Deliberative Citizens, (Non)Deliberative Politicians: A Rejoinder

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. Empirical research shows that both politicians and citizens have the capacity to deliberate when institutions are appropriate. Under optimal institutional conditions, politicians can score relatively high on measures of discourse quality derived from the ideals of deliberation as envisaged in Habermasian rational discourse. A good fraction of citizens can also approach these standards. Yet deliberation is not the only goal or the only desirable means in politics. Sometimes the institutions that further deliberation also undermine the democratic goods of responsiveness and accountability. And sometimes the institutions that allow citizens to deliberate at high-quality levels may not–and some argue should not–produce significant effects on policy outcomes.

Taking these possible trade-offs into consideration, we propose a “functional” approach to deliberation that takes the goals of deliberation in specific contexts more seriously and allows for a more nuanced reading of the empirical results. Such an approach does not see deliberation as a panacea for the ills of democracy. Rather, it takes a “need-oriented” perspective, asking when and where citizens and the political workings of democracy most need high-quality deliberation and when and where they need it less. Based on such a functional understanding of deliberation, 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 and simultaneously make it compatible with other democratic goods and ideals.

Before we take a stab at the deliberative potentials of politicians and citizens, we first need to charter some conceptual territory. Drawing from a common metaphor in institutional theory, we shall distinguish between “old” and “new” deliberation.¹ “Old” deliberation (frequently denoted as “classic” deliberation) incorporates the standards of rationality in argumentation, listening, reflection (weighing), respect, and “authenticity” in the sense that actors are oriented toward sincere understanding of others rather than toward strategic goal attainment. An underlying assumption in old deliberation is that the various deliberative ideals are fixed and work in tandem. This vision is “unitary” in that it assumes that all of the deliberative virtues will complement one another in a cohesive whole. It also assumes that these deliberative virtues will, in practice, produce an array of desirable outcomes, including epistemic advancement, ethical goals (such as mutual understanding and accommodating diversity), and individual transformation.

The “new” approach to deliberation that we propose takes a functional perspective, emphasizing that the various forms that deliberation can take should depend on the goals of that deliberation and the contexts in which it takes place. For instance, to reach deliberation's epistemic goals, a high level of justification rationality may be a key procedural requirement, whereas respect may play only a subordinate role. By contrast, if you want to achieve deliberation's ethical goals, respectful interactions likely play a larger role than rational argumentation.

The approach of new deliberation resembles political scholar Michael Saward's “shape-shifting” approach to representation.² Rather than thinking that deliberators play one distinct deliberative role at a time, we should understand them as creative actors who make productive and flexible use of various forms of deliberation depending on goals and context.

Finally, the new approach to deliberation assumes that deliberation cannot and should not play a major role in all stages of a democratic system.³ Although deliberation may be critical for producing epistemically sound policy or mutual agreement, it may be counterproductive for achieving other democratic goods, such as responsiveness, accountability, or consequentiality. As a result, new deliberation takes a need-oriented perspective on deliberation: rather than claiming that more deliberation is always good, it analyzes contexts and situations to determine where deliberation is most needed and functional for a particular democratic system.

Applying the approaches of the old and new to an analysis of deliberation among politicians and citizens leads to different evaluations of its viability and functionality. We start with deliberation in the realm of politics and then turn to deliberative potentials among citizens.

That “politics” is by its nature not deliberative is a common theme in the literature of political science. In a programmatic article, Ian Shapiro holds: “Enough of deliberation: Politics is about interests and power.”⁴ According to Shapiro, a deliberative reading of politics fails to consider conflicting interests and powerful players who have no incentives to deliberate, but will pursue their goals with coercive means. Deliberative democrats often have similar views about the possibilities for good deliberation in politics. James Fishkin and Robert Luskin, for instance, have argued that political elites tend to focus on negotiation rather than deliberation, so that their changes of position are the product of changing circumstances rather than the product of the better argument.⁵ The criticisms may be partially misplaced. Much of the criticism against the possibility of deliberation in politics is based on an analysis of Anglo-American politics and Westminster systems. It is easy to identify major deliberative failures in contemporary U.S. politics and in Westminster democracies, but different institutional setups–in combination with issue types and partisan strategies–may bring about higher levels of deliberative action in politics. Empirical findings from legislative deliberation underline that under appropriate institutional, contextual, and partisan conditions–namely, coalition settings, second chambers, secrecy, low party discipline, low issue polarization, and the strong presence of moderate parties–genuine deliberation is possible in parliaments.⁶ If favorable institutional and issue factors combine–that is, when a less-polarized issue is debated in a nonpublic second chamber of a consensus system with low party discipline–we find debates that resemble “ideal” deliberation with highly reasoned, respectful, reflective, and open-minded actors.⁷

