Deliberative democracy and the public sphere

Intuition Beginner

Suppose your city must decide whether to build a new highway. You could put it to a referendum: count the yes and no votes, and the larger side wins. That settles the question, but it leaves something out. The yes voters and the no voters never had to explain themselves to each other. Deliberative democracy begins from the feeling that this missing piece matters. A decision is more legitimate, on this view, when the people it affects have reasoned together, heard the opposing case, and had a real chance to change their minds.

The model rests on a distinctive claim. Preferences are not fixed inputs to politics, the way weights are inputs to a scale. People form and revise what they want through discussion with others. Hear why your neighbour dreads the highway, or why a shop owner welcomes it, and your own view may shift. Democracy, the deliberative theorist argues, should track that process: a collective choice deserves respect when it could be defended in a discussion open to everyone affected.

Why care about this extra layer? Because pure vote-counting can ratify ignorance, prejudice, and raw interest. A majority that has never heard the strongest case against its position has not really chosen. Deliberative theory supplies a standard for telling considered choice from mere aggregation. Its twentieth-century architects, Habermas and Rawls, gave the model its philosophical depth; later theorists, from Joshua Cohen to Chantal Mouffe, sharpened it, tested it, and in Mouffe's case attacked it. This unit follows that argument in detail.

Visual Beginner

The diagram captures the architecture Habermas calls the two-track model. Below sits the informal public sphere, the open sea of citizens, media, associations, and argument in which opinions form. Above sit the formal bodies that make binding law. The deliberative claim is that legitimacy flows upward from reasoned exchange below and back downward as law. The whole system depends on the filter working: that what rises is genuinely argument rather than money, propaganda, or unexamined passion.

Worked example Beginner

Consider the Irish Citizens' Assembly, which met from 2016 to 2018. The government convened ninety-nine members plus an independent chair, chosen by stratified random selection to mirror the adult population by age, sex, region, and social class. The assembly took up five questions across thirteen weekends, the most famous being the Eighth Amendment to the constitution, which had in effect banned abortion. At each session, members heard from legal experts, doctors, and advocates on all sides, then debated in small moderated groups and questioned witnesses directly. Finally they voted, recommending repeal of the amendment and a specific replacement framework.

What changed? Opinion moved. Members entered with mixed and often firm views; a substantial share left supporting repeal. The recommendation went to parliament, which referred the question to a national referendum. In May 2018 voters repealed the amendment by a two-to-one margin. The assembly did not make policy alone, but it reshaped what was politically possible. A subject long treated as untouchable became debatable in public because ninety-nine ordinary citizens had done the work of listening and arguing first.

What this tells us. The case shows the deliberative ideal operating inside a representative constitution rather than replacing it. Random selection supplied equality of access; balanced briefing supplied information; moderated discussion supplied the give-and-take of reasons; the referendum preserved final popular control. The assembly did not eliminate conflict. It channelled a deep moral disagreement through a procedure the losing side could recognise as fair. That is the legitimacy claim deliberative theory rests on: procedure, not unanimity, earns obedience.

Check your understanding Beginner

Formal definition Intermediate+

Deliberative democracy holds that a collective decision is legitimate insofar as it would be the object of a free and reasoned agreement among equals. The qualifier matters. The model is not the claim that citizens must literally assemble and argue before every vote. It is a standard against which actual institutions are measured: the closer a political process approximates unconstrained public reasoning, the more authority its outputs deserve ^{[Cohen 1989]}.

Three theorems of political legitimacy follow this standard in different ways, and the differences are load-bearing.

Habermas: discourse principle and the public sphere. Habermas grounds legitimacy in the discourse principle (D): only those norms are valid that could meet with the agreement of all affected parties in a practical discourse. A practical discourse is a conversation regulated by the ideal speech conditions -- symmetry among participants, inclusiveness of all affected, freedom to introduce any assertion or doubt, and sincerity of utterance ^{[Habermas 1992]}. These conditions are never fully realised; they function as a counterfactual standard. The public sphere is the social space in which such discourse is supposed to occur: a network of citizens, media, associations, and fora in which opinions form before they harden into votes. Habermas distinguishes communicative action, oriented toward mutual understanding, from strategic action, oriented toward success; legitimacy tracks the former.

