominant belief in the group. teracting the fragmentation of the public If one wishes to check the effects of con- sphere. In order to be opposed to an opin- firmation bias–a phenomenon to which ion and to contest it, it is necessary that one groups are particularly vulnerable–one be cognizant of that opinion and take it into must take proactive measures. One can, consideration. In a society in which points for example, call special attention to ar- of view are objectively diverse, the open and guments contrary to prior beliefs by high- explicit clash of opposing ideas is neither lighting them (literally) or by making them the natural state nor the sole possible condi- cognitively more salient. Not only is the tion. Another configuration is just as likely: free expression of a multiplicity of voices mutual ignorance. The German sociologist not sufficient to assure the confrontation Georg Simmel therefore argued that con- of opposing views, but the mere expression flict between social groups paradoxically of contrary arguments is not sufficient for served the cause of social integration: first, others to understand those arguments or by placing the conflicting groups into a re- consider them objectively. lationship with one another and, second, In the absence of measures that active- by exerting a pressure for unity among the ly induce individuals to pay particular at- secondary divisions within each group. We tention to evidence and points of view op- can advance a similar argument in matters posed to their own, collective deliberation of opinion. The clash of opinions unifies the will have the greatest likelihood of simply field in which beliefs confront one another, reinforcing antecedent opinions. In a polit- creating a space in which those beliefs are ical deliberation, in short, we cannot expect addressed to and respond to one another.

44 Dædalus, the Journal of the American Academy of Arts & Sciences This task of mutual addressing is harder within the political domain. Progressive Bernard when the space of opinions is fragment- blogs and forums link to other progressive Manin ed into a multitude of islets, homogenous sites but not to conservative sites, and vice within themselves but formed in conditions versa. From these islets and networks of of little communication with outsiders. like-minded individuals we can expect the Several factors–some older, some of increased effects of reinforcement and po- more recent origin–now trend in the di- larization, because, in general, interacting rection of this sort of fragmentation. First, with people of similar beliefs pushes one we have long known that people are selec- more toward the extreme positions of the tive in their choice of contacts and social views common to the group.20 relations. They tend disproportionately to Finally, the movement toward residential be in contact with people who share their segregation, which has already progressed political opinions.18 Psychologically, many in the United States and is at work today in fear the face-to-face expression of political Europe, further contributes to the fragmen- disagreement and want to avoid it as much tation of the public space of communica- as possible. tion. If it is true that opinions are strong- More recent factors also work in the di- ly correlated with sociocultural and ethnic rection of fragmentation: the develop- factors, then in a neighborhood whose in- ment of cable television and its themat- habitants share the same sociocultural or ic stations, the spread of the Internet, and ethnic profile, each person is likely, for the finally the movement toward residential most part, to encounter neighbors who and territorial segregation. Although the share the same opinions. A selective expo- effects of these transformations are still sure to similar opinions emerges de facto. difficult to estimate, they all present an Faced with these forces of fragmenta- analogous structure: people are now of- tion, only intentional collective action can fered, in multiple ways, greater oppor- be expected to produce a degree of unifica- tunities for communicating and coming tion of the public political sphere. This sec- into contact only with other individuals ond justification for the deliberate encour- like themselves. agement of adversarial political debates is Cable television and the rise of opinion- particularly salient today. based television stations (a phenomenon currently more pronounced in the Unit- True, adversarial debate is by nature re- ed States than in Europe) provide view- ductive. Faced with some political prob- ers with the possibility of receiving a high lem, the polity usually has a multitude of proportion of their information only from possible courses of action, not all of which a channel to which they feel ideologically will be mutually exclusive. Yet the reductive close. Worse still, cable TV allows individ- character of the adversarial method is also uals with little interest in politics to avoid one of its merits. It simplifies complexity, political news altogether and watch only making the choices easier to grasp. There entertainment programs.19 is no doubt, for example, that the current For its part, the Internet has dramatical- economic situation in Europe and the Unit- ly increased the number and types of peo- ed States calls for a range of measures that ple with whom one can enter into contact. are more or less intermingled and comple- But studies on the usage of the Web sug- mentary to one another. To obtain a syn- gest that contacts and links are established optic view of these measures and their re- primarily through personal affinities, and lations and to choose among them would in particular through ideological affinities require considerable cognitive effort. There

146 (3) Summer 2017 45 Political are cognitive advantages to presenting the reached the age of autonomy, parents Deliberation policy response as a choice between reduc- must justify the orders they give them. & the Adversarial ing public deficits now and maintaining or When they are not yet autonomous, chil- Principle increasing these deficits in the short term dren must obey orders simply because they to prevent further decline until the econo- are orders. So, too, if the minority mem- my has regained its normal growth rate. As bers have not had the possibility of hear- democrats, we cannot discount the value ing the reasons for the decision they must of such cognitive simplicity. Groups of ex- obey against their wishes, they are placed perts may be able to deliberate without us- in a situation of having to obey the order ing the adversarial method. But if we want simply because it is an order, or simply be- ordinary (or even well-informed) citizens cause it obtained the most votes. I do not to participate ably in collective delibera- mean to imply that the members of the mi- tion, the simplification achieved by the ad- nority will consequently be more disposed versary system is an almost indispensable to obey the decision. Sometimes justifica- instrument. tions exacerbate the opposition. But justifying orders shows greater respect for the The fourth principle in support of the ad- autonomy of those receiving them. versarial method is the value of treating the On the flip side, the reasons for not tak- minority with respect. No matter how con- ing the decision that ultimately triumphed scientiously citizens deliberate, it is likely would also have been put forward. These that disagreement will remain at the end criticisms and objections did not prevail, of the process. Decisions will therefore be but they were at least articulated and made taken by the majority. The majority of peo- public. From the majority’s perspective, ple will get to live with the decision they because it won, it will naturally think that desired; a minority of people will have to it was right; but in the process, it had to lis- live under a decision they did not support. ten to the opposition explaining their jus- To be sure, the decision itself formed the tification. The members of the majority minority: it did not exist per se before the were at minimum forced to see that there vote. But the manner of conducting the were reasons supporting the other side. It deliberation before the vote entails con- seems reasonable to think that, as a result, sequences for the treatment of those who, the majority will be less inclined to con- after the vote, will make up the minority. sider the minority as unintelligent or ill- If the deliberation has been conducted as intentioned. a debate between opposed positions, with each camp presenting its reasons in favor of Before I proceed to the practical conse- its position and criticizing those advanced quences that we can draw from my argu- by the opposition, two consequences fol- ment, I must first respond to an objection: low. After a vote has been taken, the mi- that rendering obligatory the presentation nority must obey the decision, but at least of opposing points of view in the public the reasons aiming to justify this decision sphere would require constraints on pub- will have been formulated and made public discourse and encroachments on free- lic. The minority might still refuse to lis- dom of speech. ten to these reasons seriously and in good In response to this objection I would faith, but it was given the chance to con- first suggest turning to an institution that sider them. The minority members were in the relatively recent past followed just therefore treated with the respect owed to this path: the fairness doctrine in effect in the autonomous beings. Once children have United States from 1927 to 1987. The fair-

46 Dædalus, the Journal of the American Academy of Arts & Sciences ness doctrine, implemented by the Feder- of speech in the public sphere, an inter- Bernard al Communications Commission, made it pretation that focuses on the rights of the Manin obligatory for radio and television stations receiving public. That the U.S. Supreme to give an evenly balanced presentation of Court has since rejected this interpreta- “opposing viewpoints on controversial is- tion does not mean that the arguments ad- sues of public importance.” The fairness vanced in Red Lion were objectively weak. doctrine not only imposed equality in air- They are, in any case, consistent with the time; it also required the presentation of claims of this essay. viewpoints opposed to one another. The doc- What, then, should we do in practice to trine did not apply to airtime during elec- foster the confrontation of opposing ar- toral campaigns, which was regulated on guments in today’s democracies? With- other terms. It applied instead to any ques- out claiming to provide a complete and de- tion that became the object of public con- tailed response to this question, I will con- troversy outside of electoral periods. The clude by suggesting two concrete means constitutionality of this doctrine was up- for promoting the adversarial principle in held by the U.S. Supreme Court in the fa- politics. The first is a practice yet to be in- mous 1969 decision of Red Lion Broadcast- vented, which would be implemented out- ing Co. v. FCC.21 side of electoral periods. The second con- The central argument that the Court insists of reinforcing a practice already used voked in support of the constitutionality in electoral campaigns. of the fairness doctrine was that, in regard First, my suggestion for the future. Out- to liberty of expression on the airwaves, side of electoral periods, civil-society actors it is “the right of the viewing and listen- (such as foundations or think tanks) could ing public, and not the right of the broad- organize adversarial debates on subjects of casters, which is paramount.” The Court public interest. These debates would not thus held that listeners and viewers had be regularly scheduled, but would be orga- the right to hear conflicting viewpoints in nized only when a question sparked signif- order to make up their mind on the issues: icant interest from the public (as with such “Speech concerning public affairs is more topics as nuclear energy, assisted suicide, than self-expression; it is the essence of or, in certain countries, the wearing of the self-government.”22 hijab) or when a large number of citizens The fairness doctrine was abandoned for mobilized in favor of a cause. More gener- two reasons. First, the doctrine led radio ally, these public debates would not aim to and television stations to avoid controver- replace any existing democratic practices sial subjects for the sake of not exposing (such as electoral campaigns or parlia- themselves to lawsuits claiming they had mentary debates), but would complement violated the law. Second, the question of them. what exactly constituted the opposition of Neither the exposition of conflicting one point of view against another became viewpoints nor communication across the subject of repeated litigation, and the ideological divides can be made mandatory. fcc proved unable to reduce the insecu- This does not mean that it is useless to try rity and juridical uncertainty that arose to facilitate them. Indeed, the probability on this front. Despite its eventual aban- of being confronted with opposing points donment, however, the Red Lion decision of view matters: it tends to make one’s shows that the obligation for the media to thoughts more anticipatory, careful, and present conflicting viewpoints is compat- subtle.23 In contemporary circumstances, ible with a certain interpretation of freedom this probability tends to diminish. The ac-

146 (3) Summer 2017 47 Political tive promotion of adversarial debates aims The other implication of the principle of Deliberation to counteract this pernicious tendency. relevant reasons concerns the choice of par- & the Adversarial Given that the goal of these debates would ticipants. Speakers should be permitted to Principle be to further the formulation and diffusion defend policies that promote their own in- of arguments for and against a given public terests, but only on the condition that their decision, they should be guided by the fol- interests be both publicly declared and lowing principle: speakers should defend linked to the substance of the policy they or criticize a given policy or position only recommend–not to external interests like with reference to its own merits, and not in advancing their careers or promoting objec- response to reasons external to the policy or tives that have no connection with the policy position. The arguments advanced in these under debate. Speakers should thus mainly debates should concern the advantages or be experts, representatives of associations, disadvantages–whether technical or moral activists, and persons enjoying a recognized –inherent in the decision. I call this the moral authority. Politicians might partici- principle of relevant reasons. This principle has pate so long as that participation satisfies the two implications: the first concerns simpli- principle of relevant reasons: namely, that fying the debate to one issue; the second they address themselves solely to the ques- concerns choosing the right participants. tion under debate, to the exclusion of other In order to encourage citizens to take themes in the platform of their party. Their account of and weigh the reasons for and professional and career prospects should against a given decision, each question that play no role in these debates. can be defined objectively and independent The divisions that emerge within these from other questions should be debated debates would not have a depoliticized char- separately. Multidimensionality and the acter, but they would be distinguished from bundling of different questions undermine normal partisan divisions in two respects: 1) the coherence of the arguments. they would concern each theme taken sep- To be sure, at election time, the voter will arately, as opposed to entire programmatic vote for a candidate or party that has bun- platforms and 2) they would, as much as is dled questions without an objective con- possible, be disconnected from the stakes of nection between them. Such grouping may electoral power and competition. be desirable, because it permits negotia- As mentioned, the speakers should be rep- tions between different strands of the par- resentatives of associations and of activist ty. Nevertheless, to understand the bun- movements, as well as experts and person- dling and negotiation well enough to cast ages of recognized authority. The role of an informed vote, the voter needs to have these speakers would be to present and de- thought through the different issues sepa- fend opposing policies on the topic being rately, aided by adversarial debate. debated, in a focused and well-argued way. It is probably too difficult to completely The presence of experts and persons of mor- exclude nonrelevant reasons–that is, rea- al stature in each debate would help control sons not substantively linked to the policy the quality of the arguments and supporting in question–at the moment of organizing evidence. A live audience could question the a deliberative debate. But the principle re- speakers and criticize their positions. In this mains valid: nonrelevant arguments should way, the audience could be engaged and not be sidelined as much as possible. As a conse- just confined to a passive role. Afterward, quence, each debate should focus on a spe- the debate’s organizing institution could cific theme, rather than on platforms com- post video of the event online and open dis- prising multiple dimensions. cussion forums, for further conversations.

48 Dædalus, the Journal of the American Academy of Arts & Sciences Second, my endorsement of current prac- protagonists will be advancing clear argu- Bernard tice: televised adversarial debates between ments for and against specific policies. Manin the leaders of competing parties or coali- One may object that in-person confrontations during electoral campaigns make an tions between party leaders lend too much important contribution to mass political weight to the personality of the speakers. deliberation. These debates already exist But the personalization of electoral com- in several countries. One of the oldest de- petition seems to be a permanent feature mocracies in the world, Britain, recently in- of our world. If personalities are going to troduced them, for the first time, in the 2010 play an important role anyway, a setting elections. In general, such debates are suc- in which the participants are reciprocal- cessful at drawing an audience: typically ly encouraged to bring to light the other’s several million people follow them. They defects is preferable to one in which each are thus one of the very rare occasions when leader can deploy his personal advantag- large numbers of citizens think about the es without contest. Experience shows that same subject at the same time and are con- plebiscitary leaders have little fondness scious of so being joined in common at- for contradiction; electoral debates would tention. This coordination of time and make avoiding contradiction difficult. Ad- the object of attention makes possible– versarial electoral debates are not an inno- and even generates–conversations in the vation, but they have not taken root every- places of daily life, like cafés or meetings where. Leaders sometimes forbid them. So with friends. Media coverage of debates it is not useless to defend the principle that also puts citizens in contact with conflict- underlies the existing practice. A culture ing viewpoints without imposing on them that has accepted the principle of adversari- the psychic discomfort caused by face-to- al debate–recognizing that such debate im- face political disagreement. Without doubt, proves the quality of decisions, counteracts such debates often lack substance and argu- fragmentation, facilitates the comprehen- ment. But sometimes they have a good deal sion of choices, and respects minorities– of both. And it is possible to arrange such would discredit any leader who sought to debates so as to make it more likely that the escape the test.

endnotes Author’s Note: This essay was originally written in French. I wish to thank Greg Conti for his help in turning the French into English. I am also grateful to Jane Mansbridge and James Fishkin for their comments and suggestions. 1 Thucydides The Peloponnesian War 1.48–88, 3.36–49, 6.8–26. Thucydides also presents other assembly debates using the same structure: for instance, the debate over the war in Syracuse between Hermocrates and Athenagoras (6.33–51) and the debate at Camarina between Eu- phemus and Hermocrates (6.76–88). 2 Thucydides The Peloponnesian War 3.36.6, 3.42.1, 3.44.3, 3.44.4. 3 Aristotle Politics 3.1.1275b.18–20, 3.6.1281b.31, 4.11.1297b.35–1299a.1. 4 Aristotle Rhetoric 1.1.1354b.10–1355a.11, 1.2.1358b.3. 5 Herodotus The Histories 7.10. 6 The graphè paranomôn (literally, a “suit against bills contrary to the laws”) was a prosecution against the proposer of a law that was thought to be in conflict with existing law. The nomothetai was a board of five hundred that oversaw legal reform. Its members were selected by

146 (3) Summer 2017 49 Political lot from a larger pool of citizens aged thirty years or older, themselves selected by lot from Deliberation among volunteers. & the Adversarial 7 Abrams v. United States, 250 U.S. 630-1 (1919). Principle 8 See the Asch experiments in Solomon E. Asch, “Effects of Group Pressure upon the Modifi- cation and Distortion of Judgment,” in Leadership and Men, ed. Harold Guetzkov (Pittsburgh: Carnegie Press, 1951). More recently, see Elisabeth Noelle-Neumann, The Spiral of Silence: Public Opinion–Our Social Skin (Chicago: University of Chicago Press, 1987). 9 See Serge Moscovici and Marisa Zavalloni, “The Group as Polarizer of Attitudes,” Journal of Personality and Social Psychology 12 (2) (1969): 125–135. See also Serge Moscovici, Social Influence and Social Change (London: Academic Press, 1976). 10 Charles G. Lord, Lee Ross, and Mark R. Lepper, “Biased Assimilation and Attitude Polarization: The Effects of Prior Theories on Subsequently Considered Evidence,” Journal of Personality and Social Psychology 37 (1) (1979): 2098–2109. 11 For an overview, see Matthew Rabin, “Psychology and Economics,” Journal of Economic Liter- ature 36 (1) (1998): 11–46. 12 Stefan Schulz-Hardt, Dieter Frey, Carsten Lüthgens, and Serge Moscovici, “Biased Information Search in Group Decision Making,” Journal of Personality and Social Psychology 78 (4) (2000): 655–669. 13 The importance of these results ought to be underscored. The fact that group discussion involves primarily already-shared information constitutes a major difficulty for the thesis that discussion promotes the pooling and aggregation of knowledge that is dispersed among the agents. This thesis, often mentioned today, perhaps finds its origin in Aristotle’s famous argument: “That is why the many are better judges of music and the writings of poets. Some appreciate one part, some another, and all together appreciate all.” Granted, Aristotle does not mention discussion in this context. But extending his argument to discussion settings sounds prima facie unproblematic. The experimental findings reported here show that this is not the case. Aristotle Politics 3.11.1281b.8–10. 14 Garold Stasser and William Titus, “Pooling of Unshared Information in Group Decision Making: Biased Information Sampling during Discussion,” Journal of Personality and Social Psychology 48 (6) (1985): 1467–1478; Daniel Gigone and Reid Hastie, “The Common Knowledge Effect: In- formation Sharing and Group Judgment,” Journal of Personality and Social Psychology 65 (5) (1993): 959–974; and Dennis D. Stewart and Garold Stasser, “The Sampling of Critical, Unshared Information in Decision-Making Groups: The Role of an Informed Minority,” The European Journal of Social Psychology 28 (1) (1998): 95–113. 15 Stasser and Titus, “Pooling of Unshared Information in Group Decision Making.” 16 John Stuart Mill, On Liberty, ed. David Bromwich and George Kateb (New Haven, Conn.: Yale University Press, 2003 [1859]), 91. The emphasis is mine. 17 The intervention of the “defender of the faith” (such was the title of what in ordinary language was named the “devil’s advocate”) in the canonization procedure was abolished by Pope John Paul II in 1989. 18 For example, see Diana C. Mutz and Paul S. Martin, “Facilitating Communication across Lines of Political Difference: The Role of the Mass Media,” American Political Science Review 95 (1) (2001): 97–114. 19 Markus Prior, Post-Broadcast Democracy: How Media Choice Increases Inequality in Political Involvement and Polarizes Elections (Princeton, N.J.: Princeton University Press, 2007). 20 Cass R. Sunstein, Republic.com (Princeton, N.J.: Princeton University Press, 2000). 21 Red Lion Broadcasting Co. v. FCC, 395 U.S. 367–401 (1969). 22 On this argument, see Owen Fiss, Liberalism Divided: Freedom of Speech and the Many Uses of State Power (Boulder, Colo.: Westview Press, 1996). 23 Mutz and Martin, “Facilitating Communication across Lines of Political Difference.”

50 Dædalus, the Journal of the American Academy of Arts & Sciences Deliberative Democracy as Open, Not (Just) Representative Democracy

Hélène Landemore

Abstract: Deliberative democracy is at risk of becoming collateral damage of the current crisis of representative democracy. If deliberative democracy is necessarily representative and if representation betrays the true meaning of democracy as rule of, by, and for the people, then how can deliberative democracy retain any validity as a theory of political legitimacy? Any tight connection between deliberative democracy and representative democracy thus risks making deliberative democracy obsolete: a dated paradigm fit for a precrisis order, but maladjusted to the world of Occupy, the Pirate Party, the Zapatistas, and other anti- representative movements. This essay argues that the problem comes from a particular and historically situated understanding of representative democracy as rule by elected elites. I argue that in order to retain its normative appeal and political relevance, deliberative democracy should dissociate itself from representative democracy thus understood and reinvent itself as the core of a more truly democratic paradigm, which I call “open democracy.” In open democracy, popular rule means the mediated but real exercise of power by ordinary citizens. This new paradigm privileges nonelectoral forms of representation and in it, power is meant to remain constantly inclusive of and accessible–in other words open–to ordinary citizens.

The motivating concern for this essay is the impact that the crisis of representative democracy, widely diagnosed by political commentators and democratic theorists alike, has or should have on deliberative democracy as a mainstream theory of democratic legitimacy. To the extent that the fate of deliberative democracy has become intimately intertwined with representative democracy as both a normative paradigm and a set of particular historical institutions, HÉLÈNE LANDEMORE is Associ- and to the extent that representative democracy is ate Professor of Political Science at under attack precisely for being representative and Yale University. Her publications keeping ordinary citizens at arm’s length of the real include Democratic Reason: Politics, site of decision and power, deliberative democrats Collective Intelligence, and the Rule of the Many (2013), Hume: Probabilité should be worried about the status of their theory. et choix raisonnable (2004), and Col- Deliberative democracy risks becoming collateral lective Wisdom: Principles and Mech- damage of the problems currently facing represen- anisms (ed. with Jon Elster, 2012). tative democracy.

51 Deliberative Deliberative democrats thus need to Deliberative democracy is a theory of Democracy clarify the relationship between delibera- democratic legitimacy that traces the au- as Open, Not (Just) tion and representation and, more gener- thority of laws and policies to the public Representative ally, deliberative democracy as a theory of exchange of arguments among free and Democracy legitimacy, on the one hand, and represen- equal citizens. This theory was developed tative democracy as a specific institution- in the late 1980s and 1990s as an alterna- al instantiation of democracy, on the other. tive to the then-dominant theory of aggre- This clarification should reveal that while gative democracy, whereby democratic le- the connection between deliberation and gitimacy stems simply from the proper ag- representation might indeed be essential, gregation of votes in free and fair elections at least in mass societies, the relation be- pitting various elites against one another. tween deliberative democracy as a theory The relation of deliberative democracy of legitimacy and representative democ- to representative democracy has always racy as a historical paradigm is essentially been undertheorized. Early proponents contingent: it is possible to separate the of the theory assumed direct democracy two. I suggest that deliberative democracy as their base model.1 Nothing much was is better seen as an independent theoretical supposed to change, normatively speak- module that is compatible with, and indeed ing, when deliberation took place among better suited to, a different set of institutional elected representatives rather than the peo- principles than the one called “representa- ple themselves. The legitimacy was sim- tive democracy.” I propose that deliberative ply transferred to the outcomes of the de- democracy should be made a central part of liberation among representatives, as if it a new and more attractive paradigm of de- played out as a perfect substitute for democracy, which I call open democracy. liberation among all citizens. To ensure a The first section of this essay scrutinizes seamless translation of democratic legit- the relation of deliberation and represen- imacy from the direct to the representa- tation in mainstream theories of deliber- tive context, most people resorted to the ative democracy and shows the problems then-dominant theory of representation that arise when deliberative democracy is formulated by political scientist Hannah confused or too tightly associated with rep- Pitkin in 1967. At an abstract level, repre- resentative democracy. The second section sentation is, for Pitkin, the conceptual solu- shows that representative democracy can- tion to the problem of “making present” not be salvaged as a normative model of de- that which is absent. Democratic legitima- mocracy because it fails at least three basic cy was found at the level of a representative criteria we should expect a genuinely demo- assembly making present and pursuing the cratic rule to satisfy (namely agenda-setting, interests of people who could not be pres- effective participation, and enlightened un- ent all at once. derstanding). The third section sketches out Given that direct deliberation among all an alternative: open democracy. Open de- citizens is widely assumed to be impossible mocracy is meant as a more authentically on the scale of the modern nation-state,2 democratic paradigm in which deliberation this simplifying premise of the early de- among free and equal members–the core liberative democrats was perfectly under- of deliberative democracy–is made a cen- standable, and most other deliberative the- tral institutional principle. As a result, I ar- orists took it onboard. Philosopher Jürgen gue that open democracy offers to delibera- Habermas, in a way, merely complicated tive democrats a more hospitable home than the picture by conceptualizing two kinds representative democracy. of deliberation happening in two distinct

52 Dædalus, the Journal of the American Academy of Arts & Sciences deliberative “tracks.” The first kind of de- tives allows for a reflexive delay between Hélène liberation was meant to be formal and the expression of raw judgments and pref- Landemore decision-oriented, taking place within the erences, on the one hand, and the crafting walls of Parliament. The other, taking place of policy outcomes, on the other.6 Repre- among the public, was decentralized, dis- sentation also allows a circular process of tributed, informal, and diffuse, with the as- communication between representatives sumed function of setting the agenda for and the represented. Representative de- Parliament.3 Habermas additionally posit- mocracy is, for Urbinati, a more accom- ed a porous demarcation between the two plished form of democracy than direct tracks, so as to allow for feedback loops be- democracy precisely because it allows for tween the two spheres. In so doing, he plau- a discursive exchange to occur over time sibly extended the early version of delibera- between representatives and represented. tive democracy, making it applicable to the If this account of the link between delib- actual world of representative democracies. eration and representation in mainstream More recently, however, democratic the- democratic theory is correct, what happens orists have modified and tightened the nor- to the paradigm of deliberative democracy mative link between deliberation and repre- as a theory of political legitimacy when rep- sentation to the point that one can hardly be resentative democracy itself comes under conceptualized without the other. The first attack? What happens when the relation- move has been to show that representation ship assumed and described by Habermas is, in a nutshell, the essence of democracy. between representatives and represented For politics scholar David Plotke, “repre- no longer seems a plausible or normatively sentation is democracy” in the sense that appealing theory of the way things work representative practices are always “con- and ought to work, in particular because it stitutive” of democracy.4 Representation is no longer credible that the informal pub- no longer consists primarily in making lic sphere can set up the agenda for the more present the absent, but in constructing the formal one? What happens when repre- demos and its interests. Similarly, for po- sentation no longer is democracy, as Plot- litical theorist Sofia Näsström, representa- ke has it, but becomes instead, as Rousseau tive democracy is a “tautology” because it warned long ago, its very demise? is only through representative structures There is no question, at this point, that and practices that the demos constitutes representative democracy is in the midst of itself.5 If the authors behind this so-called a serious crisis, at least if one is to judge by constructive turn are right, though, then the the recent numbers of books and articles on task of deliberation among citizens is not the topic.7 Institutionally, the symptoms are delegated to representatives only for rea- well-known: voting absenteeism, the de- sons of size and convenience. Deliberation cline of parties as vehicles for mass partici- must become the affair of representatives, pation,8 abysmal rates of approval for pol- rather than directly that of citizens them- iticians and legislatures across much of the selves, in order to be truly democratic. Western world, the rise of populist move- Political theorist Nadia Urbinati’s theory ments and the return of calls for more direct of representative democracy exemplifies forms of democracy–as are some of the a similar view. For Urbinati, deliberation causes–in the United States, a near com- among a smaller number of representa- plete lack of correlation between majority tives is not just equivalent but superior to preferences and policy outcomes when elite direct deliberation among all citizens. This preferences differ from those of the major- is because deliberation among representa- ity,9 rising economic inequalities across the

146 (3) Summer 2017 53 Deliberative Western world, and a sense that democ- This electoral and elitist character of rep- Democracy racies have been emptied of their mean- resentative democracy runs deep. It is un- as Open, 10 Not (Just) ing, if not altogether replaced by the rule surprisingly evidenced in the way politi- Representative of experts, bureaucrats, and judges.11 By cal theorists have captured the institutional Democracy contrast, both populist and authoritarian principles of representative democracy. movements are on the rise. These move- These, it turns out, fall short of being dem- ments have in common an antirepresenta- ocratic. In order to show this, I apply to tive stance that signals the problems with Bernard Manin and Nadia Urbinati’s list representative democracy and is some- of principles of representative democracy times meant to hasten its demise. the five criteria that Robert Dahl advances While it is likely that the crisis of repre- in his classic Democracy and Its Critics as the sentative democracy is in part due to exter- benchmarks of authentic democratic asso- nal factors (such as globalization and tech- ciations: 1) effective participation; 2) voting nological change or what some see as the equality at the decisive stage; 3) enlightened crisis of capitalism in the West), it can also understanding; 4) control of the agenda; plausibly be traced to more fundamental and 5) inclusion of all adults. These five cri- design flaws. To understand what may be teria are, according to Dahl, “criteria that a wrong with representative democracy per process for governing an association would se, it helps to look critically at its core prin- have to meet in order to satisfy the require- ciples, a task to which I now turn. ment that all the members are equally entitled to participate in the association’s de- Representative democracy is the para- cision about its policies.”15 digm we associate with the form of democ- Effective participation means that there racy that emerged in the eighteenth century must be a direct connection between pop- at the time of the French and American rev- ular involvement and ultimate decision- olutions. It can be defined as a regime cen- making. Voting equality at the decisive tered on the elections of elites who act as stage means, very simply, “one person, one trustees of and make decisions on behalf vote” at the stage when decisions are made of the larger population.12 In theory, repre- final. Enlightened understanding means sentation need not involve election (I will that citizens must be able to pass informed return to this point). In practice, however, judgment on the matters deliberated or elections have become part of the very defi- voted on. Control of the agenda means that nition of representative democracy, partly the set of issues deliberated on should be because the theories developed to justify it defined by the citizens themselves.16 Inclu- crucially associate popular sovereignty with sion of all adults means that all adult mem- democratic authorization, and democratic bers of the demos (itself more polemically authorization, in turn, with consent ex- defined by Dahl as the people directly af- pressed through the ballot box.13 Thus, al- fected by the laws and policy outcomes) be though democratic representation need not given a share of power. imply elections, representative democracy Now, let us use these five democratic cri- has come to mean electoral democracy.14 As teria to assess the four institutional prin- a result, a core feature of representative de- ciples that Manin has articulated as cap- mocracy is the delegation of agenda-setting, turing the core of “representative gov- deliberation, and decision-making to a sub- ernment” (historically the first version set of the polity that is distinct from ordi- of representative democracy): 1) periodic nary people and explicitly identified and elections; 2) independence of the repre- chosen as a separate elite. sentatives; 3) freedom of opinion; and

54 Dædalus, the Journal of the American Academy of Arts & Sciences 4) trial by discussion.17 The first princi- which the franchise is universal and voting Hélène ple, periodic elections, is the most cen- rights strictly equal so that, by default, cri- Landemore tral and is the one that most people asso- teria 2 and 5 (equality at the decisive stage ciate with democracy. It is a principle of and inclusion of all adults) are met. Even the authorization of representatives, re- then, or so I argue, the principles of what newed at periodic intervals. The periodic- we are now supposed to call (and are used ity is crucial in that, in theory, it ensures to calling) representative democracy still not only renewed consent and thus autho- fail three out of Dahl’s five criteria. rization, but also the accountability and re- Indeed, neither effective participation, sponsiveness of the representatives. Elec- nor agenda-setting, nor enlightened un- tions thus double as democratic principle derstanding are credibly ensured by the and accountability mechanism. The sec- mere ability to elect one’s leaders every ond principle, the relative independence four years and, in between, publicly crit- of elites from their constituents, ensures a icize their decisions from outside the sites meaningful space for the exercise of judg- of decisive power. Representative democra- ment by the representatives, who can de- cy does not, in theory, require any form of part from their constituents’ preferences popular participation besides voting and, as needed. The third principle–freedom because it also does not credibly accommo- of opinion–counterbalances the second by date, let alone commit to, agenda-setting by ensuring that representatives, despite their ordinary citizens, it even weakens voting as freedom of judgment, can be criticized for a form of effective participation. Access to their decisions and choices. Popular pres- power is only possible through becoming sure does not jeopardize representatives’ elected, a path that, even in theory, is open independence but supposedly ensures, like only to people endowed with certain qual- periodic election, a form of accountability ities and, in practice, is mostly restricted to and responsiveness, including, crucially, people with either money or connections. in the period between elections. Manin’s Nor does representative democracy require last feature of representative government or guarantee enlightened understanding on is that public decisions are subject to trial the part of citizens. On the contrary, peri- by discussion, a feature one may equate odic elections and the independence of rep- with the deliberation at the heart of de- resentatives are intended to compensate for liberative democrats’ theories. the assumed absence of popular enlight- How does this list of the established prac- enment about political issues. On certain tices of representative government fare in Schumpeterian or “realist” versions of rep- light of Dahl’s normative criteria? Arguably resentative democracy, no room is made for it satisfies none of them. First, the principle democratic deliberation among ordinary of periodic elections does not specify uni- citizens as a vehicle for individual and col- versal franchise or the principle of “one per- lective enlightenment, since the latter is son, one vote,” and is thus fully compatible seen as either pointless or even counterpro- with voting systems based on a tax thresh- ductive.18 Representative democracy, final- old and plural voting schemes. As such, ly, also allows for the possibility of a com- representative government can violate plete disconnect between the decisions of both the second and last criteria: namely, representatives and the preferences of the voting equality at the decisive stage and represented, at least to the extent that the inclusiveness. But let us assume that these critical bite of “freedom of opinion” proves principles of representative government insufficient to bind elected representatives should today be applied only to a system in to their constituents’ preferences.

146 (3) Summer 2017 55 Deliberative The only democratic credentials of rep- perspectives, and interests of the popula- Democracy resentative democracy therefore seem to tion are not only reported on, but also made as Open, Not (Just) reside with the authorization and account- present in the political sphere in a way that Representative ability supposedly ensured by the princi- reflects some minimal amount of identifi- Democracy ple of periodic elections in a context of cation and similarities between represented universal suffrage and equal voting rights. and representatives. The argument that authorization at the Because of these two addenda, Urbinati’s voting booth and accountability through picture of representative democracy is more retrospective voting amounts to genuine democratic and thus more normatively at- rule of the people may have worked in the tractive than representative democracy as it eighteenth century, when such a promise can be theorized on the basis of Manin’s his- seemed radical compared with past and ex- torical account. Urbinati’s theory, howev- isting regimes. But today, at the beginning er, accepts as a given the premise that dem- of the twenty-first century, who can still ar- ocratic representation must be electoral and, gue this with a straight face? The reality is despite the promise of a participatory model that representative government was mostly of representation, seemingly limits citi- designed to maintain the people at a safe zens’ possibility for action to judgment, crit- distance from any actual decision-making icism, and deliberation, all of them decou- power. Manin wrote that representative pled from actual decision-making power. As government, as a set of institutional prin- in Manin’s representative government, in ciples, replaced the ability to hold office Urbinati’s representative democracy, citi- that citizens enjoyed in Ancient Athens zens can protest and criticize all they want, with the mere ability to consent to power. but they are not meant to have any form of Expanding the franchise over the last two direct access to the decision-making pro- hundred years has allowed the advocates cess. Similarly, the ability to set the agen- of representative government to call it rep- da is missing from her model. Citizens can resentative democracy without altering this hope to influence the representatives’ agen- fundamental and problematic fact.19 da only through the blunt mechanism of Urbinati’s normative theory of represen- elections and the indirect pressure of pub- tative democracy arguably elevates this his- lic opinion. torical substitution (of consent for exercise Like that of other prominent deliberative of power) to the status of normative ideal. democrats, such as Habermas, Urbinati’s Urbinati’s list of principles of represen- theory assumes a reflexive and smooth cir- tative democracy includes all of Manin’s, cularity between the sphere of opinion for- embraced as normatively desirable in their mation through which ordinary citizens ex- own right, rather than merely recognized as change ideas and form views in decentral- de facto historical practices.20 But she also ized and unregulated ways and the sphere makes two crucial additions to the list: ad- of the formal will expressed by party rep- vocacy and representativity. Advocacy could resentatives and government officials. Yet be read as a stronger version of Manin’s the dichotomization between the spheres third principle of freedom of public opin- of opinion and will operates as, or at least ion in that representatives are supposed to tolerates, de facto closure of government to listen to the criticisms and views voiced in ordinary citizens. In Habermas, the “sluice” the public sphere, integrate them into their metaphor that is supposed to capture the re- reflections, and make it their duty to make lation between the two deliberative tracks those claims known and considered. Rep- (the formal and the informal) similarly resentativity means ensuring that the views, suggests a filtering mechanism separating

56 Dædalus, the Journal of the American Academy of Arts & Sciences the unstructured deliberations of the peo- deliberative democrats ensure that delib- Hélène ple from those of elected elites. In the end, erative democracy is authentically dem- Landemore such dichotomies function as a way to close ocratic if it must also be representative? off the sphere of actual power and effective One way out is to try and bypass repre- deliberation to ordinary citizens. sentation altogether by developing models The history of representative democracy of inclusive deliberation among all citizens, and its conceptual elevation to a norma- as opposed to just their representatives. The tive ideal reveal that the crucial novelty of digital revolution has created the hope that this regime is not so much the indirectness the need for representation is now over and of the rule.21 Rather, the innovation is the that all citizens can and should now deliber- regime form’s reflexiveness, and the fact ate with each other at once, online, in what that this reflexiveness is ensured by plac- can be theorized as “mass online delibera- ing agenda-setting, deliberation, and deci- tion.”22 The recent “systemic” turn in de- sion-making power in the hands of elected liberative democracy may perhaps be read elites as opposed to ordinary citizens. Rep- as a similar extension of the hope of real- resentative democracy thus marks the pas- izing direct deliberation on a mass scale.23 sage from a citizen-centric and people-cen- Such an approach has to assume either that tric model of democracy to an elite-centric the people and its interests are self-reveal- and government-centric one. This elitism ing in immediate ways or can be constructed and government centricity are present in all in nonrepresentative ways. institutional versions of representative de- Another way out–more promising, in mocracy that have evolved since the eigh- my view–is to acknowledge that democ- teenth century: parliamentary, party, and racy is always representative but that “rep- now audience democracy. These three iter- resentative democracy” as a historical par- ations marked important expansions of the adigm is but one model of indirect or (more franchise just as they maintained, and argu- aptly) deliberative and reflexive democracy. ably deepened, the rift between the people But here, too, there are two possible strate- and the class of law- and policy-makers sup- gies. One is to reclaim the concept of repre- posed to represent them. In other words, sentation and build into it new, more dem- to put it bluntly, representative democracy ocratic meanings. This is the path currently as we know it has turned out to be an ex- taken by a number of democratic theorists. clusionary paradigm, not a truly demo- Michael Saward, for example, has argued cratic one. It satisfies, at best, only two of for “making representation strange again” Dahl’s democratic criteria (inclusiveness and redefining it away from electoral au- and equality at the decisive stage), failing to thorization, as well as one-to-one or one- meet the crucial standards of effective par- to-many relationships mediated by voting ticipation, enlightened understanding, and only, and toward a pluralized understand- control of the agenda. ing of representation as “claim-making.” In the same vein, a number of democratic the- If what I just said is true, it has potentially orists have started advocating for nonelec- worrying implications for deliberative de- toral forms of democratic representation. mocracy. Deliberative democrats cannot at In theory, nothing precludes us from re- the same time claim that proper delibera- covering the term representative democracy tion is only possible, and indeed desirable, to mean a truly democratic system. But my in representative bodies and that their the- sense is that, at this point, we are better off ory of legitimacy is unaffected by the cri- starting fresh, and this is for at least four rea- sis of representative democracy. How can sons. The first is semantic. If we accept the

146 (3) Summer 2017 57 Deliberative constructive turn and the view that “rep- size that my analysis presupposes the lexical Democracy resentation is democracy,” as Plotke has priority of two higher-order principles that as Open, Not (Just) it, then the expression “representative de- should be at the core of any form of democ- Representative mocracy” is largely redundant and uninfor- racy: namely, inclusiveness and equality. Democracy mative. We need a better, more meaningful Inclusiveness means both that every adult name. A second reason to abandon repre- member of the demos is entitled to a share of sentative democracy is historical. Represen- power and that the definition of the demos tative democracy was born as an alternative itself is inclusive.26 Equality means that this to democracy: the mixed regime known as share of power must be equal for all. Con- “representative government.” It was only cretely, equality will often mean “one per- slowly and painfully (and only somewhat) son, one vote” where voting (as distinct democratized over the last two centuries, from elections) is needed. This principle of with exclusionary trends arising to combat equality also means that each voice should each move toward inclusion. Despite theo- be given the same ex ante chance of being rists’ best efforts, one can only do so much to heard where deliberation is needed. Finally, change a fundamentally elitist and antidem- equality means that each individual has the ocratic construct into one in which power same opportunity of being a representative is exercised by ordinary citizens. The third where representation is needed. These two reason is pragmatic: it is simply too difficult higher-order principles, inclusiveness and at this point to clear the name of a paradigm equality, have to be assumed as underlying that is, the world over, associated with elec- (or lexically prior to) any of the other, lower- toral (and thus partly elitist) democracy.24 order principles that follow. Finally, a fourth reason has to do with the de Building on this, I propose that the main facto association of representative democra- five institutional principles of open de- cy with the nation-state and a narrow under- mocracy are: standing of what counts as “political.” In to- 1) Deliberation day’s global age, one can argue that our un- 2) The majoritarian principle derstanding of democracy should be more 3) Complex representation ambitious, expanding both laterally (to the 4) Rotation economic sphere) and vertically (to the in- 5) Openness. ternational level).25 My own suggestion, therefore, is to move The first principle, deliberation, forms entirely past and beyond “representative de- the core of the theory of democratic legit- mocracy.” Instead, deliberative democrats imacy that deliberative democrats have should build a new paradigm that places at convincingly developed over the last thir- its core democratic deliberation as a source ty years.27 Deliberation applied in the dem- of political legitimacy, meets basic demo- ocratic context is usually defined as the cratic standards such as effective participa- public exchange of reasons among free tion, agenda-control, and enlightened un- and equals. It is, to some degree, similar derstanding, and accommodates the reali- to Manin’s and Urbinati’s “trial by discus- ties and expectations of twenty-first-century sion” principle, except that deliberation is citizens. I offer below what I take to be an at- not assumed to involve ordinary citizens tractive version of such a new paradigm of only as members of a diffuse civil society democracy: open democracy. without access to direct decision-making power (as in Manin’s, Habermas’s, or Urbi- Let me offer a list of principles for open nati’s model). In this paradigm, democratic democracy. But before I do, let me empha- deliberation, whether direct or performed

58 Dædalus, the Journal of the American Academy of Arts & Sciences through representatives, must involve or- Rotation, as a fourth principle, ensures Hélène dinary citizens. This principle helps ensure that power be made to circulate and not Landemore that the system meets Dahl’s requirement stay with any subset of the polity for longer of “enlightened understanding.” than strictly necessary. In the context of ran- The second principle is the majoritar- domly selected assemblies characteristic of ian principle. It is, strangely, the princi- lottocratic representation, periodic rotation ple that often makes most people recoil in would have the beneficial effect of impeding fear of the “tyranny of the majority.” Yet group-think, corruption, the formation of above and beyond elections, majority rule static coalitions, and the creation of a sep- or some variant of it (such as majority judg- arate class of rulers. The mandates for ran- ment)28 is also the principle most widely domly selected or elected assemblies could associated with democracy. To the extent last from a few months to a few years, but that voting is necessary to resolve disagree- this principle makes it clear that the practice ments when deliberation does not produce of politics as a profession and politicians as a consensus, a default decision rule must be a separate caste is not part of this new ideal in place. The most democratic one, barring of democracy. While there should be plen- any good countervailing arguments to pos- ty of room for expert administrators in the it voting thresholds and minority vetoes, is machinery of government, the law and pol- some version of majority rule, for which icy decisions should ultimately be vetted by both strictly procedural and epistemic rea- ordinary citizens (properly educated for and sons can be adduced.29 The majoritarian informed about the tasks at hand), not ex- principle is the only principle that prevents perts or career politicians. To the extent that the domination of any minority. open democracy may still accommodate The third principle, complex representa- elected politicians, the ideal would ensure a tion, acknowledges that delegation of au- significant turnover of the personnel occupy- thority is both unavoidable in any reasoning these elected functions, not just through ably sized polity and desirable on its own, the periodicity of elections (which, as we now insofar as it allows for the discovery, articu- know, may ensure some responsiveness and lation, and even construction of shared in- accountability but does little for actual turn- terests. In a democratic context, however, over of the political personnel) but also, for representation should not necessarily (or example, through term limits. at all) translate into electoral modes of rep- Openness, finally, is an umbrella con- resentation. Thus, the principles of open cept for both direct popular participation democracy do not explicitly include the of different types and transparency.31 Be- principle of elections because elections, far cause representation always creates the from being a, let alone the democratic prin- risk of robbing the people of the capacity ciple, are merely one selection mechanism for effective participation, agenda-setting, among others.30 Instead, lottery-based rep- and enlightened understanding, one needs resentation–or “lottocratic representation” to introduce the counterbalancing principle of the kind arguably central to Ancient of openness, in which, in the ideal, citizens Athens–becomes the default democratic can make their voices heard at any point in mode of representation, though not nec- time, initiate laws when they are not satis- essarily the only one. In some contexts, fied with the agenda set by representative self-selection–and perhaps even reinvent- authorities, and keep an informed eye on ed forms of electoral representation–may every step of the political process. also prove an appropriate form of demo- Openness thus prevents the closure and cratic representation. entrenchment of the divide between repre-

146 (3) Summer 2017 59 Deliberative sented and representative that may accom- tive democracy. Unlike representative de- Democracy pany representation. Openness means that mocracy, which is fully compatible with as Open, Not (Just) power should flow through the body poli- purely aggregative (usually Schumpeterian Representative tic, rather than stagnate with a few people. or “realist”) models of democracy, open Democracy Openness should translate into process democracy explicitly places deliberation at transparency much of the time (though not its normative core. It also acknowledges the always transparency about substance). It majoritarian principle as pointing to a cer- should also translate into a citizens’ right of tain type of democratic default rule when initiative and other modes of direct, effec- deliberation does not produce a consensus tive participation. The principle of open- and disagreement subsists. And at least at ness is uniquely enabled by late-twentieth- the theoretical level, open democracy mea- and early-twenty-first-century technolo- sures up to basic democratic criteria that gies, such as the Internet, smartphones, representative democracy fails to satisfy, and social media. It is what makes open including effective participation, agenda- democracy most distinctive. setting, and enlightened understanding. In These five institutional principles are open democracy, ordinary citizens have a meant to operate under specific and en- meaningful chance to participate in law- abling conditions: liberal and what might and policy-making. They can be chosen by more inclusively be termed “empower- lotteries to occupy a position in significant ment” rights. Such rights constrain from political assemblies, something that under the outside the five institutional principles the right implementation should happen listed above. They also enable them by en- often enough. Even if they are not selected suring that everyone, including minority by the lottery process, citizens can freely members, is given a meaningful voice in access crowdsourcing platforms through the democratic process. Such rights may which their voice can be heard and can need to translate into quotas or parity laws make a difference to the outcome. In open ensuring that deliberations take into ac- democracy, ordinary citizens are also in count minority perspectives, especially in control of the agenda, either indirectly via contexts in which systematic minorities are randomly selected assemblies or more di- at risk of exclusion. Empowerment rights rectly via procedures such as a constitution- may also translate into rights of initiative, ally entrenched citizen’s initiative or a right which allow the discontents to challenge of referral. Having control of the agenda the status quo provided they garner a min- and a say in deliberation early in the pro- imal amount of support. Finally, to count- cess in turn renders voting, typically in a er the oppressive potential of the state, em- referendum, a genuinely effective form of fi- powerment rights may translate into spe- nal say and participation. The principles of cific protection rights for whistleblowers. deliberation combined with complex rep- These empowerment rights, however, need resentation and openness thus spreads en- not amount to full-fledged countermajori- lightened understanding among citizens. tarian constraints (such as vetoes, superma- In open democracy, democracy no lon- jority thresholds, or the creation of inde- ger means merely consenting to pow- pendent courts and agencies), which would er, as it does in our current understand- impede, rather than enable the principles of ing of representative democracy. It does open democracy. not always mean holding office, as it did in ancient democracy. But it means be- Let us now review the crucial differences ing able to access and thus hold power, between open democracy and representa- whether as a simple citizen able to influ-

60 Dædalus, the Journal of the American Academy of Arts & Sciences ence the agenda of the legislative assembly periodic elections, parties, and geography- Hélène through an initiative, the content of repre- based constituencies) as is representative Landemore sentatives’ deliberations through crowd- democracy, it opens itself to entirely new sourcing platforms,32 and the outcome of applications, including in firms, online a vote in a referendum whose options were communities, and at various levels of the shaped by his views; or even more directly, international stage. Open democracy al- by being chosen to participate in a ran- lows us to reinvent democratic politics for domly selected assembly charged with set- the twenty-first century. ting the agenda or making the law. I have not mentioned in this list of prin- In order for deliberative democracy to re- ciples the nature of the relation between main relevant in the crisis of representa- representatives and represented, because tive democracy, its advocates must distance in this new paradigm, the representative themselves from the paradigm of represen- relationship should be able to take many tative democracy, at least as this essay has forms as long as it is broadly democratic reconstructed an important critical read- (a question that needs a lot more investiga- ing of it. There are probably many ways to tion than can be conducted here). Most im- rescue deliberative democracy from the cri- portant, anyone should be able to be a rep- sis of representative democracy. The strate- resentative. This is what lottocratic repre- gy pursued here is to break entirely the cur- sentation would ensure by default, though rent association between deliberative de- one could envision a reimagined elector- mocracy and representative democracy al system along the lines of what is some- by sketching a new paradigm of democra- times theorized as “delegative” or “liquid” cy that maintains deliberation among free democracy, in which people can give their and equal individuals as the core of demo- votes to anyone they like, either for a spec- cratic legitimacy, but also complicates our ified amount of time or just to work on cer- understanding of democratic representa- tain issues, with the option of recall at any tion and detaches it from electoral mecha- time and the possibility of retaining the nisms. In this new paradigm of open democ- right to direct input throughout.33 racy, deliberation, the majoritarian princi- Open democracy, finally, marks a distinct ple, complex representation, rotation, and historical stage in the unfolding of the ideal openness would bring power back to the of democracy, including deliberative de- people, instantiating the ideal of people’s mocracy. Because it is not as tied down to es- rule (demokratia) more fully than represen- tablished practices and institutions (such as tative democracy as we know it.

endnotes 1 Joshua Cohen, “Deliberation and Democratic Legitimacy,” in The Good Polity, ed. Alan Hamlin and Philip Pettit (Oxford: Basil Blackwell, 1989). Otherwise anticipated by many other authors from Aristotle to Madison and John Stuart Mill, as well as, in its contemporary version, Joseph Bessette. 2 One could argue that Bruce Ackerman and James Fishkin’s Deliberation Day, as well as constitutional moments, count as direct deliberation on a national scale, but the reality is that they are more akin to deliberation among clusters of individuals occurring across the country, with no evidence that these clusters add up to what genuine mass deliberation should look like: namely, one single, integrated conversation among all individuals gathered in the same room.