Even with this more nuanced and differentiated reading of deliberation's potential in politics, a number of challenges persist. First, a deep-seated analytical challenge claims that the very nature of politics is conflict rather than cooperation through deliberation. This “adversarial” reading of politics, which has been dominant in democratic thinking since the seventeenth century, makes any claim for deliberation in politics–even if supported by empirical data–a dubious affair.⁸

Second, compounding this analytical challenge, we lack any straightforward test that might differentiate clearly between fully deliberative actions (oriented toward the common understanding of common goals) and strategic actions (oriented only toward self-interested and conflicting goals). Assume that we find a political actor who scores high on all deliberative indicators: that is, provides extensive justifications for positions and shows respect for other positions and arguments. Although these indicators may suggest deliberative action, we cannot exclude the possibility that the actor is engaging in sophisticated “rhetorical action,” intended to manipulate an audience. Put differently, until we can read minds, we will never be able to “prove” that actors were really motivated by a logic of common understanding.

Third, the real world of politics suggests that there is no unitary core of deliberation in representative politics. Consider a comparison of deliberative behavior under the public eye and behind closed doors: public debates in parliament increase justification rationality but decrease respect, while nonpublic debates increase respect but decrease justification rationality. From the perspective of old deliberation, which implies the compatibility of all the elements of good deliberation, this juxtaposition might cast doubt on the validity of the empirical findings. One might be forced to assume that political behavior under the public eye was only deliberative rhetoric, in which the pressures of publicity force the actors to produce justifications that they do not sincerely believe. Because the norms of public debate in Western democracies generally value reasoned argument but not explicit respect for the political opponent, strategic actors will use strategies that mix justification and disrespect.

Fourth, empirical research shows that classic deliberation in politics is highly context-bound, with the conditions of good deliberation (a less-polarized issue debate in a second chamber of a consensus system behind closed doors) representing relatively rare conditions. If this is the case, political deliberation would be so unlikely as to lose its real world significance. A deliberative lens on politics might be misplaced when we consider “normal conditions” of politics: namely, strong partisan competition and high issue polarization.

Fifth, the conditions for good parliamentary deliberation create a challenge for the two other democratic goods of responsiveness and accountability. As political scientist Gerry Mackie has noted: “It is worrisome that each of the discourse-improving institutions is also one that reduces accountability of representatives to the citizenry (it's harder to know who to blame in a consensus coalition, in a presidential regime, and in a system of closed meetings, and the political elite can collude against the population).”⁹ Institutionalizing more deliberative politics in legislatures seems to imply a return to “old parliamentarism,”¹⁰ with a premodern “trustee” model of representation in which politicians have loose links to their constituents and can freely change their minds on the basis of the better argument. Yet today, at least in the United States, a majority of citizens rejects the trustee model of government and instead prefers strong government responsiveness.¹¹

Taken together, these challenges seem to underscore the criticism of deliberation as even a useful ideal for politics. From a new deliberative perspective, however, these challenges appear in a different light.

First, we need to clarify the goals of deliberation in politics before we specify what qualities we want to see in it. Focusing on legislatures, political scientists Gary Mucciaroni and Paul Quirk have made a first step in this direction.¹² They claim that the key goal of parliamentary debate should be the substantive consideration of policy issues and the related informational quality of a debate, rather than the ethical goals of deliberation. In other words, in this setting, the justificatory component of deliberation trumps respectful interactions. From this vantage point, even adversarial debating–which many deliberative democrats have placed in contradiction with deliberation -might yield an epistemic function in that the audience is provided with robust reasons for competing policy goals.¹³

Yet Mucciaroni and Quirk tend to overstretch their argument. Because both the goals of legislative deliberation and their institutional and issue contexts vary, a focus on informational quality as the sole goal of parliamentary debate at the expense of ethical goals may be misplaced. In the context of a nonparliamentary consensus system–or any other negotiation setting in politics–the ethical dimensions of respect may play an important instrumental role in facilitating the negotiation process. Respectful interactions are likely to bolster cooperative attitudes among negotiation partners. Moreover, in the context of moral and ethical issues (such as abortion) or highly divisive issues (such as conflict regulation in divided societies), an exclusive focus on informational quality may be deeply misplaced. On such issues, it is hard to say that one of the principles under dispute is more correct than the other. To regulate deep conflict in divided societies, it is sometimes necessary to concentrate on what the other side can accept rather than searching for the “truth.”