Rawls: public reason. Rawls proposes a different constraint. Citizens in a constitutional democracy hold conflicting comprehensive doctrines -- religious, moral, philosophical. The fact of reasonable pluralism is permanent; it cannot be wished away and ought not be suppressed. Citizens therefore owe one another public reason: when debating constitutional essentials and matters of basic justice in the public political forum, they should appeal to genuine political values intelligible to all, not to the authority of their own comprehensive doctrine ^{[Rawls 1993]}. Comprehensive reasons may enter, Rawls's proviso holds, provided public reasons are given in due course. The model is procedural in spirit but anchored to a shared political vocabulary rather than to the ideal speech situation.

Cohen: ideal deliberative procedure. Joshua Cohen formalises the deliberative model as an ideal deliberative procedure whose conditions any legitimate association must approximate: the procedure is the ongoing business of an independent association; its members are committed to settling decisions by reason; they are plural, with distinct preferences and convictions; participation is free and equal, with no participant coercively bound by another; and the deliberation aims at a rationally motivated consensus. On Cohen's account a decision is legitimate iff it would be the outcome of such a procedure ^{[Cohen 1989]}.

Procedural versus substantive. The accounts divide on where legitimacy is seated. A purely procedural theory says the outcome is right because it issued from a fair procedure, whatever its content. A substantive theory says the procedure is justified because it tends to reach outcomes that are independently right. Deliberative theory refuses the clean opposition: it is procedurally grounded in source -- legitimacy comes from the procedure, not from an external measure of justice -- yet the procedure's constraints (inclusion, equality, non-coercion, reason-responsiveness) guarantee a substantive floor below which outcomes cannot fall without violating the procedure itself. Estlund calls this hybrid epistemic proceduralism: a procedure is legitimate when it is epistemically the best among those that are publicly justifiable ^{[Estlund 2008]}.

Counterexamples to common slips

• "Deliberative democracy demands actual consensus." No. It demands that decisions be justifiable to those bound by them. Rawls's overlapping consensus permits citizens to endorse the same arrangement for different reasons; Habermas distinguishes a rationally motivated agreement from a mere compromise. Dissenters remain bound because the procedure was fair, not because they were persuaded.

• "Public reason bans religion from politics." No. Rawls's proviso allows comprehensive reasons, including religious ones, to be introduced provided public reasons are eventually offered. The constraint is on what justifies coercive law, not on what motivates individual citizens.

• "Deliberation is just rational talk; power does not enter." This misreads the ideal as a description rather than a standard. Habermas's ideal speech conditions exist precisely to expose where real discourse deviates through coercion, money, and propaganda. The point of the model is to diagnose such deviations, not to pretend they are absent.

Key argument -- Cohen's ideal deliberative procedure Intermediate+

Cohen's argument in "Deliberation and Democratic Legitimacy" (1989) is the canonical reconstruction of why deliberation, rather than voting alone, grounds democratic authority ^{[Cohen 1989]}. The argument runs from the nature of a democratic association to a constraint on its decision procedure.

Premise 1 (Association). A democratic association is one in which members are committed to settling their collective affairs through reasoning with one another rather than through force, tradition, or the private aggregation of interests.

Premise 2 (Pluralism). Members hold distinct preferences, convictions, and ideals; no member's view is privileged in advance.

Premise 3 (Freedom and equality). The conditions of deliberation are free (no member is coercively bound by another and all have equal access to the agenda) and equal (no participant enjoys binding authority over others).

Premise 4 (Reason-responsiveness). Outcomes are settled by the giving and weighing of reasons, where a reason is a consideration others can recognise as such, not a mere statement of interest.