146 (3) Summer 2017 61 Deliberative 3 Jürgen Habermas, Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy, Democracy trans. William Rehg (Cambridge, Mass.: The mit Press, 1996). as Open, 4 Not (Just) David Plotke, “Representation is Democracy,” Constellations 4 (1) (1997): 19–34; Sofia Näsström, Representative “Representation as Tautology,” European Journal of Political Theory 5 (3) (2006): 321–342; Nadia Democracy Urbinati, Representative Democracy: Principles and Genealogy (Chicago: University of Chicago Press, 2006), 328; Nadia Urbinati and Mark Warren, “The Concept of Representation in Contem- porary Political Theory,” Annual Review of Political Science 11 (2008): 387–412; and Lisa Disch, “Toward a Mobilization Conception of Democratic Representation,” American Political Science Review 105 (1) (2011): 100–114. 5 Näsström, “Representation as Tautology.” 6 This model departs from traditional views on the role of representatives (such as Madison’s or Burke’s) in that no assumption is made that the judgments of the representatives are in any way superior to those of the ordinary citizens. 7 Jack Hayward, ed., The Crisis of Representation in Europe (Abingdon, United Kingdom: Routledge, 1996); Colin Crouch, Post-Democracy (Cambridge: Polity, 2004); Pippa Norris, Democratic Deficit: Criti- cal Citizens Revisited (Cambridge: Cambridge University Press, 2011); Donatella della Porta, Can Democracy Be Saved? Participation, Deliberation, and Social Movements (Cambridge: Polity, 2013); Yannis Papadopoulos, Democracy in Crisis: Politics, Governance, and Policy (London: Palgrave, Mac- Millan, 2013); and Simon Tormey, The End of Representative Politics (Cambridge: Polity, 2015). 8 See Peter Mair, Ruling the Void: The Hollowing Out of Western Democracy (New York: Verso, 2013). 9 Martin Gilens and Benjamin Page, “Testing Theories of American Politics: Elites, Interest-Groups, and Average Citizens,” Perspectives on Politics 12 (3) (2014): 564–581. 10 Colin Crouch, Post-Democracy (Cambridge: Polity, 2004); and John Keane, The Life and Death of Democracy (New York, London: W. W. Norton & Company, 2009). 11 Frank Vibert, The Rise of the Unelected (Cambridge: Cambridge University Press, 2007); and Papadopoulos, Democracy in Crisis. 12 Bernard Manin, The Principles of Representative Government (Cambridge: Cambridge University Press, 1997). 13 Ibid. 14 For example, see ibid., 18; and Eric Nelson, “Prerogative, Popular Sovereignty, and the Amer- ican Founding,” in Popular Sovereignty in Historical Perspective, ed. Richard Bourke and Quentin Skinner (Cambridge: Cambridge University Press, 2016). 15 Robert Dahl, Democracy and Its Critics (New Haven, Conn.: Yale University Press, 1998), 37–38. 16 Political scientists in the twentieth century learned that power involves not only deciding (the most visible “face” of power) but also agenda-setting. See Peter Bachrach and Morton S. Baratz, “Two Faces of Power,” American Political Science Review 56 (4) (1962): 947–952; and Ste- ven Lukes, Power: A Radical View (London: Macmillan Press, 1974). 17 Bernard Manin, The Principles of Representative Government (Cambridge: Cambridge University Press, 1997). These four principles are not ideal criteria (like Dahl’s) but mere idealizations of established historical practices; thus, their fuction is essentially descriptive rather than normative. 18 See Christopher Achen and Larry Bartels, Democracy for Realists: Why Elections Do Not Produce Re- sponsive Government (Princeton, N.J.: Princeton University Press, 2016); and Ian Shapiro, “Col- lusion in Restraint of Democracy: Against Political Deliberation,” Dædalus 146 (3) (Summer 2017). 19 Manin, The Principles of Representative Government, 79. 20 Though she does not list them explicitly in her book, Urbinati mentions that she accepts all of Manin’s principles in Hélène Landemore, “Is Representative Democracy Really Demo-

62 Dædalus, the Journal of the American Academy of Arts & Sciences cratic? Interview of Bernard Manin and Nadia Urbinati, New York, April 10, 2007,” http:// Hélène www.booksandideas.net/Is-representative-democracy-really.html. Landemore 21 For example, agenda-setting in Ancient Athens, the classic archetype of “direct” democracy, was done by a randomly selected subset of citizens, the Boule–arguably a representative, rather than direct, assembly. 22 Cyril Velikanov, “Mass Online Deliberation,” unpublished article, https://www.academia.edu/ 12031548/Mass_Online_Deliberation (accessed December 25, 2016). 23 John Parkinson and Jane Mansbridge, Deliberative Systems: Deliberative Democracy at the Large Scale (Cambridge: Cambridge University Press, 2012). 24 Consider the Freedom House definition of democracy as a political system “whose leaders are elected in competitive multi-party and multi-candidate processes in which opposition parties have a legitimate chance of attaining power or participating in power.” 25 For a defense of democracy in the firm, see Hélène Landemore and Isabelle Ferreras, “Toward a Justification of the Firm-State Analogy: In Defense of Workplace Democracy,” Political Theory 44 (1) (2016): 53–81. 26 The exact criterion for inclusiveness may be too controversial to specify further, since the “all-affected” principle is often seen as too general and any restriction based on nationality, geography, or ethnicity too arbitrary. 27 By itself, deliberation is not a democratic principle. Deliberation becomes democratic when it is constrained by the underlying higher-order principles of inclusiveness and equality. Addi- tionally, deliberation must take place publicly in order to count as democratic. Embedded in the ideal of deliberation as a condition of legitimacy is another principle, orthogonal to that of equality: namely, the principle of substantive merit. Although all should have an equal chance of being heard, arguments should be judged on the merits. It is therefore legitimate that not all voices influence the outcome equally. 28 See Michel Balinski and Rida Lariki, Majority Judgment: Measuring, Ranking, and Electing (Cam- bridge, Mass.: The mit Press, 2011). In this book, Balinski and Lariki suggest substituting the traditional interpretation of majority rule as a collective ranking of the available alternatives with an interpretation of majority rule as a collective judgment or evaluation of these same alternatives (as in, typically, wine competitions). 29 I have myself offered an argument for the epistemic properties of democracy based on democratic inclusiveness as a proxy for the cognitive diversity necessary for smart deliberation. See Hélène Landemore, Democratic Reason: Politics, Collective Intelligence, and the Rule of the Many (Princeton, N.J.: Princeton University Press, 2013). 30 It is possible to imagine elections falling closer to democracy than they currently do on the con- tinuum between oligarchic and democratic device. Nonetheless, elections are a fundamentally Janus-faced selection mechanism that functions to exclude as well as include. Because their democratic credentials are only partial, elections should not be raised to the level of democratic principle. One could even imagine democracies that would do away with elections entirely, if elections proved too difficult to reconcile with the equality of opportunities to become a representative. For example, see David Van Reybrouk, Against Elections: The Case for Democracy (New York: Random House, 2016). 31 See the Obama administration’s concept of “open government.” Although the implementation fell short of the promise, the concept remains enticing as an umbrella for democratic values such as transparency, participation, and collaboration. 32 For examples of crowdsourcing as a participatory tool in policy processes, see Hélène Lande- more, “Inclusive Constitution-Making: The Icelandic Experiment,” Journal of Political Philosophy 23 (2) (2015): 166–191. 33 See Bryan Ford, “Delegative Democracy,” unpublished article (2002), http://www.bryno saurus.com/log/2002/0515-DelegativeDemocracy.pdf.

146 (3) Summer 2017 63 Inequality is Always in the Room: Language & Power in Deliberative Democracy

Arthur Lupia & Anne Norton

Abstract: Deliberative democracy has the potential to legitimize collective decisions. Deliberation’s legitimating potential, however, depends on whether those who deliberate truly enter as equals, whether they are able to express on equal terms their visions of the common good, and whether the forms and practices that govern deliberative assemblies advance or undermine their goals. Here, we examine these sources of deliberation’s legitimating potential. We contend that even in situations of apparent procedural equality, deliberation’s legitimating potential is limited by its potential to increase normatively focal power asymmetries. We conclude by describing how deliberative contexts can be modified to reduce certain types of power asymmetries, such as those often associated with gender, race, or class. In so doing, we hope to help readers consider a broader range of factors that influence the outcomes of attempts to restructure power relationships through communicative forums.

Deliberative democracy seems to offer democracy not only in our time, but in our neighborhoods. People meet as equals and reason together to find their way to a common good. We are not surprised, therefore, that deliberation is an idea with many advocates. Where people meet as equals, democracy is advanced. Where ARTHUR LUPIA, a Fellow of the people reason together, democracy is advanced. American Academy since 2007, is Deliberative democracy has the potential to legit- the Hal R. Varian Professor of Polit- imize collective decisions. Deliberation’s legitimat- ical Science and Research Profes- ing potential, however, depends on whether those sor in the Center for Political Stud- who deliberate truly enter as equals, whether they ies at the University of Michigan. are able to express on equal terms their visions of the ANNE NORTON is Professor of common good, and whether the forms and practices Political Science and Compara- that govern deliberative assemblies advance or un- tive Literature at the University dermine their goals. In this essay, we examine these of Pennsylvania. sources of deliberation’s legitimating potential. (*See endnotes for complete contributor Beneath and throughout the evaluation of delib- biographies.) erative democracy are questions about whether and

64 how language facilitates communication and imating potential of deliberative mecha- Arthur whether and how communications inform as- nisms is limited by the possibility that they Lupia & Anne sent. In attempts to measure the effective- can increase, rather than reduce, norma- Norton ness of deliberation, either theoretically tively focal power asymmetries. or empirically, it is common to reference Language and communication them- instances of consensus, compromise, or selves entail power relationships. Lan- clarifying sources of conflict as evidence guage gains meaning, and communication of success. Deliberative endeavors that fail becomes an efficient means of communi- to produce such outcomes are seen as less cating ideas, in part because language and successful. communication each build from, build on, The path to such outcomes travels and reify existing power imbalances. Lan- through sequences of communicative acts. guage issues from power, language creates These acts entail members of a society de- power, language is inseparable from pow- scribing their lifeworlds to one another. In er. Deliberative exercises that use language the deliberative ideal, participants are free and communication to produce assent and to make these descriptions without having legitimacy cannot help but produce their to filter them in ways that conform to ex- outcomes on the backs of existing pow- isting power imbalances. Participants de- er asymmetries. Even language environ- scribe their lives as they live and feel them. ments that claim to feature universal in- In the deliberative ideal, participants are clusion and procedural equality cannot be free to express their views on any social- assumed to be independent of deep and ly relevant issue. They need not subjugate potentially destructive power dynamics. themselves to dominant views of history, In what follows, we seek to inform delib- culture, and power. Through listening to eration as a means of producing legitimate these narratives, participants may come social decisions. We focus on the kinds of to an appreciation of diverse lifeworlds. power imbalances that are present in lan- Through this understanding, communi- guage and communicative practices. In so ties may come to realize shared norms and doing, we demonstrate how the language shared foundations for legitimate collec- and communication that people might use tive action. in deliberative settings carry these inequi- Deliberation’s potential to create legiti- ties to new places–even when a delibera- macy lies in its ability to limit the kinds of tor’s intention is to reduce their impact. oppression and power asymmetries pres- In our examples, language and communi- ent in other means of social decision-mak- cation pertain not only to what is formal- ing, where these other ways of “legitimat- ly written or intentionally said, but also to ing” social decisions include violence, the what is read by others when they see our edicts of oligarchs, decisions produced bodies or imagine our backgrounds. We by the power structures underlying many will argue that it is difficult or impossible modern democracies, and distributional for participants in a deliberative setting to outcomes influenced by the world’s myri- unsee what they are clearly seeing or un- ad systems of markets. For this reason, we think the meanings others communicate focus particular attention on the extent to when they present themselves. These non- which deliberative mechanisms mitigate verbal communications infuse conversa- power asymmetries. We contend that even tions and affect deliberation’s ability to in situations of apparent procedural equal- produce legitimate outcomes. ity with respect to every individual’s basic We conclude by describing how delib- right to convey their lifeworlds, the legit- erative contexts can be modified to reduce

146 (3) Summer 2017 65 Inequality certain types of power asymmetries, such gives us access to the past. Language can be is Always as those often associated with gender, race, used to categorize the present and to pro- in the Room: Language & or class. In so doing, we hope to help read- pose desired futures. Language enables peo- Power in ers consider a broader range of factors that ple to overcome the isolation integral to hu- Deliberative Democracy influence the outcomes of attempts to de- man experience. People are able to speak of velop norms or restructure power relation- their pain, their pleasure, their needs, their ships through communicative acts. hopes, and their experiences. They are free to make public their sense of things: their When communication and language are in interpretation of events, institutions, laws, the room, so are inequality and coercion. and customs. Communication. The foundation of hu- Language also constrains. We enter a man interaction. The principal means by world already named, in which meanings which we express basic emotions. Love. are attached to all that we encounter, in- Anger. Fear. The vehicle through which cluding our bodies. Language makes us we convey tales of heroes and villains. The meaningful to ourselves. We know our sex, medium through which individuals testify our race, our ancestry, our faith, and our about their vulnerabilities and adversities. politics through language. We are often giv- The means by which oppressed persons en a race (or two), an ethnicity, a class posi- seek assurance and plead for assistance. tion, a nation. Each of these comes with a When seeking to manage problems that history. Each of these comes with a mean- we as individuals cannot solve on our own, ing that predates our awareness of them. we seek communicative currencies that al- We are governed by language even when low us to discover shared histories, devel- we are silent. When we use language we op common interests, and build trust. Com- are bound by words whose meanings are munication offers a foundation from which already set, by grammatical rules and by we construct social compacts and con- other linguistic conventions.1 Philosopher tracts. These pacts set the stage for all forms of language Paul Grice observed that we of collective action and influence the terms have incentives to use terms that are eas- by which such actions are remembered. ily understood by others.2 To achieve un- Language. The languages and lexicons derstanding in the space of a single con- that we use to communicate with one an- versation, we use familiar words. We seek other are intricate human creations. They analogies, metaphors, and examples that help us organize the world for ourselves are likely to be familiar to others. Many and describe it to others. Language proof these words are well established with vides a means for categorizing worlds ob- long histories. In our quests for fluidity served and imagined. Language is, howev- and speed, we seldom take time to reflect er, our maker as well as our servant. We en- on the origins of these rules, examples, or ter a world language has made for us. Our words. Those who do, Heidegger and Ni- most intimate experiences are mediated by etzsche among them, enrich our thinking, language. but they too cannot fully comprehend the At all times, language frees and con- infinite and changing richness of a word. strains. The meaning of a word may shift as it Language frees us by allowing us to com- travels from one geographic, political, or bine its words and phrases in an infinite class site to another. Words change with number of ways. Language gives us the cause. However erudite, however careful we pacity to express diverse ideas and emo- are, we cannot fully control the meanings tions. The continuity of language over time and connotations of the words that we use.

66 Dædalus, the Journal of the American Academy of Arts & Sciences They have associations that may be unin- this topic, for example, we can consider Arthur tended by us, even unknown to us. “googling” the topic and then “emailing” Lupia & Anne Our words reflect and extend power in or “texting” what we find to others. We Norton ways that we only partially understand. can ask what has changed in jihad as the Some words and rules provide easy ways to word has traveled west. inflict cruelty. Some words and rules have Yet in this march of linguistic “progress,” a cruelty that is felt by others but hidden new meaning sometimes emerges at the ex- from us. Often that capacity for cruelty is pense of old meaning. What is lost in these available only to some. The sting of racial evolutions may be seen by imaginative in- epithets, for example, is often more severe terlocutors as rightfully subservient to new when voiced by those who occupy a high- expressions. But the old meanings we sac- er position in a racial hierarchy. rifice for contemporaneous convenience We use language to hide, as well as to re- may be lost to future selves in need of those veal, our sentiments. The purposes of some meanings. The march of meaning as mani- of these devices range from saving face to fest in language is a dialectic between pres- deflecting attention away from unattractive ent and past, between the creators of lan- elements of ourselves and toward the un- guage and those who are made in it. attractive attributes of those who threaten Most of us don’t think about communi- us. We bite our tongues at critical moments cation and language in these ways very of- not because we are uninjured, but because ten, if at all. As long as our words elicit the we hope that such patience will produce its social and cognitive reactions that we seek, own rewards. we carry on without thinking about what We are made in language–as Wittgen- meaning and power our words have con- stein and Lacan, Gadamer and Lévi-Strauss veyed. We don’t think about these things (and a host of others) recognized–but we despite the fact that every time we use lan- are also the makers of language. We coin guage, we transport myriad residues of hu- new words and phrases. For each new mix- man relationships to new places and peo- ture we create, others attend to them or ple. We reinforce social and political struc- they do not. They derive meaning or walk tures, often without willing to do so. In our away confused. They use our words to ex- use of language, we export broader, frac- press their own feelings–or they do not. If tured elements of history from the past to they do not, whatever we were trying to ex- the present. press at that moment withers away. In the In every quest to achieve fast coordina- hands of skilled or privileged communication and mutual understanding through tors, language is an instrument of incred- communication and language, we neces- ible power, yet even the most eloquent of- sarily provide new energy to a continuance ten find language inadequate.3 of the past imbalances and power asym- Because we make language, important metries that are part of every widely used attributes of language change. Grammar language. In so doing, we use words that changes. Idioms change. Some changes are assert authority. We use terms that can in- willful and deliberate: we choose not to say spire or injure. We do this even when at- a host of once common racial and ethnic tempting to find common meaning. Even epithets. Other changes are unconscious. when attempting to mitigate power rela- They are the work of practice, represent- tionships. Even when we incompletely un- ing erosions in the structure of language derstand the meaning that others derive made by the currents of speech and writ- from our words–which is to say, almost ing in the everyday. To learn more about always.

146 (3) Summer 2017 67 Inequality Language is not only a matter of words, ligion will be inflected by the hijab or the is Always spoken and heard. Language is also writ- habit the speaker wears. in the Room: Language & ten. The written is carried not only in We are often not fully conscious of the Power in words but also in other signs. The silent texts we write on our bodies as we dress, Deliberative Democracy body speaks, whether it wills that speech but we are unconscious adepts at reading or not. It speaks of its place in the social or- them. We see the people that surround us der: of race, sex, age. The black man must not as naked human beings, not even sim- speak as a black man, the white woman as ply as people inscribed with only race and a white woman. The old speak from the sex and age. We see them as members of shell of age. Some speak from the haze of social orders, clothed with information beauty. The text written on the body, read about their positions and their preferenc- from the body, may amplify or mute what es. Policemen and firemen, military offi- the speaker says, but it cannot be easily si- cers and security guards wear uniforms. lenced. We know where they work. We know the Nonverbal communication communi- Army lieutenant has taken an oath, that cates. Nonverbal expression expresses. the fireman is willing to risk his life for oth- Utterances and meanings enter the room ers. We know the captain outranks the ser- with us. They are part of the conversation, geant, though the sergeant may be older. whether they are formally recognized, whis- We know the general makes a lot more pered in the shadows, or have emerged in money than the private, that he may have others’ consciousness automatically once advised presidents, that he has power. We we are seen. Few speeches have the power know the workers at Target and McDon- of the silent body. Often the texts of race ald’s make less money than the general, and sex and age operate as supplements; that they almost certainly have less edu- they are, in Derrida’s phrase, “that which cation. We know they have less power. We adds only to replace.” The man of the Dec- conclude, on the basis of good evidence, laration of the Rights of Man is displaced that the woman in the starched cap of the by blackness, evoking not the triumph of Amish and the Mennonites is unlikely to freedom but the legacy of slavery. support abortion or military intervention. We have spoken before we speak, we The texts written in clothing have all the have been read before we write. The peo- power and imperfections, all the strategies ple who enter a room carry not only the in- and misfires, that one encounters in other scribed body, but the many texts they have uses of language. A person may try to dress written on that body: when they shaved unobtrusively and nevertheless boast of or didn’t shave, when they put on make- wealth, forgetting, for example, the famil- up, when they dressed. The people who de- iar Rolex on the wrist. But clothing–like liberate do so clothed in texts that speak every other form of speech–does not al- of their place: of their wealth or pover- ways tell the truth. One can use clothing ty, their religion, their level of education, to pass: as richer or poorer, man or wom- their regions, their preferences and poli- an, even as black or white. tics. The uniform and the political T-shirt The texts written on the body, and those carry messages, but so do headphones and that people write on themselves, enter with Birkenstocks. The clothes a speaker wears those who deliberate. One can, of course, inflect the speech. Speech about policies forbid uniforms, but clothing will remain, toward Israel carries different meanings and the physical signs of race, sex, poverty, when it comes from a body wearing a kipot age, and certain types of abuse are difficult or a kaffiyeh. Speech about freedom of re- to erase.

68 Dædalus, the Journal of the American Academy of Arts & Sciences The texts written on the body are not, lifeworlds. Everyone, in principle, having Arthur however, simply problems to be overcome, a right to speak. No one having to modi- Lupia & Anne impediments to be set aside. They can op- fy their truths for the sake of going along Norton erate–rarely, but with great force–as occa- to get along. Each person obliged to con- sions for questioning and enlightenment. front, face-to-face, the people with whom Thinking of the Declaration of the Rights they share a caucus, a district, a country. of Man and of the Citizen when the man Each person faced with a body that, like or is black and the citizen is Native Ameri- unlike their own, shares a common human- can raises questions that can spur deliber- ity and with it the whole human comple- ation. Sojourner Truth changed the debate ment. These deliberative ideas are seen as when she asked in her blackness “Ar’n’t I a way to mediate and reduce socially dam- a woman?” The visible presence of pover- aging power inequities. ty can alert the well-to-do about relations Common ideal forms of deliberation build between luck and skill in the current social not just from the premise that everyone has order. The vision can induce them to think a right to speak; they build from the premise about how they would want to distribute that speakers will actually be heard.5 In oth- power and privilege if skill-luck relations er words, participants enter not only with a turned out to be even a little bit different license to speak but also with an obligation than they imagined. to listen. The obligation is not necessarily Whether carried by language or appear- to agree, but to actively engage what all the ance, inequality is in the room even before stakeholders feel they need to say. the deliberators enter. One can always, must Hence, many deliberative ideals depend always, ask what the room says. Who is sion the implicit assumption that in the ideal, lenced or intimidated by the room?4 Who participants would have an unconstrained feels at home in the room? Is the room in capacity for attention and listening. This is a public building or a church? The ex-con- a problem in practice. Time is scarce. Atten- vict and the undocumented immigrant en- tion is limited. Sometimes people say things ter public buildings on different terms than that we have heard a hundred times before. the policeman and the public official. The We tune them out. Or they repeat them- union hall may be enemy territory for the selves. We tune them out. Or they offend businessman; the church an unsettling us. We tune them out. There is also preju- space for a Jew. Parents walking into a pub- dice. Not just what we think about when we lic school will be faced with rooms like those see certain types of people, but also what we their own children occupy, or with rooms think about when we hear certain types of that boast of riches denied to their children, words spoken in certain types of ways. In or a poverty that their children do not face. many cases, we tune those words out. The room will speak of privilege or depriva- A considerable scholarly literature on at- tion, of class and regional identities. There tention tells us what our experience has al- is no neutral, unspeaking space. Perhaps the ready shown: we ignore almost every piece harshest speech of all would be the clinical of information to which we are exposed, we sterility of a room with no chairs and white pay fleeting attention to almost every piece walls. of information to which we pay any attention, and most of the phenomena to which Which brings us to deliberation. The idea we do pay attention leave no lasting impact has such promise. The idea of forward-look- on our subsequent feelings or memories.6 ing individuals. The idea of sharing. Unfil- In other cases, we have the ability to pay tered descriptions of diverse individual attention, but we are so self-focused that

146 (3) Summer 2017 69 Inequality we “listen” only for an instrumental rea- sensuses, compromises, or agreements is Always son: to set up our own awesome response. reached after a deliberative session? Can in the Room: Language & Here, we are constantly thinking of how we say that everyone reached an identical Power in to shore up our positions. How to expand understanding about the entirety of the tes- Deliberative Democracy or protect our self-esteem. How to elevate timony that their setting allowed? No, we our social position. How to arrange the mo- cannot say that or, indeed, anything close ment in a way that will help others appreci- to it unless the deliberation was remark- ate our virtuousness in the stories that we ably short and its content was the type to will later tell others about this exchange. which people could devote undivided at- So we listen as tacticians, plotting the next tention. In all other cases, physical limits move in a game we think we already know of attention and memory prevent people how to win. from recalling all elements of a sustained Even when we are not consciously using communicative interaction. Even if peo- a conversation for the purpose of self-infla- ple remember many such elements, there tion, we can be led astray by our attempts would be questions about how heavily they to place another’s words in a context that should weight them in any post-delibera- we feel comfortable contemplating. When tive conclusions that they draw. Should the black woman speaks, the white wom- people weight all aspects of all utteranc- an may think “she is a woman like me, es equally? Should they realize that some she will be an ally” or “the black woman people take longer than others to “get to is speaking, will she reproach me?” The the point” and perhaps discount utterances black woman may speak unwillingly of of excessively wordy individuals? Should blackness, willingly of her wealth and priv- they account for the fact that some people ilege, or vice versa. The white woman may may be speaking strategically in order to be so distracted by her concern with her achieve a certain outcome where others’ own standing in the black woman’s eyes utterances are more heartfelt? that she fails to listen, fails to hear, what There are limits to what deliberative out- the black woman actually says. She may comes can tell us about what thoughts and listen, but only to hear the voice of the race feelings its participants share.7 If a deliber- while ignoring a person caught in a partic- ative proceeding goes on for too long, peo- ular mesh of structures and constraints of ple may lose hope about their ability to be which race is only a part. She may hear the heard. Others may be more likely to be- voice of the race, but if she fails to hear the come tired and less likely than others to voice of the speaker, she will have heard a recall a particular moment in a conversa- message quite different from the one the tion. Some may be hungry, have children speaker intended to convey. When we lis- waiting at home for dinner or bedtime, or ten to others, we may listen for the voice even have to go to the bathroom and “as- of the race and fail to hear the voice of pov- sent” to a particular proposition to facili- erty or the wisdom of age. We may listen tate a speedy exit. Others may by physically for the guidance of the educated and fail to or intellectually attracted to a person in the hear a more refined bigotry. We may work room and assent to a particular proposition to hear difference and fail to hear an invita- to increase the likelihood of subsequent in- tion to make common cause. We may fail teractions. None of these forces can be kept to hear the voice of one person, like and from a deliberative context. unlike all others. In some theories of deliberation, con- With these communicative dynamics straints to interpreting a post-deliberative in mind, what can we say about the con- consensus, compromise, or agreement

70 Dædalus, the Journal of the American Academy of Arts & Sciences would themselves be limited by the idea but it also excluded abolitionists. Refusal Arthur that deliberation is not a discrete event. In to accept a compromise of that kind can be Lupia & Anne some theories, perpetual openness to new politically and morally defensible. Should Norton information is a key part of the device’s nor- we commend efforts to reach a compro- mative appeal. So agreements made for the mise over segregation or apartheid? Those purpose of a fleeting convenience can be who value peace, order, and the rule of law undone. A practical problem, however, is very highly may say yes. That hierarchy of that undoing agreements that were alleged political values is not without its defects to represent the broad assent of deliberative and dangers. The civil rights movement in participants takes time to do. If such “do- the United States depended on a willing- overs” were to happen often, they could re- ness to disturb the peace. King called for duce confidence in the future force of a cur- civil disobedience and defiance of the laws rent agreement. Why invest one’s heart and that maintained an unjust racial order. De- soul into a deep conversation about how we colonization required more aggressive, should live if we are repeatedly asked to “re- even violent confrontations with law and consider” any consensus, compromise, or order. Even a tacit assumption that com- agreement that we might reach? promise is what deliberation seeks can un- To avoid attaching to deliberative out- dermine the larger democratic end of seek- comes interpretations that limits of mem- ing common understanding and the com- ory and forces of identity cannot sustain, mon good. people should enter fully conscious of their In all interpretations, moreover, we are fallibility, unsure that they understand–or also apt to overestimate our capacities for even could understand–the experiences empathy. Consider, for example, Hannah of their fellow participants. They should Arendt’s “Essay on Little Rock.” In her distrust their knowledge, their capacity for rejection of forced desegregation, Arendt empathy, and even their values. speaks for and as “the Negro mother.” Ar- They should realize that if consensus is endt’s confidence in her capacity for un- the object, certain outcomes are foreclosed derstanding and sympathy misleads her. at the outset. If agreement is the end, cer- The passage is cited now not as an instance tain positions are delegitimated at the out- of empathy or solidarity, but as evidence set. Consider a meeting that asks Scot- of the limits of her thinking. tish nationalists to join an effort to reach a Efforts to reach common ground or a consensus on how to maintain the United common understanding are seductive, Kingdom. The Scottish nationalist would particularly for Americans. We often be- be better served by a call to deliberate over lieve these are easier to reach than our his- “whether the United Kingdom should con- tory indicates. We retain a commitment in tinue in its present form” than a call to “find principle to the idea that “all men are cre- common ground for the United Kingdom.” ated equal,” that they are endowed with a They should realize that if compromise common set of rights, needs, and desires. is a desired outcome of deliberation, those Yet even if we all have the right (and the who reject compromise are excluded. Yet need) for life and liberty, even if we all have rejection and refusal may be the most use- the right (and the desire) to pursue happi- ful and honorable forms of action in some ness, we differ profoundly on what these instances. Consider the Missouri Compro- objects are and how we should be permit- mise. That compromise was predicated ted to pursue them. The belief that we un- on the imperative to maintain the Union. derstand the rights, the needs, and the in- That construction excluded secessionists, terests of those we join in discussion is un-

146 (3) Summer 2017 71 Inequality reliable at best. It may lead–it has led–to Democratic passion and will are problems is Always efforts to impose compromises that have to be solved. The liberal answers to the in the Room: Language & held us back: the three-fifths clause being problem of democracy have been rules and Power in an early and shaming example. Common representation. Deliberative Democracy ground can be rocky and shifting. Histo- Many observers fear that the great mass- ry and memory may come to reproach us es of people are incapable of deliberation. for decisions we reach together. We can- Most people, they conclude, are prone to ir- not completely avoid error and, therefore, rational fears, hatreds, appetites, and hopes. we must regard any common ground we Rules are necessary to rein them in. Rep- reach not as where we end, but as a rest- resentation moves the most important de- ing place along the way. cisions, the most technical decisions, and perhaps any decision requiring reason away Deliberation is a liberal enterptise. It ex- from the masses toward a smaller group. presses the liberal commitment to order The few, it is argued, can reason as the many and procedure. Deliberative meetings are cannot. governed by rules, procedures, and norms In liberal democratic systems, the legit- of practice. These mechanisms aim at en- imacy of the decisions of the representa- suring equality and giving everyone a hear- tives is grounded in democratic right. The ing. Those who follow the rules and ob- answer to the question “who gave them serve the conventions appear to be show- the right to decide for the people?” is “the ing a greater willingness to advance the people.” That claim is far less tenable for deliberative process, to engage with others any deliberative group making decisions and to find common ground, but it is also for others. It is still less tenable for any de- possible that they are simply better served liberative group not chosen by those they by the rules in place. Those who are most are supposed to represent. Legitimacy is willing to search for common ground may further compromised with any delibera- be those who hold a strategic advantage on tive group impeded by unseen power asym- that ground. metries in communication. The advocates Liberalism is, however, not always con- of deliberative endeavors are not always at- ducive to liberal values, and it can be very tentive to these matters. How those who much at odds with democracy. If those deliberate are chosen and how they view who deliberate and subsequently decide one another determine whether the assem- make their decisions only for themselves, bly will be liberal, liberal-democratic, or the enterprise may capture, in its form, neither, in relation to the people for whom valuable elements of liberal democracy. they speak. That is, deliberation linked with decision This matters because deliberation values is an instance of people governing them- rationality in both its forms: as reason and selves within a set of procedures (ideally, as order. For many deliberation advocates, ones they make themselves) and a com- the commitment to reason is explicit, pro- mitment to using reason to advance de- found, and made with conviction. In this mocracy. If those who deliberate decide advocacy, those who deliberate are called for others, the enterprise is troubled as all not only (and perhaps not primarily) to representation is troubled. share their lifeworlds with one another. Deliberation also reflects the liberal un- Participants are called to reason together. ease with democracy. Liberalism, like so The language of reason is always appropri- much of political thought before it, regards ate and welcome in such meetings. The lan- democratic power as a force to be managed. guage of passion is not.

72 Dædalus, the Journal of the American Academy of Arts & Sciences We believe, however, that politics re- skilled. Any resulting consensus, compro- Arthur quires more than reason alone; politics mise, or agreement will not simply emanate Lupia & Anne requires passion. It is passion that enables from equal consideration of all relevant life- Norton people to endure the “slow drilling through worlds, it will also reflect different abilities hard boards” that is the work of politics. It to use language in quests for influence. is passion that enables people to endure the Moreover, the acts in question, both the frustration of listening to views they find speaking acts and the listening acts, will be tedious or abhorrent. It is passion that en- made by people. These people will be seen ables people to convey not just the facts, before they speak and they will be interpret- but the subjective experience of a lifeworld. ed before they attempt to convey any mean- It is passion that enables people to chal- ing. We will know who enters marked with lenge settled beliefs and political conven- signs of privilege. We will know who lacks tions that they believe are unjust. If deliber- those signs. We will know who enters a fa- ation is to produce shared understandings miliar place and who enters a foreign one. with legitimating potential, if it is to pro- Appearance and words will interact. Some duce shared assent that reflects the life ex- appearances will help deliberative partic- periences of the diverse people whom such ipants recognize the diversity, glory, and endeavors are meant to represent, delibera- pain of different lives. Other appearanc- tion requires passion as well as reason. Jane es will lead deliberative participants to ig- Mansbridge’s distinction between first- nore what is being said or to substitute their and second-generation deliberative theory own privileged narrative for the one that the marks this recognition among deliberative speaker is attempting to convey. theorists themselves. Second-generation We will know things about the process. deliberativists have recognized that emo- We will know who is likely to be advan- tion and passionate intensity contain truth taged by its procedures. We will know that as well.8 assent may be the product of people holding back. People may not reveal their true mo- With these and related challenges in tivations. People may give in to power out mind, what can we read from a deliberative of desperation, fatigue, or fear. People may outcome that can legitimate a collective de- choose to remain silent in the face of history- cision? To answer this question, suppose bound and institutionally reinforced asym- that a major goal of deliberation is to convey metries. legitimacy to some socially relevant propo- For these reasons and more, we will know sitions and withhold such legitimacy from that a deliberation-generated consensus, others. Suppose, moreover, that the form of compromise, or agreement that represents deliberation is an ideal version that entails a deeply shared understanding to a clearly a universal right to participation. stated set of principles will often be obser- Let’s start with what we know. The com- vationally equivalent to a deliberative out- municative acts that precede the outcome come that is the result of all of the asymmet- will use language that conveys power. They ric and oppressive factors described above. will be used by people who are more and So a deliberatively generated outcome can less skilled in using language to acquire be normatively desirable, it can represent power. If participants are not paying close real intellectual exchange, and it can be le- attention to these skill imbalances, and if gitimating–but it is none of these things the deliberative rules are not built to miti- automatically. gate deleterious effects of such imbalances, As a result, now is an opportune mo- participants are likely to be swayed by the ment to reevaluate claims about deliber-

146 (3) Summer 2017 73 Inequality ation that gauge its effectiveness by refer- fy violent responses? Are there some state- is Always encing instances of opinion movement, ments to which a society’s best response is in the Room: Language & opinion convergence, or language-based to, at minimum, stop the conversation? Power in consensus. Such outcomes represent the What if deliberation reveals insurmount- Deliberative Democracy normative desires that have led many to be able oppositions? This discovery might interested in deliberation. The forces de- not require violence, but it might well call scribed above allow language to carry pow- for secession or partition. For any numer asymmetries to new destinations. They ber of reasons, deliberative situations can allow seemingly open and equal commu- be as coercive as violence, with the add- nicative domains to be dens of oppression. ed insult that the coerced are presumed to Language-based consensus, compromise, consent, or to have been overcome by rea- or agreement, in the presence of such forc- son. In extreme cases, this outcome, while es, becomes a limited means of conferring not entailing physical violence, would be legitimacy to collective decisions in mod- attempting to generate legitimacy on the ern societies. basis of dishonest claims about what language-based consensus, compromise, or Deliberation takes place in a communi- agreement actually means. cative forum. In such forums, participants Having now raised questions about wheth- engage in speech acts with the possibility er and how language facilitates communication of converging on shared meaning. Delib- and whether and how communications inform eration is endorsed on the basis of theories assent, we turn to two final questions that and beliefs about how these shared mean- scholars and practitioners can use to rec- ings provide individuals and societies with oncile their motives for seeking delibera- a stronger and broader moral, ethical, and tive activities with likely outcomes of those technical foundation for improving quali- attempts: ty of life. But communication and language 1) What outcomes can we actually ex- carry inequality, and the limits of human pect from deliberation? attention, patience, and self-love create or 2) Are there any conditions that would reinforce coercive conversational norms. make these outcomes more tolerable from The promise and the principal challenge the perspective of persons or populations of deliberation is that language is a weap- who are otherwise run asunder by the on that can be wielded with great force. wheels of political and social institutions? There is no way to construct a deliberative To address these questions, we begin environment in which asymmetry, power, with the recognition that deliberation is and potentially coercive flashpoints do not another way of allocating power. It privi- contribute to the outcome. If deliberation leges some interests at the expense of oth- is to be justified on the basis of its abili- ers. It is not generally neutral with respect ty to mitigate power imbalances, the do- to who wins and who loses. main of deliberative interactions must be When the social project motivating de- constrained. liberative democracy is to reduce a partic- Many people who advocate for deliber- ular set of social imbalances, the question ation take for granted that deliberation is becomes when and whether it is possible preferable to violence. But what if deliber- for deliberative participants to recognize ation simply reinforces the experience of these imbalances and design subsequent oppression? Given the examples and fac- interactions to diminish them. The power tors raised in this essay, such outcomes are imbalances that deliberation proponents imaginable. Do some uses of speech justi- believe they are stopping at a deliberative

74 Dædalus, the Journal of the American Academy of Arts & Sciences chamber’s front door will storm in through than seeking to overcome differences, it Arthur the back and take over the proceedings. might be better to enshrine them institu- Lupia & Anne To take such concerns seriously, a suffi- tionally (for example, through federalism Norton cient number of deliberative participants or concurrent majority) or to develop a mo- must share a set of values that induces dus vivendi that preserves the differences. them to be aware of the imbalances, to In this stance, we echo the second genera- try to mitigate them procedurally, and to tion of deliberative theorists who see clar- seek measures of progress that the affect- ifying conflict as an important goal of deed participants would recognize as valid. liberation and extend their view by asking If there is not a sufficient values consensus for further introspection about how agree- on the need to protect a particular popu- ment is or is not a product of the coercive lation or point of view, there will be little power of language described above. or no motive to pursue procedural change Politics entails deep value conflicts, mon- or to measure the effects of these proce- umental struggles for power, and real ques- dures on the affected. In such cases, claims tions about quality of life. To manage these of having achieved legitimacy or advanced dynamics and facilitate efficient social in- democracy would not reflect actual cir- teraction, communities seek to discover cumstances. If deliberation is to be legiti- shared values and build agreements from mating from the broadest set of perspec- these discoveries. If it is important that po- tives, then the expectation must be that the litical communities are built from honest weak can receive justification from their assessments of what their members actu- own perspectives and on their own terms. ally share, then it is important to be cog- One of the redemptive possibilities of nizant of how deliberative outcomes are language is that it enables people to trans- manufactured. In such inquiries, we can form status; to take a lower status position come closer to understanding whether de- and use it as a claim to power. Such trans- liberative outcomes are meaningful or illu- formations can produce situations when sory, sustainable or ephemeral, and, hence, formerly (or presently) less powerful peo- whether they are capable of securing legit- ple control the conversation (or seem to). imate decisions and advancing a common Thus the many complaints about political good. correctness. One may respond: “So what? It is the turn of the less powerful to exercise a control that once silenced them.” While this type of response may dismay some deliberation advocates, it should not be light- ly dismissed. Deliberate changes in who controls communication can reveal new foundations of justice that would otherwise go unspoken. Another proposal that could make deliberation’s outcomes more tolerable from the perspective of persons or populations whom political and social institutions otherwise diminish is that consensus, compromise, or agreement should not always be the aim. If differences arise, perhaps they should remain: open and acknowledged. Rather

146 (3) Summer 2017 75 Inequality endnotes is Always * in the Room: Contributor Biographies: ARTHUR LUPIA, a Fellow of the American Academy since 2007, is the Language & Hal R. Varian Professor of Political Science and Research Professor in the Center for Political Power in Studies at the University of Michigan. He is the author of Uninformed: Why People Know So Little Deliberative about Politics and What We Can Do about It (2016) and The Democratic Dilemma: Can Citizens Learn What Democracy They Need to Know? (with Mathew D. McCubbins, 1998) and editor of The Cambridge Handbook of Experimental Political Science (with James N. Druckman, Donald P. Green, and James H. Kuklinski, 2011). ANNE NORTON is Professor of Political Science and Comparative Literature at the Univer- sity of Pennsylvania. She is the author of On the Muslim Question (2013), Leo Strauss and the Politics of American Empire (2004), 95 Theses on Politics, Culture, and Method (2004), Republic of Signs: Liberal Theory and American Popular Culture (1993), Reflections on Political Identity (1988), and Alternative Amer- icas: A Reading of Antebellum Political Culture (1986). 1 For a summary and discussion with extensive references, including those in this and the following paragraphs, see Anne Norton, 95 Theses on Politics, Culture, and Method (New Haven, Conn.: Yale University Press, 2004), 12–27. 2 Paul Grice, “Logic and Conversation,” in Syntax and Semantics, Volume 3: Speech Acts, ed. Peter Cole and Jerry L. Morgan (New York: Academic Press, 1975), 41–58. See also Arthur Lupia, Uninformed: Why People Know So Little about Politics and What We Can Do about It (New York: Ox- ford University Press, 2016). 3 The literature on the politics of language is vast and rich. For a summary and discussion with extensive references, see Norton, 95 Theses on Politics, Culture, and Method, 12–27. 4 Pierre Bourdieu, Distinction: A Social Critique of the Judgement of Taste, trans. Richard Nice (Lon- don: Routledge, Kegan & Paul, 1984). 5 Arthur Lupia, Yanna Krupnikov, and Adam Seth Levine, “Beyond Facts and Norms: How Psychological Transparency Threatens and Restores Deliberation’s Legitimating Potential,” Southern California Law Review 86 (3) (March 2013): 459–493. The authors link normative theories of deliberative democracy to empirical literatures on attention limits and communication dynamics. 6 For example, see Alan Baddeley, “Working Memory: Theories, Models, and Controversies,” An- nual Review of Psychology 63 (1) (2012): 1–29; and Earl K. Miller and Timothy J. Buschman, “Work- ing Memory Capacity: Limits on the Bandwidth of Cognition,” Dædalus 144 (1) (Winter 2015): 112–122. 7 Iris Marion Young, “Communication and the Other: Beyond Deliberative Democracy,” in Democracy and Difference: Contesting the Boundaries of the Political, ed. Seyla Benhabib (Princeton, N.J.: Princeton University Press, 1996), 120–135. 8 Jane Mansbridge, “A Minimalist Definition of Deliberation,” in Deliberation and Development: Rethinking The Role of Voice and Collective Action In Unequal Societies, ed. Patrick Heller and Vijayendra Rao (Washington, D.C.: World Bank, 2014).