Second, from the perspective of new deliberation, it is not a deficiency if we cannot fully distinguish between true deliberative and strategic action. It is neither realistic nor even desirable that politicians be oriented only toward a common understanding of the common good or comply with the full list of deliberative virtues in all venues of politics.¹⁴ Under the public eye, for instance, politicians contribute to democratic goods–responsiveness and accountability–if they debate properly rather than deliberate in a way that reflects all the deliberative virtues. The standard of quality from a deliberative vantage here requires neither the deliberative virtue of reflection nor that of respect, but rather a high level of justification rationality (or what political scientist Simone Chambers has called “robust reasoning”).¹⁵ A low level of deliberative quality under conditions of publicity involves what Chambers calls “plebiscitory reasoning” in which “arguments … become shallow, poorly reasoned, pandering, or appeal to the worst we have in common.”¹⁶ By contrast, behind closed doors, where pressures of public opinion are reduced and other governing logics, deriving from the possibility of agreement, set in, we can expect other deliberative virtues–such as some open-mindedness as well as listening and respect–to flourish more fully.

Good democratic representatives should pursue not only the common good, but also the interests of their constituents when they conflict with those of constituents in other districts or parties. If we build such conflict into our ideal of politics, we should not expect the representatives to ignore or supersede their constituents' interests in deliberation. Even in the ideal, therefore, and even when politicians are motivated to find good solutions and are open to good arguments, we should expect a mix of strategic and deliberative behavior. The deliberative quality standard here should not be rational discourse but “deliberative negotiation” in which actors justify their positions extensively with respect, but are allowed to “bargain,” constrained by fairness, by making promises while abstaining from threats and strategic misrepresentation.¹⁷ Negotiations that score low on justification and respect but high on force, threats, and strategic misrepresentation would count as less deliberative or, if there were no deliberative elements, not deliberative at all. Overall, rather than searching for fully fledged deliberative actors in politics, we should desire creative political actors who can engage in deliberation when needed and where contextually possible and appropriate. In this situation, it becomes analytically less necessary to draw strict dividing lines between strategic and deliberatively authentic political actors.

The idea of a “deliberative citizen” has been met with as much skepticism as has “deliberative politics.” Drawing from his own cases of citizen participation in deliberations, political scientist Shawn Rosenberg points out that most “participants who attend a deliberation do not, in fact, engage in the give and take of the discussion.” Rather, they “offer simple, short, unelaborated statements of their views of an event.”¹⁸ Critics have also argued that classic deliberation may be undemocratic because classic forms of deliberation discriminate against already disadvantaged persons (especially people with low socioeconomic status); different speaking styles with less classically deliberative qualities often translate into a lack of influence; and group discussion–the hallmark of any deliberative event -often triggers nondeliberative group dynamics, such as group polarization, reducing the normative value of any transforms, such as opinion change, that the deliberation may produce.

As with politics, empirical research displays a different picture when it focuses on deliberative events that are well-structured to include supportive conditions such as information provision, expert questioning, and facilitator intervention. In their analysis of a transnational Deliberative Poll (the “Europolis”)–which represents a demanding setting for citizen deliberation–political scientist Marlène Gerber and colleagues found that “the standards of classic deliberation are far from being utopian standards that only very few citizen deliberators can achieve.”¹⁹ In this context, the number of participants who both provided a sophisticated justification and engaged in respectful listening is almost 30 percent. Comparing the Europolis proceedings with parliamentary debates, the former fare quite well: although the Europolis discussion groups did not match the deliberative standards under the most ideal conditions in representative politics and also had slightly lower scores on justification rationality than in the average parliamentary debate, respect levels were significantly higher than in politics under the usual political conditions of strong partisan competition and issue polarization.²⁰ Not everything was perfect in Europolis. Gerber's team found that working-class participants from Eastern as well as Southern Europe were less apt to reach most standards of high-quality deliberation (such as justification rationality, common good orientation, and respectful listening) than other participants, raising some concerns about the democratic dimensions of deliberation among citizens with culturally heterogeneous backgrounds. However, the research team did not find any indication that the different speaking styles and cultures had an impact on influence. That is, the predeliberation opinions of the highly skilled deliberators were no more likely to affect the opinion changes in the group than were the predeliberation opinions of the less highly-skilled deliberators. Nor did the more highly-skilled deliberators impose their views on other participants. Nor, again, did these highly skilled deliberators stay stuck in their positions: they showed an almost identical amount of opinion change as the lower-skilled deliberators. Finally, the evidence suggests that opinion change in Europolis can be partly attributed to a systematic, justificatory, and argument-based component, and not to undesirable group dynamics such as group polarization: well-justified arguments seemed to affect opinion change.²¹ These findings are by no means unique. In analyzing well-structured deliberative events, several independent research teams have reached strikingly similar conclusions regarding the deliberative potential of citizens, the nonviolation of democratic standards, and the systematic and justificatory basis of opinion change.²²