Conclusion (Legitimacy). A binding collective decision is legitimate iff it would be the outcome of an ideal deliberative procedure satisfying Premises 1-4, that is, iff it could be the object of a free and reasoned agreement among equals.

The argument's force lies in the conditional. Cohen does not claim that real deliberation always reaches consensus, nor that citizens must personally participate in every decision. The claim is that legitimacy is a relational property between an outcome and a procedure: the outcome is authoritative because of the procedure that would produce it, and the procedure earns its authority from the four conditions.

This is the move that lets deliberative theory answer the standing objection from social choice. Where pure preference aggregation faces Arrow-type impossibility -- no rule converts arbitrary preference profiles into a coherent collective ranking -- the deliberative model restricts the domain. Under genuine deliberation, preferences are not fixed inputs; they are revised in light of the reasons others offer. The profiles that generate Arrow's cycles are, Cohen argues, precisely the profiles that survive deliberation less often. Legitimacy is not extracted from exogenous preferences; it is produced by the procedure that shapes them.

The argument also fixes the procedural-substantive position. Cohen's procedure is procedural in source: nothing is built in that says which policy is correct. But it is substantive in content: because every affected party must be included as an equal, no outcome can ride on excluding some from the conversation. Equality, freedom, and reason-responsiveness are not outputs of the procedure; they are its entry conditions. The hybrid is exactly what later theorists -- Estlund's epistemic proceduralism, Habermas's co-originality of private and public autonomy -- refine and extend.

Bridge. Cohen's procedure builds toward the systemic question taken up in the Master tier: how does a single ideal forum scale to a mass society of millions who can never meet? The procedural-substantive hybrid appears again in the debate over whether deliberative systems carry substantive functions -- epistemic, ethical, democratic -- that no isolated mini-public can perform alone. This is exactly the tension Habermas diagnosed through the two-track model and the structural transformation of the public sphere; the central insight -- that procedure, not unanimous outcome, grounds authority -- generalises from Cohen's ideal forum to Dryzek's deliberative system, and the bridge is the claim that legitimacy tracks not a single conversation but an entire ecology of reasoning kept open across a polity.

Exercises Intermediate+

Advanced results Master

The Intermediate treatment of Cohen's ideal procedure leaves three problems open, and the contemporary literature is organised around them. The first is genealogical and sociological: where does the public sphere come from, and what undermines it? The second is systemic: how does a single ideal forum scale to a mass society? The third is empirical: when deliberative institutions are actually built, what do they reveal about the model's presuppositions?

The structural transformation of the public sphere

Habermas's earliest contribution, The Structural Transformation of the Public Sphere (1962), is not yet the discourse-ethics theory but its precondition ^{[Habermas 1962]}. The work reconstructs the rise of the bourgeois public sphere in eighteenth-century Britain, France, and Germany -- the coffee houses, salons, literary journals, and newspapers in which private persons assembled as a public to subject art and, increasingly, politics to the judgment of critical reason. Its constitutive fiction was that authority could be challenged not by appeal to rank or revelation but by appeal to arguments anyone could assess. Habermas treats this as a genuine achievement and a limited one: the bourgeois public was always restricted by property, gender, and education, even as it universalised its own normative claim.

The "transformation" is the decline. In late-nineteenth- and twentieth-century mass democracy, the public sphere was restructured by the commercial mass press, public relations, organised interest groups, and broadcast media. The citizen-as-reasoner gave way to the citizen-as-consumer; the exchange of arguments gave way to the management of opinion. Habermas names this refeudalization: the public sphere comes to resemble the representative display of pre-modern absolutism, in which authority stages itself before a public rather than submitting to its judgment. The normative core survives -- the standard by which refeudalization is diagnosed is precisely the ideal of critical-rational exchange -- but the institutions that once approximated it have eroded.