76 Dædalus, the Journal of the American Academy of Arts & Sciences Collusion in Restraint of Democracy: Against Political Deliberation

Ian Shapiro

Abstract: Recent calls to inject substantial doses of deliberation into democratic politics rest on a misdiag- nosis of its infirmities. Far from improving political outcomes, deliberation undermines competition over proposed political programs–the lifeblood of healthy democratic politics. Moreover, institutions that are intended to encourage deliberation are all too easily hijacked by people with intense preferences and abun- dant resources, who can deploy their leverage in deliberative settings to bargain for the outcomes they prefer. Arguments in support of deliberation are, at best, diversions from more serious threats to democracy, notably money’s toxic role in politics. A better focus would be on restoring meaningful competition between representatives of two strong political parties over the policies that, if elected, they will implement. I sketch the main outlines of this kind of political competition, differentiating it from less healthy forms of multiparty and intraparty competition that undermine the accountability of governments.

Advocates of political deliberation usually defend it as a collaborative activity motivated by the possibility of agreement. Even when agreement proves elusive, deliberation helps people come to grips with one another’s views, draw on their different experiences and expertise, and better understand the contours of their enduring disagreements. People’s views will be better informed, and the decisions they make will be of higher quality than if they had not deliberated. When study after study reveals most people to be appallingly ill-informed about much public policy, deliberation’s appeal seems obvious. Two IAN SHAPIRO, a Fellow of the Amer- minds are better than one, three better than two, and ican Academy since 2000, is the so on. Democracy will be improved if its decision- Sterling Professor of Political Sci- making can incorporate, and build on, the benefits ence and Henry R. Luce Director of of deliberation. Or so it is frequently claimed.1 the MacMillan Center at Yale Uni- versity. His recent books include Deliberation should not be confused with argu- Politics against Domination (2016) and ment. When people argue, there is an expectation The Real World of Democratic Theory that one of them will, or at least should, win. Even (2011). when we speak of one person making an argument,

77 Against we see this as something that stands until quire power by prevailing in a “competi- Political it is contradicted, or challenged and beat- tive struggle for the people’s vote,” gives Deliberation en by a better argument. Like the delibera- institutional expression to the argumen- tionists, proponents of argument believe it tative ideal.4 This was perhaps best exem- will enhance understanding and improve plified in the Westminster system as it ex- the quality of decisions. This was the es- isted from 1911, when the Parliament Act sence of John Stuart Mill’s defense of the stripped the House of Lords of its real pow- robust clash of opinions in On Liberty: it ers, until the late 1990s, when the Lords would lead people to hold better-informed was reformed to enhance its legitimacy as and more accurate views. Mill even went a second chamber and the Commons be- so far as to worry–needlessly, it turned gan ceding authority to European and oth- out–that as advancing science expand- er courts, the Bank of England, and indeed the realm of settled knowledge, people pendent agencies. The twentieth century’s would be deprived of argument’s benefits. middle eight decades were the heyday of No longer forced to sharpen their wits by Parliament’s supremacy within the British defending their views in the marketplace political system and of the Commons’ su- of ideas, they would become mediocre premacy within Parliament. Epitomized dullards; less able to think for themselves at Prime Minister’s Questions, the some- and more easily manipulated by others.2 times overwrought weekly gladiatorial My claim here is that the argumenta- clashes over the famous wooden despatch tive and deliberative ideals should be more boxes, it thrives on the ongoing contest be- clearly distinguished than they usually are. tween opposing policies and ideologies. They support different and incompatible Schumpeterian democracy depends on institutional arrangements. I also maintain alternation between two strong parties in that the argumentative ideal is superior be- government. The party that wins the elec- cause, when appropriately institutionalized, tion exercises a temporary power monop- it helps hold governments accountable for oly, but the loyal opposition–a govern- their actions. By contrast, the deliberative ment-in-waiting whose leaders hope to ideal cannot easily be institutionalized–and take power at the next election–continu- perhaps cannot be institutionalized at all– ally challenges its policies. This system de- because people who prefer to bargain can pends on combining first-past-the-post sin- easily abuse rules designed to promote de- gle member plurality (smp) electoral sys- liberation. But deliberation’s difficulties tems with parliamentary democracy. The run deeper. Its defenders fail to appreciate smp electoral system produces two large that, in politics, deliberation and the search parties, so long as the political makeup of for agreement are–to borrow an antitrust the constituencies more or less reflects the analogy–unhealthy forms of collusion in political makeup of the national popula- restraint of democracy. They should worry tion.5 Parliamentary systems ensure that less about voter ignorance, which, as Antho- the parties will be strong because the lead- ny Downs noted long ago, might well reflect er of the majority party is also the chief ex- sensible budgeting of scarce time, and wor- ecutive. Government and opposition clash ry more when office-seekers fail to engage in across the aisle continually, and compete robust public debates over the policies that, during elections by offering voters the dif- if elected, they will enact.3 ferent programs they plan to implement. The deliberative model, by contrast, Joseph Schumpeter’s competitive model calls for institutions that create incentives of democracy, in which governments ac- to seek agreement rather than victory–or

78 Dædalus, the Journal of the American Academy of Arts & Sciences at least agreement as a condition for vic- which gives temporary control of the gov- Ian tory. Rules that require concurrent major- ernment’s power monopoly to the majority Shapiro ities in bicameral chambers force represen- party and relies on alternation over time as tatives to find common ground when they its main mechanism of accountability, the can, and compromise when they cannot. U.S. model divides up the control of pow- Executive vetoes and supermajority pro- er on an ongoing basis. Madison’s slogan visions to override them create similar in- was that “ambition must be made to coun- centives. Proponents of deliberation often teract ambition.”8 The checks and balances find proportional representation pr( ) con- force the players in the different branch- genial for comparable reasons. Instead of es to accommodate themselves to one an- two catchall parties that must submerge other; hence its affinities with the delib- their disagreements in order to win elec- erative ideal. tions, pr leads to party proliferation, bringing a more diverse array of voices to the po- Up to a point. A major limitation of insti- litical table. In addition to the left-of-center tutions that encourage deliberation is that and right-of-center parties characteristic of they can produce bargaining instead. Ju- smp systems, in pr systems, liberals, reli- ries, for example, are traditionally subject gious groups, Greens, separatists, and na- to unanimity requirements that put pres- tionalists, among others, can all elect rep- sure on their members to talk out their dif- resentatives to the legislature to be part of ferences until they reach agreement. When the conversation. Because one party sel- this works well, it produces thorough ex- dom wins an absolute majority, coalition ploration of all the arguments and evidence government, which forces parties to seek provided by the contending parties: a post- and perhaps even manufacture common er child for the benefits of deliberation. But ground, is the norm. a jury can also be held hostage by a recal- The U.S. system is a hybrid. The smp citrant crank who has nothing better to do electoral system produces two large par- when everyone else wants to go home. His ties, but the independently elected pres- superior bargaining power and stubborn- ident weakens them, and the system of ness might enable him to extract agreement checks and balances forces consensus-seek- from the others, but this will not be delib- ing and compromise to the extent possible. erative consensus on the merits of the case. The American founders intended the Sen- What holds for juries also holds for other ate, in particular, to be a constraining body institutions that we might hope will induce made up of what Jefferson would later re- deliberation. When they produce bargain- fer to as an “aristocracy of virtue and tal- ing instead, those with the most leverage ent.” It has been heralded as such by com- will prevail. So it is that small parties often mentators dating back at least to Alexis de exert disproportionate influence over coa- Tocqueville.6 The idea that the Senate is the lition governments, U.S. Senators can use world’s greatest deliberative body, which holds and filibuster rules to thwart the will first gained currency with Daniel Web- of the majority, and various other super ster’s three-hour soliloquy in defense of majority and concurrent majority rules can the Union in 1850, has been repeated to the be deployed to similar effect. point of banality, no matter how scant its In short, deliberation requires people to connection with reality.7 I will have more to act in good faith, but it is not possible to de- say about the kind of competition the U.S. sign institutions to induce good faith. “If system fosters shortly. As a prelude to this, men were angels,” Madison wrote, “no notice that, unlike the Westminster model, government would be necessary.”9 In-

146 (3) Summer 2017 79 Against deed, when power is at stake and repre- er the changes in people’s views produced Political sentatives must answer to constituents, by Deliberative Polls and other consulta- Deliberation the impulse to bargain will likely overpow- tive mechanisms tried thus far are really im- er even genuine desires to reason collabo- provements on their pre-deliberative views ratively. In 2009, a number of centrist Re- or simply changes.13 These issues need not publican Senators showed an interest in detain us here, however, since my present working with the Obama White House for point is that–whatever its merits–institu- “cap-and-trade” legislation on toxic emis- tionalizing deliberation turns out to be an sions control. They soon bolted, however, elusive endeavor. If it is purely consultative, when confronted with Tea Party–orches- it is not clear why anyone will or should pay trated threats of primary challenges in their attention to it. Yet if rules are created to in- constituencies, should they choose to per- stitutionalize deliberation and give it real sist.10 Since power is endemically at stake decision-making teeth, they can all too easi- in politics, it seems unlikely that there will ly undermine political competition and em- be much genuine deliberation or that politi- power people with leverage to appropriate cians will resist the impulse to exploit rules them for their own purposes. that might maximize their leverage instead. An exception that proves the rule is the Schumpeter’s competitive model of de- British House of Lords. It functioned most mocracy trades on analogies between the effectively as a deliberative body after it lost political marketplace of ideas and the econ- most of its real powers in 1911. Peers who omy. Political parties are the analogues of participated were mainly public-spirited in- firms; voters mirror consumers. Schum- dividuals who specialized in particular ar- peter treats the policies that parties pro- eas and were often nonpartisan or cross- pose to enact if they become governments benchers. But the Lords has become more as the political analogues of the goods and partisan and assertive since the 1999 re- services that firms sell, and the votes that forms restored a measure of its legitimacy politicians seek as analogues of the reve- as a somewhat democratic institution, albeit nues that firms try to earn. Democratic ac- one at a considerable distance from the bal- countability is the political equivalent of lot box.11 What the Lords has gained in legit- consumer sovereignty: the party that does imacy has come at the price of diminished best at satisfying voters wins their support. effectiveness as a deliberative institution.12 Schumpeter’s illuminating analogy is The various deliberative institutions that nonetheless strained in several ways, two have been tried out or proposed in recent of which matter here. One is that political years are exclusively consultative. Deliber- parties are vying to control a monopoly, a ative Polls and citizens’ juries have no au- fact that constrains competitive possibili- thority to decide anything. They might af- ties. As I argue below, the best option is com- fect how people vote, but it is the voting petition between two large, centrally con- that will be decisive. Objects of theoreti- trolled parties. The Schumpeterian analogy cal conjecture like ideal speech situations also falters because there is no unproblem- are even more radically divorced from pol- atic equivalent of a firm’s shareholders for itics, since they depend on armchair spec- political parties. Some will single out par- ulation about what people would decide in ty members or activists as the appropriate settings that are devoid of power relation- political shareholders, but parties that em- ships. Questions can and have been raised power them run into trouble. Membership about whether such speculations add up in political parties is typically free or very to anything we should believe, or wheth- cheap, rendering them susceptible to hos-

80 Dædalus, the Journal of the American Academy of Arts & Sciences tile and anomalous takeovers, like that per- the sword fell, with each side blaming the Ian petrated by Donald Trump in the 2016 Re- other for intransigence. Perhaps it was a Shapiro publican primaries, or that which occurred cynical way for both parties to achieve cuts in the British Labour Party in the summer without being savaged by their electoral of 2016. Party leader Jeremy Corbyn lost bases. Whether due to blundering or col- a confidence vote in the Parliamentary La- lusive cynicism, the result was that every- bour Party by 172 to 40 in June, triggering one had an alibi and no one was undeni- a leadership challenge, but an easily aug- ably responsible for the outcome. Coali- mented membership nonetheless reelected tion governments live perpetually on such him as leader with 61.8 percent of the vote ambiguous terrain, undermining account- three months later.14 As this example under- ability for what governments actually do. scores, grass roots activists tend to be un- Competition enhances political account- representative of a party’s supporters in ability, but some kinds of competition are the electorate. This imbalance can be espe- better than others. As we have seen, com- cially pronounced in two-party systems, petition between representatives of two which, as I argue below, are nonetheless best parties, one of which will become the from the standpoint of robust public debate. government, enhances accountability be- Representation should be geared to max- cause they run on the platform they will be imizing the chances that public debate will judged on as governments. Moreover, the center on the policies that parties, if elect- need to sustain broad bases of voter sup- ed, will implement as governments.15 This port gives them strong incentives to advo- is why smp beats pr, and why strong, cen- cate policies that will be good for the coun- tralized parties are better than weak, decen- try as a whole, or at least for large swaths of tralized ones. Supporting a party in a mul- the population. Smaller parties represent tiparty system can help voters feel better more narrowly drawn interests: business, represented because their representatives’ organized labor, and ethnic and religious views are likely closer to their ideals than groups. This loads the dice in favor of clien- would be the case in a two-party system. telism, because politicians know that they But this is an illusion. What really matters will be held accountable for how effectively is the policies that governments will im- they advocate or bargain for their group’s plement. That cannot be known until af- interests in a governing coalition. It is better the coalition is formed, post-election. ter for parties to compete over what is best Coalition governments decrease account- for the country as a whole than to bargain ability, since different coalition members over the rents they can extract for their cli- can blame one another for unpopular pol- ents. This contrast can be overdrawn, to be icies.16 Americans got a taste of this when sure, because large catchall parties consist unusual conditions produced a cross-party of different interests among whom implicit coalition to enact the Budget Sequestra- bargains must be struck to keep them in the tion Act in August 2011, putting in place party. But that bargaining is constrained by $1.1 trillion of automatic spending cuts over the need to propound and defend platforms eight years split evenly between defense that can win support from other groups as and domestic programs, unless Congress well, otherwise they cannot hope to be- passed an alternative by January 2013. The come the government. Sword-of-Damocles proposal was widely said to be sufficiently draconian that the The sequester episode underscores the fact representatives would be forced to find a that the weakness of U.S. political parties is compromise. In the event, they did not and only partly due to republican institution-

146 (3) Summer 2017 81 Against al arrangements. Another source of party is to say nothing of the fact that in politics, Political weakness is decentralized competition, an preferences are always expressed subject to Deliberation artifact of the wrongheaded idea that local budget constraints. The intense antiregu- selection of candidates somehow makes the lation preferences of the multibillionaires process more democratic. In reality, because Charles and David Koch are massively am- of their comparatively high rates of partic- plified because their budget constraints dif- ipation, activists, whose beliefs and pref- fer vastly from those of the typical voter.20 In erences tend to be both more extreme and short, there are good reasons for the rules of more intensely held than the median voter democratic decision-making to reflect how in their constituencies, dominate primaries many people want something, rather than and caucuses. This enables them to force how intensely they want it. representatives to pursue agendas that the median voter in their district abjures, or to People have theorized about democracy serve the median voter only with the kind for millennia, yet it is only in the past few of subterfuge that might have been at work decades that the idea has gained currency behind the Budget Sequester Act. The same that democracy depends on, or at any rate is true of referenda, which sound demo- can be substantially enhanced by, deliber- cratic–“hooray for direct democracy!” ation. I have sought to show here that this –but which also enfranchise intense single- is a dubious proposition. It is hard, if not issue voters who turn out at disproportion- impossible, to create institutions that will ately high rates. Thus it was with the Brexit foster deliberation in politics, and institu- referendum in June of 2016, when a major- tions designed to do so are all-too-easily ity of those who voted produced the result hijacked for other purposes. But deliber- to leave, even though polling indicated that ation is in any case the wrong goal. Com- the median British voter favored the uk’s petition is the lifeblood of democratic pol- remaining in the European Union, as did itics, and not just because it is the mech- substantial majorities of both major parties anism by which governments that lose in the House of Commons.17 elections give up power. Institutions that Some will say that making the system re- foster competition also structure politics sponsive to voters with intense preferenc- around argument, which Mill was right es is a good thing. There is, indeed, a strand to identify as vital to the advancement of of democratic theory dating back to James knowledge and good public policy. Buchannan and Gordon Tullock’s Calculus But not any competition. The contesta- of Consent in 1962 whose proponents defend tion over governing ideas that Mill prized is vote trading and vote buying on the utilitari- best served when two large parties are con- an ground that it improves the overall social strained to compete over potential govern- utility.18 But democracy’s purpose is to man- ing programs. It is compromised by multi- age power relations, not to maximize social party competition that encourages clien- utility. The contrary view would suggest that telism, as we have seen. And it is damaged it was right for the U.S. government to aban- even more by competition within parties, don Reconstruction when Southern whites which empowers people with local agen- opposed it with greater intensity than most das and intense preferences who partici- voters favored it, and that it was right for pate disproportionately in primaries and the intense preferences of neoconserva- caucuses. This can render parties vulnera- tives who wanted the United States to in- ble to the ideological capture of candidates vade Iraq in 2003 to override those of more- by well-funded groups, as has happened numerous but less-fervent skeptics.19 This with the Tea Party in Southern and Mid-

82 Dædalus, the Journal of the American Academy of Arts & Sciences western Republican primaries since 2009. democratic politics by relegating it to a Ian But a more general problem is associated purely consultative role; but in that case, it Shapiro with local control of selection processes, is hard to see what the hype surrounding de- in which candidates find themselves com- liberation amounts to. Regardless, the most pelled to compete by promising to secure pressing political challenges in the United local goods. Once elected, they face pow- States do not result from lack of delibera- erful incentives to engage in pork barrel tion. Rather, they stem from the increas- politics with other similarly situated poli- ing subversion of democracy by powerful ticians, protecting public funding for sine- private interests since the Supreme Court’s cures and bridges to nowhere in their dis- disastrous equation of money with speech tricts. This problem is worse in districts– in Buckley v. Valeo four decades ago, and the the vast majority in the United States–that subsequent playing out of that logic in Cit- have been gerrymandered to be safe seats, izens United and subsequent decisions.21 As so that the primary is the only meaningful politicians have become increasingly depen- election. It is better for party leaders to seek dent on countless millions of dollars to gain candidates who can both win in their dis- and retain political office, those with the re- tricts and support a program that can win sources they need undermine the process by nationally. The leaders, in turn, are held manufacturing–and then manning–huge accountable by the backbenchers who re- barriers to entry, by contributing to both po- move them when they fail to deliver win- litical parties in ways that stifle competition, ning platforms. In sum, two large, centrally by capturing regulators and whole regulato- controlled parties are most likely to foster ry agencies, by giving multimillionaires and the programmatic competition that is best billionaires the preposterous advantage of for democratic politics. By contrast, multi- running self-funded campaigns, and by do- party competition encourages wholesale ing other end-runs around democratic pol- clientelism, and intraparty competition en- itics. Unless and until that challenge can be courages retail clientelism. addressed, debating what deliberation can Deliberation can be rendered harmless add to politics is little more than a waste of and perhaps, occasionally, beneficial for time.

endnotes 1 For example, see Jane J. Mansbridge, Beyond Adversary Democracy, 2nd ed. (Chicago: Universi- ty of Chicago Press, 1983); Amy Gutmann and Dennis Thompson, Democracy and Disagreement (Cambridge, Mass.: Harvard University Press, 1996); Amy Gutmann and Dennis Thompson, Why Deliberative Democracy (Princeton, N.J.: Princeton University Press, 2009); James S. Fish- kin, The Voice of the People (New Haven, Conn.: Yale University Press, 1995); James S. Fishkin, When the People Speak: Deliberative Democracy and Public Consultation (Oxford: Oxford University Press, 2011); Hélène Landemore, Democratic Reason: Politics, Collective Intelligence, and the Rule of the Many (Princeton, N.J.: Princeton University Press, 2012); Jürgen Habermas, Communica- tion and the Evolution of Society (Boston: Beacon Press, 1975); and Jürgen Habermas, The Theory of Communicative Action (Boston: Beacon Press, 1984). There are, of course, major differences among these and other theorists of deliberative democracy that do not concern me here. 2 John Stuart Mill, On Liberty, ed. David Bromwich (New Haven, Conn.: Yale University Press, 2003 [1859]), 86–120. 3 See Anthony Downs, An Economic Theory of Democracy (New York: Harper and Row, 1957), 244– 246, 266–271.

146 (3) Summer 2017 83 Against 4 Joseph Schumpeter, Capitalism, Socialism, and Democracy (London: George Allen & Unwin, 1942), Political 269. Deliberation 5 Where there is substantial regional variation, by contrast, as in India, smp systems can produce party proliferation. 6 Tocqueville described the Senate as peopled by America’s “ablest citizens”; men moved by “lofty thoughts and generous instincts.” By contrast, the House of Representatives consist- ed of “village lawyers, tradesmen, or even men of the lowest class” who were of “vulgar de- meanor,” animated by “vices” and “petty passions.” Alexis de Tocqueville, Democracy in Amer- ica, ed. J. P. Mayer (New York: Anchor Books, 1969 [1835, 1840]), 200–201. 7 Terence Samuel, The Upper House: A Journey behind Closed Doors (New York: Palgrave MacMillan, 2010), 68. 8 James Madison, “Federalist No. 51,” in The Federalist Papers, ed. Ian Shapiro (New Haven, Conn.: Yale University Press, 2009 [1787–1788]), 264. 9 Ibid. 10 Jane Mayer, Dark Money: The Hidden History of the Billionaires behind the Rise of the Radical Right (New York: Doubleday, 2016), 198–225. 11 The House of Lords Act of 1999 reduced the membership from 1,330 to 699 and got rid of all but ninety-two of the hereditary peers, who were allowed to remain on an interim basis, and an additional ten who were made life peers. On the recent evolution, see Meg Russell, The Contemporary House of Lords: Westminster Bicameralism Revived (Oxford: Oxford University Press, 2013), 13–35, 258–284. 12 For elaboration, see Ian Shapiro, Politics against Domination (Cambridge, Mass.: Harvard Uni- versity Press, 2016), 73–78. 13 See Ian Shapiro, The Real World of Democratic Theory (Princeton, N.J.: Princeton University Press, 2010), 266–271. 14 James Lyons and Sanya Burgess, “Corbyn Reelected as Labour Leader with Increased Mandate,” The Times, September 24, 2016, http://www.thetimes.co.uk/edition/news/corbyn-re-elected -as-labour-leader-with-increased-mandate-c7jqjgjm7. 15 The argument in this and the next two paragraphs will be developed more fully in Frances Rosenbluth and Ian Shapiro, Democratic Competition: The Good, the Bad, and the Ugly (New Haven, Conn.: Yale University Press, forthcoming 2018). 16 Kathleen Bawn and Frances Rosenbluth, “Short versus Long Coalitions: Electoral Accountabil- ity and the Size of the Public Sector,” American Journal of Political Science 50 (2) (2006): 251–265. 17 Charlie Cooper, “eu Referendum: Final Polls Show Remain with Edge over Brexit,” The Inde- pendent, June 23, 2016, http://www.independent.co.uk/news/uk/politics/eu-referendum-poll -brexit-remain-vote-leave-live-latest-who-will-win-results-populus-a7097261.html; and Anushka Asthana, “Parliamentary Fightback against Brexit on Cards,” The Guardian, June 26, 2016, http://www.theguardian.com/politics/2016/jun/26/fightback-against-brexit-on-cards-remain -eu-referendum-heseltine. 18 James Buchanan and Gordon Tullock, The Calculus of Consent: Logical Foundations of Constitutional Democracy (Ann Arbor: University of Michigan Press, 1962). Robert Dahl flirted with the notion that attending to intensity might be desirable from the standpoint of political stability, though he was skeptical that it could be measured. Robert Dahl, A Preface to Democratic Theory (Chicago: University of Chicago Press, 1956), 90–123. 19 For discussion of the dangers inherent in catering to intense preferences, see Shapiro, Politics against Domination, 46–61. 20 See Mayer, Dark Money, 120–158, 226–270, 354–387. 21 See Buckley v. Valeo 424 U.S. 1, 59 (1976); Citizens United v. Federal Election Commission 558 U.S. 310 (2010); and SpeechNow.org v. Federal Election Commission 599 F.3d 686, 689 (D.C. Cir. 2010). 84 Dædalus, the Journal of the American Academy of Arts & Sciences Can Democracy be Deliberative & Participatory? The Democratic Case for Political Uses of Mini-Publics

Cristina Lafont

Abstract: This essay focuses on recent proposals to confer decisional status upon deliberative mini-publics such as citizens’ juries, Deliberative Polls, and citizens’ assemblies. Against such proposals, I argue that inserting deliberative mini-publics into political decision-making processes would diminish the democratic legitimacy of the political system as a whole. This negative conclusion invites a question: which political uses of mini-publics would yield genuinely democratic improvements? Drawing from a participatory conception of deliberative democracy, I propose several uses of mini-publics that could enhance the democratic legitimacy of political decision-making in current societies.

There is a difference between a sample of several hundred speaking for the nation and the entire citizenry actually speaking for itself.1 –James Fishkin, The Voice of the People

In recent decades, deliberative democracy has become increasingly popular.2 One of the reasons for its popularity is that it offers an attractive interpretation of the democratic ideal of self-government. CRISTINA LAFONT is Professor of Philosophy at Northwestern Uni- According to the ideal of deliberative democracy, versity. She is the author of Global citizens must justify to one another–based on rea- Governance and Human Rights (2012), sons that everyone can reasonably accept–the co- Heidegger, Language and World-Disclo- ercive policies with which they must comply. To the sure (2000), and The Linguistic Turn extent that citizens can mutually justify the political in Hermeneutic Philosophy (1999) and coercion they exercise over one another, they can see coeditor of The Future of Critical The- themselves as colegislators or political equals in pre- ory: Transforming the Global Political cisely the way the democratic ideal of self-govern- and Economic Order (with Penelope 3 Deutscher, 2017) and the Habermas ment requires. The essential contribution of pub- Handbuch (with Hauke Brunkhorst lic deliberation to democratic legitimacy is that it and Regina Kreide, forthcoming). enables citizens to endorse the laws and policies to

85 The which they are subject as their own. In the only reaching better outcomes but, above Democratic absence of a commitment to mutual justi- all, convincing those who will be bound by Case for Political fication, citizens cannot meaningfully see them that this is indeed the case by provid- Uses of themselves as participants in collective self- ing mutually acceptable reasons. Therefore, Mini-Publics rule, but instead come to see themselves as improving the quality of deliberation in the coerced into compliance by others. Indeed, processes of opinion and will formation in by adding a requirement of public justifi- which citizens participate is an equally non- cation, the deliberative model provides a negotiable aim for the realization of delib- way in which citizens might prevent po- erative democracy. Institutional proposals litical domination by consolidated major- for realizing deliberative democracy must ities. They can engage in public delibera- be assessed by their promise to enhance tion in order to show that their proposals the democratic legitimacy of the political are supported by better reasons and hold system in which they will be implemented out hope that the force of the better argu- from both the deliberative and the partici- ment may move other citizens to change patory perspective.6 their political preferences.4 The claim that Taking this double criterion as a guide better reasons (and not just a higher num- can be helpful in examining proposals for ber of votes) lend legitimacy to the out- democratic innovation that focus on polit- comes of democratic decisions is crucial to ical uses of deliberative mini-publics, in- the idea of mutual justification as a criteri- cluding citizens’ juries, consensus confer- on of democratic legitimacy, distinguish- ences, Deliberative Polls, and citizens’ as- ing deliberative democracy from other semblies.7 Many deliberative democrats conceptions of democracy. enthusiastically endorse the proliferation The idea of mutual justification helps in of mini-publics as a way to lead current understanding the internal connection be- democratic societies closer to the ideal of tween the different political values essen- a deliberative democracy.8 Some authors tial to deliberative democracy. On the one are more daring than others: among cur- hand, given that the epistemic quality of po- rent proposals there is a split between those litical deliberation has a direct impact on who endorse conferring decisional status the legitimacy of its outcomes, improving on mini-publics directly, so that their rec- the quality of deliberation is a nonnegotia- ommendations would be taken up by the ble aim for the realization of deliberative relevant political authorities without any democracy. The more informed, impartial, need to ask for ratification by the citizenry mutually respectful, and open to counter- (such as through elections or a referen- arguments participants are in deliberation, dum) and those who hesitate to go as far the more likely it is that they will reach sub- as to hand over actual political power (like stantively better political decisions, such as of legislation or constitutional interpreta- those supported by the better reasons. On tion) to mini-publics.9 It is easy to see what the other hand, since the justification of drives the push toward the most ambitious political decisions to those subject to them option. A key reason to favor the institu- has a direct impact on their democratic le- tionalization of deliberative mini-publics gitimacy as well, it is not sufficient that po- is that their recommendations are of better litical decisions be substantively good ac- deliberative quality and thus would lead to cording to someone or other. They must better outcomes. They reflect the delibera- be endorsed by those who will be bound tive transformation of raw, uniformed pub- by them: that is, the citizenry in question.5 lic opinion into considered public opinion. The point of democratic deliberation is not However, if the citizenry must ultimately

86 Dædalus, the Journal of the American Academy of Arts & Sciences accept or reject the mini-publics’ recom- deliberative democracy: namely, their high Cristina mendations, and (as may often be the case) deliberative quality and their democratic Lafont they are not aware of the mini-publics’ de- representativeness. Deliberative mini-pub- liberations or reasons, then the decision lics provide a space for high-quality face- will in fact be based on their raw, uniformed to-face deliberation in which participants opinions, canceling out the potential gains receive balanced information on some im- of using mini-publics. If giving mini-pub- portant political issue, are exposed to a vari- lics some decisional status is normatively ety of relevant social perspectives, and have desirable at all, then they should be allowed the opportunity to weigh the pro and con to make the decisions in question. There arguments in order to reach a considered does not seem to be a lot of space for hesi- judgment. Participants are randomly select- tation at that point. ed among ordinary citizens and, as a conse- In light of the general enthusiasm among quence, their initial raw opinions on the is- deliberative democrats about the potential sues in question can be quite uninformed, benefits of inserting mini-publics in the po- perhaps even biased or manipulated. How- litical process, it is becoming increasingly ever, the filter provided by the deliberative harder to see the motivations of those who experience enables them to reach consid- hesitate. In what follows, I would like to ered judgments on the issues in question. contribute to this debate by offering some In fact, their views are often significantly arguments from the other side. First, adopt- transformed. Thus, it is plausible to assume ing a participatory perspective, I argue that, that inserting mini-publics into the politi- whatever the benefits of conferring deci- cal process would lead to substantively bet- sional status on mini-publics may be, they ter outcomes. Still, quality deliberation has are unrelated to democratization. Whether nothing to do with democracy per se. It is or not they would increase the deliberative the representativeness of mini-publics that quality of the political system as a whole, makes them democratically significant, as they would diminish their democratic le- compared with other deliberative forums. gitimacy.10 However, the point of the argu- Participants in mini-publics are randomly ment is not to reject the use of mini-pub- selected among ordinary citizens precisely lics altogether or to claim that they cannot with the purpose of getting a representative genuinely contribute to democratization. sample of the population. Although differ- To the contrary, showing that conferring ent types of mini-publics reach that goal to decisional status on mini-publics would different degrees, I will focus on Delibera- not be a democratic improvement is only a tive Polls, since it is generally considered the first step in addressing the question of when gold standard in terms of achieving repre- and how mini-publics could be used to lead sentativeness.11 to such improvement. I address this ques- As with all other types of mini-publics, tion in a second step by exploring different the idea behind Deliberative Polling is to uses of mini-publics that could improve the take a relatively small group, which every- democratic legitimacy of current societies. one had an equal chance to be a part of, and provide it with good conditions for deliber- Among the many political innovations ating over some relatively short period of developed in recent decades, deliberative time.12 The techniques of stratified random mini-publics are particularly attractive to sampling used in Deliberative Polling offer deliberative democrats. The reasons have scientific support for the claim that the or- to do with two features of mini-publics that dinary citizens who participate in the de- are of special significance to the ideal of a liberative experience are an accurate mirror

146 (3) Summer 2017 87 The of the population as a whole; consequently, epistemic and the democratic quality of po- Democratic their views, interests, and values reflect litical outcomes. Case for Political those of the people. What is so interesting Unfortunately, I think that this impres- Uses of in recreating a microcosm of the people is sion is an illusion. Here is the argument in Mini-Publics that precisely to the same extent that their a nutshell. Proposals to confer decisional initial judgments reflect the raw and unin- status on mini-publics can be justified by formed public opinion that can be captured epistemic considerations concerning the by regular polls, it is plausible to claim that better quality of their outcomes. Howev- their judgments after the deliberative expe- er, this line of argument, which is based on rience reflect what the public would think if the filter claim, offers no basis for justify- they were informed and had the opportuni- ing the mirror claim. Alternatively, propos- ty to deliberate about the matter. As James als to confer decisional status on mini-pub- Fishkin has put it, “deliberative polling has lics can be justified by democratic consider- a strong basis for representing the consid- ations concerning their representativeness. ered judgments of the people.”13 However, this line of argument, which is This explains why mini-publics are so based on the mirror claim, offers no ba- fascinating for deliberative democrats, sis for justifying the filter claim. Since de- for they offer precisely the combination fending the ideal of deliberative democracy of deliberative filter and democratic mir- requires justifying both claims, neither of ror that the ideal of a deliberative democ- these lines of argument is viable for vindi- racy requires. In endorsing this political in- cating deliberative democracy. If one fol- novation, deliberative democrats can avoid lows the first line of argument, the proposal having to choose one or the other as alter- collapses into a special version of elite con- native conceptions of democracy do. Elite ceptions of democracy (that is, a more egal- conceptions of democracy choose the fil- itarian version of blind deference to experts ter over the mirror. They promise better than the standard variety), whereas if one political outcomes, but at the price of tak- follows the second line of argument, the ing decision-making away from the people proposal collapses into a special version of and placing it in the hands of experts and pluralist democracy (that is, a stronger ver- political elites. By contrast, pluralist con- sion of procedural majoritarianism than the ceptions of democracy choose the mirror standard variety). over the filter. They promise to leave decision-making in the hands of the people, but The first of two possible defenses of em- at the price of endorsing majoritarian pro- powered mini-publics is the epistemic defense. cedures that are insensitive to the quality Following this line of defense, proposals of citizens’ preferences and can therefore to confer decisional status on mini-pub- lead to unreasonable outcomes (whenev- lics must justify the choice of this partic- er those preferences are uninformed, self- ular institution vis-à-vis other alternatives interested, biased, manipulated, and so on epistemic grounds. Whether mini-pub- on). The contrast to these unattractive al- lics in particular are preferable to potential ternatives explains the motivation behind alternatives depends on whether their pe- proposals to confer decision-making au- culiar feature of “mirroring the people” has thority to mini-publics. Since mini-pub- some superior epistemic value over other lics combine the deliberative filter with the features of alternative institutions that, pre- democratic mirror, conferring decisional cisely because they do not have to mirror the status on them seems to offer a unique op- people, may yield higher epistemic payoffs. portunity to simultaneously improve the It is always possible that, by offering high-

88 Dædalus, the Journal of the American Academy of Arts & Sciences er levels of expertise, diversity, or impar- worrisome to democrats. For if decision- Cristina tiality (whichever features matter most in al status should be conferred to mini-pub- Lafont each case from a substantive point of view), lics in virtue of the better epistemic qual- alternative institutions could lead to even ity of their considered judgments when better outcomes. On complex political is- compared with the raw opinions of the sues, it would seem that actual experts with actual people, then it is not clear how one deep knowledge about the issues in ques- follows the argument up to precisely that tion would be a better option than a ran- point of better epistemic quality and then dom group of laypeople who had a few days’ avoids drawing the full consequences that worth of training to form their opinions. Al- seem to follow. Whether or not delibera- most any other group of experts would of- tive mini-publics deserve a special hearing ten have stronger epistemic credentials. At in the political system, the biggest concern the very least, it would seem to be an open, with this line of argument is the obvious empirical question which group of experts implication that the raw voice of the actu- would be best in each particular case, deal people “is not a voice that by itself depending on the issue at hand. But setting serves any special hearing.”15 If this is the aside technocratic alternatives, let us focus case, then one wonders what justifies demon possible variations in the configuration ocratic elections, which give the strongest of mini-publics. possible hearing to that voice by letting the If what matters is the epistemic quali- actual people make crucial political deci- ty of their outcomes, why limit the pool sions with no deliberative filter whatsoev- of participants to citizens of the polity in er (by secret ballot). If the voice of the actu- question? Why not add epistemic diver- al people does not deserve any special hear- sity to the sample by including some for- ing, why let them vote? Indeed, it is upon eigners (whether laypeople or experts) if the basis of precisely this line of argument that would lead to better outcomes?14 Or that defenders of elite conceptions of de- why not increase impartiality by exclud- mocracy conclude that the actual people ing from the sample any citizens likely to should never be allowed to make substan- be biased with regard to the political issue at tive political decisions; they should only be hand? This is the normal procedure in jury allowed to elect officials among the com- selection, for instance, where no attempt is peting political elites and keep them ac- made to “mirror the people” precisely in or- countable through the threat of removing der to reach better outcomes (more impar- them from office.16 Taking this line of ar- tial decisions). There are endless variations gument seriously would suggest more am- for improving the quality of the deliberative bitious proposals for reform. Why not use filter. The point is simply that it would be deliberative mini-publics to make all politi- extraordinarily serendipitous if “mirroring cal decisions that are currently made by the the people” would invariably be the supe- actual people in democratic societies (such rior option among all possible alternatives as general elections, referenda, and popular from the strictly epistemic point of view of initiatives)? If institutionalizing mini-pub- the quality of outcomes. lics for making some political decisions is a Proving this claim seems like a tall or- net improvement in the deliberative qual- der. But even assuming for the sake of ar- ity of the political system as a whole, then gument that the challenge could be met and it would seem to follow that the more de- that mini-publics would always turn out to cisions made by mini-publics and the few- be epistemically superior to other alterna- er by the actual people, the more the polit- tives, this line of argument should still be ical system would improve.