Overall, these findings contradict popular assumptions regarding citizen capacity and the inevitability of undemocratic deliberative structures. Just as in legislative politics, much seems to depend on the institutional setup: if an institution is explicitly geared toward dialogue and deliberation, many well-known psychological biases tend to be reduced or wither away. Political scientist Kimmo Grönlund and colleagues, for example, varied discussion rules in an experiment on citizen deliberation on the future of the Swedish language in Finland. Their main finding was that discussion with a facilitator and deliberative norms reversed tendencies to group polarization, whereas “free” discussion without a facilitator and explicit deliberative norms–as is implemented in most psychological experiments–produced the undesired polarization patterns described by Cass Sunstein and others.

One outstanding question involves the policy impact of citizen deliberation, for example, in randomly selected “mini-publics.” Much depends on the sincerity of the intention to implement and the capacity to implement of the authorizing entity. In Canada, political scientist Genevieve Fuji Johnson has found that the interests of the authorizing bodies play a dispositive role in the implementation of the results of citizen deliberations.²³ Using Deliberative Polls, Fishkin has documented real effects on policy when the authorizers intended such effects or were extremely open to them. Yet as John Dryzek has noted: “Direct influence on and in policy making is a hard test for mini-publics to pass. While examples exist of influence and impact, they are outnumbered by cases where a mini-public is established but turns out to have little or no effect on public decision-making.”²⁴

A new study has also suggested that the more deliberative mini-publics are, the less likely they are to influence policy.²⁵ This study documents that mini-publics with low representativeness and low deliberative quality are most likely to produce important policy effects.²⁶ Yet if deliberation is not tied to decision-making, it loses its democratic character. As Mark Warren has put it: “Political processes that fail to enable this moment of constitution [that is, a system's capacity of making binding decisions] also disempower the people as a collective agent and thus undermine the normative point of inclusion and collective will formation.”²⁷

Deliberative mini-publics also raise important questions of democratic legitimacy. Cristina Lafont has argued that mini-publics reach conclusions for reasons that most ordinary voters are not likely to fully appreciate, which, in turn, creates a fundamental challenge for their legitimacy as policy-making tools.²⁸

From the perspective of new deliberation, these problems appear in a different light. A key mistake in our view is to take a “totalizing view” of citizen deliberation and deliberative mini-publics and expect that a single institution can achieve all of deliberation's goals at once.²⁹ In our view, this “unitary” vision should be replaced by a differentiated vision that takes the deliberative needs of different political contexts into account. Such a differentiated vision would specify different functions for deliberative mini-publics in different parts of any political system.

First, we need to consider which political systems can be well-served by mini-public input, and why. Political scientist Archon Fung, for example, has pointed out how patronage systems in Latin America have hollowed out their procedures of representative democracy. When competitive elections do not advance the collective goods that citizens want and need, then it is sensible to hand over policy-making activities to citizens who can produce these goods more effectively.³⁰ In another intriguing (and perplexing) example, authoritarian regimes may find that deliberative mini-publics yield distinct benefits for elites. They generate information about society and policy, co-opt dissent and maintain social order, and enable leaders to deflect responsibility onto the mini-public processes and thus avoid blame.³¹ In certain townships in China, administrators have commissioned well-structured deliberative mini-publics and then implemented their recommendations. One could imagine such processes either making later electoral democracy more possible or, to the contrary, undermining citizen demand for democracy. Similar demands among existing authorities for high-quality citizen feedback also exist in Western democracies. Baden-Württemberg in Germany is a good example. The massive protests surrounding the “Stuttgart 21” project to rebuild a train station in the central city challenged the functioning of traditional representative politics and consequently triggered several democratic innovations in order to reduce the disconnect between representative politics and citizen views. The Green-Left (now Green-Black) government introduced and institutionalized forums for citizen participation and deliberation, subsequently taking up the policy recommendations of those forums.

But not all political systems require the input of deliberative mini-publics. In the Swiss polity, extended direct-democratic mechanisms create feedback from the public. Politicians learn from both negative and positive votes in referenda, even if the exact reasoning behind the voting decisions is not always clear. Over time, this system has led to relatively good anticipations of what the “median voter” may desire, rendering additional input from deliberative mini-publics less necessary. Surely, more deliberative median voters might decide differently in direct democratic voting–and perhaps in less populist ways–compared with nondeliberative median voters.³² But if the goal of deliberative mini-publics is just more feedback to politics, then a fully fledged direct-democratic system like Switzerland might provide a sufficient route to achieve this goal.

