Week 8: Deliberation, sortition, and mini-publics

In plain words

Week 8 of the My Vote Forward seminar studies deliberation: people reasoning together before deciding. Students read Gutmann and Thompson’s case for deliberative democracy, the Stanford Deliberative Democracy Lab’s description of Deliberative Polling, studies of the British Columbia and Irish citizens’ assemblies, Hélène Landemore’s proposal for “open mini-publics,” and research on whether group discussion pushes people toward extremes.

A citizens’ assembly picks ordinary people at random, gives them balanced information and time, and asks them to recommend something. Supporters say this beats a quick poll. Critics say a small group cannot speak for everyone, and that the wider public may reject what it decides, as happened in British Columbia. This week weighs both sides.

Defined terms

• Deliberation: discussion in which participants exchange reasons and consider others’ views before reaching a judgment. See deliberation.
• Deliberative democracy: in Gutmann and Thompson’s formulation, the idea that “citizens or their representatives owe each other mutually acceptable reasons for the laws they enact.”
• Deliberative Poll: a method in which a random sample is polled, deliberates with balanced materials and experts, and is polled again. See deliberative poll.
• Sortition: selection of public decision-makers by lot. See sortition.
• Mini-public: a small, randomly or near-randomly selected body of citizens convened to deliberate on a question. See citizens’ assembly.
• Group polarization: the tendency of a group’s view after discussion to move further in the direction its members already leaned.

Learning objectives

By the end of week 8, a participant should be able to:

1. State the core claim of deliberative democracy and its main rivals.
2. Describe how a Deliberative Poll works and what it is designed to measure.
3. Summarize the British Columbia Citizens’ Assembly and the Irish Citizens’ Assembly, including their outcomes.
4. Explain Landemore’s argument for open mini-publics.
5. Explain group polarization and why it is the strongest empirical objection to deliberation.
6. Evaluate whether a deliberative body’s conclusion should count as “the public’s” view.

Core readings

1. Amy Gutmann and Dennis Thompson, Why Deliberative Democracy? (Princeton University Press, 2004), chapter 1, “What Deliberative Democracy Means”.⁠¹
2. James S. Fishkin, When the People Speak: Deliberative Democracy and Public Consultation (Oxford University Press, 2009), selected chapters; with Stuart Weir’s review in openDemocracy.⁠²
3. Deliberative Democracy Lab, Stanford University, “What is Deliberative Polling?”.⁠³
4. Mark E. Warren and Hilary Pearse, editors, Designing Deliberative Democracy: The British Columbia Citizens’ Assembly (Cambridge University Press, 2008), introduction.⁠⁴
5. David M. Farrell, Jane Suiter, and Clodagh Harris, “’Systematizing’ Constitutional Deliberation: The 2016 to 2018 Citizens’ Assembly in Ireland,” Irish Political Studies 34(1), 2019, pp. 113 to 123.⁠⁵
6. Hélène Landemore, Open Democracy: Reinventing Popular Rule for the Twenty-First Century (Princeton University Press, 2020), introduction.⁠⁶
7. Lyn M. Van Swol, “Extreme Members and Group Polarization,” Social Influence 4(3), 2009, pp. 185 to 199.⁠⁷

Recommended: Cass R. Sunstein, “The Law of Group Polarization,” Journal of Political Philosophy 10(2), 2002, pp. 175 to 195.⁠⁸

Lecture notes

Why deliberation enters the course here

Deliberative democracy is the subject of week 8, and it answers a problem the course has built up over seven weeks. Opinion is constructed on the fly (week 6), shaped by media and elites (week 7), stratified by resources (week 3), and hard to aggregate (week 5). Deliberative democrats respond that the problem lies partly in measuring raw, unconsidered opinion. If citizens were given time, balanced information, and the chance to reason with people who disagree, their judgments might be more informed, more stable, and more worthy of weight. Week 8 asks whether that response works, and at what cost.

Gutmann and Thompson: reasons owed

Amy Gutmann and Dennis Thompson describe deliberative democracy as “the idea that citizens or their representatives owe each other mutually acceptable reasons for the laws they enact.” Their book asks “what exactly is deliberative democracy” and “why is it more defensible than its rivals,” and applies the theory to cases including bioethics, health care, truth commissions, educational policy, and decisions to declare war, addressing how citizens and representatives can make justifiable decisions “in the face of the fundamental disagreements that are inevitable”.⁠¹

The key move is from aggregation to justification. An aggregative view, like the rankings studied in week 5, counts preferences. A deliberative view asks whether a decision can be justified to those bound by it with reasons they could accept. These are different standards, and a decision can satisfy one and fail the other. The site’s chapter on opinion and choice discusses how stated preferences and considered judgments differ.