146 (3) Summer 2017 89 The Now, for those who might find these kin’s mirror claim. After all, for all their Democratic consequences worrisome, the alternative deficiencies, no one questions that partic- Case for Political line of defense, the democratic defense, may ipants in mini-publics are more represen- Uses of seem more promising. Instead of focusing tative of the people as a whole than partic- Mini-Publics on the filter claim and thereby jeopardizing ipants of other political institutions (in- the ability to hold on to the mirror claim, cluding judges, experts, political elites, and from a democratic point of view, it is more bureaucrats). This is why many authors as- appealing to focus on the mirror claim and sume that conferring decisional status on see whether the filter claim can be retained mini-publics would be a net democratic as well. Proposals to confer decisional sta- improvement for the political system, at tus on mini-publics could be defended on least compared with the alternative of lim- the basis of democratic considerations of iting that status to less representative in- representativeness while also getting the stitutions.18 extra boost that their deliberative quality These considerations point to the sec- provides for free, so to speak. It could be ond line of argument mentioned above. argued that, even if some alternative insti- The case for conferring decisional status tutions could offer a better deliberative fil- on mini-publics can be based on demo- ter and thus lead to better outcomes, since cratic considerations of representative- they will be less democratically representa- ness. It can be argued that citizens should tive than mini-publics, the latter win by de- trust mini-publics’ decisions, but not be- fault simply by virtue of the democratic val- cause their participants would always be the ue expressed by the mirror claim. But even if most reliable group to make the best deci- conferring decisional status on mini-publics sion–a claim most likely to be false. Rath- is not the highest epistemic improvement er, citizens should trust mini-publics’ de- possible, so the argument goes, whichever cisions because their participants are like modest improvement over the status quo them.19 They should trust mini-publics their deliberative filter offers is an addition- precisely because they are a mirror of the al benefit that also counts in their favor. To people. Therefore, their considered opin- show why this line of argument does not ion is likely to reflect what they themselves work, we need to examine the mirror claim would have concluded had they participat- in depth. ed. This view is often associated with a se- As mentioned, one of the most interest- lection model of representation in contrast ing features of mini-publics is their statis- to a sanction model.20 Within the sanction tical representativeness. In particular, De- model, representatives are expected to ac- liberative Polls seem most able to avoid curately track the attitudes and views of the problems of self-selection that plague their constituents or face the sanction of other types of mini-publics. Many authors not being reelected, whereas in the selec- therefore consider them to be strongest in tion model, constituents choose represen- representativeness. This is not to deny that tatives with views and objectives largely from an empirical perspective the actual aligned with their own, giving representa- accuracy of the stratified random selec- tives self-motivated, exogenous reasons to tion techniques used in Deliberative Poll- do what their constituents want. As polit- ing can be called into question in specif- ical scientists Warren Miller and Donald ic cases.17 But for the purposes of my ar- Stokes put it, they choose representatives gument, let’s assume that methodological who “so share their views that in following improvements could satisfactorily solve their own convictions they do their con- these problems and we can grant Fish- stituent’s will.”21

90 Dædalus, the Journal of the American Academy of Arts & Sciences Let’s examine the mirror claim in detail. ipants cannot assume that the conclusions Cristina As just mentioned, the argument support- reached by the mini-public reflect what Lafont ing the mirror claim is that we should confer they would have thought if they had par- decisional status on mini-publics because ticipated. For, in principle, the opposite is their participants are like us. There are sev- equally possible. After all, the participants eral claims involved in this argument. Par- in the minority have reviewed the same in- ticipants in mini-publics are like us in the formation and deliberated as much as the sense that they are ordinary citizens and others while reaching the opposite conclu- thus, in contrast to politicians, lobbyists, sion. Even if citizens can trust that all par- and other political actors, are unlikely to ticipants were genuinely interested in fig- have hidden agendas or conflicts of inter- uring out what is best for the polity, they est in their deliberations about the public know that in pluralistic democracies there interest. We can trust them as our represen- is ongoing contestation over a variety of so- tatives in the sense that we do not need to cial, moral, ethical, religious, and economic monitor them or threaten them with sanc- views and values, which bears significantly tions because they are independently motion political questions and policy objectives. vated to figure out what is best for the poli- The selection model of representation ty. But, in line with the selection model of seems plausible at a smaller scale. Citizens representation, participants in mini-pub- can trust some political party, civil soci- lics are supposedly like us in a stronger ety organization, or individuals who share sense: namely, they share our interests, val- their interests, values, and policy objectives. ues, and policy objectives.22 This is why we But for that same reason, it would not make are supposed to trust them. Not only do we sense for them to also trust those political not need to threaten them with sanctions to parties, organizations, and individuals that keep them accountable, but we can also as- defend the contrary views, values, and poli- sume that their recommendations coincide cy objectives. If I trust Oxfam’s recommen- with what we would have thought if we had dations on poverty relief, I cannot also trust participated. For that reason, we should en- the opposite recommendations of, say, the dorse their recommendations as our own, Chamber of Commerce. If I trust Planned for example, when we decide how to vote Parenthood’s recommendations on wom- on referenda.23 en’s reproductive health, I cannot also trust Now, whereas the first mirror claim the opposite recommendations of the Pro- seems plausible, the second seems prob- Life Action League. Since I cannot simulta- lematic. Given how much ethical and po- neously trust the conflicting views, values, litical disagreement there is among citizens and policy objectives of all these different in pluralistic societies, the stronger mirror actors, I cannot trust the recommendations claim can hardly be true of a genuinely rep- of the majority of the sample without first resentative sample of the population. The knowing whether they have taken the side more diverse evaluative perspectives (con- in the political spectrum that I would have cerning need interpretations, value orienta- taken if I had participated. tions, comprehensive views, and so on) are Of course, if the materials and delibera- included in the sample, as they should be, tions are made public, citizens can always the less sense it makes for nonparticipant find out whether this is the case. But once citizens to assume that their interests, val- they do, they will no longer be trusting the ues, and political objectives will invariably mini-public. They will be trusting them- coincide with those of the majority of the selves. More important, many of them sample regardless of the issue. Nonpartic- will find out that the majority of the sam-

146 (3) Summer 2017 91 The ple is not like them, since they actually op- This is precisely the problem: the discon- Democratic pose their views, values, and policy objec- nect between the views of the mini-pub- Case for Political tives on the issue in question. At this point, lic and those of the actual people, induced Uses of the line of argument based on the mirror by the effective intervention of the delib- Mini-Publics claim predictably collapses. For the fact erative filter, undermines the mirror claim that the random sample is a microcosm of at the post-deliberative stage. As political the people taken collectively means that, scholar John Parkinson points out, by be- for contested issues, there will be a major- coming better informed and having rea- ity defending one view and a minority de- soned about the issues in question, par- fending the opposite view, and therefore it ticipants in the mini-public have become cannot be true of all people taken individ- more like experts on those issues than or- ually that the majority of the sample is like dinary citizens.26 them. But if they are not, in which sense are Before the deliberative experience, the they their representatives? If the majority mirror claim could plausibly be granted. of the sample is neither like them nor ac- Assuming the selection process were suc- countable to them, why should those citi- cessfully conducted, it seems trivially true zens trust them? It seems that the argument to claim that participants in the mini-pub- is not that citizens should trust the majori- lic were like the people in the sense that ty of the sample because they are like them, the views of the random sample would ac- but because they are like the majority of the curately reflect the views of the popula- people.24 But are they? At this point, it be- tion as a whole. This is why regular polls comes clear why this line of argument can- can be used (more or less reliably) to track not get the filter claim for free. In fact, the the views of the people despite the fact that filter claim undermines the mirror claim. only a handful of randomly selected citizens are actually interviewed. However, The key question here is whether the coun- once the deliberative filter is added, which terfactual or the actual people should rule. One is the very purpose of organizing delibera- of the main attractions behind the proposal tive mini-publics, the views of participants of conferring decisional status on mini-pub- undergo significant, at times drastic, trans- lics is precisely that their considered opin- formations. But, precisely for that reason, it ions are often different (and presumably would be a clear case of usurpation to claim better) than the raw opinions of the actu- that the voice of the mini-publics’ partici- al people. If they were not, there would not pants is the voice of the people at the post- be much of a point in conducting Delibera- deliberative stage, especially in those cas- tive Polls instead of regular polls. As Fish- es when they are on record as dissenting kin has argued: from the people. The populist temptation The thoughtful and informed views created to speak for the people is common among in the experiment [of Deliberative Polling] political actors of all kinds, but the dissimi- are not widely shared because the bulk of the larity between them and the people in real- public is still, in all likelihood, disengaged ity helps to undermine such claims. By con- and inattentive precisely because it is sub- trast, the similarity between mini-publics’ ject to all of the limitations . . . that routinely participants and the people at the initial apply to the opinions of citizens in the large- stage makes their dissimilarity at the empowered stage harder to spot. As such, they scale nation-state. Deliberative Polling over- 27 comes those conditions, at least for a time, could become the ultimate usurpers! for a microcosm, but leaves the rest of the pop- Deliberative democrats simply cannot ulation largely untouched.25 have it both ways. If the voice of the mini-

92 Dædalus, the Journal of the American Academy of Arts & Sciences publics deserves a special hearing, it is pre- We can easily see why such a politician or a Cristina cisely because it is not the voice of the ac- party, particularly one out of government, Lafont tual people. But because it is not, proposals can have political advantage to make from to confer decisional status on them cannot denouncing the existing, relatively lenient be justified on grounds of democratic repre- pattern of sentencing, calling for heavier sentativeness. Democracies are stuck with sentences, even perhaps for capital punish- the people they have, so political improvement. They can activate a politics of passion ments can count as democratic only if they in which they appear as the only individual take the people along instead of trying to or the only group really concerned about the bypass them by appealing to some favored sort of horrible crime in question. They can proxy. Political innovations can count as call into existence what Montesquieu called democratic only if they aim to transform a tyranny of the avengers, letting loose a rule the interests, views, and policy objectives of kneejerk emotional politics that works sys- of the actual people, so that the people can tematically against the common good. How continue to identify with the policies to might this sort of affront to deliberative de- which they are subject and endorse them as mocracy be rectified? Once again, the only their own, instead of being simply coerced hope would seem to lie in depolitization. It into compliance. This is what the demo- would require parliament to appoint a com- cratic ideal of self-rule requires. But before mission representative of relevant bodies of I analyze possible democratic contributions expertise and opinion, as well as of the people of mini-publics in that participatory sense, as a whole, to oversee criminal sentencing.29 let me add some argumentative support to the participatory case against empowered In the example, Pettit assumes that access mini-publics with the help of an example. to information about the adverse consequences of higher sentencing would move In his paper “Depoliticizing Democracy,” participants in the mini-public to reject ma- political theorist Phillip Pettit discuss- nipulative proposals of politicians, where- es an interesting example in support of as nonparticipants would be easily manip- one of his proposals for institutionalizing ulated to embrace higher sentences, even to mini-publics.28 We are asked to imagine a endorse capital punishment. This is why he polity in which a relatively mild sentenc- proposes the shortcut of mini-publics as the ing regime is working so well that impris- best solution to the problem. Instead of tak- onment is not often imposed. But it could ing the long road of providing the informa- happen that some convicted offender who tion to the citizenry so that they eventually received a light sentence (like community make up their minds on whether to oppose service) commits some horrific crime that higher sentences, he proposes institution- would not have happened if the offender alizing a mini-public as part of a commis- had been put in prison. In that context, pol- sion in charge of overseeing criminal sen- iticians looking for reelection can take ad- tencing. It seems to me that the example is vantage of the passions of the citizenry and plausible only if one assumes that there is ask for tougher sentencing in order to make no such thing as settled political views in their political opponents look weak and not a polity. Whereas it is easy to see how the sufficiently concerned, even if tougher sen- example would work in a country like the tencing would not serve the common good United States, where the death penalty is at all (it might increase rather than diminish not a settled issue, it is hard to imagine that the crime rate or be too expensive). Pettit it would work in a European country. To the explains: extent that rejection of the death penalty is

146 (3) Summer 2017 93 The a settled political view for an overwhelm- in pluralistic democracies. The collectivist Democratic ing majority of European citizens, it seems use of the expression suggests a kind of ho- Case for Political that no amount of political manipulation mogeneity among the citizenry that neither Uses of exercised upon an allegedly inattentive cit- exists nor is desirable in democratic societ- Mini-Publics izenry would succeed in bringing it back.30 ies committed to the maintenance of free If we compare these two hypothetical cas- institutions. Mini-publics are no exception. es, it seems to me that, contrary to Pettit’s Their members tend to disagree in their conclusion, informing the citizenry about considered opinions. Thus, even at their the political issue in question so that it be- best, what the outcomes of mini-publics comes settled is the only way a polity can reflect is not the considered opinion of the successfully shield itself from political ma- people, but the considered opinion of the nipulation. By contrast, taking the shortcut majority of the people. This is particularly of informing the members of a mini-pub- clear in the case of Deliberative Polls. Since lic while bypassing the citizenry as a whole participants are under no pressure to come would only delay the settling of the issue to an agreement on some collective opinion and thereby leave an open flank for political or recommendation, Deliberative Polling manipulation, backlash, and resentment. In reflects the real-world composition of ma- sum, the shortcut to better outcomes is the jority and minority opinions on the political long (participatory) road. issue in question. But even if one recogniz- Now, in order to take this participatory es that mini-publics’ outcomes only reflect perspective for evaluating proposals to in- the considered opinion of the majority of stitutionalize mini-publics seriously, we the people, it is still easy to see what is spe- have to enlarge the scope of analysis in both cial about them. They reflect what the ma- the temporal and spatial dimensions. We jority of the citizenry would think if they need to adopt a diachronic perspective in were informed and had the opportunity to order to assess their potential effects, not form a considered opinion on the political just at the particular moment in which a issue in question. policy decision is made but over time as If we adopt a participatory perspective, well. And we need to adopt a holistic per- what possible use could this information spective that takes into account the effects have for the citizenry? There are two as- of using mini-publics in the deliberative pects of this information that are demo- system as a whole.31 However, since the par- cratically significant: namely, that these are ticipatory perspective is citizen-centered, “considered judgments” and not just raw not system-centered, the potential effects preferences or uninformed opinions, and in the ongoing public debate among the cit- that they are the considered judgments of izenry are of special normative significance “the majority of the population.” But before to the analysis.32 I analyze these features of mini-publics, let me mention that in order to be of any use As we have seen, the mirror claim is an es- to the citizenry, citizens would need to be sential element in democratic defenses of familiarized with the mini-publics’ work- proposals to institutionalize mini-publics. ings, so that they would understand the However, in order to avoid the problemat- political significance they reveal. Different ic ambiguities we analyzed in the previous types of deliberative mini-publics have dif- section, it is important to avoid identify- ferent characteristics, but for simplicity of ing mini-public participants with “the peo- exposition, I will take Deliberative Polls as ple.”33 Speaking of “the people” in the sin- the paradigmatic example. My analysis of gular is always problematic, but particularly possible functions that mini-publics could

94 Dædalus, the Journal of the American Academy of Arts & Sciences perform in the political system does not as- aligned, the stakes could not be higher in Cristina sume that citizens would need to know all political struggles for shaping what counts Lafont the details about the workings of different as the majority opinion in a political com- types of mini-publics, but it does assume munity. It is in the context of this struggle that citizens would have become sufficient- that the information provided by mini-pub- ly familiar with them as to be aware of at lics acquires its political significance. least the following features: The alignment or misalignment be- 1) The techniques of stratified random tween majority opinion, public policies, sampling help ensure diversity and inclu- and mini-publics’ recommendations offers sion (that is, presence and voice of mar- a way of organizing the potential political ginalized social groups). This gives a high- uses of the latter so that their benefits or er level of representativeness to mini-pub- drawbacks can be better assessed. I distin- lics than almost any other political forum in guish the following four general catego- which the presence and voice of powerful ries under which the many potential uses social groups tend to predominate. of mini-publics can be subsumed: contes- 2) The random selection of participants tatory, vigilant, anticipatory, and empow- among ordinary citizens prevents co-op- ered. My brief analysis, however, does not tion by politicians or capture by organized aim to cover the innumerable applications interest groups. It helps ensure the political of mini-publics currently under discussion independence and impartiality of partici- in the vast empirical literature on applied pants and increases the chances that their deliberative democracy or to answer em- deliberations are oriented toward the public pirical questions of institutional design for interest. each type of mini-public use.34 My aims 3) The provision of information helps se- are more modest: I analyze some possible cure balanced briefing materials as well as political uses of mini-publics from the per- the inclusion of all relevant social perspec- spective of a participatory conception of tives. The presence of trained moderators deliberative democracy in order to identi- facilitates mutual deliberation, helps weigh fy the specific democratic values that could the pros and cons of different proposals, be served in each case, while offering a few and prevents collective deliberation from examples of how the relevant political ac- being hijacked. This allows participants to tors could best engage them. reach considered judgments on the political issues in question. First is the contestatory use of mini-publics. This special combination of features One reason to insert mini-publics into the justifies the claim that the conclusions of political process is the expectation that the mini-publics reflect the considered judg- majority opinion reached after deliberation ments of the majority of the population. by the mini-public will differ from the ma- Now, citizens do not need to believe that jority opinion of the population on the po- the considered judgments of the majority litical issue at hand. Discussions of this type are always right in order to appreciate the of mismatch tend to focus on the difference political significance of the majority opin- in the deliberative quality of the outcome. ion in democratic societies. For any politi- However, in my view, the fact that the dif- cal issue that can be legitimately decided by ference concerns the majority opinion has majority rule, the opinion of the decisional even greater significance. To the extent that majority determines the policies to which the political decisions in question are sup- all citizens are subject. Since majority opin- posed to be made by majority rule, show- ion and actual policies are supposed to be ing that considered majority opinion differs

146 (3) Summer 2017 95 The from current majority opinion gives minori- out what would happen if the general pub- Democratic ties a powerful tool to challenge consolidat- lic or the powerful groups that define the Case for Political ed majorities in their own terrain. It is one majority culture could actually listen to the Uses of thing for a minority to simply claim that needs, views, and arguments of minorities Mini-Publics they are right and the majority is wrong. It and marginalized groups. is quite another to provide some indepen- Assuming the general public is aware dent evidence indicating that the majority of the unique features of the venue that of a representative sample of the popula- mini-publics provide, mini-publics could tion came to endorse their view after hav- be used by organized social groups in their ing been properly informed. The fact that political struggles to contest the views of the minority view became a majority view consolidated majorities on specific politi- under these circumstances can be a power- cal issues. The more the mini-publics’ opin- ful political tool. In the context of a politi- ions differ from actual majority opinion, cal struggle on the contested political issue the more this should signal to the public the in question, the independent evidence pro- need to examine the available information vided by mini-publics could help minorities and the relevant perspectives so as to scru- challenge consolidated majorities and hold tinize their soundness and their potential them to account. The use of mini-publics for need for revision. This could lead to more political and legal contestation can thereby nuanced positions on polarizing issues or it serve the important function of protecting could prompt a general reconsideration of the democratic value of “non-tyranny,” to popular but unjust views held by consoli- use Fishkin’s expression.35 dated majorities. However, this is not to As mentioned above, a distinctive and suggest that the public should take the evi- valuable feature of mini-publics is the bet- dence provided by the mini-publics’ opin- ter ability to secure effective inclusion of ions as decisive or authoritative. The func- marginalized voices and social perspec- tion of mini-publics should not be to shut tives. By virtue of achieving higher statis- down political debate but, to the contrary, tical representativeness, mini-publics of- to reignite and facilitate the ongoing pub- fer a mirror of the people that is unmatched lic debate on contested political issues.38 by mirrors offered by other institutions in Mini-publics can enrich those wider de- the political system (from the judiciary to bates by enhancing the voices of silenced or the legislature, the media, and the public marginalized groups and perspectives in the sphere, among others), which tend to be public sphere. Precisely because the recom- highly exclusionary and therefore reflect mendations of the mini-public differ from back a distorted image of the people. Even actual public opinion, the distinction sig- in democratic societies, it is hard to ensure nals the need to transform public opinion effective inclusion in public political debate accordingly. This means that political ac- or in voting, given the disenfranchisement tors must address the mini-publics’ recom- of marginalized groups and the difficulties mendations to both officials and the pub- of providing a proper hearing to their in- lic with the aim of shaping ongoing politi- terests and views.36 Even if new venues for cal debate in the public sphere. citizen participation are created, self-selec- Mini-publics could be inserted in the po- tion, which tends to favor the wealthy and litical process not only for the purposes of educated, can worsen rather than improve political but also legal contestation. Of the the underrepresentation of the powerless many possibilities here, let me mention and marginalized.37 Thus, even democrat- two. Civil society groups could include the ic political systems lack venues for finding recommendations of mini-publics when fil-

96 Dædalus, the Journal of the American Academy of Arts & Sciences ing amicus curiae briefs to the Supreme Court mini-publics was based on cases when the Cristina as independent evidence for challenging the mini-publics’ recommendations differed Lafont assumption that raw public opinion actu- from the actual majority opinion on some ally reflects views “deeply rooted in the political issue. The driving idea was that the country’s history and traditions.”39 The more mini-publics’ recommendations dif- evidence in question should not be taken as fer from actual public opinion, the more this authoritatively settling the issue. Still, the should signal to the public the need to re- special features of mini-publics (their in- examine the available information and re- dependence, impartiality, representative- consider the soundness of the views and ness) confer on their recommendations a arguments supported by the majority cul- status of independent evidence that no oth- ture on the issue in question. But perhaps er evidence that parties may provide from even more significant are cases when the like-minded sources (interested groups or mini-public’s recommendations coincide organizations) can match. If mini-publics with the majority opinion but differ from are working as intended, outside parties can existing policy. This mismatch should sig- do nothing to influence the outcome. nal to the public the need to scrutinize the Stronger forms of institutionalization political system. The more mini-publics’ could also be beneficial. For example, it recommendations are aligned with public could become standard practice that, in opinion, but differ from the actually enact- cases involving suspect classifications of ed policies, the more this signals to the pub- groups with a history of discrimination, lic that the political system is not proper- which trigger a higher level of scrutiny, ly responsive to their views, interests, and some form of mini-public is routinely con- policy objectives. The evidence provided by vened to provide the Supreme Court with mini-publics could draw additional support additional information on what the consid- from the general public toward social and ered majority opinion of the country may political groups mobilized against which- be at a given time. Again, there is no need ever forces are impeding the proper flow of to claim that this information should be au- influence between the enacted policies and thoritative about the right way to interpret the processes of citizen opinion and will for- constitutional rights. The considered opin- mation in which citizens participate. By en- ion of the majority may still be unduly hos- hancing the responsiveness of the political tile toward protecting the rights of unpop- system to the interests, views, and policy ular minorities. But the information may objectives of the citizenry, such critical or nonetheless be valuable as an indication of vigilant uses of mini-publics would serve how far the considered judgment of the ma- the important political function of enhanc- jority is moving in a particular direction.40 ing democratic control. Whereas the con- Precisely because mini-publics would not testatory uses would strengthen political have decisional status, the political contes- equality in the horizontal dimension (be- tation that is likely to surround the interpre- tween socially powerful citizens and less tation of their opinions by different politi- powerful or marginalized citizens), the vig- cal groups would not be detrimental, espe- ilant uses of mini-publics would strengthen cially if it manages to spark a broader debate political equality in the vertical dimension in the public sphere as well, which impor- (between ordinary citizens and political tant Supreme Court cases tend to do. officials). One could also use mini-publics to en- I now turn to the vigilant uses of mini-pub- hance the agenda-setting power of ordi- lics. The analysis of contestatory uses of nary citizens, giving them more effective

146 (3) Summer 2017 97 The influence in the selection of policy objec- extremely powerful to the citizenry. It can Democratic tives to which the political system must re- effectively counteract arguments to the ef- Case for Political spond. Citizens could be regularly polled to fect that the majority’s support for some Uses of rank important political issues that need to popular policy is due to the citizenry’s lack Mini-Publics be tackled, and then mini-publics could be of information or familiarity with the com- convened to make recommendations con- plexity of the problems involved; or that it cerning the top-ranked issues. This process is due to irresponsible wishful thinking that would provide public visibility to the issues fails to take into account the potential con- in question, and this would be particularly sequences, legal constraints, or any other helpful concerning political issues that relevant dimensions that only experts (but elected officials may see as intractable or not ordinary citizens) can fully grasp. Pop- not worth confronting.41 Because officials ularity for self-defeating policy objectives have little incentive to tackle such issues, is not unheard of, as when citizens favor they are therefore likely to remain forever both expanding public services and lower- unresolved, even if the overwhelming ma- ing taxes at the same time. When this is the jority of citizens agree on what the right po- case, following the political will of the ma- litical solution would be. Think of the pol- jority could be extremely harmful. In such icy proposals for enforcing background public political debates, the contribution checks on gun sales in the United States, of a mini-public could be invaluable to the which are supported by 85 percent of the citizenry. It would force the political sys- population, but cannot make it through the tem to provide the needed information so legislature. that participants in the mini-public could Situations of political gridlock or the cap- engage in an independent examination of ture of political institutions by powerful in- the soundness of the arguments in ques- terest groups provide one of the key motivation. Whatever the mini-publics’ conclu- tions behind proposals to confer decision- sions may be, the public availability of these al status on mini-publics: they can get done arguments would be a great improvement what the legislature (perhaps even the ju- over the status quo. Indeed, for ordinary cit- diciary) is demonstrably unable to do. In izens, it would be a win-win situation. If the the context of his proposal to create a pop- arguments were right, they would have in- ular branch of government modeled on dependent evidence that might lead them mini-publics, law scholar Ethan Leib has to change their political opinions accord- argued that empowered mini-publics could ingly instead of having to blindly trust the make an essential contribution in situations bare assertions of potentially self-interested when citizens are frustrated by the legisla- parties. If the arguments were wrong, this ture’s unwillingness to take action or when would strengthen the ability of ordinary cit- legislatures find themselves unable to reach izens to pressure the relevant political ac- a reasonable compromise.42 Mini-publics tors into action by removing their demon- without decisional status would seem to strably unsupported excuses for inaction. make no contribution at all. If the citizenry already overwhelmingly endorses some Let’s turn to anticipatory uses of mini- political solution, organizing a mini-public publics. So far I have considered two differ- is likely only to reinforce the opinion the cit- ent forms of misalignment between major- izenry already holds, and thus would seem ity opinion, public policies, and mini-pub- to fulfill no function at all. lics’ opinions. But another form of mis- However, the fact that the mini-public of- alignment can be even more worrisome fers a considered majority opinion can be from a democratic perspective: when the

98 Dædalus, the Journal of the American Academy of Arts & Sciences public has no opinion at all about the polit- Under current conditions of global- Cristina ical issues in question. This type of discon- ization, inserting anticipatory mini-pub- Lafont nect does not have to be problematic. For lics into transnational political processes low-stakes issues that are technical in na- could have, in my opinion, the highest de- ture or serve merely an administrative pur- mocratizing impact.45 From a participato- pose, there may be no need at all for citi- ry perspective, the function of mini-pub- zens even to form an opinion on the poli- lics would not be to directly shape the pol- cies in question. But it is worrisome when icies in question, but instead to enhance the the public does not know anything about visibility of what is at stake so as to enable policies or legal developments that can neg- public debate among citizens. Their prima- atively impact their well-being or their fun- ry role would not be to recommend some damental rights. policies over others, but rather to acquire Such public ignorance can have various sufficient information so as to be able to roots. The policies in question may con- identify, among the various policies under cern technological innovations with unpre- consideration, those whose potential im- dictable consequences, with the public un- pact on citizens’ well-being, fundamen- aware of what may be at stake. For exam- tal rights, and interests is so high that the ple, think of new gene editing technologies public needs to know about them in or- such as crispr, which may permanently al- der to collectively determine in public deter the human genome.43 Or the public may bate which priorities, interests, and values be ignorant because the political decisions should guide the political decisions in ques- in question are migrating beyond nation- tion. By anticipating what citizens would al borders. International trade agreements think if they knew more about what is at are a paradigmatic example. Although they stake in political decisions that, for a vari- can have a tremendous impact on the do- ety of reasons, fall under the radar of the mestic economy of a country and its abil- public sphere–and by providing public vis- ity to protect the fundamental rights of its ibility to those decisions in which the stakes citizens, they are negotiated beyond nation- are so high that the citizenry should not real borders, often by the executive branch of main ignorant–mini-publics would fulfill government, without strong oversight by the crucial political function of enhancing the legislature, under the unilateral influ- democratic control.46 Instead of becoming ence of powerful lobbies, and surrounded another shortcut for bypassing the citizen- by secrecy. In the absence of public politi- ry, mini-publics could be deployed against cal debate and proper media coverage, most many of the existing shortcuts in order to citizens do not even know that they should force the political system to take the long know about the political decisions in ques- road of properly involving the citizenry. tion, given what is at stake. Disguised as remote foreign relations matters, transna- Finally, let me briefly address whether a tional agreements are not perceived by the participatory conception of deliberative de- citizenry as affecting domestic policy, with mocracy can ever endorse the use of empow- potentially severely harmful consequences ered mini-publics. This is a complex issue that would be quite hard to reverse, given that I cannot properly take on here. But to the number of countries involved. Transna- avoid possible misunderstandings, let me tional negotiations lack the visibility in the clarify that I do not take my argument to domestic public sphere needed to generate the conclusion that all uses of empowered a political debate in which citizens could ei- mini-publics would necessarily be demo- ther endorse or reject such policies.44 cratically suspect or illegitimate. Empow-

146 (3) Summer 2017 99 The ering mini-publics in connection with or in ical institutions that, for good reasons, are Democratic the form of an institution, like Bruce Ack- not themselves tied to direct forms of citi- Case for Political erman and James Fishkin’s Deliberation zen ratification (such as the judiciary). In Uses of Day, could be highly desirable from a par- such a case, although the inclusion of em- Mini-Publics ticipatory perspective.47 I also do not rule powered mini-publics may not increase the out the possibility of legitimate uses of em- democratic quality of the political system powered mini-publics that may not be di- as a whole, it may not decrease it either.48 rectly tied to referenda or some other form And if their use were recommended on oth- of citizen ratification. Indeed, empowered er grounds, then for all I have argued here, mini-publics could be inserted in the politi- there may be no reason to oppose them. cal process to share power with other polit-

endnotes 1 James S. Fishkin, The Voice of the People (New Haven, Conn.: Yale University Press, 1997), 44. 2 For surveys, see James Bohman, “Survey Article: The Coming of Age of Deliberative Democ- racy,” Journal of Political Philosophy 6 (4) (1998): 400–425; and Simone Chambers, “Delibera- tive Democratic Theory,” Annual Review of Political Science 6 (2003): 307–326. For recent collections that include different theoretical approaches, see Seyla Benhabib, ed., Democracy and Differ- ence (Princeton, N.J.: Princeton University Press, 1996); Samantha Besson and José Luis Martí, eds., Deliberative Democracy and its Discontents (Aldershot, United Kingdom: Ashgate, 2006); James Bohman and William Rehg, eds., Deliberative Democracy (Cambridge, Mass.: The mit Press, 1999); Jon Elster, Deliberative Democracy (Cambridge: Cambridge University Press, 1998); James S. Fishkin and Peter Laslett, eds., Debating Deliberative Democracy (Oxford: Blackwell, 2003); and Stephen Macedo, ed., Deliberative Politics (Princeton, N.J.: Princeton University Press, 1999). 3 As Amy Gutmann puts it, “the legitimate exercise of political authority requires justification to those people who are bound by it, and decision making by deliberation among free and equal citizens is the most defensible justification anyone has to offer for provisionally settling controversial issues.” See Amy Gutmann, “Democracy, Philosophy, and Justification,” in Democracy and Dif- ference, ed. Benhabib, 344. For similar renderings of this basic idea by other deliberative democrats, see Joshua Cohen, “Deliberation and Democratic Legitimacy,” in The Good Polity, ed. Alan Hamlin and Philip Pettit (Oxford: Blackwell, 1989), 17–34; Joshua Cohen, “Reflections on De- liberative Democracy,” in Philosophy, Politics, Democracy, ed. Joshua Cohen (Cambridge, Mass.: Harvard University Press, 2009), 330; Jürgen Habermas, Between Facts and Norms, trans. William Rehg (Cambridge, Mass.: The mit Press, 1996), 110; John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971), 227; John Rawls, Political Liberalism (New York: Columbia University Press, 1993), 137; and Philip Pettit, Republicanism (Oxford: Oxford University Press, 1997), 184. I analyze the intricacies of the commitment to mutual justification at the center of the ideal of deliberative democracy in Cristina Lafont, “Is the Ideal of Deliberative Democracy Coherent?” in Deliberative Democracy and its Discontents, ed. Besson and Martí, 3–26. 4 See Cohen, “Deliberation and Democratic Legitimacy”; Jon Elster, “The Market and the Fo- rum,” in The Foundations of Social Choice Theory, ed. Jon Elster and Aanund Hylland (Cambridge: Cambridge University Press, 1986), 103–132; Bernard Manin, “On Legitimacy and Political De- liberation,” Political Theory 15 (3) (1987): 338–368; David Miller, “Deliberative Democracy and Social Choice,” Political Studies 40 (1) (1992): 54–67; Cass Sunstein, “Preferences and Politics,” Philosophy and Public Affairs 20 (1) (1991): 3–34; and John S. Dryzek, Deliberative Democracy and Be- yond (Oxford: Oxford University Press, 2000). 5 As Dryzek puts it, political “outcomes are legitimate to the extent they receive reflective assent through participation in authentic deliberation by all those subject to the decision in question.” John S. Dryzek, Foundations and Frontiers of Deliberative Governance (Oxford: Oxford Uni-

100 Dædalus, the Journal of the American Academy of Arts & Sciences versity Press, 2010), 23. See also Habermas, Between Facts and Norms, 110; and James S. Fishkin, Cristina When the People Speak (Oxford: Oxford University Press, 2009), 46. Lafont 6 Proposals must be democratically legitimate from both the participatory and the deliberative perspective. This does not mean that they must always improve the political system in both dimensions, only that they cannot threaten one dimension of legitimacy for the sake of enhancing the other. I analyze the tensions between deliberation and participation in different conceptions of deliberative democracy in Cristina Lafont, “Deliberation, Participation, and Democratic Legitimacy: Should Mini-Publics Shape Public Policy?” Journal of Political Philosophy 23 (1) (2015): 40–63. 7 On citizens’ juries, see Anna Coote and Jo Lenaghan, eds., Citizens’ Juries: Theory into Practice (Lon- don: Institute for Public Policy Research, 1997); Ned Crosby and Doug Nethercut, “Citizens Juries: Creating a Trustworthy Voice of the People,” in The Deliberative Democracy Handbook, ed. John Gastil and Peter Levine (San Francisco: Jossey-Bass, 2005), 111–119; and Graham Smith and Corinne Wales, “Citizens’ Juries and Deliberative Democracy,” Political Studies 48 (1) (2000): 51–65. On consensus conferences, see Peter Dienel, Die Planungszelle. Der Bürger als Chance (Wies- baden, Germany: Westdeutscher Verlag, 2002); and Simon Joss and John Durant, eds., Public Participation in Science: The Role of Consensus Conferences in Europe (London: Science Museum, 1995). On deliberative polls, see James S. Fishkin, Democracy and Deliberation (New Haven, Conn.: Yale University Press, 1991); Fishkin, The Voice of the People; and Fishkin, When the People Speak. For a detailed catalog of the different institutional innovations currently available, see Gastil and Levine, eds., The Deliberative Democracy Handbook. For an analysis of different types of mini-publics, see Matthew Ryan and Graham Smith, “Defining Mini-Publics,” in Deliberative Mini-Publics: Involving Citizens in the Democratic Process, ed. Kimmo Grönlund, André Bächtiger, and Mija Setälä (Colchester, United Kingdom: ecpr Press, 2014), 9–26. 8 The view is too popular to provide an exhaustive list, but for a few examples, see endnote nine below. For some critical voices, see Simone Chambers, “Rhetoric and the Public Sphere: Has Deliberative Democracy Abandoned Mass Democracy?” Political Theory 37 (3) (2009): 323–350; Lafont, “Deliberation, Participation, and Democratic Legitimacy”; Carole Pateman, “Partic- ipatory Democracy Revisited,” Perspectives on Politics 10 (1) (2012): 7–19; and John Parkinson, Deliberating in the Real World (Oxford: Oxford University Press, 2006). 9 For examples of the first kind, see Hubertus Buchstein, “Reviving Randomness for Political Ra- tionality: Elements of a Theory of Aleatory Democracy,” Constellations 17 (3) (2010): 435–454; James S. Fishkin and Robert C. Luskin, “Broadcasts of Deliberative Polls: Aspirations and Ef- fects,” British Journal of Political Science 36 (1) (2006): 184–188; Archon Fung, “Minipublics: Delib- erative Designs and Their Consequences,” in Deliberation, Participation, and Democracy, ed. Shawn W. Rosenberg (New York: Palgrave Macmillan, 2007), 161, 165; Eric Ghosh, “Deliberative De- mocracy and the Countermajoritarian Difficulty: Considering Constitutional Juries,” Oxford Journal of Legal Studies 30 (2) (2010): 327–359; Robert E. Goodin and John S. Dryzek, “Delibera- tive Impacts: The Macro-Political Uptake of Mini-Publics,” Politics and Society 34 (2) (2006): 225; Ethan Leib, Deliberative Democracy in America: A Proposal for a Popular Branch of Government (Univer- sity Park: Penn State University Press, 2004); Sanford Levinson, “Democracy and the Extend- ed Republic: Reflections on the Fishkian Project,” The Good Society 19 (1) (2010): 66; Philip Pet- tit, “Depoliticizing Democracy,” in Deliberative Democracy and its Discontents, ed. Besson and Martí, 93–106; Philip Pettit, On the People’s Terms (Cambridge: Cambridge University Press, 2013); and William Talbott, Human Rights and Human Well-Being (Oxford: Oxford University Press, 2010). For examples of the second kind, see James S. Fishkin, “Deliberation by the People Themselves: Entry Points for the Public Voice,” Election Law Journal 12 (4) (2013): 496–507; and Christopher Zurn, Deliberative Democracy and the Institutions of Judicial Review (Cambridge: Cambridge Universi- ty Press, 2007). For an intermediate option that would leave it up to citizens whether to blindly trust the mini-publics’ recommendations, see Michael MacKenzie and Mark Warren, “Two Trust-Based Uses of Minipublics in Democratic Systems,” in Deliberative Systems, ed. Jane Mans- bridge and John Parkinson (Cambridge: Cambridge University Press, 2012), 95–124. For the sake of simplicity, I will refer to mini-publics with the power to make binding political decisions as “empowered mini-publics.”

146 (3) Summer 2017 101 The 10 My argument here focuses on democratic representation and builds on a more general analy- Democratic sis of democratic legitimacy that I offer in Lafont, “Deliberation, Participation, and Demo- Case for cratic Legitimacy,” but cannot reproduce here. Political Uses of 11 See Jane Mansbridge, “Deliberative Polling as the Gold Standard,” The Good Society 19 (1) Mini-Publics (2010): 55–62. 12 See Fishkin, Democracy and Deliberation; Fishkin, The Voice of the People; and Fishkin, When the People Speak. 13 Fishkin, When the People Speak, 28. 14 For instrumental justifications of democracy based on the value of epistemic diversity, see, for example, Hélène Landemore, Democratic Reason: Politics, Collective Intelligence, and the Rule of the Many (Princeton, N.J.: Princeton University Press, 2013). 15 Fishkin, “Deliberation by the People Themselves,” 504. 16 For classic defenses of this view, see Joseph A. Schumpeter, Capitalism, Socialism, and Democracy (New York: Harper & Row, 1942); and William H. Riker, Liberalism Against Populism (Long Grove, Ill.: Waveland Press, 1982). For a recent defense, see Richard A. Posner, Law, Pragma- tism, and Democracy (Cambridge, Mass.: Harvard University Press, 2003). 17 I discuss this issue in Lafont, “Deliberation, Participation, and Democratic Legitimacy” 49. 18 This assumption is particularly visible in proposals to institutionalize mini-publics for constitutional review. For example, see Ghosh, “Deliberative Democracy and the Countermajor- itarian Difficulty”; Horacio Spector, “Judicial Review, Rights, and Democracy,” Law and Phi- losophy 22 (3) (2003): 285–334; Horacio Spector, “The Right to a Constitutional Jury,” Legis- prudence 3 (1) (2009): 111–123; and Christopher Zurn, “Judicial Review, Constitutional Juries, and Civic Constitutional Fora: Rights, Democracy and Law,” Theoria: A Journal of Social and Po- litical Theory 58 (127) (2011): 63–94. Not to mention proposals for creating a “popular” branch of government modeled on mini-publics; see Leib, Deliberative Democracy in America. 19 It is important to notice that this line of argument does not fit well with the epistemic strategy that focuses on outcome considerations. From a strictly epistemic point of view, there is no reason to assume that “the people” are always or even often likely to reach the substantively best decisions. Think of all the important decisions, including the judicial, medical, economic, and scientific, that no one would propose being made by democratic referendum. So even if mini-publics reliably indicate the considered opinion of the majority of the population (and assuming they do), that still says nothing about whether those opinions are likely to be substantively correct. Indeed, given the drastic differences in considered public opinion on contested political issues among all countries of the world, they cannot all be right. If we take into account the temporal dimension, it is even more obvious how much considered public opinion on contested political issues has changed over time in all countries. Adopting this expanded perspective makes it entirely clear that the justification of the mirror claim depends on democratic, not epistemic, considerations. It assumes that the citizenry as a whole in each country is the constituent power: that is, has the legitimate authority to make the decisions in question regardless of whether it makes the right or wrong decisions. Under the democratic assumption of the right to self-government, the question then becomes whether the people should defer their decisional authority to mini-publics in some cases and, if so, why. 20 For an in-depth analysis of the selection model of representation, see Jane Mansbridge, “A ‘Selection Model’ of Political Representation,” Journal of Political Philosophy 17 (4) (2009): 369– 398. For the contrast between the selection and sanction models regarding mini-publics such as Deliberative Polls, see Mansbridge, “Deliberative Polling as the Gold Standard.” For an analysis of the contrast between these two models of representation under the rubrics “responsive” and “indicative,” see Philip Pettit, “Representation, Responsive and Indicative,” Constellations 17 (3) (2010): 426–434.

102 Dædalus, the Journal of the American Academy of Arts & Sciences 21 Warren E. Miller and Donald E. Stokes, “Constituency Influence in Congress,” American Po- Cristina litical Science Review 57 (1) (1963), as quoted in Mansbridge, “A ‘Selection Model’ of Political Lafont Representation,” 371. 22 As Mansbridge indicates concerning the alignment of objectives between agent and principal according to the selection model, “the alignment of objectives can take place not only on the high ground of similar understandings of what is best for the nation as a whole but also on what is best for particular individuals or communities such as farmers, miners, or inner city residents.” See Mansbridge, “A ‘Selection Model’ of Political Representation,” 380. 23 For a defense of such trust-based uses of mini-publics, see MacKenzie and Warren, “Two Trust-Based Uses of Minipublics in Democratic Systems.” 24 I cannot think of any interpretation of the selection model of representation in which it would be plausible to claim that citizens should trust the considered opinion of a majority of random others. I analyze the difficulties of this claim in Lafont, “Deliberation, Participation, and Democratic Legitimacy,” 54–57. But whether or not this view of representation could be considered plausible, the problem in our context is that the modified mirror claim on which it is based is false. 25 Fishkin, When the People Speak, 28. The emphasis is mine. 26 See Parkinson, Deliberating in the Real World, 82. 27 See David Owen and Graham Smith, “Deliberation, Democracy and the Systemic Turn,” Jour- nal of Political Philosophy 23 (2) (2015): 213–234. 28 See Pettit, “Depoliticizing Democracy,” 54–55. Pettit’s proposal leaves open whether to confer decisional status on mini-publics or to leave the ultimate control over them to Parliament. Either way, the innovation would bypass the citizenry, which is my focus here. 29 Ibid. 30 Since nothing turns on the specific example of a settled political issue, those with doubts about how settled the death penalty is in European countries can substitute it with any other example they consider settled, like burning offenders at the stake. 31 For an overview of different versions of the deliberative system approach, see Mansbridge and Parkinson, Deliberative Systems. 32 For an analysis of the differences between a system-centered and a citizen-centered interpretation of the deliberative systems approach, see Owen and Smith, “Deliberation, Democracy and the Systemic Turn,” 213–234. 33 This tendency is particularly visible in Leib’s proposal for a popular branch of government modeled on mini-publics, in which the voice and will of “a group of stratified random samples of laymen” is routinely identified with the voice and will of “the people.” See Leib, Deliberative Democracy in America, 72, 66. 34 For overviews of empirical applications of mini-publics, see Grönlund, Bächtiger, and Setälä, eds., Deliberative Mini-Publics; and Graham Smith, Democratic Innovations: Designing Institutions for Citizen Participation (Cambridge: Cambridge University Press, 2009), 72–110. For a comparative empirical analysis of the potential impacts of mini-publics in different kinds of states, see Dryzek, Foundations and Frontiers of Deliberative Democracy, 155–176. 35 See Fishkin, When the People Speak, 60–64. 36 See Iris Young, Inclusion and Democracy (Oxford: Oxford University Press, 2000). 37 See Jane Mansbridge, Beyond Adversary Democracy (Chicago: Chicago University Press, 1980). 38 For a defense of this claim in the context of an interesting analysis of different uses of mini-publics, see Nicole Curato and Marit Böker, “Linking Mini-Publics to the Deliberative System: A Re- search Agenda,” Policy Sciences 49 (2) (2016).

146 (3) Summer 2017 103 The 39 I am using here the formulation of the “deep roots test” that the U.S. Supreme Court uses as Democratic its substantive due process standard. For a critical analysis of that standard, see John C. Toro, Case for “The Charade of Tradition-Based Substantive Due Process,” New York University Journal of Law Political & Liberty 4 (2) (2009): 172–208. Uses of Mini-Publics 40 The level of empowerment of mini-publics in this context could be increased. For example, it could be required that the Supreme Court takes up their recommendations in the legal reasoning justifying its decisions and offers an explicit, reasoned justification whenever it rules against them. I mention this intermediate possibility not as a proposal I endorse, but simply to indicate that political empowerment comes in degrees; so for any possible use of mini-publics, the level of empowerment can range from the weakest option of conferring upon them a merely nonbinding and advisory role to the strongest possible option of conferring upon them the binding power to make final decisions unchecked by the citizenry or by any other political institution. Opposing the strongest form of empowerment, as I do, does not require endorsing the weakest form as the only legitimate option. 41 The latter include political issues in which elected officials have a clear conflict of interest, such as choosing among electoral systems or drawing electoral boundaries. Regarding these kinds of questions, vigilant uses of mini-publics would strengthen popular oversight of public officials especially if they were empowered to require public officials to appear before them to testify. See Goodin and Dryzek, “Deliberative Impacts: The Macro-Political Uptake of Mini-Publics,” 235–236; Dryzek, Foundations and Frontiers of Deliberative Democracy, 169; and John Ferenjohn, “Conclusion: The Citizens’ Assembly Model,” in Designing Deliberative Democracy, ed. Mark E. Warren and Hilary Pearse (Cambridge: Cambridge University Press, 2008), 192–213. 42 See Leib, Deliberative Democracy in America, 62. 43 See, for example, Jeffrey Perker, “crispr/cas Faces the Bioethics Spotlight,” BioTechniques 58 (5) (2015): 223–227. 44 Visibility should not be confused with transparency. Even when the information in question is publicly available, this may still be useless to the citizenry if its importance is not visible in the public sphere so that it can generate public awareness and political debate. On the crucial difference between transparency and visibility, and the special importance of the latter, see Stefan Rummens, “Staging Deliberation: The Role of Representative Institutions in the Deliberative Democratic Process,” Journal of Political Philosophy 20 (1) (2012): 29–41. 45 Many different institutional processes could accomplish these ends. One would require legislative standing committees overseeing major transnational agreements to convene some form of mini-public in advance of important binding decisions. Their empowerment could vary from merely indicating whether or not public debate is needed to setting the agenda on the specific issues in need of public debate (such as identifying specific environmental or ethical concerns and establishing proper priorities in light of significant trade-offs). 46 I take the idea of anticipatory uses of mini-publics from MacKenzie and Warren, “Two Trust- Based Uses of Minipublics in Democratic Systems.” However, my participatory interpretation of this use differs from theirs in that I do not consider it to be a trust-based use. Rather than the public simply entrusting mini-publics with the task of reaching a considered public opinion on the political issues in question so that these opinions may then be communicated to executive agencies or other public officials, from a participatory perspective, the function is instead to identify the issues about which the public needs to collectively form a considered public opinion, and communicate this information to both public officials and the citizenry. 47 See Bruce Ackerman and James S. Fishkin, Deliberation Day (New Haven, Conn.: Yale Uni- versity Press, 2004). However, this is not to say that all such uses would always be desirable, since other considerations may speak against them. For example, Christopher Zurn proposes to empower mini-publics for certifying popular amendment proposals and to require De- liberation Days for ratification or rejection by the citizenry. See Zurn, Deliberative Democracy and the Institutions of Judicial Review, 336; and Zurn, “Judicial Review, Constitutional Juries, and Civic Constitutional Fora.” This type of proposal may have impeccable participatory creden-

104 Dædalus, the Journal of the American Academy of Arts & Sciences tials, but it may raise concerns regarding political stability because it offers no criteria to lim- Cristina it what can and cannot be up for amendment. For a criticism along these lines, see Fishkin, Lafont “Deliberation by the People Themselves,” 506. 48 I am thinking here of proposals for empowering mini-publics in the context of constitutional review. See, for example, Ghosh, “Deliberative Democracy and the Countermajoritarian Dif- ficulty”; Spector, “Judicial Review, Rights, and Democracy”; Spector, “The Right to a Con- stitutional Jury”; and Zurn, “Judicial Review, Constitutional Juries, and Civic Constitutional Fora.” I have serious doubts that any of the proposals currently under discussion meet these criteria, but it cannot be ruled out a priori that some modified proposal could meet them.