The structural-transformation thesis is the source of the recurring worry that deliberative democracy presupposes a sociological condition -- a functioning public sphere -- that modern societies do not reliably possess. Empirical work on fragmentation, polarisation, and platform-mediated discourse is continuous with this diagnosis, though the causal story is contested.

From forum to system: the deliberative systems turn

If Cohen's ideal procedure describes a single forum, the question of how it scales to millions drove the "deliberative systems" turn (Mansbridge et al. 2010, 2012; Dryzek 2010) ^{[Dryzek 2010] [Mansbridge 2012]}. The central claim is that deliberation should be analysed at the level of the whole polity, not the single institution. A deliberative system is an ecology of sites -- parliaments, courts, mini-publics, social movements, media, everyday talk -- that, taken together, perform deliberative functions no single site performs alone.

Mansbridge and colleagues identify a set of functions a healthy system performs: an epistemic function (it processes information and reasons), an ethical function (it promotes mutual respect and the reflection of preferences), and a democratic function (it inclusively distributes voice and resists domination). The systemic move reframes several standing problems. Apparent non-deliberation in one site -- a protest slogan, a judicial ruling -- may contribute to deliberation in the system as a whole. The test shifts from "did this forum deliberate well?" to "does the system, across its sites, sustain deliberative functions over time?"

The turn is not without cost. Systemic claims are harder to falsify: if any local failure can be redeemed by appeal to system-level function, the standard risks becoming elastic. Dryzek, who develops the most developed systemic account in Foundations and Frontiers of Deliberative Democracy (2010), responds that the test is comparative and consequentialist -- a system is more or less deliberative, and the standard gains bite from the gradients it reveals ^{[Dryzek 2010]}. The unresolved question is whether systemic deliberation can specify the mechanisms by which diffuse public opinion in the informal sphere actually binds formal decision-making, rather than merely coexisting with it.

The empirical record of mini-publics

Deliberative mini-publics -- citizens' assemblies, deliberative polls, planning cells, consensus conferences -- are the most developed institutional expression of the deliberative model, and Fishkin's programme has produced the most systematic evidence ^{[Fishkin 2009]}. A deliberative poll convenes a stratified random sample, briefs it with balanced materials, exposes it to small-group moderated discussion and questioning of experts and politicians, and then surveys opinion before and after. The design isolates deliberation as a variable: what changes when citizens are given information, time, and the chance to argue?

The consistent empirical finding is that opinion does change, often substantially, and that participants report higher political efficacy, knowledge, and tolerance. Fishkin interprets this as evidence of what the public would conclude under favourable conditions. Citizens' assemblies -- British Columbia (2004) on electoral reform, Ireland (2012, 2016-2018) on constitutional questions, France's Convention Citoyenne pour le Climat (2019-2020), the UK Climate Assembly (2020) -- show the design applied to binding or quasi-binding decisions, with the Irish abortion sequence the clearest case of a mini-public reshaping a national outcome.

Three criticisms structure the empirical debate. First, the scaling problem (Exercise 3): the deliberation of a few hundred people may not extrapolate to a mass public. Second, the influence gap: mini-publics recommend, but governments decide whether to listen, and recommendations are frequently diluted or ignored -- the French climate convention's outputs were substantially modified before legislation. Third, internal dynamics: deliberation can produce polarisation, groupthink, and moderator effects (Sunstein on enclave deliberation), which qualify the uniformly positive early findings. The literature has moved toward design research: which features of selection, briefing, and facilitation produce robust rather than fragile opinion change?

Synthesis. The three results cohere around the central insight that deliberation is irreducibly procedural yet irreducibly systemic: Cohen's ideal procedure generalises from a single forum to Mansbridge's ecology of sites only because the substantive entry conditions -- equality, inclusion, reason-responsiveness -- must hold somewhere in the system if they hold nowhere. Putting these together, the structural-transformation thesis builds toward the contemporary diagnosis that mass democracies may possess the formal machinery of deliberation while lacking the public sphere that supplies it with reasoned inputs; the bridge is that legitimacy, on the deliberative account, is not a property of a vote but of an entire process of public reason kept open across a polity and across time, and this is exactly what mini-publics attempt to restore when the ambient public sphere has decayed.