Deliberative Polling

James Fishkin’s Deliberative Polling turns the theory into a method. The Stanford Deliberative Democracy Lab describes the process: “A random, representative sample is first polled on the targeted issues. After this baseline poll, members of the sample are invited to gather at a single place for a weekend in order to discuss the issues. Carefully balanced briefing materials are sent to the participants and are also made publicly available.” Participants then engage with competing experts and political leaders, and are polled again. The Lab states that Fishkin “originated the concept of Deliberative Polling in 1988”.⁠³

The method’s rationale is a contrast with ordinary polls, which the Lab describes as capturing opinion from people who have “little reason to confront trade-offs or invest time and effort in acquiring information or coming to a considered judgment”.⁠³ A Deliberative Poll is meant to show what the public would think under better conditions. Stuart Weir’s review of When the People Speak summarizes Fishkin’s design: participants should be “properly representative of the population as a whole,” “fully and impartially briefed,” have “genuine opportunities to hear and question a variety of experts,” and be able to exchange views before reaching conclusions, which “can then reasonably be taken to be representative of the whole population if everyone experienced the same quality of preparation and debate.” Weir also names the central political problem: “the wider public will not necessarily accept the conclusions, or not enough of them will (as in British Columbia’s review of electoral reform)”.⁠²

British Columbia: authority without acceptance

The British Columbia Citizens’ Assembly is the case that tests Weir’s warning. Warren and Pearse’s volume describes it: in 2004, the provincial government “created an assembly of 160 near-randomly selected citizens to assess and redesign the province’s electoral system.” The editors describe it as “the first time a citizen body has had the power to reform fundamental political institutions,” and note that the model was “replicated elsewhere in Canada and in the Netherlands”.⁠⁴ The site’s British Columbia case traces what happened when its recommendation went to voters. For this lecture, the important point is structural: a deliberative body can reach a considered judgment, and the non-deliberating public, whose consent is needed, can still decline to adopt it. That gap is not a malfunction. It is the core question about the authority of mini-publics.

Ireland: deliberation within a system

Ireland offers a different relationship between assembly and public. David Farrell, Jane Suiter, and Clodagh Harris describe Ireland as “something of a trail-blazer in the use of deliberative methods in the process of constitutional review,” the first case where the process was used a second time: the Citizens’ Assembly of 2016 to 2018 “followed upon the Convention on the Constitution (2012 to 2014).” They argue that the creation of two mini-publics in quick succession, and “their significant role in supporting key referendums for constitutional change that followed (marriage equality in 2015 and abortion in 2018),” suggests “a degree of ‘systemization’ of deliberation in the Irish process of constitutional review”.⁠⁵

Two cautions keep this case in proportion. First, as the site’s Ireland case documents, decisions about topics, wording, and follow-up were made by politicians, so the assemblies’ influence passed through representative institutions. Second, not every later Irish process succeeded. The site’s Ireland case covers the full record from 2012 to 2023, including a later assembly that preceded referendums that failed. Students should resist reading Ireland as proof that assemblies “work”; it is evidence that they can matter when embedded in a system that takes them up.

The site also covers the French Citizens’ Convention on Climate, the Ostbelgien permanent citizens’ dialogue, and Iceland’s crowdsourced constitution, each of which shows a different way the link between deliberation and decision can hold or break.

Landemore: open mini-publics

Hélène Landemore’s Open Democracy pushes the argument from consultation toward governance. Her publisher describes her claim that representative democracy has become closed to ordinary people, and her proposal to rebuild democracy around randomly selected bodies: she favors “representing and being represented in turn” over direct democracy, and proposes an “open mini-public,” a jury-like body of randomly chosen citizens that defines laws and policies in connection with the wider public. The aim, in the publisher’s summary, is “placing ordinary citizens, rather than elites, at the heart of democratic power”.⁠⁶

Landemore’s position connects to week 2. Random selection secures a strong form of descriptive representation without elections, but it gives up the formal authorization and accountability that elections provide. Critics ask to whom a randomly selected body answers, and how the excluded majority can control it. Supporters reply that elected bodies are also unaccountable in practice, and that lot avoids the resource biases of week 3.