146 (3) Summer 2017 105 Deliberative Citizens, (Non)Deliberative Politicians: A Rejoinder

André Bächtiger & Simon Beste

Abstract: Are citizens or politicians (more) capable of deliberation, and when should they be willing to do so? In this essay, we first show that both politicians and citizens have the capacity to deliberate when institutions are appropriate. Yet high-quality deliberation sometimes collides with democratic principles and ideals. Therefore, we employ a “need-oriented” perspective, asking when and where citizens and the political workings of democracy need high-quality deliberation and when and where this is less the case. On this account, we propose a number of institutional interventions and reforms that may help boost deliberation in ways that both exploit its unique epistemic and ethical potential while simultaneously making it compatible with democratic principles and ideals.

When political scientists and political analysts are asked whether there is potential for deliberation in our contemporary political systems, the answer is usually negative. The standard argument is that politicians do not want to deliberate and citizens are not able to do it. Some deliberative democrats have given this argument a slightly different spin, claiming that although we should not hold high hopes for deliberation in the power-riddled realm of electoral politics, citizens have a latent deliberative potential that appropriate institutions (especially deliberative mini-publics) can unleash. In this essay, we argue that both answers are wrong. ANDRÉ BÄCHTIGER is Chair of Empirical research shows that both politicians and Political Theory at the Depart- citizens have the capacity to deliberate when insti- ment of Social Sciences at the Uni- tutions are appropriate. Under optimal institution- versity of Stuttgart, Germany. al conditions, politicians can score relatively high on SIMON BESTE is a Ph.D. candi- measures of discourse quality derived from the ide- date at the University of Lucerne, als of deliberation as envisaged in Habermasian ra- Switzerland. tional discourse. A good fraction of citizens can also (*See endnotes for complete contributor approach these standards. Yet deliberation is not biographies.) the only goal or the only desirable means in politics.

106 Sometimes the institutions that further de- It also assumes that these deliberative vir- André liberation also undermine the democratic tues will, in practice, produce an array of Bächtiger & Simon goods of responsiveness and accountability. desirable outcomes, including epistemic Beste And sometimes the institutions that allow advancement, ethical goals (such as mu- citizens to deliberate at high-quality levels tual understanding and accommodating may not–and some argue should not–pro- diversity), and individual transformation. duce significant effects on policy outcomes. The “new” approach to deliberation that Taking these possible trade-offs into con- we propose takes a functional perspective, sideration, we propose a “functional” ap- emphasizing that the various forms that proach to deliberation that takes the goals deliberation can take should depend on of deliberation in specific contexts more se- the goals of that deliberation and the con- riously and allows for a more nuanced read- texts in which it takes place. For instance, ing of the empirical results. Such an ap- to reach deliberation’s epistemic goals, a proach does not see deliberation as a pana- high level of justification rationality may cea for the ills of democracy. Rather, it takes be a key procedural requirement, whereas a “need-oriented” perspective, asking when respect may play only a subordinate role. and where citizens and the political work- By contrast, if you want to achieve delib- ings of democracy most need high-quali- eration’s ethical goals, respectful interac- ty deliberation and when and where they tions likely play a larger role than rational need it less. Based on such a functional un- argumentation. derstanding of deliberation, we propose a The approach of new deliberation re- number of institutional interventions and sembles political scholar Michael Saward’s reforms that may help boost deliberation in “shape-shifting” approach to representa- ways that both exploit its unique epistem- tion.2 Rather than thinking that delibera- ic and ethical potential and simultaneously tors play one distinct deliberative role at a make it compatible with other democratic time, we should understand them as cre- goods and ideals. ative actors who make productive and flex- ible use of various forms of deliberation Before we take a stab at the deliberative depending on goals and context. potentials of politicians and citizens, we Finally, the new approach to delibera- first need to charter some conceptual ter- tion assumes that deliberation cannot and ritory. Drawing from a common metaphor should not play a major role in all stages of in institutional theory, we shall distinguish a democratic system.3 Although delibera- between “old” and “new” deliberation.1 tion may be critical for producing epistem- “Old” deliberation (frequently denoted ically sound policy or mutual agreement, as “classic” deliberation) incorporates the it may be counterproductive for achieving standards of rationality in argumentation, other democratic goods, such as respon- listening, reflection (weighing), respect, siveness, accountability, or consequenti- and “authenticity” in the sense that actors ality. As a result, new deliberation takes are oriented toward sincere understanding a need-oriented perspective on delibera- of others rather than toward strategic goal tion: rather than claiming that more delib- attainment. An underlying assumption in eration is always good, it analyzes contexts old deliberation is that the various deliber- and situations to determine where delib- ative ideals are fixed and work in tandem. eration is most needed and functional for This vision is “unitary” in that it assumes a particular democratic system. that all of the deliberative virtues will com- Applying the approaches of the old and plement one another in a cohesive whole. new to an analysis of deliberation among

146 (3) Summer 2017 107 Deliberative politicians and citizens leads to different system with low party discipline–we find Citizens, evaluations of its viability and functional- debates that resemble “ideal” deliberation (Non) Deliberative ity. We start with deliberation in the realm with highly reasoned, respectful, reflective, Politicians: of politics and then turn to deliberative po- and open-minded actors.7 A Rejoinder tentials among citizens. Even with this more nuanced and differentiated reading of deliberation’s po- That “politics” is by its nature not delib- tential in politics, a number of challenges erative is a common theme in the litera- persist. First, a deep-seated analytical chal- ture of political science. In a programmatic lenge claims that the very nature of politics article, Ian Shapiro holds: “Enough of de- is conflict rather than cooperation through liberation: Politics is about interests and deliberation. This “adversarial” reading of power.”4 According to Shapiro, a deliber- politics, which has been dominant in dem- ative reading of politics fails to consider ocratic thinking since the seventeenth cen- conflicting interests and powerful play- tury, makes any claim for deliberation in ers who have no incentives to deliberate, politics–even if supported by empirical but will pursue their goals with coercive data–a dubious affair.8 means. Deliberative democrats often have Second, compounding this analytical similar views about the possibilities for challenge, we lack any straightforward test good deliberation in politics. James Fish- that might differentiate clearly between ful- kin and Robert Luskin, for instance, have ly deliberative actions (oriented toward the argued that political elites tend to focus common understanding of common goals) on negotiation rather than deliberation, and strategic actions (oriented only toward so that their changes of position are the self-interested and conflicting goals). As- product of changing circumstances rather sume that we find a political actor who than the product of the better argument.5 scores high on all deliberative indicators: The criticisms may be partially misplaced. that is, provides extensive justifications for Much of the criticism against the possibili- positions and shows respect for other posi- ty of deliberation in politics is based on an tions and arguments. Although these indi- analysis of Anglo-American politics and cators may suggest deliberative action, we Westminster systems. It is easy to identi- cannot exclude the possibility that the ac- fy major deliberative failures in contempo- tor is engaging in sophisticated “rhetori- rary U.S. politics and in Westminster de- cal action,” intended to manipulate an au- mocracies, but different institutional set- dience. Put differently, until we can read ups–in combination with issue types and minds, we will never be able to “prove” partisan strategies–may bring about high- that actors were really motivated by a log- er levels of deliberative action in politics. ic of common understanding. Empirical findings from legislative deliber- Third, the real world of politics suggests ation underline that under appropriate in- that there is no unitary core of deliberation stitutional, contextual, and partisan condi- in representative politics. Consider a com- tions–namely, coalition settings, second parison of deliberative behavior under the chambers, secrecy, low party discipline, low public eye and behind closed doors: public issue polarization, and the strong presence debates in parliament increase justifica- of moderate parties–genuine deliberation tion rationality but decrease respect, while is possible in parliaments.6 If favorable in- nonpublic debates increase respect but de- stitutional and issue factors combine–that crease justification rationality. From the is, when a less-polarized issue is debated in perspective of old deliberation, which im- a nonpublic second chamber of a consensus plies the compatibility of all the elements

108 Dædalus, the Journal of the American Academy of Arts & Sciences of good deliberation, this juxtaposition the trustee model of government and in- André might cast doubt on the validity of the em- stead prefers strong government respon- Bächtiger 11 & Simon pirical findings. One might be forced to siveness. Beste assume that political behavior under the public eye was only deliberative rhetoric, Taken together, these challenges seem to in which the pressures of publicity force underscore the criticism of deliberation the actors to produce justifications that as even a useful ideal for politics. From a they do not sincerely believe. Because the new deliberative perspective, however, norms of public debate in Western democ- these challenges appear in a different light. racies generally value reasoned argument First, we need to clarify the goals of de- but not explicit respect for the political op- liberation in politics before we specify what ponent, strategic actors will use strategies qualities we want to see in it. Focusing on that mix justification and disrespect. legislatures, political scientists Gary Muc- Fourth, empirical research shows that ciaroni and Paul Quirk have made a first classic deliberation in politics is highly con- step in this direction.12 They claim that the text-bound, with the conditions of good de- key goal of parliamentary debate should be liberation (a less-polarized issue debate in a the substantive consideration of policy is- second chamber of a consensus system be- sues and the related informational quality hind closed doors) representing relatively of a debate, rather than the ethical goals of rare conditions. If this is the case, political deliberation. In other words, in this setting, deliberation would be so unlikely as to lose the justificatory component of deliberation its real world significance. A deliberative trumps respectful interactions. From this lens on politics might be misplaced when vantage point, even adversarial debating– we consider “normal conditions” of poli- which many deliberative democrats have tics: namely, strong partisan competition placed in contradiction with deliberation– and high issue polarization. might yield an epistemic function in that Fifth, the conditions for good parliamen- the audience is provided with robust rea- tary deliberation create a challenge for the sons for competing policy goals.13 two other democratic goods of responsive- Yet Mucciaroni and Quirk tend to over- ness and accountability. As political scien- stretch their argument. Because both the tist Gerry Mackie has noted: “It is worri- goals of legislative deliberation and their some that each of the discourse-improving institutional and issue contexts vary, a fo- institutions is also one that reduces ac- cus on informational quality as the sole goal countability of representatives to the citi- of parliamentary debate at the expense of zenry (it’s harder to know who to blame in ethical goals may be misplaced. In the con- a consensus coalition, in a presidential re- text of a nonparliamentary consensus sys- gime, and in a system of closed meetings, tem–or any other negotiation setting in and the political elite can collude against politics–the ethical dimensions of respect the population).”9 Institutionalizing more may play an important instrumental role deliberative politics in legislatures seems to in facilitating the negotiation process. Re- imply a return to “old parliamentarism,”10 spectful interactions are likely to bolster with a premodern “trustee” model of rep- cooperative attitudes among negotiation resentation in which politicians have loose partners. Moreover, in the context of mor- links to their constituents and can free- al and ethical issues (such as abortion) or ly change their minds on the basis of the highly divisive issues (such as conflict regu- better argument. Yet today, at least in the lation in divided societies), an exclusive fo- United States, a majority of citizens rejects cus on informational quality may be deeply

146 (3) Summer 2017 109 Deliberative misplaced. On such issues, it is hard to say liberation. Even in the ideal, therefore, and Citizens, that one of the principles under dispute is even when politicians are motivated to find (Non) Deliberative more correct than the other. To regulate good solutions and are open to good argu- Politicians: deep conflict in divided societies, it is some- ments, we should expect a mix of strategic A Rejoinder times necessary to concentrate on what the and deliberative behavior. The deliberative other side can accept rather than searching quality standard here should not be ratio- for the “truth.” nal discourse but “deliberative negotiation” Second, from the perspective of new de- in which actors justify their positions ex- liberation, it is not a deficiency if we can- tensively with respect, but are allowed to not fully distinguish between true deliber- “bargain,” constrained by fairness, by mak- ative and strategic action. It is neither re- ing promises while abstaining from threats alistic nor even desirable that politicians and strategic misrepresentation.17 Negotia- be oriented only toward a common under- tions that score low on justification and re- standing of the common good or comply spect but high on force, threats, and strate- with the full list of deliberative virtues in gic misrepresentation would count as less all venues of politics.14 Under the public deliberative or, if there were no delibera- eye, for instance, politicians contribute to tive elements, not deliberative at all. Over- democratic goods–responsiveness and ac- all, rather than searching for fully fledged countability–if they debate properly rather deliberative actors in politics, we should de- than deliberate in a way that reflects all the sire creative political actors who can engage deliberative virtues. The standard of quali- in deliberation when needed and where ty from a deliberative vantage here requires contextually possible and appropriate. In neither the deliberative virtue of reflection this situation, it becomes analytically less nor that of respect, but rather a high level necessary to draw strict dividing lines be- of justification rationality (or what political tween strategic and deliberatively authen- scientist Simone Chambers has called “ro- tic political actors. bust reasoning”).15 A low level of deliberative quality under conditions of publicity The idea of a “deliberative citizen” has involves what Chambers calls “plebiscitory been met with as much skepticism as has reasoning” in which “arguments . . . become “deliberative politics.” Drawing from his shallow, poorly reasoned, pandering, or ap- own cases of citizen participation in deliber- peal to the worst we have in common.”16 By ations, political scientist Shawn Rosenberg contrast, behind closed doors, where pres- points out that most “participants who at- sures of public opinion are reduced and oth- tend a deliberation do not, in fact, engage in er governing logics, deriving from the pos- the give and take of the discussion.” Rath- sibility of agreement, set in, we can expect er, they “offer simple, short, unelaborated other deliberative virtues–such as some statements of their views of an event.”18 open-mindedness as well as listening and Critics have also argued that classic deliber- respect–to flourish more fully. ation may be undemocratic because classic Good democratic representatives should forms of deliberation discriminate against pursue not only the common good, but also already disadvantaged persons (especially the interests of their constituents when they people with low socioeconomic status); dif- conflict with those of constituents in oth- ferent speaking styles with less classically er districts or parties. If we build such con- deliberative qualities often translate into a flict into our ideal of politics, we should not lack of influence; and group discussion– expect the representatives to ignore or su- the hallmark of any deliberative event– persede their constituents’ interests in de- often triggers nondeliberative group dy-

110 Dædalus, the Journal of the American Academy of Arts & Sciences namics, such as group polarization, reduc- highly skilled deliberators were no more André ing the normative value of any transforms, likely to affect the opinion changes in the Bächtiger & Simon such as opinion change, that the delibera- group than were the predeliberation opin- Beste tion may produce. ions of the less highly-skilled deliberators. As with politics, empirical research dis- Nor did the more highly-skilled delibera- plays a different picture when it focuses on tors impose their views on other partici- deliberative events that are well-structured pants. Nor, again, did these highly skilled to include supportive conditions such as in- deliberators stay stuck in their positions: formation provision, expert questioning, they showed an almost identical amount of and facilitator intervention. In their analy- opinion change as the lower-skilled delib- sis of a transnational Deliberative Poll (the erators. Finally, the evidence suggests that “Europolis”)–which represents a demand- opinion change in Europolis can be partly ing setting for citizen deliberation–politi- attributed to a systematic, justificatory, and cal scientist Marlène Gerber and colleagues argument-based component, and not to un- found that “the standards of classic deliber- desirable group dynamics such as group po- ation are far from being utopian standards larization: well-justified arguments seemed that only very few citizen deliberators can to affect opinion change.21 These findings achieve.”19 In this context, the number of are by no means unique. In analyzing well- participants who both provided a sophisti- structured deliberative events, several in- cated justification and engaged in respect- dependent research teams have reached ful listening is almost 30 percent. Compar- strikingly similar conclusions regarding the ing the Europolis proceedings with parlia- deliberative potential of citizens, the non- mentary debates, the former fare quite well: violation of democratic standards, and the although the Europolis discussion groups systematic and justificatory basis of opinion did not match the deliberative standards change.22 under the most ideal conditions in repre- Overall, these findings contradict pop- sentative politics and also had slightly low- ular assumptions regarding citizen capac- er scores on justification rationality than in ity and the inevitability of undemocratic the average parliamentary debate, respect deliberative structures. Just as in legisla- levels were significantly higher than in pol- tive politics, much seems to depend on itics under the usual political conditions of the institutional setup: if an institution strong partisan competition and issue po- is explicitly geared toward dialogue and larization.20 Not everything was perfect in deliberation, many well-known psycho- Europolis. Gerber’s team found that work- logical biases tend to be reduced or with- ing-class participants from Eastern as well er away. Political scientist Kimmo Grön- as Southern Europe were less apt to reach lund and colleagues, for example, varied most standards of high-quality deliberation discussion rules in an experiment on citi- (such as justification rationality, common zen deliberation on the future of the Swed- good orientation, and respectful listening) ish language in Finland. Their main find- than other participants, raising some con- ing was that discussion with a facilitator cerns about the democratic dimensions of and deliberative norms reversed tenden- deliberation among citizens with cultur- cies to group polarization, whereas “free” ally heterogeneous backgrounds. Howev- discussion without a facilitator and explic- er, the research team did not find any in- it deliberative norms–as is implemented dication that the different speaking styles in most psychological experiments–pro- and cultures had an impact on influence. duced the undesired polarization patterns That is, the predeliberation opinions of the described by Cass Sunstein and others.

146 (3) Summer 2017 111 Deliberative One outstanding question involves the light. A key mistake in our view is to take Citizens, policy impact of citizen deliberation, for a “totalizing view” of citizen deliberation (Non) Deliberative example, in randomly selected “mini-pub- and deliberative mini-publics and expect Politicians: lics.” Much depends on the sincerity of the that a single institution can achieve all of A Rejoinder intention to implement and the capacity deliberation’s goals at once.29 In our view, to implement of the authorizing entity. In this “unitary” vision should be replaced by Canada, political scientist Genevieve Fuji a differentiated vision that takes the delib- Johnson has found that the interests of the erative needs of different political contexts authorizing bodies play a dispositive role in into account. Such a differentiated vision the implementation of the results of citizen would specify different functions for de- deliberations.23 Using Deliberative Polls, liberative mini-publics in different parts Fishkin has documented real effects on pol- of any political system. icy when the authorizers intended such ef- First, we need to consider which political fects or were extremely open to them. Yet as systems can be well-served by mini-pub- John Dryzek has noted: “Direct influence lic input, and why. Political scientist Ar- on and in policy making is a hard test for chon Fung, for example, has pointed out mini-publics to pass. While examples exist how patronage systems in Latin America of influence and impact, they are outnum- have hollowed out their procedures of rep- bered by cases where a mini-public is estab- resentative democracy. When competi- lished but turns out to have little or no ef- tive elections do not advance the collec- fect on public decision-making.”24 tive goods that citizens want and need, A new study has also suggested that the then it is sensible to hand over policy-mak- more deliberative mini-publics are, the less ing activities to citizens who can produce likely they are to influence policy.25 This these goods more effectively.30 In another study documents that mini-publics with intriguing (and perplexing) example, au- low representativeness and low delibera- thoritarian regimes may find that deliberative quality are most likely to produce im- tive mini-publics yield distinct benefits for portant policy effects.26 Yet if deliberation elites. They generate information about so- is not tied to decision-making, it loses its ciety and policy, co-opt dissent and main- democratic character. As Mark Warren has tain social order, and enable leaders to de- put it: “Political processes that fail to en- flect responsibility onto the mini-public able this moment of constitution [that is, a processes and thus avoid blame.31 In cer- system’s capacity of making binding deci- tain townships in China, administrators sions] also disempower the people as a col- have commissioned well-structured de- lective agent and thus undermine the nor- liberative mini-publics and then imple- mative point of inclusion and collective will mented their recommendations. One could formation.”27 imagine such processes either making lat- Deliberative mini-publics also raise im- er electoral democracy more possible or, to portant questions of democratic legitimacy. the contrary, undermining citizen demand Cristina Lafont has argued that mini-publics for democracy. Similar demands among reach conclusions for reasons that most or- existing authorities for high-quality citi- dinary voters are not likely to fully appre- zen feedback also exist in Western democ- ciate, which, in turn, creates a fundamen- racies. Baden-Württemberg in Germany tal challenge for their legitimacy as policy- is a good example. The massive protests making tools.28 surrounding the “Stuttgart 21” project to From the perspective of new delibera- rebuild a train station in the central city tion, these problems appear in a different challenged the functioning of traditional

112 Dædalus, the Journal of the American Academy of Arts & Sciences representative politics and consequently trust-based function works in practice: the André triggered several democratic innovations more voters knew about the randomly se- Bächtiger & Simon in order to reduce the disconnect between lected British Columbia Citizen Assembly Beste representative politics and citizen views. and Irish Citizen Convention–such as their The Green-Left (now Green-Black) gov- recruitment mechanisms or their freedom ernment introduced and institutionalized from partisan instructions–the more likely forums for citizen participation and delib- they were to vote for the mini-public’s poli- eration, subsequently taking up the policy cy recommendation in the later citizen ref- recommendations of those forums. erendum. But not all political systems require the Deliberative mini-publics can also func- input of deliberative mini-publics. In the tion as schools of deliberation and democ- Swiss polity, extended direct-democrat- racy. In today’s fragmented and mediatized ic mechanisms create feedback from the societies, truly dialogical opportunities public. Politicians learn from both nega- have become rare for ordinary citizens. Yet tive and positive votes in referenda, even psychologists argue that, in the formation if the exact reasoning behind the voting of considered opinions, dialogue is much decisions is not always clear. Over time, more effective than simply listening to ar- this system has led to relatively good antic- guments.34 Deliberative mini-publics en- ipations of what the “median voter” may able ordinary citizens to enter into reasoned desire, rendering additional input from political dialogue on important questions. deliberative mini-publics less necessary. In the U.S. context, political scientist Law- Surely, more deliberative median voters rence Jacobs and colleagues have found that might decide differently in direct demo- those who regularly participate in struc- cratic voting–and perhaps in less popu- tured public discussions have a higher pro- list ways–compared with nondeliberative clivity to connect with elites, engage in civic median voters.32 But if the goal of deliber- voluntary activities, and participate in elec- ative mini-publics is just more feedback to toral politics.35 In short, even when deliber- politics, then a fully fledged direct-demo- ative events do not directly influence poli- cratic system like Switzerland might pro- cy, they may nonetheless produce a demo- vide a sufficient route to achieve this goal. cratic and deliberative “culture,” which–as Second, we need to consider the func- we shall detail below–may be essential for tions of mini-publics beyond direct pol- the renewal of our contemporary political icy uptake. One prominent example is a systems. “trust-based” function, which can inform citizens’ own later deliberations. Such func- We think that deliberation brings some- tions do not replace citizen input, as La- thing unique to democracy. It promotes font argues, but instead supplement and both epistemic advancement, through ar- inform it. The idea behind the trust-based gument and reasoning, and mutual under- function is that the (large) majority of non- standing and accommodation among di- deliberating citizens can trust the judg- verse actors, through respectful interac- ments of the (small) minority of deliber- tion. With these goals in mind, most past ating citizens because that small minori- research on deliberation has taken a strong ty, selected randomly for a deliberative reformist perspective. By contrast, the re- mini-public, does not have to follow par- cent systemic approach–which has had a tisan logics of electoral representation and significant effect on current thinking about can focus instead on common concerns.33 deliberation and deliberative democracy– Some empirical evidence indicates that this seems to have left behind these reformist

146 (3) Summer 2017 113 Deliberative goals. The systemic approach asks us to faction than competitive Westminster sys- Citizens, evaluate the deliberative system as a whole, tems, even in times of major political cri- (Non) Deliberative suggesting that systemic mechanisms may sis (as in the aftermath of the 2008 finan- Politicians: sometimes be at work, in which compo- cial crisis). A Rejoinder nents in a deliberative system may correct A second possible institutional reform for each other’s deliberative (and democrat- in politics might involve strengthening the ic) deficiencies.36 Although we agree that bodies that Dryzek has called the “Cham- deliberative “wrongs” can sometimes pro- bers of Reflection.”38 Empirical research duce deliberative and democratic “rights,” suggests that second chambers, such as the we think that such correcting mechanisms House of Lords in England and the second are increasingly hollowed out in contempo- chamber in Switzerland, inject a modicum rary times. Mediatization, for instance, sys- of deliberation and reflection into the po- tematically undermines the deliberative ca- litical process, even in times of increasing pacities of political elites, forcing them to political polarization.39 Elected second follow media logics and engage in “plebi- chambers have the advantage that they scitory reasoning.” Increasing party po- are accountable to their constituents. But larization, especially in the United States, a well-functioning political system will try has severely reduced the potential for mak- to balance the different needs of represen- ing respectful compromises. These devel- tative politics, including both strong re- opments make it even more important to sponsiveness and reflective deliberation. think of smart interventions and reforms One way to achieve the combination of to existing institutional settings so that the these two ends may be a clever division of unique contribution of deliberation to de- labor between different political venues: mocracy can be realized. the partisan and the reflective. If the two In the political sphere, one might imag- venues are institutionally nested (for in- ine institutional reforms toward more ne- stance, if one arena is constrained at least gotiation, similar to what we find in consen- in part by the decisions of the other), this sus democracies. Introducing proportional nesting may advance the realization of the representation (pr) electoral systems, for deliberative goods of epistemic advance- example, makes it easier for several parties ment and mutual understanding without to form, which thereby forces the parties to bypassing the other democratic goods of enter into coalitions and negotiations with responsiveness and accountability. other parties. This requirement for negoti- A third possible institutional reform is ation, in turn, can involve deliberative ele- simply enriching existing political systems ments. In recent years, negotiated systems with a plethora of democratic innovations. have sparked interest among political the- This supplementary approach is happening orists.37 Denmark provides an interesting worldwide, but had a particular grip in Lat- case. Here, an inclusive “negotiated” but in America and in polities deeply affected also “authoritative” political system pro- by the financial crisis in 2008, such as Ice- duces high-quality governance outputs; in- land, Ireland, and Spain. In all of these poli- triguingly (but perhaps not surprisingly), ties, various actors have started a number of Denmark is also one of the few countries deliberative initiatives, ranging from direct where the recommendations of one form of citizen involvement to new party architec- deliberative mini-publics (“consensus con- tures and new direct communication links ferences”) have found their way into legis- between representatives and the represent- lation. Overall, well-functioning negotiated ed.40 We think that the positive effects of systems involve much higher citizen satis- these innovations can be increased when

114 Dædalus, the Journal of the American Academy of Arts & Sciences institutional designers develop a higher Another challenge to institutional de- André sensitivity to context, asking in each setting sign derives from institutional interac- Bächtiger & Simon what goals they want to realize and what tions and details. First, institutions are fre- Beste conditions they confront. They should re- quently nested in other institutions. For in- alize that there is no one-size-fits-all in or- stance, the deliberative potential of second ganizing deliberative events. Depending on chambers is affected by the overall system the goals of deliberation (such as epistemic architecture: if the larger system is domi- advancement and accommodation), differ- nated by partisan advocacy–as in Austra- ent forms of recruitment and communica- lia and the United States–then the deliber- tion are necessary. For instance, if we want ative potential of second chambers is limit- to promote the interests of disadvantaged ed. Second, institutional details matter. In and disaffected groups, we should over- coalition government, for example, recent recruit members of such groups in order to research has shown that constructive po- enhance their public standing and adopt a litical action may be strongly influenced by cooperative communication format in or- the coalition composition. Political scien- der to enhance their deliberative influence. tists Stephen Fisher and Sara Hobolt have But if we want to achieve epistemic goals, provided empirical evidence that when a then random selection of participations coalition government is composed of two and more contestatory forms of engage- parties, the head of government’s party is ment might be more effective.41 subject to greater punishment and reward Overall, we should note that institution- from voters in their retrospective voting in al reforms do not always work in straight- the next election than the other coalition forward ways. pr electoral systems and party. Conversely, when a coalition govern- their resulting coalition arrangements ment has more than two parties, the effects may tend to enhance constructive negoti- of retrospective voting on any of the parties ations, but in some conditions (like in Is- is substantially reduced. Accountability rael) this institutional arrangement does seems to be reduced because it is harder to not conduce to cooperation. With regard know whom to blame. On the other hand, to coalition systems, rational choice the- the space for deliberative engagement may orists have long argued that coalition set- be enhanced, since it is more difficult for tings entail mixed-motive games. On the parties to use political successes for parti- one hand, coalition parties have reason to san electoral advantages. Details also mat- cooperate with their partners to pursue ter for deliberative mini-publics. Political successful common policies. On the oth- scientist Lucio Baccaro and colleagues have er hand, each party faces strong incentives shown that an institutional detail, such as to move policy in ways that appeal to their asking participants to justify their positions party members and to the constituencies before making a choice, can have major ef- on which the party relies for support.42 Co- fects on the dynamics of deliberative pro- alition arrangements thus do not automat- cess and the subsequent outcomes.44 ically produce high-quality deliberation. In short, we argue for a new contextual Rather, as empirical research has shown, and functional approach to deliberation. the deliberative capacity of coalition ar- Both electoral politics and citizen partici- rangements is strongly affected by partisan pation can become more deliberative, of- strategies and motivations.43 This power- ten without undermining other democratic ful effect of partisan variables underlines values. What is needed, what will work, that there is, at least at the moment, a clear and what innovations make sense depend limit to the effects of institutional design. on the details and the context.

146 (3) Summer 2017 115 Deliberative endnotes Citizens, * (Non) Contributor Biographies: ANDRÉ BÄCHTIGER is Chair of Political Theory at the Department of Deliberative Social Sciences at the University of Stuttgart, Germany. His research has been published by Cam- Politicians: bridge University Press and in such journals as the British Journal of Political Science, European Journal A Rejoinder of Political Research, Journal of Political Philosophy, Political Studies, and Acta Politica. He is coeditor of the Oxford Handbook of Deliberative Democracy (with John Dryzek, Jane Mansbridge, and Mark Warren, forthcoming). SIMON BESTE is a Ph.D. candidate at the University of Lucerne, Switzerland. He has published articles in such journals as Acta Politica, Journal of Public Deliberation, Swiss Political Science Review, and Journal of Legislative Studies. 1 Jane Mansbridge, James Bohman, Simone Chambers, et al., “The Place of Self-Interest and the Role of Power in Deliberative Theory,” Journal of Political Philosophy 18 (1) (2010): 64–100; and André Bächtiger and John Parkinson, Mapping and Measuring Deliberation: Micro and Macro Knowledge of Deliberative Quality, Dynamics and Contexts (Oxford: Oxford University Press, forthcoming). 2 Michael Saward, “Shape-Shifting Representation,” American Political Science Review 108 (4) (2014): 723–736. 3 Mark E. Warren, “When, Where and Why Do We Need Deliberation, Voting, and Other Means of Organizing Democracy? A Problem-Based Approach to Democratic Systems,” Amer- ican Political Science Review 111 (1) (2017): 39–53. 4 Ian Shapiro, “Enough of Deliberation: Politics is about Interests and Power,” in Deliberative Politics: Essays on Democracy and Disagreement, ed. Stephen Macedo (Oxford: Oxford University Press, 1999), 28–38. 5 James S. Fishkin and Robert Luskin, “Experimenting with a Democratic Ideal: Deliberative Polling and Public Opinion,” Acta Politica 40 (3) (2005): 284–298. 6 Jürg Steiner, André Bächtiger, Markus Spörndli, and Marco R. Steenbergen, Deliberative Poli- tics in Action: Analyzing Parliamentary Discourse (Cambridge: Cambridge University Press, 2004); André Bächtiger and Dominik Hangartner, “When Deliberative Theory Meets Political Sci- ence: Theoretical and Methodological Challenges in the Study of a Philosophical Ideal,” Polit- ical Studies 58 (2010): 609–629; and Kevin M. Esterling, “‘Deliberative Disagreement’ in U.S. Health Policy Committee Hearings,” Legislative Studies Quarterly 36 (2) (2011): 169–198. For instances of deliberation in the U.S. Senate in a less polarized era, see Joseph Bessette, The Mild Voice of Reason (Chicago: University of Chicago Press, 1994). 7 Steiner et al., Deliberative Politics in Action, 111 ff. 8 Jane Mansbridge, Beyond Adversary Democracy (Chicago: University of Chicago Press, 1980). 9 Gerry Mackie, “Deliberation, but Voting Too,” in Approaching Deliberative Democracy: Theory and Practice, ed. Robert Cavalier (Albany: State University of New York Press, 2012), 75–104. 10 Bernard Manin, The Principles of Representative Government (Cambridge: Cambridge University Press, 1997). 11 Lawrence R. Jacobs and Robert Y. Shapiro, Politicians Don’t Pander: Political Manipulation and the Loss of Democratic Responsiveness (Chicago: University of Chicago Press, 2000). 12 Gary Mucciaroni and Paul J. Quirk, Deliberative Choice: Debating Public Policy in Congress (Chicago: University of Chicago Press, 2006). 13 Bernard Manin, “Political Deliberation & the Adversarial Principle,” Dædalus 146 (3) (Sum- mer 2017). 14 Amy Gutmann and Dennis Thompson, The Spirit of Compromise: Why Governing Demands It and Campaigning Undermines It (Princeton, N.J.: Princeton University Press, 2014).

116 Dædalus, the Journal of the American Academy of Arts & Sciences 15 Simone Chambers, “Measuring Publicity’s Effect: Reconciling Empirical Research and Nor- André mative Theory,” Acta Politica 40 (2) (2005): 255–266. Bächtiger & Simon 16 Ibid., 257. Beste 17 Mark E. Warren and Jane Mansbridge, “Deliberative Negotiation,” in Negotiating Agreement in Politics: Report of the Task Force on Negotiating Agreement in Politics, ed. Jane Mansbridge and Cathie Joe Martin (Washington, D.C.: American Political Science Association, 2013), 86–120. 18 Shawn W. Rosenberg, “Citizen Competence and the Psychology of Deliberation,” in Delib- erative Democracy: Issues and Cases, ed. Stephen Elstub and Peter McLaverty (Edinburgh: Edin- burgh University Press, 2014), 98–117. 19 Marlène Gerber, André Bächtiger, Susumu Shikano, Simon Reber, and Samuel Rohr, “Delib- erative Abilities and Influence in a Transnational Deliberative Poll (EuroPolis),” British Jour- nal of Political Science (forthcoming). 20 See also Seraina Pedrini, “Deliberative Capacity in the Political and Civic Sphere,” Swiss Po- litical Science Review 20 (2) (2014): 263–286. 21 See also Marlène Gerber, André Bächtiger, Irena Fiket, Marco R. Steenbergen, and Jürg Steiner, “Deliberative and Non-Deliberative Persuasion: Opinion Change in a Pan-European Delib- erative Poll (Europolis),” European Union Politics 15 (2014): 410–429. 22 Alice Siu, Look Who’s Talking: Examining Social Influence, Opinion Change and Argument Quality in De- liberation (Ph.D. diss., Stanford University, 2009); Alice Siu, “Deliberation & the Challenge of Inequality,” Dædalus 146 (3) (Summer 2017); and Sean J. Westwood, “The Role of Persua- sion in Deliberative Opinion Change,” Political Communication 32 (4) (2015). 23 Genevieve Fuji Johnson, Democratic Illusion: Deliberative Democracy in Canadian Public Policy (To- ronto: University of Toronto Press, 2015). 24 John S. Dryzek, Foundations and Frontiers of Deliberative Governance (Oxford: Oxford University Press, 2010). 25 Graham Smith, Robert C. Richards, Jr., and John Gastil. “The Potential of Participedia as a Crowdsourcing Tool for Comparative Analysis of Democratic Innovations,” Policy and Internet (forthcoming). 26 Nicole Curato and Marit Böker, “Linking Mini-Publics to the Deliberative System: A Research Agenda,” Policy Sciences (forthcoming). 27 Warren, “When, Where and Why Do We Need Deliberation, Voting, and Other Means of Organizing Democracy?” 45. 28 Cristina Lafont, “Deliberation, Participation and Democratic Legitimacy: Should Delibera- tive Mini-Publics Shape Public Policy?” Journal of Political Philosophy 23 (1) (2015): 40–63. 29 Edana Beauvais and André Bächtiger, “Taking the Goals of Deliberation Seriously: A Differ- entiated View on Equality and Equity in Deliberative Designs and Processes,” Journal of Public Deliberation 12 (2) (2016). 30 Archon Fung, “Democratic Theory and Political Science: A Pragmatic Method of Construc- tive Engagement,” American Political Science Review 101 (3) (2007): 443–458 31 Baogang He and Mark E. Warren, “Authoritarian Deliberation: The Deliberative Turn in Chi- nese Political Development,” Perspectives on Politics 9 (2) (2011): 269–289. 32 André Bächtiger, Marco Steenbergen, Thomas Gautschi, and Seraina Pedrini, “Deliberation in Swiss Direct Democracy: A Field Experiment on the Expulsion Initiative,” NCCR Democracy Newsletter, February 2011. 33 Michael K. MacKenzie and Mark E. Warren, “Two Trust-Based Uses of Minipublics in Dem- ocratic Systems,” in Deliberative Systems: Deliberative Democracy at the Large Scale, ed. John Parkin- son and Jane Mansbridge (Cambridge: Cambridge University Press, 2012), 95–124.

146 (3) Summer 2017 117 Deliberative 34 Hugo Mercier, “The Argumentative Theory: Prediction and Empirical Evidence,” Trends in Citizens, Cognitive Science (forthcoming). (Non) Deliberative 35 Lawrence R. Jacobs, Fay Lomax Cook, and Michael X. Delli Carpini, Talking Together: Public Politicians: Deliberation and Political Participation in America (Chicago: University of Chicago Press, 2009). A Rejoinder 36 Jane Mansbridge, James Bohman, Simone Chambers, et al., “A Systemic Approach to Delib- eration,” in Deliberative Systems, ed. Parkinson and Mansbridge, 1–26. 37 Jane Mansbridge, “On the Importance of Getting Things Done,” Political Science & Politics 45 (1) (2012): 1–8; and Warren et al., “Deliberative Negotiation,” 86–120. 38 John Dryzek, “The Forum, the System, and the Polity: Three Varieties of Democratic Theory,” Political Theory (forthcoming). 39 Dominik Wyss, Simon Beste, and André Bächtiger, “A Decline in the Quality of Debate? The Evolution of Cognitive Complexity in Swiss Parliamentary Debates on Immigration (1968– 2014),” Swiss Political Science Review 21 (4) (2015): 636–653. 40 Michael Neblo, Kevin M. Esterling, and David M. J. Lazer, Directly Representative Democracy for a 21st Century Republic (Cambridge: Cambridge University Press, 2017). 41 Beauvais and Bächtiger, “Taking the Goals of Deliberation Seriously.” 42 Larry W. Martin and Georg Vanberg, “Coalition Government and Political Communication,” Political Research Quarterly 61 (3) (2008): 502–516. 43 Bächtiger and Hangartner, “When Deliberative Theory Meets Political Science.” 44 Lucio Baccaro, André Bächtiger, and Marion Deville, “Small Differences That Matter: The Impact of Discussion Modalities on Deliberative Outcomes,” British Journal of Political Science 46 (3) (2016): 551–566.

118 Dædalus, the Journal of the American Academy of Arts & Sciences Deliberation & the Challenge of Inequality

Alice Siu

Abstract: Deliberative critics contend that because societal inequalities cannot be bracketed in deliberative settings, the deliberative process inevitably perpetuates these inequalities. As a result, they argue, deliberation does not serve its theorized purposes, but rather produces distorted dialogue determined by inequalities, not merits. Advocates of deliberation must confront these criticisms: do less-privileged, less-educated, or perhaps illiterate participants stand a chance in discussions with the more privileged, better educated, and well spoken? Could their arguments ever be perceived or weighed equally? This essay presents empirical evidence to demonstrate that, in deliberations that are structured to provide a more level playing field, inequalities in skill and status do not translate into inequalities of influence.