Full proof set Master

The signature results of the deliberative model are reconstructive arguments rather than theorems in the mathematical sense. The four propositions below fix the logical structure of the model's central claims and expose their points of vulnerability.

Proposition 1 (Legitimacy criterion). Under Cohen's procedure, a binding collective decision is legitimate if and only if it would survive an ideal deliberation among free and equal affected parties.

Argument. Let $O$ be a binding decision and let $D$ denote an ideal deliberative procedure satisfying the association, pluralism, free-and-equal, and reason-responsiveness conditions. Cohen's theory identifies the normative predicate legitimate with the relational predicate would-be-the-outcome-of$(O,D)$. The forward direction holds because, given Premises 1-4 of the key argument, authority derives solely from the procedure; if $O$ would not survive $D$, no independent source of authority remains to legitimate it. The reverse direction holds because the procedure is the only legitimating source: whatever $D$ would produce is, by construction, legitimate. The biconditional follows because the two predicates are stipulated as co-extensive by the theory. The result is non-vacuous only because the procedure's entry conditions constrain the admissible $O$: outcomes incompatible with free-and-equal standing cannot be the output of $D$, which is why the criterion is procedural in source yet substantive in content. $\square$

Proposition 2 (Procedural-substantive floor). Cohen's procedure guarantees a substantive floor: no outcome that violates the equal-standing condition can be legitimate, regardless of the size of the supporting majority.

Argument. Let $O$ be a decision that subordinates a category of participants -- excluding them from the agenda, denying them standing to challenge reasons, or binding them coercively without justification. By the free-and-equal condition, such participants are parties to the ideal deliberation. By reason-responsiveness, every party may challenge any consideration. An $O$ that depended on their exclusion could not be reached through a procedure in which they participate as equals; their continued objection, if reason-grounded, cannot be overridden within the procedure without violating the condition that authorises it. Therefore $O$ is not the would-be outcome of $D$, and by Proposition 1 it is not legitimate. Majority size is irrelevant because the procedural predicate fails before aggregation. $\square$

Proposition 3 (Rawls-Habermas convergence). Rawls's public reason and Habermas's discourse ethics impose the same justificatory constraint -- that coercive norms be defensible to all those subject to them -- despite their different foundations.

Argument. Habermas's discourse principle (D) declares a norm valid iff all affected could agree to it in a practical discourse under ideal speech conditions. Rawls's public reason requires that constitutional essentials and matters of basic justice be justifiable by political values all reasonable citizens can recognise. Both impose a universality test on justification: a reason that could not in principle be accepted by those bound by the norm cannot ground its legitimacy. The difference is foundational: Habermas derives the test from the presuppositions of communicative action, Rawls from the conditions of fair cooperation under reasonable pluralism. The convergence is at the level of the constraint, not its grounding. Where they diverge is scope and mechanism: Rawls scopes the constraint to constitutional essentials and polices the content of public justification via the proviso; Habermas extends it across normative claims and polices the procedure of discourse. The shared structure explains why each can incorporate the other's insights -- Habermas's translation requirement echoes Rawls's proviso, and Rawls's overlapping consensus echoes the rationally motivated agreement Habermas opposes to mere compromise. $\square$

Proposition 4 (Mouffe's residual). Even under fully satisfied ideal speech conditions, a rationally motivated consensus does not entail the absence of conflict; residual antagonism is irreducible.