Group polarization: the strongest objection

The most serious empirical objection to deliberation is that discussion can make groups more extreme. Cass Sunstein’s “law of group polarization” is the standard statement of the concern, and a 2021 article in Análise Social describes the “so-called ‘law of group polarization’” as one of two influential objections that cast “serious doubts on the value of democratic deliberation”.⁠⁹ The concern has experimental support. Lyn Van Swol had participants rate their opinion on a 1 to 9 scale, then discuss in three-person groups. “Groups did polarize on their consensual opinion from the mean of individuals’ pre-group opinions.” Interestingly, “groups containing a more extreme member did not polarize more than groups not containing an extreme member,” and members who used more extreme ratings “talked more and took more turns”.⁠⁷

Deliberative designers respond that structured deliberation differs from unstructured discussion: random selection mixes people who disagree, balanced materials and competing experts counter one-sided arguments, and trained facilitators manage who speaks. Whether these features prevent polarization is an empirical question, and students should look for evidence on both sides rather than assume either answer. The site’s chapter on polarization covers the broader evidence on how Americans’ attitudes toward the other party have changed.

Deliberation and the priorities question

The deliberative tradition offers a pointed challenge to any priorities channel. A ranking collected online in a few minutes captures exactly the unconsidered opinion that Fishkin’s method was designed to improve on. A deliberative democrat might argue that a small, deliberating random sample would produce more meaningful priorities than a large opt-in ranking. A defender of mass participation would reply that a few hundred people, however well briefed, cannot claim to speak for millions who did not deliberate, as British Columbia showed. The seminar does not resolve the dispute. It notes that the two approaches could be combined, with each reported for what it is.

Where open participation and deliberation could meet

Hypothetical example: a city runs an open priorities survey and, in parallel, a randomly selected panel of 60 residents who meet over three weekends with balanced briefings. The open survey ranks public safety first; the panel, after deliberation, ranks housing first and safety third, citing evidence presented on how the city’s budget is divided. How should the city report this? One approach presents the panel as a correction of the survey. Another presents both as different measurements, one of quick opinion among volunteers and one of considered judgment among a random sample, and lets officials and the public weigh them. The second approach is more honest about what each measures, and it avoids the British Columbia problem of presenting a small body’s judgment as the public’s. It also exposes officials to a hard choice that neither method can make for them.

Discussion questions

1. Gutmann and Thompson say citizens owe each other “mutually acceptable reasons.” Is a ranked list of priorities a reason? What would be needed to turn it into one?
2. A Deliberative Poll shows what a sample thinks after deliberation. Should that count more than what the public thinks without deliberation? Why?
3. British Columbia’s assembly reached a considered recommendation that the wider public did not adopt. Who was right to decide?
4. In Ireland, politicians made the decisions about topics, wording, and follow-up. Does that strengthen or weaken the assemblies’ democratic credentials?
5. Landemore proposes governing through randomly selected bodies. How would you hold such a body accountable?
6. Van Swol found groups polarized after discussion. What design features, if any, would you trust to prevent that?

Assignment

Mini-public design memo (1,500 words). Design a randomly selected citizen body to advise a named local government on its budget priorities. Specify selection (including stratification and compensation), size, agenda-setting, information and expert balance, facilitation, decision rule, and how its output would relate to an open priorities channel and to elected officials. Address group polarization, the British Columbia acceptance problem, and agenda power from week 7. Label all numbers as hypothetical.

Evidence

Gutmann and Thompson, Warren and Pearse, Farrell and coauthors, and Van Swol are quoted from published abstracts or book descriptions. Deliberative Polling is described from the Stanford Deliberative Democracy Lab page. Fishkin’s design is described from Stuart Weir’s openDemocracy review. Landemore’s argument is described from the Princeton University Press page. The Análise Social description of the “law of group polarization” is from that article’s abstract. The Sunstein essay itself was verified bibliographically; no claim relies on its text. All sources were opened on the review date.

Competing interpretations

Deliberative democrats argue that structured deliberation produces better-informed and more justifiable judgments and that sortition avoids the resource biases of elections and voluntary participation. Participatory and populist critics argue that a small body cannot substitute for mass participation and that its conclusions lack legitimacy if the public rejects them. Realists in the tradition of Achen and Bartels (week 1) doubt that deliberation changes the identity-driven basis of political judgment. Social psychologists who study group polarization question whether discussion reliably improves judgment. Practitioners point to Ireland as evidence of influence and to British Columbia and other cases as evidence of its limits.

Application to this framework

My Vote Forward’s priorities and opinions are collected individually, without deliberation. This week’s readings suggest that such data should not be presented as considered public judgment, and that a framework serious about quality might pair open participation with a deliberative sample. The framework’s participation over time chapter could be evaluated on whether it allows for that pairing.

Unresolved questions

• Do structured deliberative designs reliably avoid group polarization, and under what conditions?
• What mechanisms, if any, can close the gap between a mini-public’s recommendation and the wider public’s acceptance?
• Could an open priorities channel and a deliberative sample be reported side by side without one being treated as a correction of the other?