When we think of the greatest orators, we often see men. In many developed democracies, those men are also likely to be white, educated, and privileged– men who had better opportunities from birth. It would be natural to expect these same privileged men to dominate in deliberation; indeed, we have all seen this kind of discursive domination in our own lives. Thus, many critics of deliberation have identified societal inequalities in deliberative settings, from town meetings to the jury room. Compared with the vote, which is explicitly structured to foster equality, deliberation seems destined to perpet- uate existing societal inequalities, and perhaps foster greater inequality. These criticisms apply to many ALICE SIU is Deputy Director of contexts; yet in deliberative settings structured to the Center for Deliberative Democ provide a more level playing field, we do not find racy at Stanford University. Her empirical evidence to support these claims. Since work on Deliberative Polling has the critics’ claims are empirical, it is necessary to ex- been published in The Search for De- amine them empirically. This essay provides empir- liberative Democracy in China (2006), ical evidence to demonstrate that inequality is not a Governance Reform under Real-World Conditions: Citizens, Stakeholders, and necessary attribute of deliberation. Voice (2008), and prominent jour nals such as the British Journal of Po- Deliberative theorists contend that forums for litical Science. public deliberation provide opportunities for citi-

119 Deliberation & zens to engage each other in thoughtful dis- competition in terms of them versus us, in the Challenge cussion; in such settings, they may share which we are good and honest and they are of Inequality competing views and, over time, develop untrustworthy and evil. The rampant use reasoned and considered opinions.1 John of such strategies in political campaigns Stuart Mill argued that taking part in pub- and politics in general has strengthened lic functions, such as small town offices or populist movements and created more di- jury duty, serves as a school of public spirit.2 visiveness in society. Offe’s remedies in- In the case of juries, people, privileged or clude strengthening the people’s voice not, would engage in deliberations togeth- through various forms of participation er to decide the fate of others.3 and improving the public’s will formation More recently, German sociologist and through deliberation. He argues that delib- philosopher Jürgen Habermas has envi- eration, through reason-giving, listening, sioned the larger public sphere as a space in and respecting, could bring forth positive which “private people com[e] together as effects, such as more informed opinion, a public.”4 That public sphere can serve as internal efficacy, and the ability to widen a place for “critical public debate,”5 where social inclusion. Offe also suggests using public opinion can form and to which all “randomness” in the composition of par- citizens could have access, with freedom to ticipants in deliberation to ensure diver- discuss and gather as desired. In his essay sity and inclusive representation. Delib- “Further Reflections on the Public Sphere,” eration, he argues, can offer a path to re- Habermas discusses how the public debate storing liberal democracy, especially if that could be carried through in a variety of to- deliberation can be institutionalized. day’s civil society organizations, including Many deliberative critics find the aims volunteer organizations, churches, and ac- and aspirations of deliberation too lofty. ademic institutions.6 Such organizations, They have argued that, given societal in- he argues, are capable of helping the pub- equalities, deliberation in practice does lic engage in the kind of debate and discus- not come close to reaching those ideals. sion that could produce reflective opinion The critical theorist Nancy Fraser has ar- formation. gued, for instance, that Habermas’s public Claus Offe has also argued that delib- sphere is incapable of “bracketing” societal eration could help rehabilitate liberal de- inequalities: that is, to neutralize inequali- mocracy. He has noted the abundance of ty in a deliberative setting.8 Even if people literature on the crisis of democracy and voluntarily agree to participate and delib- even “the end” of democracy.7 Defining erate “as if” they were equals, it is simply the basic elements of liberal democracy as not possible to impose deliberative equali- “stateness” (the state’s ability to maintain ty on a social base of inequality.9 The com- the allegiance of its population and to exe- mon societal behaviors of men and women cute central functions), rule of law, politi- play out in deliberative settings. Men tend cal competition, and accountability, he has to interrupt women, speak longer and more illuminated the inherent threats to these often than women, and ignore women in key elements. For example, although po- deliberations.10 Men are also more likely to litical competition allows for parties and be assertive, while women are more tenta- contenders to have a clear and legitimate tive and accommodating.11 Further, Fraser winner, the process of competition has is concerned that group deliberation often created an electoral need to portray candi- transforms what were individual opinions dates through their “personalities.” It has into one single group opinion.12 Societal in- also encouraged politicians to treat their equalities then not only unbalance the dis-

120 Dædalus, the Journal of the American Academy of Arts & Sciences cussion; they may also, through a dynamic The most common point of reference for Alice of interrupting and silencing, create inac- deliberation is the jury. Accordingly, many Siu curate impressions of group opinion. Fi- scholars have used juries (usually mock ju- nally, certain speech styles, characteristic ries) to examine group dynamics, behav- of the dominant, diminish the value of oth- iors, and decision-making mechanisms.19 er’s opinions.13 The inherent inequalities Correlational evidence from jury studies in our society cause deliberation to bene- has supported deliberative critics. Most of fit more dominant individuals and groups, the jury research on gender has found that while disadvantaging minority individuals men participate significantly more than and groups. women.20 Men are also more likely to ref- Political theorist Iris Marion Young has erence facts, dispute facts, and discuss or- similarly argued that “speech that is as- ganizational matters, whereas women are sertive and confrontational is here more more likely to discuss the consequences of valued than speech that is tentative, ex- verdicts. Significant differences also appear ploratory, or conciliatory. In most actu- between the more- and less-educated ju- al situations of discussion, this privileges rors. More-educated jurors not only par- male speaking styles over female.”14 More- ticipate more than less-educated jurors, but over, if Habermas were right and the es- also discuss legal and factual issues more of- sence of good deliberation were “the force ten. But discussing more factual issues does of the better argument” restricted to rea- not necessarily mean that the more-edu- son alone, then deliberation would dis- cated jurors bring forth more facts. Re- advantage people who typically use oth- searchers have found that the number of er forms of argumentation, such as telling facts shared by the more- and less-educat- stories and sharing experiences.15 Partici- ed jurors differed only marginally in favor pants who are less privileged may be par- of the more educated.21 ticularly likely to share their arguments in Research on juries has also shown a rela- the form of stories. Thus, even if partici- tionship between juror occupation and in- pants from varying socioeconomic classes come and the likelihood of being selected did deliberate together, there would be no as foreperson.22 Across the jury research, guarantee that they would actually listen the foreperson was almost always male and to and understand one another.16 usually was more educated, had a higher in- A further argument against deliberation come, and had a higher-status occupation.23 stresses the undue influence of those who Typically, income and occupation are high- dominate deliberations verbally.17 Because ly correlated with education, and the rela- there is limited time in any given delibera- tionship is no different in juries. Jurors with tion, the participants who use the most time higher-status occupations and higher in- and are most articulate in sharing their opin- come have been shown to participate in de- ions are likely the most successful in per- liberations more than their counterparts.24 suading their group members. The evidence Further, men with higher-status occupa- from jury studies shows both how the more tions and higher income participated more socially advantaged members speak more than women who possessed the same cri- and how the sheer quantity of remarks from teria.25 The evidence from jury studies sug- participants is correlated with influence on gests why deliberative critics fear the nega- fellow jurors.18 Given these findings, the re- tive consequences of introducing more cit- sults of deliberations should favor the more izen deliberation into democracy. privileged, contrary to the aims and aspira- Until recently, there has been little em- tions of deliberative theorists. pirical evidence on forms of citizen delib-

146 (3) Summer 2017 121 Deliberation & eration other than juries and mock juries. event. All small groups were recorded and the Challenge This essay presents one of the most com- transcribed for qualitative and quantita- of Inequality prehensive analyses of deliberation to date, tive analyses. All projects also had control using both quantitative and qualitative cor- groups that were surveyed at the same time relational evidence drawn from five na- as the deliberative groups but did not re- tionally representative Deliberative Poll- ceive the deliberative “treatment.”27 ing projects in the United States, consist- To directly address concerns regarding of four online Deliberative Polls and ing participation inequality in deliber- one face-to-face Deliberative Poll. Across ation, the analyses examined the num- these five projects, the study included 1,474 ber of words spoken, minutes used, and participants and ninety-nine small groups. statements made by each person in each Typically, empirical results, if available, de- small group discussion. The expectation rive from much smaller samples with few- from jury studies was that economically er participants. This study made possible and educationally privileged white males a more systematic and thorough analysis. would dominate the time in the deliber- Each of these five projects covered a dif- ations, leaving less time for others in the ferent discussion topic, including U.S. for- group to participate. In these Deliberative eign policy (online in 2002 and face-to- Polls, however, there were no statistically face in 2003), primary elections (online in significant differences between the partic- 2004), general elections (online in 2004), ipation levels of men and women: in to- and health care and education (online in tal minutes used, total statements made, 2005).26 The four online projects were con- or total words spoken. Income, age, and ducted through voice-only, like a typical race produced some statistically signifi- phone conference but with the addition cant differences, but not in a consistent of web-based software that enabled par- pattern of dominance. Those with high- ticipants to see their own and other group er incomes took more time in the discus- members’ icons on screen and queue up to sion and used more words, but spoke less speak. This software also allowed moder- frequently.28 Comparing participants over ators to mimic in-person moderation by and under fifty years old yielded statistical- interrupting speakers with pause/mute ly significant differences only on the topics buttons and moving people around in the of health care and education. Even here, al- queue. Instead of a full-day or weekend de- though those over fifty years old contribut- liberation, the discussions in the online ed significantly more statements and used Deliberative Polls were spread out over more time, they did not use more words a few weeks, on a weekly or biweekly ba- in the discussions. Like age, the results for sis, with each session taking approximate- race, coded as white and nonwhite, were ly 1.25 to 1.5 hours. Like the in-person de- mixed. In the discussion of the candidates liberations, the online sessions included in the presidential primaries, white partic- plenary sessions in which experts and/or ipants contributed statistically significant- policy-makers would answer questions de- ly more statements than nonwhite partici- veloped by participants in the small group pants, but did not use more words or time. discussions. Instead of a live plenary ses- In the discussion of issues relating to health sion, the plenary sessions were prerecord- care and education, however, nonwhite ed and then listened to together as a group. participants used more time, words, and In total, the small group and plenary ses- statements than the white participants. sion discussions took between five to six In short, in these five Deliberative Polls, hours, similar to a one-day face-to-face the more-privileged participants did not

122 Dædalus, the Journal of the American Academy of Arts & Sciences consistently dominate the deliberations. ileged participants. If, after deliberation, Alice Only one Deliberative Poll produced an in- participants in a small group moved closer Siu stance in which one group used more time, to the opinions of the more privileged, the words, and statements. In this instance, movement was coded as a 1; if further away, the demographic variable was race. But in a 0. Across all five Deliberative Polls, on av- this case, nonwhite participants contribut- erage, the proportion of small groups that ed more than white participants in the dis- moved closer to the opinions of the more cussions of both health care, with a specific privileged was about 0.500–a coin flip. focus on the rising costs of health care and For this analysis, “more privileged” universal health care, and education, with means being either male, more highly ed- a specific focus on educational standards, ucated (having an undergraduate four-year testing, No Child Left Behind, vouchers, college degree or beyond), with a higher in- and charter schools. In this deliberation, on come ($60,000 or higher annual income), average across the small groups, nonwhite or white. On average, across the five De- participants used twenty minutes each of liberative Polls, the proportion moving to- the allotted deliberation time, while white ward the average male opinion was 0.515; participants used thirteen minutes each. toward the average higher-educated opin- (This difference is statistically significant ion 0.542, toward the average higher-in- at the highest level: p=0.000. Each small come opinion 0.526, and toward the aver- group session allowed for between fifty- age white opinion 0.484. The range of opin- five to seventy minutes of discussion time.) ion change toward the more privileged in Nonwhite participants collectively used an these five Deliberative Polls was fairly nar- average of 2,587 words, whereas white par- row, with the exception of the movement ticipants used 1,742 words (p=0.002). Non- toward the higher educated. The range for white participants also contributed more the proportion of small groups moving to- statements than white participants: an av- ward the opinions of the more highly edu- erage of sixteen versus twelve statements, cated was 0.448 to 0.714, toward the opin- respectively. In short, the nonwhite partic- ions of males 0.438 to 0.558, toward the ipants made more individual statements, opinions of those with more income 0.479 using more words and time per statement. to 0.617, and toward the opinions of whites Although the participation measures 0.438 to 0.563. In short, we see no consis- show no consistent domination by the tent movement toward the more privileged more-privileged participants, that may be in these Deliberative Polls. because more-privileged participants have We may now ask whether increased lev- more experience in deliberative settings, els of participation themselves might in- and thus their contributions influenced fluence participation. That is, regardless of their group members’ opinions more ef- socioeconomic status, does simply partici- fectively, bringing opinions closer to their pating more in the deliberation cause group own. The analysis used to test for this pos- members to move toward your views? To sibility did so by quantifying the pre- and answer this question, the analyses applied post-deliberation opinions of participants “participation weights” to each participant, to see whether their opinions moved after based on the amount of time, number of deliberation and, if so, in what direction. words, and number of statements used. Us- The analyses examined the proportion of ing these participant weights, the regression participants in each small group that moved model used three explanatory variables: the either closer or further away (in a binary in- weighted opinion of each small group (with- dicator) from the opinions of more-priv- out the individual’s own opinion), each in-

146 (3) Summer 2017 123 Deliberation & dividual’s opinion, and each individual’s ments with one reason; and 3) statements the Challenge knowledge. The dependent variable was in- with two or more reasons. For example, of Inequality dividual level change in opinion. (The ex- the simple statement “I support free trade” planatory variables of the weighted opin- would be coded as a statement without rea- ion of the group and each individual’s opin- sons. The statement “free trade is harmful ion were measured before deliberation; the because it takes away jobs from the Unit- variable of each individual’s knowledge was ed States” would be coded as a statement measured after deliberation.) with one reason. The statement “the gov- If the standard social patterns revealed ernment should consider universal health in the jury studies held up, the regression care because millions of Americans are un- analyses should have yielded a statistical- insured and governments in other countries ly significant and positive association be- provide universal health care for their citi- tween the weighted group mean opinion zens” would be coded as giving two or more (group mean with participation weights ap- reasons. plied) and individual change in opinion. In It turns out that gender, income, and po- these five Deliberative Polls, however, none litical affiliation did not predict reasoned of the coefficients were statistically signif- arguments in these five small group delib- icant. Only two of the five coefficients were erations. The significant explanatory vari- in the hypothesized positive direction.29 In ables were predeliberation knowledge, ed- short, participants with higher participa- ucation, race, and age. But in the case of the tion scores did not influence outcomes any social variables–education, race, and age– more than participants with lower partici- the association did not fall in the predicted pation scores.30 direction. Predeliberation knowledge did The final piece of empirical evidence perform as predicted. As one might ex- from these data speaks to the possibility, pect, the small groups with more partici- suggested by some critics of deliberation, pants with higher predeliberation knowl- that those who are less privileged will be less edge scores produced the greatest num- likely to deliberate by giving explicit justifi- ber of reasoned arguments. These groups cations and reasons (“argument quality”). contributed more statements with two or Some earlier experiments interviewed ex- more reasons, more statements with one perts and nonexperts to determine the abil- reason, and fewer statements with no rea- ity of persons in these two groups to offer sons. Education, however, revealed a pat- and counter arguments. These experiments tern contrary to prediction. The more- found that most interviewees, both expert educated small groups offered fewer rea- and nonexpert, were capable of defend- sons in their arguments. That is, they con- ing their arguments.31 Other experiments, tributed fewer statements with two or however, have found that citizens who had more reasons, fewer statements with one previously engaged in political conversa- reason, and more statements with no rea- tion were more likely to offer reasoned ar- sons. Thus, although the more highly ed- guments in deliberation.32 ucated groups were more likely to con- Using the transcripts of these five Delib- tribute opinions to the deliberations, they erative Polls, the analysis examined how were not as likely to give the reasons for well participants defended their views ac- these opinions. Race, too, showed a pat- cording to whether or not they offered rea- tern contrary to expectations: the less priv- sons or justifications for their statements. ileged offered more reasons in the deliber- The amount of justification was categorized ations. The small groups with more white as 1) statements without reasons; 2) state- participants contributed significantly few-

124 Dædalus, the Journal of the American Academy of Arts & Sciences er statements with one reason and more social patterns that we might expect from Alice statements without any reasons. Groups inequalities in the world around us are, Siu with more whites also contributed fewer to some degree, negated. The more privi- statements with two or more reasons, but leged do not consistently dominate delib- this difference was not statistically signif- erations, nor are their opinions more in- icant. Finally, age also performed against fluential than their fellow group members. traditional privilege-based predictions, al- The participants who speak most frequent- though only one of these associations was ly and for the longest duration in the con- statistically significant. The younger small versations also have no greater influence groups (that is, having the greatest number over the rest of the group. The idea that of participants under fifty years old) were the more privileged will be more capable more likely to contribute statements with of engaging in the reasoned exchange of two or more reasons. They were not sig- justification is also, in this setting, not true. nificantly more likely to offer one reason What characteristics of these delibera- or no reasons. Summing up, although pre- tions made them differ from the predictions deliberation knowledge did perform as ex- of deliberation’s critics? The relevant char- pected, in these five Deliberative Polls, the acteristics are clear and replicable. These groups with more white, highly educated, deliberations began with a selection from and older participants did not provide as the citizenry that was closer to random than many reasons for their arguments as the most such endeavors, such as jury studies, groups with more nonwhite, less-educat- and far more reflective of the citizenry than ed, and younger participants. If offering the typical citizen forum, which attracts reasons is the quintessential Habermas self-selected persons, often activists or the ian characteristic of good deliberation, it extremely interested. This deliberative de- seems that the less traditionally privileged sign included sending the participants, be- groups in these Deliberative Polls acted in fore they arrived, informational materials the most Habermasian manner. that were balanced and agreed upon by both It is possible that those who believe their sides of the issue at stake. Because the delib- statements carry more weight because of erations were special occasions and the cit- their social position feel less need to offer izens involved were aware of being person- reasons. Those with less power, converse- ally selected for participation, they typically ly, might believe that they are required to read some of the materials they were sent, furnish more justifications for their views. thus becoming aware, more or less, of in- For the same reason, the number of words formation on the side of the issue that they uttered might not necessarily be a sign of might usually oppose. During the delibera- power. Often the most powerful in a meet- tive process, participants also engaged with ing speaks least. These possibilities are open experts on both sides, an experience that to future investigation. At the moment, we again exposed them not only to much more can say only that on the obvious and surface information than they usually would have dimensions–those that earlier studies of ju- access to, but particularly to information ries have measured and that have informed on the side opposing their predisposition. the conclusions of earlier critics of delibera- The small group deliberation took place in tion–the predicted patterns do not emerge. two stages, one of which had time dedicat- ed to preparing questions for the experts in The empirical evidence presented in this the plenary sessions. Since not all questions essay has demonstrated that, when citizen could be asked, small groups had to careful- deliberations are well structured, the many ly consider which questions would be most

146 (3) Summer 2017 125 Deliberation & useful in helping them form their own opin- This examination of the empirical evi- the Challenge ions. They indirectly learned about one an- dence from five projects–with 1,474 partic- of Inequality other’s views through this common task, ipants and ninety-nine small groups–is, to but the process was designed so that they my knowledge, the largest study of the in- were focused on developing their group’s ternal dynamics of deliberation ever under- questions and were not likely to confront taken. It measured class, race, gender, and one another on opposing views at this stage. previous knowledge, and traced the effects Finally, and perhaps most important, small of these variables on several forms of par- group discussion moderators were trained ticipation and influence. The conclusion– explicitly to solicit opinions from every- that in fairly easily replicable circumstanc- one in the group, encourage all participants es, the expected forms of privilege have lit- to speak, and maintain a tone of courtesy tle effect on participation and no effect on and nonconfrontation in pursuit of mutu- influence–ought to undermine the auto- al understanding. With these conditions in matic association of deliberation and in- place, the predicted dynamics of domina- equality. tion by the more privileged did not appear.

endnotes 1 Amy Gutmann and Dennis Thompson, Why Deliberative Democracy? (Princeton, N.J.: Princeton University Press, 1996); James Fishkin, Democracy and Deliberation: New Directions for Democratic Re- form (New Haven, Conn., and London: Yale University Press, 1991); James Fishkin, The Voice of the People: Public Opinion and Democracy (New Haven, Conn., and London: Yale University Press, 1995); and James Fishkin, When the People Speak: Deliberative Democracy and Public Consultation (Ox ford: Oxford University Press, 2009). 2 John Stuart Mill, Considerations on Representative Government (New York: Prometheus Books, 1991 [1861]), 79. 3 Ibid. 4 Jürgen Habermas, The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society, trans. Thomas Burger (Cambridge, Mass.: The mit Press, 1989), 27. 5 Ibid., 52. 6 Jürgen Habermas, “Further Reflections on the Public Sphere,” in Habermas and the Public Sphere, ed. Craig Calhoun (Cambridge, Mass.: The mit Press, 1996), 421–461. 7 Claus Offe, “Crisis and Innovation of Liberal Democracy: Can Deliberation Be Institution alised?” Czech Sociological Review 47 (3) (2011): 447–473. 8 Nancy Fraser, “Rethinking the Public Sphere: A Contribution to the Critique of Actually Ex isting Democracy,” Social Text 25/26 (1990): 55–80; and Nancy Fraser, “Transnationalizing the Public Sphere: On the Legitimacy and Efficacy of Public Opinion in a Postwestphalian World,” in Identities, Affiliations, and Allegiances, ed. Seyla Benhabib, Ian Shapiro, and Danilo Pe tranovic (Cambridge: Cambridge University Press, 2007), 45–66. 9 Fraser, “Rethinking the Public Sphere.” 10 Ibid. For more recent work on these behaviors, see Christopher Karpowitz and Tali Mendle berg, The Silent Sex: Gender, Deliberation, and Institutions (Princeton, N.J.: Princeton University Press, 2014). 11 Ibid. 12 Ibid.

126 Dædalus, the Journal of the American Academy of Arts & Sciences 13 Iris M. Young, “Communication and the Other: Beyond Deliberative Democracy,” in Democracy Alice and Difference, ed. Seyla Benhabib (Princeton, N.J.: Princeton University Press, 1996), 60–74. Siu 14 Ibid. 15 Ibid. 16 Ibid; and Iris M. Young, Inclusion and Democracy (New York: Oxford University Press, 2000). 17 Lynn M. Sanders, “Against Deliberation,” Political Theory 25 (3) (1997): 347–373. 18 Ibid. 19 Phoebe C. Ellsworth, “Are Twelve Heads Better Than One?” Law and Contemporary Problems 52 (1989): 205–224; Reid Hastie, Steven Penrod, and Nancy Pennington, Inside the Jury (Cam bridge, Mass.: Harvard University Press, 1983); Valerie P. Hans and Neil Vidmar, Judging the Jury (New York: Plenum Publishing Corporation, 1986); Rita M. James, “Status and Compe tence of Jurors,” The American Journal of Sociology 64 (6) (1959): 563–570; Charlan Nemeth, Jef fery Endicott, and Joel Wachtler, “From the ’50s to the ’70s: Women in Jury Deliberations,” Sociometry 39 (4) (1976): 293–304; and Fred L. Strodtbeck and Richard D. Mann, “Sex Role Differentiation in Jury Deliberations,” Sociometry 19 (1956): 3–11. 20 Hans and Vidmar, Judging the Jury. 21 Hastie, Penrod, and Pennington, Inside the Jury. 22 Fred L. Strodtbeck, Rita M. James, and Charles Hawkins, “Social Status in Jury Deliberations,” American Sociological Review 22 (6) (1957): 713–719. 23 Hastie, Penrod, and Pennington, Inside the Jury; and Hans and Vidmar, Judging the Jury. 24 Ibid. 25 Ibid. 26 In respective order: James Fishkin, “Realizing Deliberative Democracy: Virtual and Face to Face Possibilities,” Electronic Working Papers Series, Paper No. deg 007 (Vancouver: W. Maurice Young Centre for Applied Ethics at the University of British Columbia, 2005); ibid; Center for Deliberative Democracy, “pbs Deliberation Day Shows Americans’ Views on Iraq, Trade,” January 24, 2004, http://cdd.stanford.edu/2004/pbs-deliberation-day-shows-americans -views-on-iraq-trade/; Center for Deliberative Democracy, “pbs Deliberation Day on Nation al Security and the Global Economy,” October 16, 2004, http://cdd.stanford.edu/2004/pbs -deliberation-day-on-national-security-and-the-global-economy/; and Center for Deliberative Democracy, “By the People: A National Deliberation on Healthcare and Education,” September 6, 2005, http://cdd.stanford.edu/2005/by-the-people-a-national-deliberation-on-healthcare -and-education/. 27 The data sets for these analyses are available to researchers at the Center for Deliberative De mocracy, https://cdd.stanford.edu/. This research is part of a larger project that will compare online and face-to-face deliberations. In the analyses that follow, the groups reported on aver aged a composition of 52 percent men, 37 percent high income, and 18 percent nonwhite, with an average age of forty-eight years old. As a point of comparison, the 2004 National Election Study’s group was 47 percent male, 40 percent high income, and 28 percent nonwhite, with an average age of forty-seven years old. 28 Of the two Deliberative Polls that were analyzed, income was unavailable for one. Therefore, this result is based on only one Deliberative Poll. 29 A further note on the regression analyses: The analyses were conducted on two of the five De liberative Polls, because the transcripts for three of the Deliberative Polls did not have indi vidual identifiers, which made it too difficult to identify participants. Without identifiers, the regression analyses were not possible. And as mentioned, of the two Deliberative Polls ana lyzed, income was unavailable for one. Therefore, one Deliberative Poll had three explana tory variables for the regression analyses and another had two explanatory variables for the

146 (3) Summer 2017 127 Deliberation & analyses, making a total of five explanatory variables or five coefficients from the regression the Challenge analyses. of Inequality 30 The analyses were also conducted using unweighted variables. The results were similar, in dicating that participation levels are not a predictor of opinion changes. 31 Deanna Kuhn, The Skills of Argument (New York: Cambridge University Press, 1991). 32 Joseph N. Cappella, Vincent Price, and Lilach Nir, “Argument Repertoire as a Reliable and Valid Measure of Opinion Quality: Electronic Dialogue during Campaign 2000,” Political Com- munication 19 (1) (2002): 73–93.

128 Dædalus, the Journal of the American Academy of Arts & Sciences Deliberative Democracy in the Trenches

Cass R. Sunstein

Abstract: In the last decades, many political theorists have explored the idea of deliberative democracy. The basic claim is that well-functioning democracies combine accountability with a commitment to reflection, information acquisition, multiple perspectives, and reason-giving. Does that claim illuminate actual practices? Much of the time, the executive branch of the United States has combined both democracy and deliberation, not least because it has placed a high premium on reason-giving and the acquisition of necessary information. It has also contained a high degree of internal diversity, encouraging debate and disagreement, not least through the public comment process. These claims are illustrated with concrete, if somewhat stylized, discussions of how the executive branch often operates.

In the last decades, a large number of political theorists have explored the idea of deliberative democracy.1 The basic claim is that well-functioning democracies combine accountability with a commitment to reflection and reason-giving. They do not merely respond to popular pressures and majority senti- ment. They also try to “refine and enlarge the public view” through acquisition of relevant information, attention to multiple perspectives, and careful deliberation in the public sphere.2 Versions of this claim have been impressively elaborated by many people, including Joseph Bessette (who originally CASS R SUNSTEIN . , a Fellow of coined the term), Jürgen Habermas, Amartya Sen, the American Academy since 1992, is the Robert Walmsley University Jane Mansbridge, James Fishkin, and the team of 3 Professor at Harvard University. Amy Gutmann and Dennis Thompson. From 2009 to 2012, he was Ad- The idea of deliberative democracy might focus on ministrator of the White House the internal operation of government, with an em- Office of Information and Regu- phasis on how the legislative, executive, and judi- latory Affairs. He is the author of, ciary branches speak with one another. It could take most recently, #Republic: Divided more or less populist forms, focusing on deliberation Democracy in the Age of Social Media (2017), Conspiracy Theories and Oth- among citizens themselves, or between citizens and er Dangerous Ideas (2014), and Why public officials. And while citizen-centered concep- Nudge? The Politics of Libertarian Pa- tions focus on widespread participation, drawing on ternalism (2014). the idea of town meetings, we can also find concep-

129 Deliberative tions of deliberative democracy that em- would be lovely to think so). Moreover, a Democracy in phasize reason-giving by elected leaders. system of deliberative government need the Trenches My goal in this essay is to explore the op- not be democratic at all. It might be un- eration of deliberative democracy in the democratically deliberative. Such a gov- trenches–not as a set of abstract ideals, ernment could be run by a set of experts, but as concrete practices. My central ques- with different perspectives, who would tion is relatively narrow: can deliberative exchange information and ideas, without democracy be found within the executive paying much attention to the public. We branch of the U.S. government? In im- could also stress one or another compo- portant ways, I will suggest, it can be, or nent of the term. A deliberative democra- at least there have been periods in which it cy would emphasize the importance of re- has flourished.4 When it is working well– flection and reason-giving. A deliberative and it often is–the executive branch places democracy would stress the importance of a large premium not only on accountabil- popular control. Or the two values could ity, but also on the exchange of informa- be given equal weight (though it is not en- tion and reason-giving within the feder- tirely clear what that would mean).5 al government, between that government No one doubts that, in the United States, and states and localities, and between that the executive branch is accountable for its government and diverse citizens. This dis- decisions and subject to democratic con- cussion draws a great deal on my own ex- straints. The president is elected, and his perience from 2009 to 2012, when I served basic convictions and proposals are a large as administrator of the White House Of- part of what accounts for his position in fice of Information and Regulatory Affairs the White House. When a president wants (oira), and also from 2013 to 2014, when I something to be done, it is often because served as a member of the President’s Re- most or at least many people want it to be view Group on Intelligence and Commu- done, though this is not always the case. nications Technologies. It is important to Elections to one side, many of the presi- acknowledge that the executive branch dent’s decisions, and those of people who can take different forms, with a stron- work for him, are subject to intense public ger or weaker focus on deliberation, and scrutiny. Accountability looms especially I shall have something to say about vari- large in the period right after an election, ability over time as well. when the new administration is inclined to ask, “What were our campaign promises?” The notion of deliberative democracy The same kind of accountability also looms has two components, and we could easily large right before elections, including both imagine different emphases, or one with- presidential reelections and the midterms. out the other. A system of purely majori- White House staffers and members of a tarian democracy could require a high lev- president’s cabinet do not want to endan- el of accountability while placing little or ger the electoral prospects of their boss. Ex- no premium on deliberation. Call this non- ecutive branch officials are also reluctant to deliberative democracy. Perhaps account- undermine the campaign efforts of legisla- ability would be a sufficient safeguard for tors within the president’s own party, and what matters, whether it is welfare, liber- certainly do not want to risk losing one of ty, or some other value. At least if we think the houses of Congress.6 In either case, the that the views of majorities have strong executive branch will be subject to a con- epistemic credentials, a nondeliberative tinuing process of careful public scrutiny, democracy might work well (or at least it at least for its most important decisions.

130 Dædalus, the Journal of the American Academy of Arts & Sciences I want to emphasize here a different tional biases. They might be mired in ex- Cass R. point: within the executive branch, de- isting practices. They might be (and some- Sunstein liberation about policies has often been a times are) resistant to significant change. fact of daily life. When the system is work- They tend to be Burkeans, wedded to tradi- ing well (and I will leave that important tions, sometimes displaying an acute form qualification implicit for most of the re- of status quo bias. But they also have an im- maining discussion), it is the best place mense stock of knowledge. in government to see deliberative democ- With respect to deliberation, the central racy in practice. And critically, its discus- point is that these officials will also work sions are largely substantive and technical and exchange facts and views with numer- rather than political, at least if we under- ous people within the executive branch, at stand the term political to refer to attention least on the most significant questions. For to electoral considerations, to the views multiple issues, this process of interagen- of various interest groups, or to issues of cy collaboration is formalized and routin- fundraising and campaign finance. For the ized.9 With respect to both domestic and most part (though not always), such po- international affairs, deliberation typically litical considerations are entirely irrele- takes something like the following (high- vant, and the exchange of reasons about ly stylized) form: participants in an inter- different policies has often been the coin agency process, including representatives of the realm. There is intense focus on con- of various parts of government, work to- sequences: What would this policy do? gether on some issue, whether short-term What are the alternatives? Would they be (in need of resolution within, say, three worse, or better, in terms of their effects?7 weeks) or long-term (not requiring reso- Of course, in some cases, political con- lution for many months). straints matter. They then become part of Sometimes these discussions take a deliberative process, in the sense that po- months or more, and can have a high degree litical deliberation is the art of the possi- of intensity and animation. Diverse people, ble. If a proposal must be enacted by Con- with different knowledge and perspectives, gress, the executive branch will think hard are frequently involved. In one discussion, about what is most likely to receive con- there might be participants from the Na- gressional support. But there is a great deal tional Economic Council, the Council of that the executive branch can or must do Economic Advisers, the Office of Manage- on its own, and when this is so, the role ment and Budget, the Department of the of politics weakens and often evaporates. Treasury, the Office of the United States In my own experience, and much of the Trade Representative, the Department of time, substance often turns out to be all State, and the Department of Energy. The that matters.8 participants might be “policy” officials; The process of deliberation involves di- some might have been chosen by the presi- verse people with a great deal of knowledge. dent and confirmed by the Senate. But those Within any cabinet-level department, there officials will be staffed and, to some extent, are numerous experts who have been work- guided by people without any evident polit- ing on the relevant issues for many years ical affiliation; they are specialists and tech- and through multiple administrations. nocrats. Of course, it is also true that policy Most of them do not care at all about elec- officials, and not their staffs, are entitled to tions, politics, or interest groups; they are make the ultimate call. policy specialists, not political animals. To At the same time, the participants will be sure, they might well have their institu- have distinctive “equities,” understood as

146 (3) Summer 2017 131 Deliberative perspectives and inclinations that grow out energy, and the deputy director of the Office Democracy in of special concerns and roles. The Office of of Management and Budget (who might, as the Trenches Management and Budget, for example, will the highest-ranking White House official, be particularly concerned about budgetary run the meeting and have special authori- implications and might well be focused on ty over resolution of the issue). After that, the possibility of excessive costs to taxpay- the issue might be resolved, or it might be ers. Meanwhile the Office of the United elevated to a “principals’ committee,” con- States Trade Representative will focus on sisting of cabinet-level officials. the implications for international trade, If the issue is a very important one, the and the Environmental Protection Agency principals’ committee might be chaired by (epa) will typically be concerned with the one of the highest-ranking officials in the effects on clean air and water. Because of White House, such as the national security their own equities, participants are unlike- adviser or the chief of staff, who is often, ly to be silent if the issue raises serious con- next to the president, the most important cerns from the standpoint of their office. person within the executive branch of the As a result, a great deal of information is government, effectively in charge of the likely to be exchanged. A failure to include cabinet. The principals’ committee might someone with a relevant perspective, or an well be able to resolve the question; many utter disregard of what they have argued, issues are settled at this level. But if there is counts as a “process foul”: a violation of an internal division, or if the issue is very the internal morality of executive branch important, it is likely to go to the presi- operations. dent. The president sometimes resolves It is true that the public might not know issues on the basis of some kind of paper what perspectives have been represent- briefing, which ends by asking him for a ed in a process. Even more important, it decision. Or he might resolve an issue as would also be extravagant to say that “all” a result of, or in, a meeting, in which com- relevant perspectives have been includ- peting perspectives are explored in con- ed. (To avoid absurdity, a judgment about siderable detail. In some cases, a principal, the meaning of “all” would require nor- having been badly outnumbered in a prin- mative criteria, which might well be con- cipals’ committee, requests an individual tested.) Some perspectives will undoubt- meeting with the president, so as to ensure edly lack representation in any such pro- that he hears all relevant arguments and cess, and that might be a serious problem. ultimately makes the decision personally. The only point is that the range of views Stylized and brief though it is, this ac- is very wide, and the construction of the count should be sufficient to suggest that, executive branch is such that many com- within the executive branch, there has typ- peting perspectives, with diverse priorities ically been a great deal of deliberation, and and concerns, will be heard. it often involves people with diverse per- After the participants are done with their spectives and high levels of technical ex- own process of deliberation, they might be pertise. Everyone within the executive able to resolve the issue in a way that is es- branch works for the president, of course, sentially final. But if that issue is very im- but there is often a surprising level of het- portant, or if agreement proves impossible, erogeneity and disagreement that has to it might be “elevated” to some kind of “dep- be worked through, typically as a result uties’ committee,” consisting, for example, of substantive exchanges that place a high of the deputy secretary of state, the deputy premium on acquisition of relevant infor- secretary of defense, the deputy secretary of mation.10 On important environmental

132 Dædalus, the Journal of the American Academy of Arts & Sciences questions, for example, there might well ments being forged through substantive Cass R. be differences in the views of the Environ- arguments. And notably, for this decision, Sunstein mental Protection Agency, the Council on politics–understood as electoral consid- Environmental Quality, the National Eco- erations, the views of interest groups as nomic Council, the Council of Economic such, or possible press reactions–did not Advisers, and the Office of Management play the slightest role in determining the and Budget; some process has to be used working group’s substantive choices.12 to work out different perspectives and un- Here, then, was a practice of deliberative derlying disagreements. democracy. It was democratic because the To take an example with which I am ultimate decision was under and by appoin- familiar: In 2009–2010, an interagency tees of an elected official, the president. It working group produced a “social cost of was deliberative for the reasons I have giv- carbon,” meaning the economic cost of a en. And while the ultimate product has cer- ton of carbon emissions, suitable for use tainly been subject to reasonable dispute, in regulatory impact analyses.11 The group it seems fair to say that the effort was both included representatives of the Council of reasonable and highly professional. In this Economic Advisers, the Council on Envi- respect, the process was hardly unique.13 ronmental Quality, the Department of Ag- riculture, the Department of Commerce, With respect to the regulatory process, the Department of Energy, the Depart- the system of internal review takes a some- ment of Transportation, the Environmen- what different but also standard form, one tal Protection Agency, the National Eco- that has been worked out over several de- nomic Council, the Office of Energy and cades.14 For example, suppose that the epa Climate Change, the Office of Manage- wishes to issue a new regulation involving ment and Budget, the Office of Science and particulate matter. If the regulation is sub- Technology Policy, and the Department of mitted to the Office of Information and the Treasury. Members of this group, like Regulatory Affairs (and it almost certainly members of countless others, had differ- would have to be),15 it will be scrutinized by ent information and different equities. numerous offices within the Executive Of- The epa, for example, is a crucially im- fice of the President, including the Office portant participant in discussions of the of Management and Budget, the Domes- social cost of carbon, and it sees envi- tic Policy Council, the National Economic ronmental protection as its major equity, Council, the Council of Economic Advis- while the Office of Science and Technol- ers, the Office of the Vice President, and the ogy Policy also knows a great deal about Office of the Chief of Staff. If it has inter- the underlying science (but might have a national implications, it will be scrutinized somewhat different perspective from the as well by the Department of State, the Na- epa). The Department of Commerce seeks tional Security Council, and the Office of to promote commercial activity; the De- the United States Trade Representative. partment of Energy has a great deal of ex- Within the Executive Office of the Presi- pertise on the effects of carbon emissions; dent, the initial comments will likely come and the Council of Economic Advisers and from staff, not from high-level officials. The the National Economic Council have ex- principal focus will be intensely substan- pertise on the appropriate discount rate. tive rather than political, and at most stages, The efforts of the working group involved and often all of them, political consider- the aggregation of a great deal of scientific, ations–including reactions of interest economic, and legal expertise, with agree- groups, congressional committees, or the

146 (3) Summer 2017 133 Deliberative media–will not be raised at all except in- suggest that some of the policy choices Democracy in sofar as they suggest legitimate substantive are wrong. Perhaps the economic analy- the Trenches questions and concerns. (When they bear sis will be seen, by someone in the Coun- directly on the merits, those questions and cil of Economic Advisers, to contain a se- concerns can turn out to be quite impor- rious mistake. Perhaps the benefits or the tant.) If political considerations are raised, costs will appear to have been inflated. (As for example, it might be to make relevant administrator of the Office of Informa- communications and legislative affairs of- tion and Regulatory Affairs, my own po- fices aware of what is coming. The White sition was that, in the face of reasonable House Office of Legislative Affairs might disagreement, the views of the Council of have to manage congressional questions, Economic Advisers are presumed to be au- certainly on high-visibility matters, and it thoritative on technical economic issues, is important to ensure that it is prepared. just as the views of the Department of Jus- Draft rules, both proposed and final, are tice are presumed to be authoritative on certainly subject to scrutiny by other de- technical legal issues. Of course, there can partments within the executive branch, at be a back-and-forth on such issues.) Any least when they raise issues within the le- analysis of benefits and costs will likely be gal authorities or policy-making expertise seen and scrutinized by numerous people. of those departments. For example, if a reg- With respect to the law, here is a relevant ulation has implications for the energy sup- fact, based on my own experience, and cast- ply, it will be assessed by the Department ing light on the operation of deliberative de- of Energy, which will have information re- mocracy in the trenches: General counsels lating the risk of energy price increases or within agencies are usually excellent, but power outages. The Departments of Com- in at least some cases, their legal judgments merce and Treasury might be involved, es- are influenced by the substantive goals and pecially if the regulation raises economic hopes of their own cabinet secretaries. Law- issues. To the extent that there are labor im- yers in other parts of government–the De- plications, the Department of Labor will partment of Justice, the White House Coun- comment. If agriculture is affected, the sel’s Office, the Office of Management and Department of Agriculture will comment Budget General Counsel–often have great- as well. And for regulations with environ- er objectivity even if they have less special- mental implications, the Department of ized expertise. What is true of legal issues the Interior might also be involved. With- can be true of policy questions as well, in- in the agencies, it is important to see that cluding predictions of likely consequences the initial analysis is typically done by peo- (such as costs and benefits). ple with no political affiliation: they will Frequently, issues and concerns can be be civil servants, specialists in the issues at worked out at the staff level, as a result of hand. (Hence again we are speaking here of brief or extended substantive conversation. deliberative democracy.) oira staff will convene staff-level discus- With respect to regulations, this process sions, and most of the issues are indeed re- of internal scrutiny can be intense. Issues solved in that way, whether they involve of policy and law might receive detailed economics, policy, or law. But here as well, attention. Perhaps people will disagree. issues might be “elevated.” For example, an There might be legal objections from law- assistant secretary of one department might yers within the Department of the Trea- engage with the assistant secretary at the sury or the Department of Justice. Some- rule-making agency and with the oira’s one in the Department of Energy could deputy administrator to explore interagen-

134 Dædalus, the Journal of the American Academy of Arts & Sciences cy concerns. Sometimes the issue will be that comments are invited on a wide range Cass R. raised with the oira administrator him- of choices that have been made in a pro- Sunstein self. If (and this is quite rare) agreement posed rule, and also on alternatives to those is not possible at that level, further discus- choices. Agencies often think, and the oira sions will be required, with ultimate res- often urges them to think, that their own olution by a group of principals or (this is judgments are provisional and that the role very rare indeed) by the president person- of the comment process is to learn whether ally. If the attorney general has a clear view or not they are right. For that learning to oc- on the law, for example, that will ordinarily cur, the public must be asked to comment dispose of legal questions, just as the presi- on the provisional choices and on alterna- dent’s science adviser will have a great deal tives to them. It is not much of a stretch to of authority on issues of pure science. see the inspiration for this form of deliberative democracy in the work of economist The regulatory process is not only an in- Friedrich Hayek and, in particular, his em- ternal one: it involves citizens, not merely phasis on the dispersed nature of knowl- public officials. In this respect, the process edge in society.16 Of course, Hayek was not of deliberative democracy, in the trenches, a great fan of the modern regulatory state– has more than an indirect democratic ped- he liked markets, not regulators–but his igree. (Now we are speaking of delibera- work on widely dispersed information has tive democracy, at least in the sense that the helped spur the effort, in both Democratic public is both engaged and influential.) For and Republican administrations, to go out- regulations, public comment is usually inside of government to learn from others. volved, and it can make a large difference. To be sure, the public as a whole does not The Administrative Procedure Act calls for comment on proposed rules; only certain a process of “notice and comment” on pro- members do. In this light, it is reasonable posed rules, which means that agencies to ask whether the comment process re- must ordinarily submit those rules to the flects a harmful kind of skew. Here is one public and take comments on their con- concern: Comments often come from well- tents. This is central to the process of de- organized interest groups with resources liberative democracy; it enables a dialogue that can support a team of experts who are between citizens and public officials. willing and able to attack what the govern- Among law professors and political sci- ment seeks to do. If the issue involves the entists, public comments are often thought environment, those who own and man- to be irrelevant, a kind of outlet, display, age power plants may be in the best posi- or show, not much affecting what gov- tion to engage in advocacy, running num- ernment actually does. Nothing could be bers and making claims of policy and law further from the truth. Public comments that are self-serving and wrong, but likely sometimes make a large difference in the to make officials nervous. If the issue in- content of what emerges from the nation- volves regulation of the transportation sec- al government, whether the issue involves tor, the airline and railroad industries will climate change, health reform, occupation- predictably make a series of objections, and al safety, or homeland security. Time and because of the incentives and immense skill again, proposed rules are changed as a result of those who work on behalf of those in- of what government learns from citizens. dustries, the objections might end up con- The strong institutional inclination of the vincing those who work for the public. If Office of Information and Regulatory Af- so, we have a form of “capture,” not in the fairs is to make it entirely clear to the public simplest form, and not through anything

146 (3) Summer 2017 135 Deliberative that is easily described as corruption, but in ideals, but much of the time it is worthy of Democracy in epistemic terms: public officials learn from the idea of deliberative democracy. the Trenches those who speak, and those who speak are These points suggest strong reasons to re- likely to have both money and self-interest ject the view, offered energetically by some at stake. law professors, that courts should be less This concern can hardly be ruled out in willing to defer to executive action when the abstract. If it accurately depicts real- that action is not a product of the auton- ity, we have a cruel parody of the ideal of omous decision-making of the particular deliberative democracy. Moreover, no one agency involved, but of numerous officials should doubt that comments are most like- within the executive branch.21 Put to one ly to come from those with resources and side the fact that courts will not ordinarily organization. The extent of epistemic cap- know about the internal process of delib- ture, if any, will vary with context and time. eration and will not be able to sort out the But the concern should not be overstated.17 precise role of various officials. The much Often resources and organization come deeper problem is that this view has things from more than one side. Environmen- exactly backwards. If an agency is acting tal groups, for example, often have a great on its own, there might well be reason to deal to say, and their experts are also well- worry about myopia, mission orientation, trained; the same is true for civil rights or- and tunnel vision, potentially compromis- ganizations and labor unions. In addition, ing the ultimate judgment. If multiple of- and even more important, the government ficials are involved, there are of course no has strong “filters” by which to test the guarantees, but the risks are reduced by the plausibility and reasonableness of public safeguards provided by multiple perspec- comments. Self-serving claims about eco- tives. The case for judicial deference to ex- nomics, policy, and law are often easy to ecutive action is far stronger if the action is dismiss. It is true that the public comment supported and produced by numerous offi- process can suffer from a kind of epistem- cials, and not only by the rule-making agen- ic skew, but it is also true that as a result of cy. That process of support, and that kind of what is learned, outcomes are both more production, ensure more in the way of both democratic and more deliberative than they deliberation and democracy. would be otherwise. One reason for the great length of final The picture I have presented might well rules is that their preambles engage with be an idealized one. Not every executive comments, frequently in considerable de- branch, and not every issue in the execu- tail. And in many cases, public comments tive branch, is the same. Some people will help produce substantial changes. Some- be deeply skeptical about any picture of times agencies learn that their proposals the Obama administration as embodying need to be withdrawn.18 Sometimes they an appealing form of deliberative democ- learn that a fundamentally different ap- racy. Others will point to their own pre- proach, saving costs, is best.19 Sometimes ferred examples, real or imagined, in which they learn that a more expansive approach, it seems misleading, incomplete, or worse increasing benefits, is justified.20 A great to depict the executive as an embodiment deal of deliberation thus occurs between of deliberative democracy. And whatever public officials and citizens, not only as a re- we think of the Obama administration, it sult of meetings, but perhaps most funda- is easy to find or to envision other adminis- mentally through the process of public com- trations, past, present, or future, that draw ment. It might not live up to the very highest that characterization into serious doubt.