Argument. Let $D$ be an ideal deliberation satisfying symmetry, inclusiveness, freedom, and sincerity. Mouffe's claim, drawing on the Schmittian thesis that collective identities are constituted through relation to an adversary, is that the formation of a "we" presupposes a "they." Suppose $D$ reaches a consensus on norm $N$. The consensus shows that the parties converged on $N$; it does not show that the identities, affects, and interests that structured the deliberation have dissolved. A new question -- or a shift in the constitution of the collective subjects themselves -- can re-expose antagonism that the consensus on $N$ suspended but did not eliminate. Therefore consensus on $N$ is compatible with the persistence of conflict at the level of the political. The deliberative model's reply, available within its own resources, is that Proposition 4 does not refute Proposition 1: legitimacy tracks the procedure for this decision, not the permanent pacification of all disagreement. Mouffe's stronger claim -- that the consensus ideal is undesirable because it suppresses legitimate conflict -- is a normative thesis about what democracy is for, not a refutation of the procedural criterion. $\square$

Connections Master

• Democratic theory: participation, deliberation, and representation 20.07.01 is the direct prerequisite and the survey within which this unit deepens one strand. The overview distributes attention across representative, participatory, deliberative, radical, and epistemic families; the present unit isolates the deliberative family and develops its full architecture -- Habermas's public sphere, Rawls's public reason, Cohen's procedure, the systemic and empirical turns -- at a depth the overview cannot accommodate. The distinction between procedural and substantive legitimacy introduced there is resolved here into epistemic proceduralism.

• Political philosophy foundations 20.14.01 supplies the conceptual background against which the deliberative model is defined: the distinction between legitimacy and justice, the problem of authority and consent, and the standing of pluralism. The deliberative turn is best read as a specific answer to the general question of what makes coercive collective power rightful, and Cohen's procedure is intelligible only against the contractualist and procedural-justice framework established there.

• Theories of justice: Rawls, Nozick, and fairness 20.02.01 connects through Rawls directly. Public reason and the overlapping consensus are the political-liberal completion of the theory of justice as fairness, extending the original position's fairness constraint from the choice of basic principles to the justification of coercive law under actual pluralism. The Rawls of Political Liberalism presupposes the Rawls of A Theory of Justice; the deliberative reading of public reason is one of the principal lines of interpretation of that later work.

• Epistemology: knowledge, justification, and truth 20.01.01 connects through the epistemic dimension of deliberation. The claim that public reason tends toward better outcomes than preference aggregation is an epistemological claim about the conditions under which collective judgment is reliable, and it presupposes the treatment of justification, testimony, and rational revisability developed in the epistemology unit. Estlund's epistemic proceduralism, which licenses deliberative legitimacy by its epistemic properties, is the explicit bridge.

Historical & philosophical context Master

The deliberative turn is a late-twentieth-century event. Through the 1950s and 1960s democratic theory had moved toward two dominant positions: the empirical-realist model of Schumpeter and Dahl, which defined democracy as competitive elite struggle for the popular vote, and the aggregative model of social choice, which treated votes as fixed preference signals to be processed by a rule. Neither assigned intrinsic importance to public reasoning. Habermas's Strukturwandel der Offentlichkeit (1962) ^{[Habermas 1962]} and the Theory of Communicative Action (1981) ^{[Habermas 1981]} prepared the alternative -- a theory in which validity, including political validity, is constituted through unconstrained communication -- but the normative programme was not yet articulated as a theory of democracy.

The articulation came in the late 1980s. Joshua Cohen's "Deliberation and Democratic Legitimacy" (1989) ^{[Cohen 1989]}, presented at the workshops that produced the first collections on the topic, gave the model its canonical formal statement: the ideal deliberative procedure as the source of legitimacy. The same year, the first English translation of Habermas's 1962 book made the public-sphere thesis available to a broad Anglophone audience and crystallised research on its historical and empirical dimensions. Bernard Manin, Elster, and the contributors to the 1989 and 1997 collections consolidated "deliberative democracy" as a self-conscious research programme.

The high statements of the position followed in the early 1990s. Habermas's Faktizitat und Geltung (1992; Between Facts and Norms, 1996) ^{[Habermas 1992]} offered the most comprehensive philosophical defence, integrating discourse ethics with a sociology of law and the two-track model of informal public sphere and formal decision-making bodies. Rawls's Political Liberalism (1993) ^{[Rawls 1993]} introduced public reason as the constraint appropriate to pluralist constitutional democracy, locating deliberation within political liberalism rather than discourse ethics.