136 Dædalus, the Journal of the American Academy of Arts & Sciences To put the point most vividly: imagine process of deliberation. If such officials Cass R. your least favorite presidential candidate are enthusiastic about renewable fuels and Sunstein of the past few elections; now imagine he want to maximize their use, it is useless to is president. Perhaps he has terrible but emphasize that executive officials are lis- fixed convictions and is unwilling to listen tening to one another and to the public. to reason; perhaps he does not much care This essay was originally written long about the facts; perhaps he is indifferent before the 2016 presidential election, and to public comments; or perhaps he is un- at least in the early months of the admin- duly influenced by well-organized interest istration of President Donald J. Trump, groups. During a Republican administra- many people believed that the executive tion many years ago, I emailed some ten- branch was not working in a highly delib- tative suggestions about how to deal with erative fashion. Critical observers think climate change to a high-level public offi- that the Trump administration is some- cial, who was (and is) a committed conser- times or often bypassing the time-honored vative and was (and is) also a good friend. processes sketched here, in which diverse My suggested approach did not involve the people explore, in great detail, policy op- imposition of high costs. I was puzzled to tions and the substantive arguments for receive no answer (though after my own ex- and against them. Whether or not that is perience in government, learning about the so, it must be acknowledged that the ar- potentially acute risks of using email, my guments I am making here depend on a puzzlement dissipated). When I next saw picture of the executive branch that may him at the White House, he came right up to not always be accurate. I believe that it is me and said: “Cass, you have absolutely no indeed accurate under most Democratic idea how conservative my colleagues are!” and Republican presidents, certainly out- Note, however, that even if the president side of the context of the most politicized were your least favorite candidate, many de- questions (and frequently enough, in that cisions of the executive branch would not context as well). But under any president, be likely to be affected. They would involve the influence of interest groups cannot be relatively routine (even if important) deci- discounted, and the risk of politicized de- sions, and they would be settled by some- cision-making or excessive domination by thing like the process I have described here. antecedent convictions is well above zero. But it must be acknowledged that with such a president, or anyone with the characteris- The theoretical literature on deliberative tics described above, deliberative democra- democracy has made significant contribu- cy would work far less well in the trenches, tions not only to theory but to practice as at least on high-profile questions, on which well. My main goal here has been to make relevant interest groups are able to exert some progress in understanding delibera- their influence, or on which the anteced- tive democracy in the trenches, in part by ent convictions of the president, and of his offering an account of practices that I wish, high-level advisers, are fixed and firm. in retrospect, I had known before starting If such officials believe that climate to work in the federal government. change is a myth, a technical process on The major lesson is that, much of the the social cost of carbon is unlikely to go time, the executive branch itself com- well. If such officials favor stringent regu- bines both democracy and deliberation, lation of ozone, mercury, and particulate and places a high premium on reason-giv- matter–whatever the facts show–deci- ing and the acquisition of information. sions will not reflect a well-functioning Perhaps surprisingly, it often contains a

146 (3) Summer 2017 137 Deliberative high degree of internal diversity, encour- But no one should doubt that it is possible Democracy in aging debate and disagreement, not least to operate national institutions in a way the Trenches through the public comment process. that insists not only on accountability, but I have not argued that this basic picture also on careful considerations of the hu- holds in all times and places. Some execu- man consequences of potential courses of tive branches are different from others, and action. When the executive branch is work- even within a year or a month, internal pro- ing well, that possibility is not fanciful; cesses can be different from one another. it is a lived reality.

endnotes 1 See, for example, Jon Elster, ed., Deliberative Democracy (Cambridge: Cambridge University Press, 1998); Jürgen Habermas, Between Facts and Norms (Cambridge, Mass.: The mit Press, 1998); and Amy Gutmann and Dennis Thompson, Why Deliberative Democracy (Princeton, N.J.: Princeton University Press, 2004). 2 See James Madison, “The Same Subject Continued: The Union as a Safeguard against Domestic Faction and Insurrection,” The Federalist Papers No. 10, November 23, 1787, accessed at https:// www.congress.gov/resources/display/content/The+Federalist+Papers#TheFederalistPapers-10. 3 See Joseph Bessette, The Mild Voice of Reason (Chicago: The University of Chicago Press, 1994); Habermas, Between Facts and Norms; Amartya Sen, Development As Freedom (Oxford: Oxford Uni- versity Press, 1999); Jane Mansbridge, Beyond Adversary Democracy (Chicago: University of Chi- cago Press, 1980); James Fishkin, Democracy and Deliberation (New Haven, Conn.: Yale Univer- sity Press, 1993); and Gutmann and Thompson, Why Deliberative Democracy. 4 A great deal of valuable information to this effect can be found in Michael Morrell, The Great War of Our Time: The CIA’s Fight Against Terrorism from Al Qa’ida to ISIS (New York: Twelve, 2015). 5 This is the tendency in Fishkin, Democracy and Deliberation. 6 A small story: When I was in the government, I helped write a “checklist” for regulatory impact analyses to promote accountability and make it easier to produce such analyses. (See Office of Information and Regulatory Affairs, “Agency Checklist: Regulatory Impact Anal- ysis,” October 28, 2010, https://obamawhitehouse.archives.gov/sites/default/files/omb/ inforeg/regpol/RIA_Checklist.pdf.) Within the Executive Office of the President, some people, especially those with “Hill experience,” were actually concerned: Perhaps the checklist could have adverse political effects? Perhaps it should not be released before the midterms? These questions seemed to me absolutely absurd, and eventually the checklist was released, midterms or no midterms. Within an hour after the Democrats lost the House in 2010, in a crushing defeat, a cabinet official emailed me: “It was the checklist.” (I wrote back immediately: “That saved the Senate.”) 7 See ibid. 8 A little story from my own experience: On a difficult regulatory issue, a very sharp and highly influential member of the White House communications staff weighed in heavily on the substance. In my view, that person’s conclusions were, in this particular case, entirely without merit; they had a political motivation and lacked evidentiary support. I communicated my concerns (less than politely, I fear) to the relevant person. The response was as unexpected as it was priceless, roughly: “I am just a political hack. Ignore me! Do the right thing.” 9 See Morrell, The Great War of Our Time; and Cass R. Sunstein, Valuing Life: Humanizing the Regu- latory State (Chicago: The University of Chicago Press, 2014). 10 I am bracketing questions raised by the independent regulatory commissions, such as the Federal Reserve Board and the Federal Communications Commission.

138 Dædalus, the Journal of the American Academy of Arts & Sciences 11 See Interagency Working Group on Social Cost of Carbon, U.S. Government, “Technical Sup- Cass R. port Document: Social Cost of Carbon for Regulatory Impact Analysis Under Executive Or- Sunstein der 12866,” February 2010, https://www.epa.gov/sites/production/files/2016-12/documents/ scc_tsd_2010.pdf. See also the discussion in Cass R. Sunstein, “On Not Revisiting Official Dis- count Rates: Institutional Inertia and the Social Cost of Carbon,” American Economic Review 104 (5) (2014). 12 This is the conclusion of the U.S. Government General Accountability Office, Regulatory Impact Analysis: Development of Social Cost of Carbon Estimates (Washington, D.C.: U.S. Government Ac- countability Office, July 2014), http://www.gao.gov/assets/670/665016.pdf. I can attest to the correctness of that conclusion. At various stages, of course, public comments were also considered on the merits. See, for example, Interagency Working Group on Social Cost of Car- bon, U.S. Government, “Response to Comments: Social Cost of Carbon for Regulatory Im- pact Analysis Under Executive Order 12866,” July 2015, https://obamawhitehouse.archives .gov/sites/default/files/omb/inforeg/scc-response-to-comments-final-july-2015.pdf. 13 For similar examples, see The Office of Management and Budget, “Circulara -4,” September 17, 2003, https://www.transportation.gov/sites/dot.gov/files/docs/OMB%20Circular%20 No.%20A-4.pdf. The analysis of the value of a statistical life is available at U.S. Department of Transportation, Office of the Secretary of Transportation, “Revised Departmental Guid- ance 2014: Treatment of the Value of Preventing Fatalities and Injuries in Preparing Econom- ic Analysis,” memorandum, June 17, 2015, https://www.transportation.gov/sites/dot.gov/ files/docs/VSL2015_0.pdf. 14 For a detailed account, see Cass R. Sunstein, “The Office of Information and Regulatory Af- fairs: Myths and Realities,” Harvard Law Review (126) (2013). 15 See Executive Order 12866, incorporated in The White House, Office of the Press Secretary, “Ex- ecutive Order 13563–Improving Regulation and Regulatory Review,” January 18, 2011, https:// obamawhitehouse.archives.gov/the-press-office/2011/01/18/executive-order-13563-improving -regulation-and-regulatory-review. 16 See Friedrich Hayek, “The Use of Knowledge in Society,” American Economic Review 4 (1945). 17 See also Daniel Naurin, Deliberation behind Closed Doors: Transparency and Lobbying in the European Union (Colchester, United Kingdom: ecpr Press, 2007). Naurin found that, in the European Union, business lobbyists who acted under closed-door conditions in their relationships with the European Commission had realized “that in order to promote their interests they have to argue carefully with reference to public interests and ideals rather than bargain from self- interest” and that “the industry lobbyists studied here sounded better, with respect to self- interest, behind closed doors than in public” due to constituency pressure toward a self-interested stance in public settings. 18 See, for example, U.S. Department of Labor, “U.S. Department of Labor’s osha Withdraws Pro- posed Interpretation on Occupational Noise,” osha National News Release, January 19, 2011, https://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=NEWS_RELEASES� &p_id=19119. 19 See, for example, Environmental Protection Agency, Effluent Limitations Guidelines and New Source Performance Standards for the Airport Deicing Category, Federal Register vol. 77, no. 95, May 16, 2012, http://www.gpo.gov/fdsys/pkg/FR-2012-05-16/pdf/2012-10633.pdf. 20 See, for example, U.S. Food & Drug Administration, “fda Finalizes Menu and Vending Ma- chine Calorie Labeling Rules,” fda news release, November 25, 2014, http://www.fda.gov/ NewsEvents/Newsroom/PressAnnouncements/ucm423952.htm (last archived February 28, 2017). 21 See Daniel A. Farber and Anne Joseph O’Connell, “The Lost World of Administrative Law,” Texas Law Review 92 (2014).

146 (3) Summer 2017 139 Applying Deliberative Democracy in Africa: Uganda’s First Deliberative Polls

James S. Fishkin, Roy William Mayega, Lynn Atuyambe, Nathan Tumuhamye, Julius Ssentongo, Alice Siu & William Bazeyo

Abstract: Practical experiments with deliberative democracy, instituted with random samples of the public, have had success in many countries. But this approach has never before been tried in Sub-Saharan Africa. Reflecting on the first two applications in Uganda, we apply the same criteria for success commonly used for such projects in the most advanced countries. Can this approach work successfully with samples of a public low in literacy and education? Can it work on some of the critical policy choices faced by the public in rural Uganda? This essay reflects on quantitative and qualitative results from Uganda’s first Delibera- tive Polls. We find that the projects were representative in both attitudes and demographics. They produced substantial opinion change supported by identifiable reasons. They avoided distortions from inequality and polarization. They produced actionable results that can be expected to influence policy on difficult choices.

The last two decades have seen a great rise in interest in deliberative democracy, in both theory and At the Center for Deliberative De 1 mocracy at Stanford University, practice. In political theory, this “deliberative turn” JAMES S. FISHKIN is Director and has largely supplanted the previous enthusiasm for ALICE SIU is Deputy Director. “participatory democracy,” a change sometimes decried by advocates of the latter. Participatory de- At Makerere University, Uganda, ROY WILLIAM MAYEGA is a Lectu mocracy generally relies on self-selected mass par- rer, LYNN ATUYAMBE is Associate ticipation. In development contexts, an iconic form Professor, NATHAN TUMUHAMYE is the “participatory budgeting” practiced in Porto is Director of the Eastern Africa Alegre, Brazil.2 By contrast, the form of deliberative Resilience Innovation Lab, JULIUS democracy that we will discuss here emphasizes de- SSENTONGO is a Research Fellow, signs that promote both the representativeness and and WILLIAM BAZEYO is Dean the thoughtfulness of public participation. Instead of the School of Public Health. of mobilizing as many people as possible, the idea (*See endnotes for complete contributor is to foster thoughtful weighing of the arguments biographies.) for and against policy alternatives by representa-

140 tive microcosms of the public.3 The num- cy, through randomly selected microcosms, Fishkin, bers who participate may be smaller than to produce public input for policy-making Mayega, Atuyambe, in mass participatory institutions, but the in Africa. Tumuhamye, conclusions offered can represent the pub- Ssentongo, Siu & lic’s considered judgments. Billions of people around the world live Bazeyo In practice, this kind of deliberative de- in poverty and deprivation. Development mocracy has found applications in various efforts to assist them increasingly invoke parts of the world with designs that foster the idea that the people should be consult- public input for actual policy-making. The ed. Those who might be affected by poli- designs vary, but they generally attempt to cies should in some way have a voice about facilitate the discussion of competing rea- them. Some argue that policies the public sons for policy alternatives in a context in can accept will be more effective. Others which members of the public can become argue that long-term development will be more informed about the issues in question. more sustainable in open and inclusive so- The more rigorous versions carefully select cieties in which people participate.8 the participants by recruiting a microcosm But how is this to be accomplished? or “mini-public” of the relevant popula- There are various approaches. Some efforts tion through random sampling.4 The ba- engage stakeholders or policy experts who sic idea is that if the sample is representa- speak on behalf of the people. Some take tive and the participants deliberate under decisions to the people themselves in self- good conditions for considering the issues, selected forums or meetings. Sometimes re- then the results should represent what the searchers employ focus groups and key in- public would think were it to engage with formant interviews to get voices from the the issues under similarly good conditions. people and from those who might have rel- This strategy makes deliberative democra- evant local knowledge. cy a practical and implementable theory, at However, stakeholders or policy experts least for the policy issues selected. may turn out to have different views from How widely can this approach be ap- those of the people themselves. Self-selected plied? There have been successful cases in forums are inevitably unrepresentative and postconflict situations and instances of eth- usually dominated by those especially mo- nic division.5 There have also been success- tivated to turn out. Further, self-selected fo- ful cases across multiple linguistic barriers, rums to discuss the distribution of benefits as when a Europe-wide sample deliberated are likely to foster mobilization for the benin Brussels in twenty-two languages with efits, rather than deliberation about the gen- simultaneous interpretation.6 There have eral good of the community. For example, even been successful cases in authoritari- in the famous “participatory budgeting” in an systems lacking electoral competition.7 Porto Alegre, a practice now spread around But never before have there been applica- the world, self-selected groups mobilize tions of deliberative democracy with ran- for specific benefits but the broader popu- dom samples of the public in Africa, where lation is not well represented.9 The question populations with low literacy and low levels we explore here is whether the move from of education often face extraordinary pol- participatory to deliberative democracy, a icy challenges. Is it applicable in such con- move made prominent in democratic theo- texts? Or is deliberative democracy just an ry,10 can be retraced in the practice of public approach for advanced countries with high- consultation in developing countries. More ly educated populations? We reflect here on specifically, is it practical to consult popula- a pilot effort to apply deliberative democra- tions in developing countries through delib-

146 (3) Summer 2017 141 Applying erating microcosms selected through ran- cy, rather than having free rein to promote Deliberative dom sampling? their own views, might the efforts be more Democracy in Africa: There have been a few efforts to incor- successful? Until now, the basic idea of de- Uganda’s First porate elements of deliberative democ- liberating microcosms chosen by random Deliberative racy in public consultations in Africa. In sampling had not been tested in Africa. We Polls Benin, for example, politics scholar Leon- report on such an effort here. ard Wantchekon has reported on a field experiment looking at how the discussion of While the idea of deliberative democ- different kinds of platforms (clientelist or racy has acquired many enthusiasts over public-policy related) just before an elec- the last two decades, it has also attracted tion affected voting.11 In Príncipe and São criticisms. Some of those criticisms might Tomé, political scientist Macartan Hum- plausibly define barriers to applying the ap- phreys and colleagues have found in a de- proach in developing countries. Consider liberative democracy experiment “robust three. First, deliberative democracy is of- evidence that the influence of leaders on ten criticized as an elite form of democratic the outcome of deliberation is extremely practice. Even the term was coined in a dis- strong, with leadership effects accounting cussion of James Madison’s theory of rep- for a large share of the variation in views resentatives who would “refine and enlarge elicited across the country.”12 Their find- the public views by passing them through ing that discussion leaders effectively de- the medium of a chosen body of citizens.”14 termine the outcome would obviously un- The refinement derived both from the pro- dermine the aspiration for using genuine cess of weighing arguments on the merits deliberation by the public to influence pub- in the legislative assembly and from the se- lic policy. lection process for selecting the “best” rep- However, neither of these studies in- resentatives. The term in its first coinage volved random samples of the public. Rath- thus applied to deliberations among the er, they involved random assignment of highly educated and supposedly especial- self-selected participants to different treat- ly virtuous, who would choose in the in- ments. And the project that seems to have terests of the public good for the rest of us. been a self-conscious application of delib- In the last two decades, the term has been erative democratic theory (the São Tomé adapted, at least for practical applications, and Príncipe experiments) apparently gave to deliberations by the people themselves moderators a great deal of flexibility and and especially, for our purposes, by random discretion to present their own opinions and and representative samples. But the ques- advocate for them. Moderators appear to tion remains whether ordinary citizens have used that discretion freely, resulting in who vary widely in education and exper- the apparent distortion of the outcomes to tise can usefully weigh the competing ar- conform to their views. As in other applica- guments at issue in actual policy choices. tions of deliberative democracy, the precise Some critics even question whether ordi- institutional design can be consequential. nary citizens in developed countries such as It remains to be seen what would happen the United States have the capacity to weigh with a deliberative design closer to the mi- competing policy arguments.15 From such crocosms that have been applied thus far in a perspective, it would seem even less plau- developed countries.13 If deliberators were sible that participants who lack education recruited through random sampling rath- or even basic literacy could do so. er than self-selection and if the moderators Second, arguments “against delibera- were strictly constrained to avoid advoca- tion” have found a footing in normative

142 Dædalus, the Journal of the American Academy of Arts & Sciences theory, building on the jury literature.16 In sion to move toward extremes. Building on Fishkin, particular, there is the worry that advan- earlier work on the “risky shift,” Cass Sun- Mayega, Atuyambe, taged groups will dominate the delibera- stein and various colleagues hypothesized a Tumuhamye, tions and impose their views on everyone “law of group polarization.”20 On an issue Ssentongo, Siu & else. In some juries, the men dominate, or for which there is a midpoint, if most par- Bazeyo the more educated, or those who have high- ticipants in the discussion are to the right er social status. Juries are the most studied of the midpoint, then there will be move- deliberative institution, and while one can ment away from the midpoint to the right. make a case that juries do fairly well at not But if most participants are to the left of the reflecting the power relationships found midpoint, then there will be movement to in society,17 any pattern of domination by the left. This polarization occurs, the argu- the more advantaged raises questions about ment goes, because of two factors: an “im- whether people are really deliberating on the balance in the argument pool” and “a social merits, rather than deferring to those who comparison effect.” If the group is mostly may be seen to have more competence or on one side, then more of the arguments authority.18 To the extent that the delibera- voiced are likely to be on that side. Hence tions are not genuine decisions on the mer- the tendency to move to a “more extreme” its, the institution loses its claim to deter- position away from the midpoint. Second, mine legitimate outcomes. Of course, ju- as people pick up on the conclusions of ries are generally deciders of fact, not policy, others, they will feel social pressure to con- but the worries about juries have inspired form to the dominant position.21 More re- concerns about the broader use of citizen cently, Sunstein has added a third argument. deliberation for policy questions. If delib- Those who feel “tentative” in their views erating microcosms or mini-publics are may choose initial moderation out of un- distorted by deference to the advantaged, certainty, but these “tentatives” are more or by the ability of the advantaged to im- easily swayed by the other two factors to pose their will on the other participants, conform to the apparently dominant ar- then the ideal of deliberation, nicely cap- guments.22 tured in German philosopher Jürgen Haber- The vulnerability of deliberative discus- mas’s famous phrase as the “unforced force sion to polarization is likely a matter of in- of the better argument,” would be under- stitutional design. While found in jury-like mined.19 In the context of developing coun- experiments conducted by Sunstein and his tries, this worry might well be exacerbated. collaborators, this pattern has not applied Those few participants who are well educat- universally to deliberating microcosms ed, or who have high status for other rea- chosen by random sampling. For example, sons, might have a great advantage over the we have not found it in Deliberative Polls.23 rest of the participants who lack basic edu- If designs have elements of balance and cation and preparation for the discussions. confidentiality, those elements may well de- The less well educated might defer to the feat the imbalance in the argument pool and advantaged, thus distorting the process. the social comparison effect. The design of Hence this critique, often applied generally Deliberative Polls includes elements of bal- to applications of deliberative democracy, ance, such as balanced briefing materials, would seem to pose a special challenge in balanced plenary session panels, and mod- developing countries. erators who are trained to draw attention to A third critique has centered on what the competing sides of the argument in the has come to be known as polarization, or briefing materials. It also ensures confiden- the tendency of groups engaging in discus- tiality for the final considered judgments

146 (3) Summer 2017 143 Applying by collecting them in confidential ques- this nonpolarizing result with samples of Deliberative tionnaires. The facilitators are trained to more-educated and -literate populations, Democracy in Africa: bring out minority opinion and to set a tone indicating that the postulated “law of group Uganda’s First for respecting the opinion-givers equally. polarization” is not an inexorable law, even Deliberative These elements appear to protect at least though it appears in jury-like designs.25 The Polls this kind of deliberative microcosm from dynamic among largely nonliterate and un- the polarization pattern.24 educated samples has only just been tested But what would happen with far less edin research. ucated and literate respondents? None of the cases thus far have involved samples The Deliberative Polling projects in Ugan- mostly composed of less-literate respon- da took place in the Mount Elgon region dents with low levels of education. It is in two districts, Bududa and Butaleja, both an open question whether the avoidance troubled by frequent environmental di- of the “law of group polarization” might sasters (floods and rock slides). Each dis- apply in such contexts. It is easy to see trict has about two hundred thousand in- why this might be a problem. The balance habitants, mostly working in subsistence achieved through written briefing docu- farming. The areas are characterized by ments and the ability to weigh arguments low levels of education and high popula- from competing experts on either side of tion density. The average population den- the issues in the plenary sessions might be sity in Uganda is 195 persons per square undermined for nonliterate and low edu- kilometer, but in the Mount Elgon region, cation respondents. Further, the less liter- the average population density is 950 per- ate might more easily defer to group pres- sons per square kilometer. This popula- sures, allowing the social comparison ef- tion density puts pressure on the subsis- fect to determine the outcome and moving tence farming and prevents investment in the mean of the group to more extreme po- education. The low education levels are es- sitions. Perhaps less literate respondents pecially pronounced for women and girls, will tend to be more “tentative” and de- many of whom get no formal education at fer to those with considered judgments or all. The common environmental disasters with higher social status. Such speculations lead to periodic evacuations and issues of imply that in the development context, the resettlement. “law of group polarization” may well prove The Deliberative Polling efforts were led to be more of a challenge. by a team from Makerere University, more What is at stake here? Why is polariza- specifically, the East Africa Lab in the Resil- tion a challenge for applications of delib- ient Africa Network sponsored by usaid erative democracy? If there were a pre- and housed in the Makerere School of Pub- dictable pattern of group psychology mov- lic Health. The Stanford Center for Delib- ing arguments to more extreme positions, erative Democracy, also with usaid sup- then it would be hard to argue that the report, provided assistance at each stage of sults were really the product of participants the two projects. weighing the issues carefully on the merits. The Lab ran focus groups and key infor- But if roughly half the time the groups move mant interviews in the two communities further from the midpoint and roughly half to identify challenges faced by the two dis- the time they move toward the midpoint, tricts. The project also convened an advi- then the potential dynamic toward group sory group, including academics, key gov- polarization would have been stopped. As ernment officials (both local and national), noted, Deliberative Polls have produced and ngos to provide further input. The ad-

144 Dædalus, the Journal of the American Academy of Arts & Sciences visory group is listed in the report, available when done well, offer the public’s impres- Fishkin, online.26 sion of sound bites and headlines. They do Mayega, Atuyambe, The advisory group developed an agenda not reflect what the public would think if Tumuhamye, with specific policy options in three broad it actually thought in depth about the is- Ssentongo, Siu & areas. Because of the region’s recurrent en- sues. Deliberative Polling is a method that Bazeyo vironmental disasters, the agenda focused offers representative and informed opin- on three related topics: resettlement man- ion. It offers a road map to the policies agement, land management, and popula- the public would accept upon reflection, tion pressure. The committee identified and for what reasons. It can also indicate policy options under each of these head- those policies the public would have reser- ings that might feasibly be implemented in vations about, and for what reasons. Bududa and Butaleja. These options were the principal topics in the pre- and post- How should we evaluate these first De- deliberation questionnaires. liberative Polls in Africa? First, is the sam- Deliberative Polling assesses the repre- ple representative? We can compare the sentative opinions of a population, both be- participants (those who take the initial sur- fore and after it has had a chance to think vey and attend the deliberations) with the about an issue and discuss it in depth. The nonparticipants (those who take the sur- idea is to gather a representative sample and vey and do not attend the deliberations). engage it in transparently favorable condi- The comparisons should include both de- tions for considering the pros and cons of mographics and attitudes. The idea is to re- competing policy options. Most citizens, cruit a microcosm of the viewpoints and most of the time, in most countries around interests of the community. Voter lists, the world, do not spend much effort consid- census data, and random digit dialing have ering public policy questions in depth. The all been used in other countries to provide premise is that when policy options are im- the sampling frame. What approach might portant for a community, then public con- work in rural Uganda, where the data for sultations about them should be represen- such approaches are flawed and where the tative of the population and thoughtfully cell phone coverage is unreliable? based on the best information available. Second, do the opinions change? Ulti- Hence the need for recruiting a random mately we are interested in the final con- sample and engaging it in good conditions sidered judgments of the sample, regard- for considering the issues and the argu- less of whether they stay the same or move ments for and against various policy op- away from where they began. But if Delib- tions. erative Polls rarely yielded significant net The method offers certain advantages change, then few consultations would go over other methods of public consultation. to the trouble of creating these balanced Self-selected town meetings are unlikely and informed discussions. It would be eas- to be representative because they involve ier just to do conventional polling. Hence, only those who feel strongly enough to at- statistically significant net change indi- tend. Focus groups cannot be used to rep- cates that something is happening when resent opinion because they are too small citizens deliberate. to be statistically meaningful. Rather, they Third, are there identifiable reasons for are useful for uncovering the way the pub- the final judgments? Does the process lic frames an issue as a step in facilitating produce considered judgments that peo- more systematic research. Conventional ple reach on the basis of having considered polls, while potentially representative competing arguments?

146 (3) Summer 2017 145 Applying Fourth, does the process avoid the dis- had no education and 57 percent had only Deliberative tortions we have already identified as po- primary education. Eighty-seven percent of Democracy in Africa: tentially undermining the deliberative pro- the Bududa participants and 86 percent of Uganda’s First cess? Two distortions have been especially the Butaleja participants were farmers. As Deliberative prominent in the literature: group polar- best we can judge, it was an excellent sam- Polls ization and domination by the more ad- ple with one serious distortion: an overrep- vantaged. These potential distortions pose resentation of men.28 Yet, as we will see be- a challenge to deliberation in that they low, issues of gender and the interests of would appear to offer explanations for the women were reasonably well represented in results independent from the merits of the the discussions, despite the underrepresen- arguments. Rather, they would render the tation of women among the participants. results an artifact of group psychology or of The pre- and post-deliberation question- the domination of the more advantaged or naires were administered in individual in- educated. As mentioned above, if these dis- terviews taking approximately thirty-five tortions pose a challenge in the most devel- to forty minutes each. The use of oral in- oped countries with highly educated popu- terviews combined with video briefings al- lations, they are even more likely to occur lowed the nonliterate to respond and par- in developing countries whose populations ticipate in the process. We will return to have low education levels. the questions of whether the participa- The participants were recruited through a tion was relatively equal and whether the random selection of households and a ran- groups were able to avoid the predictable dom selection within the households. In distortions in group discussion. Bududa, there were 210 initial interviews, with only eleven refusals. Of those 210 ini- In both communities there were thirty-six tial interviews, 201 completed the full two policy options posed for deliberation cov- days of deliberation. Counting the eleven ering three topics: resettlement, land man- refusals in the total, the response rate for agement, and population pressure. Upon the actual event was about 91 percent, an first contact at home and at the end of the extraordinarily high level for surveys by in- weekend, the participants were asked to ternational standards and especially among rate the thirty-six options in importance on processes requiring two days of discussion. a scale from zero (extremely unimportant) In Butaleja, there were 232 initial inter- to ten (extremely important), with five in views, again with only eleven refusals. Of the middle. Here we will report the percent- those 232 who took the initial interview, 217 age saying simply that an option was “im- completed the full two days of deliberation. portant” (a rating above five); the means Counting the eleven refusals in the denomi- of the respondents’ ratings appear in Ta- nator, the response rate in Butaleja is 89 per- bles 1 and 2. cent, also an extraordinary level of partic- In Bududa, the rating of eleven of the thir- ipation. Why this high level? The projects ty-six policy options changed significantly had strong buy-in from community lead- after deliberation; four other options had ers and local authorities. The topic was one changes that were marginally significant of great interest to the communities. And (see Table 1). The changes were mostly in an honorarium and transport costs helped the direction of increased support for what make participation attractive.27 became the most favored options.29 Some Ten percent of the Bududa participants options started high and went significant- had no education, and 58 percent had only ly higher. Before deliberation, 76 percent of primary education. For Butaleja, 8 percent respondents viewed the rezoning of high-

146 Dædalus, the Journal of the American Academy of Arts & Sciences Table 1 Fishkin, Bududa: Significant Policy Changes for Participants Mayega, Atuyambe, Tumuhamye, Question/Issue T1 T2 T2-T1 P-value Ssentongo, Siu & 1. Rezone high-risk areas for no settlement. 0.766 0.840 0.075 0.001*** Bazeyo 4. Give support to the host families for helping 0.685 0.759 0.073 0.003*** those who move. 5. Strengthen the local disaster management 0.760 0.827 0.068 0.002*** committees. 6. Raise funds to support the work of the local 0.646 0.766 0.119 0.000*** disaster management committees. 8. Build peri-urban centers where people can 0.752 0.810 0.058 0.013*** resettle. 9. Make sure new peri-urban centers are nearby 0.812 0.843 0.031 0.108* so people can farm. 13. Ensure that the early warning system works 0.715 0.756 0.041 0.076* with the local disaster committees. 16. Communities should manage the wetlands 0.602 0.671 0.069 0.015** during the dry season. 17. Communities should create more rice 0.529 0.631 0.102 0.003*** schemes, but not in the wetlands. 20. Communities should be responsible for 0.585 0.674 0.089 0.002*** desilting riverbeds. 21. Government should assist communities in 0.528 0.580 0.052 0.076* desilting riverbeds. 22. Communities should build sanitation drains 0.836 0.872 0.036 0.074* for the reduction of malaria. 27. The government should raise narrow bridges. 0.811 0.858 0.047 0.015** 29. Communities should build ladders in the 0.420 0.483 0.063 0.053** highlands where there are not roads. 34. Families should consider their resources in 0.740 0.797 0.058 0.011*** planning the size of their families.

Note: T1 denotes before deliberation; T2 denotes after deliberation; T2-T1 denotes after deliberation minus before deliberation; P-value denotes statistical significance. Proposals were rated on a scale of 0 to 10, with 0 be ing extremely unimportant, 10 being extremely important, and 5 being the midpoint. Data are the means of re spondents’ ratings. In the significance column, * indicates a P-value of 0.10 or below, ** 0.05 or below, and *** 0.01 or below. risk areas for no settlement as important; saster management committees jumped post-deliberation, 85 percent viewed it as from 58 percent to 79 percent. There was important. Before deliberation, 67 percent also significantly increased importance of respondents viewed supporting host given to proposals involving community families to help those who move as impor action: to create more rice schemes (but tant; post-deliberation, 78 percent viewed not in the wetlands), to manage irrigation it as important. Some of the changes were for cultivation (from 48 percent to 57 per- large: the perceived importance of raising cent), for taking responsibility to desilt the funds to support the work of the local di- riverbeds (from 52 percent to 64 percent),

146 (3) Summer 2017 147 Applying and for building sanitation drains to reduce communities should be responsible for Deliberative malaria (from 87 percent to 94 percent). On desilting riverbeds (from 55 percent to 42 Democracy in Africa: the subject of family planning, there was a percent). Discussions revealed a growing Uganda’s First significant increase in support for the no- awareness of the machinery and scale of Deliberative tion that families should consider their re- work required to get these tasks done. Polls sources in planning the size of their fami- In the family planning area, there was an lies. The endorsement of this proposal in- increase in support for the government en- creased from 76 percent to 87 percent. forcing the minimum age for marriage of The online appendix lists the top prior- eighteen years from the already-high level ities after deliberation for Bududa and for of 87 percent to 94 percent. Butaleja. In Bududa, after deliberation, the The online appendix shows the priorities top priority of all thirty-six proposals was after deliberation for Butaleja. Government that the community should encourage girls assistance in drilling for clean water tops as well as boys to go to school. This proposal, the list with 98.6 percent of participants en- which began with very high support (96 dorsing its importance. The second high- percent) ended with virtually unanimous est priority post-deliberation was that the support (99 percent). The online appendix community should encourage girls to go to includes transcript excerpts exhibiting the school as well as boys. As in Bududa, this reasoning in support of the top priorities.30 moved from 97.4 percent before deliberation to about the same level as the top pri- The Butaleja deliberation also produced ority (98.6 percent) after. significant changes on eleven policy attitudes. These changes are depicted in Table Our third question, whether the final con- 2. Some of these changes show interesting sidered judgments seem to reflect reasoned reversals with deliberation. All are signifi- deliberation, gets ample support from the cant at the 0.05 level or better. transcript excerpts detailed in the online ap- Rezoning high-risk areas for no settle- pendix. Consider the top priorities. ment began with only 46 percent of respon- In Bududa, the top priority after deliber- dents endorsing it as important before de- ation was that the community should en- liberation; but after deliberation, the level courage girls to go to school as well as boys. rose twenty points to 67 percent. Support Education can reduce the outcome of girls for an early warning system using text mes- getting pregnant and married at too young saging went down from 60 percent to 42 an age. With schooling, they may be able percent, while support for an early warn- to go to technical schools and find jobs. ing system using sirens went up from 79 The second top priority was creating more percent to 92 percent. We think that the Health Center 2s (local clinics) in small vil- unreliability of electric power for charging lages. Currently, the distance to health cen- and the unreliability of the cell connec- ters is too great for many community mem- tions moved people to support sirens as a bers to receive treatment for emergency ill- more dependable system than text messag- nesses. The local clinics could provide a first ing. While there was an increase in support response even if they do not offer all the for communities to manage the wetlands equipment and services found in the larger during the dry season (from 70 percent to hospitals. Moreover, in other discussions, 82 percent), there was a drop in support for community health centers were cited as of- the idea that communities should maintain fering support for family planning, meaning the water channels during the wet season that some women were reluctant to pursue (from 78 percent to 67 percent) and that family planning assistance because medical

148 Dædalus, the Journal of the American Academy of Arts & Sciences Table 2 Fishkin, Butaleja: Significant Policy Changes for Participants Mayega, Atuyambe, Tumuhamye, Question/Issue T1 T2 T2-T1 P-value Ssentongo, 1. Rezone high-risk areas for no settlement. 0.553 0.670 0.116 0.000*** Siu & Bazeyo 3. Resettle with host families in a low-risk area 0.563 0.626 0.063 0.017** when there is a disaster. 11. Early warning system should use sirens. 0.761 0.821 0.061 0.008*** 12. Early warning system should use text 0.628 0.525 -0.103 0.000*** messages. 14. Plant trees to protect the river banks. 0.833 0.869 0.036 0.049** 16. Communities should manage the wetlands 0.687 0.736 0.048 0.041** during the dry season. 18. Communities should maintain water channels 0.749 0.657 -0.092 0.000*** during the wet season. 20. Communities should be responsible for 0.593 0.513 -0.080 0.006*** desilting riverbeds. 21. Government should assist communities in 0.846 0.874 0.028 0.091* desilting the riverbeds. 33. The government should enforce the minimum 0.840 0.881 0.041 0.032** age requirement for marriage of eighteen years. 37. Which option do you prefer? Spend money 0.612 0.462 -0.150 0.000*** on more roads and fewer bridges, or spend money on more bridges and fewer roads (on a scale of 1 to 7, with 1 showing preference for more roads and fewer bridges and 7 more bridges and fewer roads).

help was too far away. The third top prior- was encouraging girls as well as boys to go ity in Bududa was creating one-classroom to school, for reasons very much like those schools for elementary education in remote in Bududa. The third priority was the gov- areas. The distances to school now are too ernment building roads in remote areas so far for children to walk, so many receive no residents could bring their produce to mar- elementary education at all. Concern that ket. The local communities did not have the girls may be sexually assaulted if they have resources to build the roads themselves. to walk long distances to school also moti- Another way to explore the reasons sup- vates parents to keep their children at home. porting the final ratings of the policy op- In Butaleja, the top priority was govern- tions is to run regressions to isolate the le- ment assistance in drilling for clean water. vers of opinion change. The questionnaire, Participants were very concerned about the covering thirty-six policy options plus de- disease risks of dirty water, but they needed mographics and other questions, did not help with drilling. The second top priority have much room for explanatory variables.

146 (3) Summer 2017 149 Applying However, it did probe some basic values, al- ther to the right. If this were a consistent Deliberative lowing us to make connections in the re- pattern for the issues, then it would under- Democracy in Africa: gressions between those values and the pol- mine the claim that participants were de- Uganda’s First icy options. The online appendix illustrates liberating on the merits. For example, in a Deliberative some of these connections. The floods and study of polarization in group discussions Polls rock slides, for example, periodically threat- in two locations in Colorado, researchers en the basic well-being of the communities. found 80 percent of the group issue combi- The disasters threaten the order and securi- nations (the movements of small groups on ty required for people to prosper economi- a given issue) polarizing in this way in Boul- cally and maintain their access to basic ne- der (moving left politically) and 93 percent cessities. Hence, support for the early warn- in Colorado Springs (moving right).31 ing system in Butaleja is associated in the In the Uganda projects, there were four- regression analysis with basic values such as teen groups in Bududa and fifteen groups in making sure everyone has clean air and wa- Butaleja. Both projects used the same ques- ter and promoting economic growth. These tionnaire with thirty-six policy options. two values plus the importance of educa- Hence, there were 504 group issue combi- tion are associated in Bududa with sup- nations (potential group movements on the port for new infrastructure, such as build- specified issues) in Bududa and 540 group ing roads in remote areas, building bridges, issue combinations in Butaleja. In contrast and raising narrow bridges. The link be- to the results predicted by polarization tween valuing education and supporting theory, only 54 percent of the group issue travel infrastructure may reflect that the combinations polarized in Bududa and 51 difficulties of travel pose a major impedi- percent in Butaleja.32 That is, about half ment to education. The values of economic the time the groups moved away from the growth and clean air and water are also as- midpoint in the hypothesized direction and sociated in the regression analysis with the about half the time they moved toward it in policy of offering more education for fam- the opposite direction. We see no evidence ily planning, probably because large family of a “law of group polarization” distorting size in these communities impedes eco- these deliberations. nomic growth and better access to the ne- The second distortion is domination cessities of life. by the more advantaged: do they impose Both the transcripts and the regressions their views on the others? In rural Uganda, provide evidence that the final considered where many of the respondents lack educa- judgments were supported by the reason- tion, this risk was of special concern. Using ing of participants grappling with trade- the group issue combinations, we began by offs and priorities. looking at the starting points of the more privileged. If they were dominating the dis- Turning to our fourth major issue, did cussions and imposing their views on every- the process avoid the distortions that have one else, then the group issue combinations plagued some other group discussions? should move in the direction of the views The two we focused on are polarization held by the privileged or advantaged. We and domination by the more advantaged. looked at three possible patterns of dom- Polarization, as we discussed earlier, is ination: males imposing their views on fe- the idea that on a given issue, if a group males, older participants imposing their starts out to the left of the midpoint, it will views on younger ones, and the more edu- move further to the left. If it starts out to cated imposing their views on the less ed- the right of the midpoint, it will move fur- ucated. No such patterns arose. In Bududa,

150 Dædalus, the Journal of the American Academy of Arts & Sciences the group issue combinations moved in the both the local health clinics and the schools Fishkin, direction of the males only 21 percent of the as close to the communities as possible. Al- Mayega, Atuyambe, time, in the direction of the older respon- though the trade-off between distance and Tumuhamye, dents only 47 percent of the time, and in the quality for schools and for health care gener- Ssentongo, Siu & direction of the more educated only 24 per- ates arguments on both sides, the district of- Bazeyo cent of the time. In Butaleja, the group is- ficials found that, deliberating together, the sue combinations moved in the direction people from these communities had reached of the males only 25 percent of the time, in a different decision from the one they had the direction of the older respondents only made. The officials were, however, recep- 53 percent of the time, and in the direction tive to the idea that unless the health cen- of the more educated only 42 percent of the ters were located close to the villages, many time.33 In these contexts, then, we see no people would not get critical health care, in- evidence that the advantaged are impos- cluding family planning, and unless there ing their views on others. Instead, all sec- were school facilities close to the villages, tors seem to be learning from each other, many people would not get elementary ed- sharing arguments, and coming to conclu- ucation. These burdens would fall especially sions about what should be done. on the women, for family planning, and on the girls, for education. The results of our studies in Bududa and Government officials and other policy- Butaleja, Uganda, reveal two cases of rep- makers can weigh these trade-offs by them- resentative and thoughtful deliberation ex- selves if they so decide. But if they want pressing the considered views of the com- policies that are sustainable because the munities. Rather than self-selected group people can buy into them, then they need meetings or stakeholder consultations, to hear from the people. To date, the only these two projects show a way for the public practical method for getting that public in- to provide input directly on what they con- put in both a representative and informed sider to be the most urgent issues. The re- way is through the kind of deliberative pro- sults have already been remarked upon by cess outlined here. local and national officials in Uganda and The challenges to applying deliberative in the donor communities that hope to see democracy in Africa have previously been many of the policies implemented. Consider thought overwhelming. Yet the Uganda two examples of useful input: the policies projects have already helped inspire subse- on schools and health care centers. District quent Deliberative Polls in Ghana (in Tama- officials had previously been closing sec- le), in Senegal (in an area near Dakar), and ondary health centers (or failing to rebuild in Tanzania (on a national level). All of these them after natural disasters) in order to con- deliberative projects have featured high par- solidate health care provision in bigger and ticipation rates and intense deliberation. better centers. But these larger facilities are The complex story of the policy impacts of fewer in number, requiring many people to these projects will require separate analysis. travel much farther. Officials had also been In the meantime, these first projects stand consolidating the schools to make them big- as demonstrations that it is entirely feasible ger and better, again increasing travel time to consult populations in Africa in a repre- for those who attend. The idea of small one- sentative and thoughtful way about the poli- room schools in more remote villages to pro- cies affecting their communities. We need vide elementary education, particularly to not leave it only to stakeholders and elites girls, was not on the agenda. Yet these de- to speak for the people. With the right deliberations highlight the merits of placing sign, the people can speak for themselves.

146 (3) Summer 2017 151 Applying endnotes Deliberative * Democracy Contributor Biographies: JAMES S. FISHKIN, a Fellow of the American Academy since 2014, is in Africa: Director of the Center for Deliberative Democracy, the Janet M. Peck Chair in International Uganda’s First Communication, Professor of Communication, and Professor of Political Science (by cour Deliberative tesy) at Stanford University. He is the author of When the People Speak: Deliberative Democracy Polls and Public Consultation (2009), Deliberation Day (with Bruce Ackerman, 2004), and The Voice of the People: Public Opinion and Democracy (1995). ROY WILLIAM MAYEGA is a Lecturer in the Department of Epidemiology and Biostatistics at the Makerere University School of Public Health, Uganda. He is a Ph.D. student in Medical Science on a joint program between Karolinska Institutet, Sweden, and Makerere University. He has published articles in PLOS ONE and Diabetes Research and Clinical Practice. LYNN ATUYAMBE is Associate Professor in the Department of Community Health and Be havioral Sciences at the Makerere University School of Public Health, Uganda. He has writ ten for such journals as PLOS ONE, BMC Medical Education, and Pan African Medical Journal. NATHAN TUMUHAMYE is a Research Fellow and Director of the Eastern Africa Resilience Innovation Lab at the Makerere University School of Public Health, Uganda. He has written for such journals as BMC Public Health, PLOS Currents, and Health. JULIUS SSENTONGO is a Research Fellow at the Makerere University School of Public Health, Uganda, and a practicing physician. His research has been published in PLOS Currents: Disas- ters and by the Eastern Africa Resilience Innovation Lab. ALICE SIU is Deputy Director of the Center for Deliberative Democracy at Stanford Univer sity. Her work on Deliberative Polling has been published in The Search for Deliberative Democ- racy in China (2006), Governance Reform under Real-World Conditions: Citizens, Stakeholders, and Voice (2008), and prominent journals including British Journal of Political Science. WILLIAM BAZEYO is Dean of the Makerere University School of Public Health, Uganda, where he is also Associate Professor of Occupational Medicine. He is Director of the Centre for Tobacco Control in Africa. He has published articles in such journals as BMC Psychiatry, PLOS Currents, and Annals of Global Health. Authors’ Note: This essay was made possible by support from the United States Agency for In ternational Development (usaid). It was prepared jointly by the Center for Deliberative De mocracy (cdd) at Stanford University and the Resilient Africa Network Eastern Africa Resil ience Innovation Lab at Makerere University. The contents are the responsibility of the authors and do not necessarily reflect the views of usaid or the United States Government. We also acknowledge the contribution of the advisory group members, especially those from the local governments of Bududa and Butaleja. Data and materials from the projects discussed here are available from the Center for Deliberative Democracy at Stanford University and from the East Africa Lab in the Resilient Africa Network at Makerere University. The authors would like to thank Kathleen Giles of the cdd for her tremendous contributions to this project. 1 For some collections that gather the most influential of these discussions, from various per spectives, regarding deliberative democracy, see James Bohman and William Rehg, eds., Deliber- ative Democracy: Essays on Reason and Politics (Cambridge, Mass.: The mit Press, 2007); Jon Elster, ed., Deliberative Democracy (Cambridge: Cambridge University Press, 2008); and James S. Fish kin and Peter Laslett, eds., Debating Deliberative Democracy (Oxford: Blackwell Publishing, 2003). The phrase “deliberative turn” was coined by John Dryzek in his Deliberative Democracy and Be- yond: Liberals, Critics, Contestations (Oxford: Oxford University Press, 2003). 2 Carole Pateman, “Participatory Democracy Revisited” Perspectives on Politics 10 (1) (March 2012): 7–19. 3 If and when it is possible to engage the entire public on an issue, that also exemplifies deliber ative democracy by the people. For a proposed method of doing so, see Bruce Ackerman and James Fishkin, Deliberation Day (New Haven, Conn.: Yale University Press, 2004). Most of the focus of recent empirical work on public deliberation has, however, been on the microcosmic

152 Dædalus, the Journal of the American Academy of Arts & Sciences or mini-public strategy. Other work has addressed the quality of deliberation in democratic Fishkin, legislatures; see Jürg Steiner, André Bächtiger, Markus Spörndli, and Marco R. Steenbergen, Mayega, Deliberative Politics in Action: Analyzing Parliamentary Discourse (Cambridge: Cambridge University Atuyambe, Tumuhamye, Press, 2004). Ssentongo, 4 For an overview, see Kimmo Grönlund, André Bächtiger, and Maija Setälä, eds., Deliberative Siu & Mini-Publics: Involving Citizens in the Democratic Process (Wivenhoe Park, United Kingdom: ecpr Bazeyo Press, 2014). 5 Robert C. Luskin, Ian O’Flynn, James S. Fishkin, and David Russell, “Deliberating across Deep Divides,” Political Studies 62 (1) (2014): 116–135. For the case of public deliberation about the condition of the Roma, see Center for Deliberative Democracy, “Deliberative Polling on Pol icies toward the Roma in Bulgaria,” http://cdd.stanford.edu/dp-locations/europe/bulgaria/. 6 James S. Fishkin, Robert C. Luskin, and Alice Siu, “Europolis and the European Wide Public Sphere: Empirical Applications of a Counter-Factual Ideal,” European Union Politics 15 (3) (2014): 328–351. 7 See James S. Fishkin, Baogang He, Robert C. Luskin, and Alice Siu, “Deliberative Democracy in an Unlikely Place: Deliberative Polling in China,” British Journal of Political Science 40 (2) (2010): 435–448. For the more general issue, see Baogang He and Mark Warren, “Authoritarian De liberation: The Deliberative Turn in Chinese Political Development,” Perspectives on Politics 9 (2) (2011): 269–289. 8 Joseph E. Stiglitz, “Participation and Development: Perspectives from the Comprehensive De velopment Paradigm,” Review of Development Economics 6 (6) (2002): 163–182. See also Amartya Sen, Development as Freedom (New York: Alfred Knopf, 1999). 9 See Toward a More Inclusive and Effective Participatory Budget in Porto Alegre, World Bank Report No. 40144-BR (Washington, D.C.: World Bank, 2008). For the critique of representativeness, see volume 1 of the report, 27–30. 10 For a critique of the move from participatory to deliberative democracy, see Pateman, “Par ticipatory Democracy Revisited.” 11 Leonard Wantchekon, “Clientelism and Voting Behavior: Evidence from a Field Experiment in Benin,” World Politics 55 (3) (2003): 399–422. 12 Macartan Humphreys, William A. Masters, and Martin E. Sandbu, “The Role of Leaders in Democratic Deliberations: Results from a Field Experiment in São Tomé and Príncipe,” World Politics 58 (4) (2006): 583–622, esp. 604. 13 The Ugandan projects reported on here did not collect separate survey data from the moder ators. But in a more recent Deliberative Poll in Tanzania with the Center for Global Devel opment, we did collect such data and intend to report on it separately. The Deliberative Poll process trains moderators not to offer their own substantive views during deliberations. 14 The quotation is from James Madison in Federalist No. 10. The term “deliberative democracy” was coined by Joseph Bessette in discussing Madison’s theory of deliberation by representatives (not deliberation by the people themselves). See Joseph M. Bessette, The Mild Voice of Reason: Deliberative Democracy and American National Government (Chicago: University of Chicago Press, 1994). 15 Russell Hardin asks whether “deliberative democracy is the ‘democracy’ of elite intellectuals.” Russell Hardin, “Deliberation: Method, Not Theory,” in Deliberative Politics: Essays on Democracy and Disagreement, ed. Stephen Macedo (New York: Oxford, 1999), 103–119. Shawn W. Rosenberg ar gues that deliberation even among the educated in the United States is “problematic,” conclud ing from his case study in California that “even in what might be considered the favorable case of highly educated, empowered adults raised in a liberal democratic polity, the deliberation very rarely was rational, reasonable, or critically reflective in the ways assumed by deliberative dem ocrats.” See Shawn W. Rosenberg, ed., Can the People Govern (New York: Palgrave, 2007), 157. 16 Lynn Sanders, “Against Deliberation,” Political Theory 25 (3) (1997): 347–364; and Iris Marion Young, Inclusion and Democracy (Oxford: Oxford University Press, 2000).

146 (3) Summer 2017 153 Applying 17 In the landmark study comparing views of judges with actual jury verdicts, Harry Kalven and Deliberative Hans Zeisel found considerable agreement between the judges and juries. See Harry Kalven Democracy and Hans Zeisel, The American Jury (Boston: Little Brown, 1966). in Africa: Uganda’s First 18 For classic accounts of gender and class domination, see Fred L. Strodtbeck, Rita M. James, and Deliberative Charles Hawkins, “Social Status and Jury Deliberations,” American Sociological Review 22 (1957): Polls 713–719; and Reid Hastie and Steven D. Penrod, Inside the Jury (Cambridge, Mass.: Harvard Uni versity Press, 1983). 19 Jürgen Habermas, Between Facts and Norms, trans. William Rehg (Cambridge, Mass.: The mit Press, 1998), 306. 20 Cass R. Sunstein, “The Law of Group Polarization,” in Debating Deliberative Democracy, ed. James S. Fishkin and Peter Laslett (Oxford: Blackwell Publishing, 2003), 80–101. 21 For an overview of these arguments and an extensive bibliography documenting their occur rence, see Cass R. Sunstein, Going to Extremes (Oxford: Oxford University Press, 2009). 22 Cass R. Sunstein and Reid Hastie, Wiser: Getting Beyond Groupthink to Make Groups Smarter (Boston: Harvard Business Review Press, 2015), 84. 23 See, for example, Robert C. Luskin, James Fishkin, and Roger Jowell, “Considered Opinions: Deliberative Polling in Britain,” British Journal of Political Science 32 (3) (2002): 455–487; Fishkin et al., “Deliberative Democracy in an Unlikely Place”; and Alice Siu, “Look Who’s Talking: Examining Social Influence, Opinion Change, and Argument Quality in Deliberation” (Ph.D. diss., Stanford University, 2008). 24 For more on this, see Alice Siu, “Deliberation & the Challenge of Inequality,” Dædalus 146 (3) (Summer 2017). 25 See David Schkade, Cass R. Sunstein, and Reid Hastie, “When Deliberation Produces Extrem ism,” Critical Review: A Journal of Politics and Society 22 (2–3) (2010): 227–252. 26 See Deliberative Polling in Uganda: A Report on the Projects in Bududa and Butaleja Districts, Uganda (Stan ford, Calif., and Kampala, Uganda: Center for Deliberative Democracy at Stanford University and Resilient Africa Network Eastern Africa Resilience Innovation Lab at Makerere University, 2014), Appendix Table 1, “Members of the Advisory Group,” 30, https://cdd.stanford.edu/2014/ a-report-on-the-projects-in-bududa-and-butale-districts-uganda/. 27 Participants were paid the equivalent of about seven dollars a day as an honorarium. 28 Both the Bududa and Butaleja samples were overrepresented by men. Participants in Bududa were 58.7 percent male and in Butaleja were 66 percent male. The interviewers randomly se lected the participant within each household, but they could only do so based on the list of in habitants provided by the person first contacted. Despite this limitation, the interests of wom en seem to have been well represented in the discussions, as should be clear from the priori ties after deliberation on the education of women and on population pressure. 29 We will discuss these results in percentages; the means can be seen in Table 1 and Table 2. 30 The online appendix for this article is available at Center for Deliberative Democracy, “De liberative Polling in the Bududa and Butaleja Districts of Uganda,” https://cdd.stanford.edu/ 2014/deliberative-polling-in-the-bududa-and-butaleja-districts-of-uganda/. For the top-ten priorities after deliberation in Bududa and Butaleja, see Tables H and I. The transcript excerpts are included in Tables J and K. 31 Shkade, Sunstein, and Hastie, “When Deliberation Produces Extremism,” 232. 32 See Tables D and E in the online appendix, https://cdd.stanford.edu/2014/deliberative-polling -in-the-bududa-and-butaleja-districts-of-uganda/. 33 See Tables F and G in ibid.

154 Dædalus, the Journal of the American Academy of Arts & Sciences Authoritarian Deliberation in China

Baogang He & Mark E. Warren

Abstract: Authoritarian rule in China increasingly involves a wide variety of deliberative practices. These practices combine authoritarian command with deliberative influence, producing the apparent anomaly of authoritarian deliberation. Although deliberation and democracy are usually found together, they are distinct phenomena. Democracy involves the inclusion of individuals in matters that affect them through distributions of empowerments like votes and rights. Deliberation is the kind of communication that involves persuasion-based influence. Combinations of command-based power and deliberative influence– like authoritarian deliberation–are now pervading Chinese politics, likely a consequence of the failures of command authoritarianism under the conditions of complexity and pluralism produced by market- oriented development. The concept of authoritarian deliberation frames two possible trajectories of political development in China. One possibility is that the increasing use of deliberative practices stabilizes and strengthens authoritarian rule. An alternative possibility is that deliberative practices serve as a leading edge of democratization.

Over the last several decades, authoritarian regimes in Asia have increasingly experimented with public consultation, political participation, and even deliberation within controlled venues.1 China is a particularly important example: though it remains an authoritarian regime, governments, mostly at the local level, have employed a wide variety of participatory practices that include consultation and deliberation.2 In the 1980s, leaders began to introduce direct elections at the village level. Other innovations have followed, BAOGANG HE is the Alfred Deakin including approval and recall voting at the local lev- Professor and Chair in International el, participatory budgeting, deliberative forums, De- Relations at Deakin University, liberative Polls, public hearings, citizen rights to sue Australia. the state, initiatives to make government informa- MARK E. WARREN is the Harold tion public, and acceptance of some kinds of auton- and Dorrie Merilees Chair in the omous civil society organizations. Although very un- Study of Democracy at The Univer- even, many of these innovations appear to have gen- sity of British Columbia, Canada. uinely deliberative elements: that is, they involve the (*See endnotes for complete contributor kinds of talk-based politics that generate persuasive biographies.) influence, from which political leaders take guidance,

155 Authoritarian and upon which they rely for the legitima- these dynamics by stylizing two possible Deliberation cy of their decisions.3 Curiously, these prac- trajectories of political development. One in China tices are appearing within an authoritari- possibility is that deliberative mechanisms an state led by a party with no apparent in- could provide stability for authoritarian- terest in regime-level democratization. We ism in ways that would make it compati- call this paradoxical phenomenon authori- ble with complex, decentered, multi-actor tarian deliberation. market societies. Another possibility, less We make three broad claims. The first is likely at the moment but possible in the fu- oriented toward democratic theory. We ar- ture, is that if the regime were increasing- gue that authoritarian deliberation is theo- ly to rely on deliberative influence for its retically possible: it combines authoritar- legitimacy, it might find itself locked into ian distributions of the power of decision incremental advances in democratic em- with deliberative influence. powerments. Under this scenario, democ- Our second claim characterizes China’s ratization would be driven by problems of regime type as deliberative authoritarianism: a governance and led by the current experi- regime style that makes common use of au- ments in deliberation, as opposed to regime thoritarian deliberation. But why would an change following the more familiar “liber- authoritarian regime resort to deliberative al” model in which independent social forc- practices? Our broad hypothesis is func- es propel regime-level democratization– tional: problems of governance in complex, the pattern most frequent in the democratic multi-actor, high-information, and high- transitions of the last several decades. resistance environments give elites incentives to rely on popular input and even pop- In 2012, Xi Jinping assumed office as presi- ular deliberation, especially when they be- dent of the People’s Republic of China and lieve they can use these instruments to pro- general secretary of the Central Committee vide the kinds of proximate and specific of the Communist Party of China (ccp). Xi’s responsiveness that co-opt popular orga- leadership has reversed much of the liber- nizing and substitute for democratic em- alization of the past several decades. Under powerments. These arrangements can pro- Xi, the ccp has increased authoritarian conduce a unique relationship between au- trols and Party discipline and has height- thoritarianism and deliberation. Such func- ened pressure on dissidents, universities, tionally driven deliberative developments and public spaces. Chinese foreign policy is can be found in several nations other than increasingly aggressive. Xi has also sought China: governments in developed democ- to reassert civilian control over the Peo- racies have been innovating with new ple’s Liberation Army. He has embarked on forms of participatory and deliberative a strong anticorruption campaign, proba- governance over the last few decades in re- bly motivated by concerns that corruption sponse to many of the same kinds of pres- is a kind of “slow political suicide” of the sures.4 What distinguishes China is that regime itself, and certainly aimed at more governance-driven deliberative politics is control over quasiautonomous political developing in the absence of regime-level power centers. Xi is using increasingly au- democratization.5 thoritarian controls to modernize the fi- Our third broad claim is that the contra- nancial sector, to continue to reform state- dictory features of authoritarian delibera- owned enterprises, and, more generally, to tion identify the dynamic qualities of Chi- modernize the economy so that it contin- nese political development that most in- ues to perform well. These developments terest democratic theorists. We illustrate are not entirely surprising: the legitimacy

156 Dædalus, the Journal of the American Academy of Arts & Sciences of the ccp depends heavily on economic public policy, as has always been empha- Baogang He performance, which in turn depends on re- sized by the ccp’s Mass Line; 2) democracy & Mark E. moving roadblocks to growth–including is a way of ensuring that expertise is includ- Warren entrenched and often corrupt interests–as ed in public policies; 3) consultative democ- well as developing the institutions of a mod- racy is a key resource for developing legiti- ern market economy. The general pattern macy for Party leadership; 4) consultative remains that of regime-level authoritarian- democracy is a way of ensuring social har- ism with no apparent signs of regime-level mony by providing places for the people’s democratization. problems and demands to be heard and Yet although the ccp under Xi has increas- channeled into the political system; 5) the ingly cracked down on “foreign ideas” in long-term goal is to develop not just consul- politics–liberal democracy and multipar- tative democracy in a few places, but rather ty democracy in particular–one such idea a “multi-institutional” and “complete sys- has gained influence. Theccp continues to tem of consultative democracy”; and 6) the develop and deepen what they call xie shang ultimate goal of developing consultative de- min zhu, varyingly translated as “consulta- mocracy is to ensure min zhu: “the people tive democracy” or “deliberative democ- are the masters.” racy.” Except when referring directly to ccp These central directives are both a re- documents, here we will use the term delib- sponse to governance challenges and an erative democracy, in keeping with the mean- incorporation of considerable political in- ing of xie shang, which combines xie (doing ventiveness, particularly at the local lev- things together, cooperation, and harmoni- el. Local governments in China face an in- zation) with shang (talk, dialogue, consulta- creasing number of petitions and social tion, and discussion). So democracy (min conflicts, as well as challenges from com- zhu) is modified byxie shang: discussing is- plex issues. The Beijing government, for ex- sues in the spirit of doing things together. ample, now receives more than one thou- In November 2013, the Party Central sand petititons each day! To manage the Committee held its Third Plenum of the social conflicts these petitions represent, lo- Eighteenth National Congress, in which cal governments have been introducing the deliberative democracy was given official ideas and practices of deliberative democ- encouragement in the form of a directive racy, such as citizens’ juries. From 2014 to to lower levels of government–as is often 2016, Baogang He took several trips to Bei- the ccp’s style of rule. The mention of de- jing, Xiamen, Hangzhou, Shanghai, Guang- liberative democracy (officially, “socialist dong, Zhejiang, Hebei, and Henan to inves- consultative democracy”) in the Third Ple- tigate the recent trends in deliberative pol- num document was no accident, as it was itics over the last few years. Interestingly, followed by documents from the Central He found that the ccp’s program of “social- Committee on February 9, 2015, with direc- ist consultative democracy” appears to be tions for “Strengthening Socialist Consulta- proceeding, even as authoritarian controls tive Democracy,” and on June 25, 2015, out- are increasing. lining the role of the Chinese People’s Po- First, several organizations specifically litical Consultative Meeting in furthering designed for public deliberation have been deliberative democracy. Six ideas were es- set up. An empowered Deliberative Poll pecially prominent in these directives: 1) on local budgeting was held in Wenling in consultative democracy is an ordered way 2005. The process was so popular that it is of absorbing wisdom and strength from the now institutionalized; and studies suggest Chinese people to improve governance and that it not only represents a high-quality de-

146 (3) Summer 2017 157 Authoritarian liberative process, but is also quite demo- ing citizens from other professional bodies Deliberation cratic, owing to representation through its like social psychology. The city guarantees in China near-random selection process.6 A similar that petitioners have the right to choose ju- Deliberative Poll was held in the Puxi Dis- ries from this pool to consider petitions. It trict of Shanghai in 2015.7 The Haicang Dis- has also developed a new practice of mov- trict of Xiamen established a center for pub- ing public deliberation from official offices lic deliberation that organizes and executes to the site of the dispute to help jurists bet- all local deliberative forums. Aitu County in ter understand the issues. Jinling Province set up the “People’s Arbi- Third, the topics discussed are increas- tration Center” through which citizens can ingly substantive. Ten years ago, the issues call for a public hearing. This center over- put up for public deliberation were compar- saw a much-discussed live telecast of the atively insubstantial, such as tourist devel- public debate between villagers and local opment or developing cultural signage for leaders on the issue of compensation. a city. When Baogang He proposed a public Second, some procedures that empower forum, like citizens’ juries, to deal with the citizens to participate in deliberative pro- petition issue in 2005 in one Beijing work- cesses remain in place or have been further shop, it was immediately dismissed as “too improved. In theory and often in practice, idealistic”: the petition issue was viewed citizens have entitlements such as access to by officials as sensitive and complicated, so information and rights to agenda-setting. much so that it fell into the zone of national For example, petitioners can call for pub- security concerns. Over the last few years, lic hearings in Changshan, Hunan Prov- however, important issues like land appro- ince, or Haining, Zhejiang Province. In the priation, building demolition, and compen- Ronggui neighborhood government of the sation have been hotly debated in public fo- Shuide District in Guangdong Province, all rums. There are other indications that local social policies must be proposed and dis- governments are beginning to use deliber- cussed through a citizen committee before ative forums to manage increasing num- being submitted to the Party Committee bers of petitions from citizens. Local gov- for further consideration. In 2013, Yanjin ernments in Huizhou, Changsha, Huzhou, County in Yunnan Province introduced a and Aitu have started to organize citizens’ new budgeting process in which both ran- juries to examine petition claims. Huizhou domly selected citizens and elected repre- successfully organized a modified version sentatives are able to make new proposals of Deliberative Polling to solve the “mar- about the budget, with majority rule used ried-out” women’s petitions for equal dis- to decide the result. One procedure intro- tribution of village wealth.8 Haining has duced in Haining in Zhejiang Province in developed and improved a set of concrete 2014, required the immediate release of the procedures of citizens’ juries to deal with results of votes cast by citizen jurists on the a series of the petition claims in 2014. Cit- spot. Moreover, citizen jurists can vote on izens’ juries introduced in Aitu County in whether a governmental organization has Jilin Province between 2011 and 2015 have done an adequate job or whether the peti- substantively reduced the number of petitioners in a dispute have legitimate reasons tioners. This causal effect is also indicat- for their petitions. Haining has established ed by public deliberation in Wenling and a pool of one hundred jurists comprising Huizhou. But there is not yet consensus on forty ordinary citizens, twenty locally elect- this issue, with some arguing that public deed people’s deputies, nine lawyers, nine liberation may increase the number of pe- mediators, and six social workers, includ- titioners.

158 Dædalus, the Journal of the American Academy of Arts & Sciences The Chinese Legal Database (Peking Uni- deliberation requires protection from co- Baogang He versity Law Database, Beida Fabao) provides ercion, economic dependency, and tradi- & Mark E. another source of evidence. Documents be- tional authority if deliberative influence Warren tween 2001 and 2016 show a rapid increase is to function as a means of resolving con- in the numbers of provisions on public flict and legitimizing collective decisions. hearings in municipalities, provincial capi- Democratic institutions usually provide tals, and major cities up through 2010, after these protections by limiting and distrib- which the numbers plateau, though at a rela- uting power in ways that provide both the tively high level (Table 1). These documents spaces and the incentives for persuasion, vary from informing the citizens’ right to argument, expressions of opinion, and hold public hearings, to organizing, improv- demonstration. These protected spaces en- ing, and establishing procedures for public able the formation of preferences, enable hearings, issuing public announcements on legitimate bargains, and, sometimes, pro- public hearings, and reporting the results. duce consensus. Because democracy im- Experiments with public deliberation in plies inclusion, collective decisions with- China appear to be increasingly genuine, out it–no matter how deliberative–are substantive, inclusive, and often impres- likely to be experienced by the excluded as sive. But their contributions to regime de- illegitimate impositions. Although highly mocratization remain an open question. imperfect, established democracies have, The ccp continues to control these pro- in addition to their elected representative cesses. Political elites typically define per- bodies, a high density of institutions that missible spaces by issue, scope, and level of generate relatively deliberative approach- jurisdiction. Questions of representative es to politics, such as politically oriented inclusion, especially through elections, but media, law courts, advocacy groups, ad hoc even within deliberative forums that seek committees and panels, and universities descriptive representation, are often sub- with long-standing traditions of academic merged. That is, the pattern does not ap- freedom. Whatever their other differenc- pear to be one of increasingly democratic es, all theories of deliberative democracy deliberation, but rather one in which an in- presuppose a close and symbiotic relation- creasingly authoritarian regime is also mak- ship between democratic institutions and ing greater use of deliberative mechanisms. deliberation.9 Our challenge here is to make sense of this The clear and robust connection between seemingly paradoxical development. democracy and deliberation has led democratic theorists to ignore the difficult prob- The combination of authoritarian control lem of identifying deliberative influence and deliberative mechanisms is not as par- under authoritarian circumstances. To be adoxical as it might seem once we sort out sure, authoritarian regimes are, on aver- our terms of analysis. Among other things, age, unfriendly to deliberative approaches democracy involves the inclusion of indi- to conflict. Decision-making is closed and viduals in matters that potentially affect strict limits are placed on spaces of pub- them, realized through equal distributions lic discourse, such as the press, publish- of empowerments in votes, the opportunity ing houses, the Internet, advocacy groups, for voice, and related rights. Deliberation is and universities. Authoritarian rulers typi- a mode of communication involving argu- cally command; they do not invite the peo- ment and reasoning that generate persua- ple to deliberate. sion-based influence. In many ways, “de- Yet democracy is contingently, rath- liberation” requires “democracy.” Good er than necessarily, linked to deliberative

146 (3) Summer 2017 159 Authoritarian Table 1 Deliberation The Number of Official Documents on Public Hearings in Selected Years, 2001–2016 in China

Year 2001 2005 2006 2008 2010 2011 2012 2013 2014 2015 2016

Number of 193 945 1,282 1,389 1,833 1,645 1,533 1,426 1,457 1,523 1,476 Documents

Source: Data compiled using Peking University Law Database, Beida Fabao.

devices and mechanisms. In theory, de- to the communication: the addressees are liberation can occur under authoritarian persuaded by the claims put to them. conditions when rulers decide to use it as Democracy, in many ways, favors per- a means to acquire information by which suasive influence over other ways of get- to form policies and to gain approval from ting things done, but its root meaning is those affected without giving up powers of rule by the people. Democracy empowers decision. To identify the theoretical possi- those potentially affected by collective de- bility of deliberative politics under author- cisions so they can influence those deci- itarian conditions, we define deliberation sions. The standard means of empower- as a persuasive influence generated by the ment include the rights and opportunities give and take of reasons. Here we follow to vote for political representatives in com- the sociologist Talcott Parsons’s concep- petitive elections and, on occasion, to vote tion of influence as “the capacity to bring directly for policies, as in the case of refer- about desired decisions on the part of the enda or town meetings. In addition, dem- other social units without directly offering ocratic means of empowerment include them a valued quid pro quo as an induce- representative oversight and accountabili- ment or threatening them with deleterious ty bodies; the rights to speak, write, and be consequences.”10 Thus, we understand de- heard; rights to information about public liberation broadly as any act of communi- matters; rights to associate for the purpos- cation that motivates others through per- es of representation, petition, and protest; suasion “without a quid pro quo”: that is, and due process rights against the state and in ways that are not reducible to threats or other powerful bodies.12 coercion, economic incentives, or sanc- Such empowerments can, of course, be tions based on tradition or religion, nor, we highly institutionalized as part of competi- would add, the result of deceit or manipula- tive electoral systems. But democratic em- tion. Persuasion, in this sense, can include powerments can also appear more gener- bargains and negotiations, assuming that ically in nonelectoral contexts. For exam- the procedures can be justified by reference ple, freedom of information legislation in to claims to fairness or other normative va- virtually all the developed democracies en- lidity claims.11 In contrast, commands are ables citizens to monitor public bureaucra- backed by implied threats, quid pro quos, or cies within the appointed parts of the po- the authority of position or tradition. Com- litical system. mands convey information, but the motivation for obeying the command is extrinsic to Although both democratic and authori- the communication. Deliberation, in con- tarian regimes make use of persuasive influ- trast, generates motivations that are intrinsic ence, in a democracy, citizens usually have

160 Dædalus, the Journal of the American Academy of Arts & Sciences the powers necessary to introduce delibera- agencies exist, there is no effective sepa- Baogang He tive claims into almost any issue at any lev- ration of power within governments and & Mark E. el of government. In authoritarian regimes, no independent oversight bodies (except Warren political elites decide the subject and place where the judicial system operates with in- of deliberative processes. In China, elites creasing autonomy).14 Under President Xi, constrain public deliberation to the prob- discussion of “constitutional” or “liberal” lems of governance they choose; they seek democracy is forbidden. The Chinese state to avoid spillover into nonapproved arenas still maintains a Leninist political structure. and topics. Despite regime control over the Democracy, Premier Wen Jiabao remarked domains and agendas of public delibera- about ten years ago, is “one hundred years tion, Chinese citizens have limited kinds away”–possible only when China becomes of democratic empowerments within spe- a “mature socialist system.”15 cific domains of governance, ranging from Thus, although we agree with political the negative powers of protest and obstruc- scientist Minxin Pei’s observation that de- tion to the positive powers of some kinds of mocratic change has stalled in China and voice (in organized deliberative forums), is now likely reversed at the regime level, citizen rights (like property rights), ac- when we look below the regime level, countability (like the right to vote on the where we would normally expect democ- performance of village officials), and vot- ratization, we find significant changes in ing (like village elections, intraparty elec- governance, producing a regime that com- tions, and some direct voting for policies). bines authoritarian control of agendas with In China, public deliberations 1) are usually just enough democratization to enable con- more local than national; 2) favor issues re- trolled deliberation.16 While many estab- lated to municipal governance and econom- lished democracies are seeing the emer- ic performance; and 3) channel demand gence of governance-level deliberative bo- into Party-controlled forms of represen- dies–China is not unique in this respect17– tation. These limited governance-focused what distinguishes China is that these empowerments do not add up to regime de- modes of participation are evolving in the mocratization. Rather, they contribute to absence of regime-level democratization. an overall pattern of authoritarian deliberation by empowering some domain-limited Why would elites in an authoritarian re- and scope-limited forms of voice. They also gime decide to devise and encourage new produce functioning pockets of democra- deliberative practices and institute any cy constrained by geographical scope, poli- low-level democracy, even a highly con- cy, and modes of representation.13 The constrained version? We should not rule out junction of these resources with domain normative motivations embedded in po- constraints maps the spaces of authoritar- litical culture. The post-Maoist, neo-Con- ian deliberation that have been emerging fucian culture of China imposes moral rein China. sponsibilities on elites that are not trivial.18 China lacks, of course, the major insti- But even where such motivations exist, they tutions of electoral democracy, such as in- would need to correspond with the strate- dependent political organizations, autono- gic interests of powerful elites and with es- mous public spheres, independent oversight tablished institutions in order for such prac- and separations of powers, open-agenda tices to evolve. From a strategic perspective, meetings, and, most notably, multiparty the ccp is gambling that opening some con- elections. Although divisions of power strained participatory spaces will channel among layers of government and between political demand into venues the Party can

146 (3) Summer 2017 161 Authoritarian control, containing popular protest and de- Party members with senior positions and Deliberation mands for regime-level democratization. privileges. But simply owing to their num- in China Behind this gamble is a functionalist sto- bers, neither the exit nor the loyalty strategy ry, which, in its broad outlines, is common can be applied to the hundreds of millions to developing contexts. The strategic condi- of ordinary Chinese citizens who are quite tions for deliberative experimentation were capable of collective forms of dissent. Sup- probably the result of decisions in the late pression is always possible and is used selec- 1970s to justify the continuing rule of the tively against those dissidents who have po- ccp as necessary for economic develop- litically mobilized potential or capacity but, ment, in the face of disintegrating ideologi- as with all overtly coercive tactics, overuse cal justifications. Opening China to market- produces diminishing returns. In the case oriented development introduced three of China, suppression risks undermining conditions under which deliberation could the growing openness that supports its de- become necessary to maintain ccp rule: 1) velopment agenda, as well as drawing inter- increasing complexity of governance; 2) in- national attention that may also have eco- creasing numbers of veto players as a conse- nomic consequences. Thus, voice is the re- quence of pluralized control over economic maining option for controlling dissent and resources; and 3) changing popular expecta- maintaining order. tions, especially within the growing middle Second, deliberative mechanisms can class, driven by increasing levels of educa- produce information about society and pol- tion and contact with the West. So although icy, thus helping to avoid mistakes in gov- popular deliberative influence may be most ernance. Authoritarian regimes face a di- reliably generated under democratic condi- lemma with regard to information: Under tions, elites may have incentives to generate conditions of rapid development, authori- deliberative influence even without the in- tarian methods are often at odds with the centives provided by democratic empow- information resources necessary to govern. erments. As economist Albert Hirschman Elites need information not only about op- famously noted, the limited options for exit erational and administrative matters, but under one party rule are more likely to in- also about the preferences of citizens and crease internal pressures for voice.19 other actors. Command-based methods, These functional demands do not entirely however, limit communication and ex- explain authoritarian deliberative respons- pression, while increasing the incentives for es. But they do suggest a series of hypothe- subordinates to acquire and leverage infor- ses as to why authoritarian regimes might mation. Controlled deliberation is one re- adopt deliberative mechanisms. sponse to this dilemma. First, and arguably most important, de- Third, deliberation can provide forums liberative mechanisms can co-opt dissent for business in a marketizing economy. In and maintain social order. In the context China, market-style economic develop- of Hirschman’s typology of exit, voice, and ment is greatly increasing the number and loyalty, the ccp faces functional limits with independence of business stakeholders two of the three possible means of con- with independent economic control over trolling dissent. Currently, the ccp controls not only new investment, but also tax pay- much high-profile political dissent with an ments, which can make up the bulk of rev- exit strategy, allowing dissidents to emi- enues for many local governments.20 Pres- grate to the United States and other coun- sures for deliberation thus often come from tries to minimize their domestic impact. In- an increasingly strong business sector. Con- ternally, the ccp purchases the loyalty of sultations among public and private inter-

162 Dædalus, the Journal of the American Academy of Arts & Sciences ests have become increasingly institution- rule of the ccp, producing a new form of au- Baogang He alized–a process reminiscent of the origins thoritarianism and 2) deliberative influence & Mark E. of many legislative assemblies in England tends to undermine the power of authori- Warren and Europe, in which the middle classes tarian command, thus serving as a vector of bargained with their monarchs for liberty democratization. These two tendencies are and political voice in exchange for their tax currently bridged by limiting the scope and revenues.21 domain of both deliberation and democra- Fourth, public deliberative processes can cy so they can coexist with regime-level au- protect officials from charges of corruption thoritarianism. In the short term, we expect by increasing credible transparency. When deliberative authoritarianism to prevail. But local government revenues depend on busi- deliberation-led democratization could be ness, officials are usually regarded as cor- a longer-term possibility. rupt, and not only by the public, but often With the first possibility–deliberative by their superiors as well. Officials can learn authoritarianism–deliberative influence to use transparent and inclusive delibera- will increasingly function to stabilize au- tive decision-making to avoid or at least re- thoritarian rule.23 Under this scenario, au- duce accusations that their decisions have thoritarian political resources are used to been bought by developers and other busi- mobilize deliberative mechanisms. Delib- ness elites.22 erative influence is constricted in scope and Fifth, in situations in which decisions are agenda, and removed from political move- difficult and inflict losses, deliberative pro- ments and independent political organiza- cesses enable leaders to shift responsibili- tions. Deliberative experiments are local- ty onto the process and thus avoid blame. ized and skillfully managed so as to prevent In China, the elites are recognizing that “I them from expanding beyond particular decide” implies “I take responsibility.” But policy areas, levels of government, or re- “we decide” implies that the citizens are also gions. By this logic, if deliberation is suc- responsible, thus providing (legitimate) po- cessful at demobilizing opposition and gen- litical cover for officials who have to make erating administrative capacity, it could tough decisions. enable the ccp to avoid regime-level de- In summary, deliberative processes can mocratization. Authoritarian rule would generate legitimacy when ideological sourc- undergo some important transformations, es of legitimacy are declining for the ccp, but these would fall far short of regime-level and development-oriented policies are democratization. The current nascent form creating winners and losers. Legitimacy of deliberative authoritarianism in China is a political resource that even authoritar- would develop into a more consistent and ian regimes must accumulate to reduce the sophisticated type of rule, under which costs of conflict and enforcement. cruder exercises of power would be gradually replaced with more limited, subtle, and Our argument so far has been that the ap- effective forms. Political legitimacy would parently puzzling combination of author- be produced by means of deliberative con- itarian rule and deliberative devices and sultations, locale by locale and policy by mechanisms is conceptually possible and policy, as a complement to the kind of per- empirically extant in the Chinese case. Yet formance legitimacy that depends on con- the Chinese case also highlights two very tinuing economic development. different possible developmental trajecto- With the second possibility, contemplat- ries of deliberative authoritarianism: 1) deed by an increasing number of Chinese in- liberative politics effectively strengthen the tellectuals and local officials, deliberative

146 (3) Summer 2017 163 Authoritarian institutions developing within authoritar- citizen expectations. So too, democratic Deliberation ian ones will gradually democratize the re- institutions are easier for regimes to ini- in China gime. New institutions would overlay old tiate than to retract. Once the state grants ones for the intended purpose of enhanc- the people voice and rights, they become ing their effectiveness but, at the same part of the culture of expectations and time, would also transform their character transform supplicants into citizens. in democratic directions.24 If this trajecto- Third, deliberation tends toward insti- ry were to materialize, it would be unique: tutionalized decision-making procedures. we know of no examples of regime democ- The more deliberation is regularized, the ratization as a consequence of progressive- greater the pressures for it not to be dis- ly institutionalized deliberation. It is, never- continued. Trends toward institutionaliza- theless, a possibility. Although democracy tion can be driven by elite desires to retain and deliberation are distinct phenomena, control of political demand by channeling they are, as we have pointed out, structur- it into scope-specific and domain-specific ally related. Democratic empowerments– venues. But they can also be driven by cit- such as the rights of voting, association, izen expectations that, once established, and free speech–provide the space with- elites will find difficult to reverse. in which persuasion, argument, opinion, Fourth and finally, the logic of delibera- and demonstration can form preferences, tive inclusion eventually leads to voting. Po- enable negotiated bargains, and produce litical elites in China often refer to the rela- consensus. Democracy enables delibera- tionship between deliberation and consen- tion. But can deliberation enable democ- sual decision-making. This relationship is racy? Possibly. Deliberation provides legit- consistent with authoritarian deliberation. imacy only if it has the space and inclusive- Yet when interests conflict, even after delib- ness to generate actual influence.25 Under eration, elites may find it difficult to claim this scenario, four mechanisms could re- that their preferred decisions are the result sult in transformations in the form of rule. of “consensus,” thus eroding the legitima- First, deliberative legitimacy tends to- cy of command authoritarianism. It is in- ward the inclusion of all the people affect- creasingly common for leaders in China to ed by it. When other sources of legitimacy respond to deliberation that results in the fail–ideology, traditional deference, or eco- clarification of conflict by holding votes in a nomic benefits–deliberation provides an public meeting, by submitting decisions to alternative means of generating legitimacy. the community via referenda, or by defer- However, this legitimacy is “usable” by the ring to voting by the deputies of local peo- state only when 1) those whose cooperation ple’s congresses. the state requires are included in the deliberations, either directly or through repre- Our argument should not be viewed as sentation mechanisms, and 2) the partic- a prediction that if China democratizes, it ipants believe they have had influence. As will be governance-driven and delibera- the methods of obstruction (both rights- tion-led. Our argument is both more mod- based and protest-based) and exit are wide- est and speculative: by conceptualizing au- ly available in China, elites have incentives thoritarian deliberation and exemplifying to expand empowerments to those affect- its existence in China, we identify a poten- ed by policies so as to enable more engaged, tial trajectory of democratization that is less disruptive interactions with citizens. conceptually possible and normatively sig- Second, experiences of consultative and nificant. By distinguishing between demo- deliberative engagement tend to change cratic empowerments and deliberative in-

164 Dædalus, the Journal of the American Academy of Arts & Sciences fluence, we can focus on the legitimacy- and toward the transformative practices Baogang He producing capacities of deliberation. In out of which democratic innovations arise, & Mark E. so doing, we hope to push the democratic wherever they might develop. Warren imagination beyond familiar institutions

endnotes * Contributor Biographies: BAOGANG HE is the Alfred Deakin Professor and Chair in International Relations at Deakin University, Australia. He is the author of In Search of a People-Centric Order in Asia (2016), Contested Ideas of Regionalism in Asia (2016), and Governing Taiwan and Tibet: Democratic Approaches (2015). MARK E. WARREN is the Harold and Dorrie Merilees Chair in the Study of Democracy at The University of British Columbia, Canada. He is the author of Democracy and Association (2001) and coeditor of The Oxford Handbook of Deliberative Democracy (forthcoming) and Designing Deliberative Democracy (with Hilary Pearse, 2008). Authors’ Note: This essay updates and builds on the arguments in Baogang He and Mark E. Warren, “Authoritarian Deliberation: The Deliberative Turn in Chinese Political Develop- ment,” Perspectives on Politics 9 (2) (2011): 269–289. 1 Jennifer Gandhi, Political Institutions under Dictatorship (New York: Cambridge University Press, 2008); Garry Rodan and Kanishka Jayasuriya, “Beyond Hybrid Regimes: More Participation, Less Contestation in Southeast Asia,” Democratization 14 (5) (2007): 773–794; and Larry Dia- mond, “Elections Without Democracy: Thinking About Hybrid Regimes,” Journal of Democ- racy 13 (2) (2002): 21–35. 2 Baogang He, “Participatory and Deliberative Institutions in China,” in The Search for Deliber- ative Democracy in China, ed. Ethan Leib and Baogang He (New York: Palgrave, 2006); Man- oranjan Mohanty, George Mathew, Richard Baum, and Rong Ma, eds., Grassroots Democra- cy in India and China (Thousand Oaks, Calif.: Sage Publications, 2007); Andrew J. Nathan, “Authoritarian Resilience,” Journal of Democracy 14 (1) (2003): 6–17; and Suzanne Ogden, In- klings of Democracy in China (Cambridge, Mass.: Harvard University Press, 2002). 3 Leib and He, eds., The Search for Deliberative Democracy in China; Shangli Lin, “Deliberative Pol- itics: A Reflection on the Democratic Development of China,” Academic Monthly (Shanghai) 4 (2003): 19–25; He, “Participatory and Deliberative Institutions”; and Ogden, Inklings of De- mocracy. 4 Mark E. Warren, “Governance-Driven Democratization,” in Practices of Freedom: Decentred Governance, Conflict and Democratic Participation, ed. Steven Griggs, Aletta J. Norval, and Hendrik Wagenaar (Cambridge: Cambridge University Press, 2014), 38–59. 5 See Gerald E. Frug, “Administrative Democracy,” The University of Toronto Law Journal 40 (3) (1990): 559–586; and Carl J. Bellone and George Frederick Goerl, “Reconciling Public Entre- preneurship and Democracy,” Public Administration Review 52 (2) (1992): 130–134. 6 James Fishkin, Baogang He, Bob Ruskin, and Alice Siu, “Deliberative Democracy in an Unlike- ly Place: Deliberative Polling in China,” British Journal of Political Science 40 (2) (2010): 435–448. 7 Fuguo Han and Kaiping Zhang, “Self-Interest versus Altruism: Deliberative Participatory Budgeting in Urban China,” working paper, 2016. 8 Baogang He, “Deliberative Democracy and Deliberative Governance: Towards Constructing a Rational and Mature Civil Society,” Open Times 4 (2012): 23–36. 9 James Bohman, “Survey Article: The Coming of Age of Deliberative Democracy,” Journal of Political Philosophy 6 (4) (1998): 400–425; Simone Chambers, “Deliberative Democracy The- ory,” Annual Review of Political Science 6 (2003): 307–326; Joshua Cohen, “Procedure and Sub-

146 (3) Summer 2017 165 Authoritarian stance in Deliberative Democracy,” in Democracy and Difference: Contesting the Boundaries of the Deliberation Political, ed. Seyla Benhabib (Princeton, N.J.: Princeton University Press, 1996), 95–119; Amy in China Gutmann and Dennis Thompson, Democracy and Disagreement (Cambridge, Mass.: Harvard University Press, 1996); Jürgen Habermas, Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy, trans. William Rehg (Cambridge, Mass.: The mit Press, 1996); Mark E. Warren, “Deliberative Democracy,” in Democratic Theory Today: Challenges for the 21st Century, ed. April Carter and Geoffrey Stokes (Cambridge: Polity Press, 2002); and Mark E. Warren, “Democracy and the State,” in The Oxford Handbook of Political Theory, ed. John Dryzek, Bonnie Honig, and Anne Phillips (Oxford: Oxford University Press, 2006), 382–399. 10 Talcott Parsons, The System of Modern Societies (Englewood Cliffs, N.J.: Prentice-Hall, 1971), 14. 11 Habermas, Between Facts and Norms; and Mark E. Warren and Jane Mansbridge, “Deliberative Negotiation,” in Political Negotiation: A Handbook, ed. Jane Mansbridge and Cathie Jo Martin (Washington, D.C.: Brookings Institution Press, 2015), 141–196. 12 Robert A. Dahl, On Democracy (New Haven, Conn.: Yale University Press, 1998). 13 Baogang He, “Participatory and Deliberative Institutions in China,” in Leib and He, eds., The Search for Deliberative Democracy in China. 14 Randall Peerenboom, China’s Long March toward Rule of Law (Cambridge: Cambridge University Press, 2002). 15 Scott McDonald, “Wen: China Democracy 100 Years Off,” Time Magazine, March 1, 2007, http://www.time.com/time/world/article/0,8599,1594010,00.html. See also Bruce Gilley, China’s Democratic Future: How It Will Happen and Where It Will Lead (New York: Columbia Uni- versity Press, 2004). 16 Baogang He, Rural Democracy in China (New York: Palgrave/Macmillan, 2007), chap. 13. 17 Archon Fung, “Varieties of Participation in Complex Governance,” Public Administration Re- view 66 (1) (2006): 66–75; and Warren, “Governance-Driven Democratization.” 18 Baogang He, “Four Models of the Relationship between Confucianism and Democracy,” Journal of Chinese Philosophy 37 (1) (2010): 18–33; and Baogang He, “Deliberative Culture and Politics: The Persistence of Authoritarian Deliberation in China,” Political Theory 42 (1) (2014): 58–81. 19 Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (Cambridge, Mass.: Harvard University Press, 1970), 83–85. 20 Bruce J. Dickson, Red Capitalists in China: The Party, Private Entrepreneurs, and Prospects for Political Change (Cambridge: Cambridge University Press, 2003); and Gilley, China’s Democratic Future. 21 Robert H. Bates, “The Economics of Transitions to Democracy,” PS: Political Science and Poli- tics 24 (1) (1991): 24–27. 22 He, “Participatory and Deliberative Institutions in China.” 23 See Nathan, “Authoritarian Resilience”; and Aviezer Tucker, “Pre-Emptive Democracy: Oli- garchic Tendencies in Deliberative Democracy,” Political Studies 56 (1) (2008): 127–147. 24 See Ogden, Inklings of Democracy, 257; and Kathleen Thelen, “How Institutions Evolve: Insights from Comparative Historical Analysis,” in Comparative Historical Analysis in the Social Sciences, ed. James Mahoney and Dietrich Rueschemeyer (Cambridge: Cambridge University Press, 2003). 25 John S. Dryzek, “Democratization as Deliberative Capacity Building,” Comparative Political Studies 42 (11) (2009): 1379–1402.

166 Dædalus, the Journal of the American Academy of Arts & Sciences Board of Directors Nancy C. Andrews, Chair of the Board Jonathan F. Fanton, President Diane P. Wood, Chair of the Council; Vice Chair of the Board Alan M. Dachs, Chair of the Trust; Vice Chair of the Board Geraldine L. Richmond, Secretary Carl H. Pforzheimer III, Treasurer Kwame Anthony Appiah Louise H. Bryson Ira Katznelson Nannerl O. Keohane John Lithgow Cherry A. Murray Venkatesh Narayanamurti Don M. Randel Larry Jay Shapiro Natasha Trethewey Pauline Yu Louis W. Cabot, Chair Emeritus

Inside back cover: (Top) Audience member asks a question on the International Day of the Girl. © 2013 by Ryan Rayburn / World Bank. Creative Commons Attribution- NonCommercial-NoDerivs 2.0 Generic (CC BY-NC-ND 2.0). (Middle) Small group discussion about education in Singapore. © 2011 by Preetam Rai. Creative Com- mons Attribution 2.0 Generic (CC BY 2.0). (Bottom) Community members in Flint, Michigan, gather on February 6, 2016, to discuss the lead contamination crisis. © 2016 by J. Carl Ganter / Circle of Blue.

Summer 2017