The position was immediately contested. Iris Marion Young's "Communication and the Other" (1996, expanded 1997) ^{[Young 1997]} argued that the deliberative ideal's privileging of orderly rational argument systematically excluded forms of expression -- testimony, greeting, rhetoric, narrative -- associated with marginalised groups, and that inclusion required broadening the communicative forms that count as deliberation. Mouffe's The Democratic Paradox (2000) ^{[Mouffe 2000]} pressed the agonistic critique from post-structuralist premises: the political is constitutively conflictual, and the consensus ideal mistakes the management of antagonism for its rational overcoming.

The empirical and systemic phases followed. Fishkin's Democracy and Deliberation (1991) and When the People Speak (2009) ^{[Fishkin 2009]} operationalised deliberation in the deliberative poll and produced the first systematic evidence of opinion change under deliberative conditions. Dryzek's Deliberative Democracy and Beyond (2000) and Foundations and Frontiers of Deliberative Democracy (2010) ^{[Dryzek 2010]} absorbed the critiques and moved toward a discursive-systems account, and the Mansbridge-led deliberative-systems framework (2010, 2012) ^{[Mansbridge 2012]} shifted the unit of analysis from the single forum to the polity-wide ecology of deliberation. Estlund's Democratic Authority (2008) ^{[Estlund 2008]} fixed the procedural-substantive hybrid as epistemic proceduralism. The Irish citizens' assemblies of 2012-2018 made the model a live instrument of constitutional reform rather than a purely theoretical proposal.

Bibliography Master

Foundational and primary:

1. Habermas, J. — The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society (MIT Press, 1989; German orig. Strukturwandel der Offentlichkeit, 1962).

2. Habermas, J. — The Theory of Communicative Action, 2 vols. (Beacon, 1984, 1987; German orig. 1981).

3. Habermas, J. — Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy (MIT Press, 1996; German orig. 1992).

4. Rawls, J. — Political Liberalism (Columbia University Press, 1993; expanded ed. 2005).

5. Cohen, J. — "Deliberation and Democratic Legitimacy", in A. Hamlin & P. Pettit (eds.), The Good Polity (1989); repr. in J. Bohman & W. Rehg (eds.), Deliberative Democracy (MIT Press, 1997).

Deliberative theory and systems:

6. Dryzek, J. — Deliberative Democracy and Beyond: Liberals, Critics, Contestations (Oxford University Press, 2000).

7. Dryzek, J. — Foundations and Frontiers of Deliberative Democracy (Oxford University Press, 2010).

8. Mansbridge, J., Bohman, J., Chambers, S., Dryzek, J., et al. — "A Systemic Approach to Deliberative Democracy", in J. Mansbridge & J. Parkinson (eds.), Deliberative Systems (Cambridge University Press, 2012).

9. Gutmann, A. & Thompson, D. — Democracy and Disagreement (Harvard University Press, 1996); Why Deliberative Democracy? (Princeton University Press, 2004).

10. Elster, J. (ed.) — Deliberative Democracy (Cambridge University Press, 1998).

Critique and agonism:

11. Mouffe, C. — The Democratic Paradox (Verso, 2000).

12. Young, I. M. — "Communication and the Other: Beyond Deliberative Democracy", in S. Benhabib (ed.), Democracy and Difference (1996); revised in Intersecting Voices (Princeton University Press, 1997).

13. Estlund, D. — Democratic Authority: A Philosophical Framework (Princeton University Press, 2008).

Empirical and institutional:

14. Fishkin, J. — When the People Speak: Deliberative Democracy and Public Consultation (Oxford University Press, 2009).

15. Goodin, R. E. & Spiekermann, K. — An Epistemic Theory of Democracy (Routledge, 2018).