Second, we need to consider the functions of mini-publics beyond direct policy uptake. One prominent example is a “trust-based” function, which can inform citizens' own later deliberations. Such functions do not replace citizen input, as Lafont argues, but instead supplement and inform it. The idea behind the trust-based function is that the (large) majority of non-deliberating citizens can trust the judgments of the (small) minority of deliberating citizens because that small minority, selected randomly for a deliberative mini-public, does not have to follow partisan logics of electoral representation and can focus instead on common concerns.³³ Some empirical evidence indicates that this trust-based function works in practice: the more voters knew about the randomly selected British Columbia Citizen Assembly and Irish Citizen Convention–such as their recruitment mechanisms or their freedom from partisan instructions–the more likely they were to vote for the mini-public's policy recommendation in the later citizen referendum.

Deliberative mini-publics can also function as schools of deliberation and democracy. In today's fragmented and mediatized societies, truly dialogical opportunities have become rare for ordinary citizens. Yet psychologists argue that, in the formation of considered opinions, dialogue is much more effective than simply listening to arguments.³⁴ Deliberative mini-publics enable ordinary citizens to enter into reasoned political dialogue on important questions. In the U.S. context, political scientist Lawrence Jacobs and colleagues have found that those who regularly participate in structured public discussions have a higher proclivity to connect with elites, engage in civic voluntary activities, and participate in electoral politics.³⁵ In short, even when deliberative events do not directly influence policy, they may nonetheless produce a democratic and deliberative “culture,” which–as we shall detail below–may be essential for the renewal of our contemporary political systems.

We think that deliberation brings something unique to democracy. It promotes both epistemic advancement, through argument and reasoning, and mutual understanding and accommodation among diverse actors, through respectful interaction. With these goals in mind, most past research on deliberation has taken a strong reformist perspective. By contrast, the recent systemic approach–which has had a significant effect on current thinking about deliberation and deliberative democracy–seems to have left behind these reformist goals. The systemic approach asks us to evaluate the deliberative system as a whole, suggesting that systemic mechanisms may sometimes be at work, in which components in a deliberative system may correct for each other's deliberative (and democratic) deficiencies.³⁶ Although we agree that deliberative “wrongs” can sometimes produce deliberative and democratic “rights,” we think that such correcting mechanisms are increasingly hollowed out in contemporary times. Mediatization, for instance, systematically undermines the deliberative capacities of political elites, forcing them to follow media logics and engage in “plebiscitory reasoning.” Increasing party polarization, especially in the United States, has severely reduced the potential for making respectful compromises. These developments make it even more important to think of smart interventions and reforms to existing institutional settings so that the unique contribution of deliberation to democracy can be realized.

In the political sphere, one might imagine institutional reforms toward more negotiation, similar to what we find in consensus democracies. Introducing proportional representation (PR) electoral systems, for example, makes it easier for several parties to form, which thereby forces the parties to enter into coalitions and negotiations with other parties. This requirement for negotiation, in turn, can involve deliberative elements. In recent years, negotiated systems have sparked interest among political theorists.³⁷ Denmark provides an interesting case. Here, an inclusive “negotiated” but also “authoritative” political system produces high-quality governance outputs; intriguingly (but perhaps not surprisingly), Denmark is also one of the few countries where the recommendations of one form of deliberative mini-publics (“consensus conferences”) have found their way into legislation. Overall, well-functioning negotiated systems involve much higher citizen satisfaction than competitive Westminster systems, even in times of major political crisis (as in the aftermath of the 2008 financial crisis).

A second possible institutional reform in politics might involve strengthening the bodies that Dryzek has called the “Chambers of Reflection.”³⁸ Empirical research suggests that second chambers, such as the House of Lords in England and the second chamber in Switzerland, inject a modicum of deliberation and reflection into the political process, even in times of increasing political polarization.³⁹ Elected second chambers have the advantage that they are accountable to their constituents. But a well-functioning political system will try to balance the different needs of representative politics, including both strong responsiveness and reflective deliberation. One way to achieve the combination of these two ends may be a clever division of labor between different political venues: the partisan and the reflective. If the two venues are institutionally nested (for instance, if one arena is constrained at least in part by the decisions of the other), this nesting may advance the realization of the deliberative goods of epistemic advancement and mutual understanding without bypassing the other democratic goods of responsiveness and accountability.

A third possible institutional reform is simply enriching existing political systems with a plethora of democratic innovations. This supplementary approach is happening worldwide, but had a particular grip in Latin America and in polities deeply affected by the financial crisis in 2008, such as Iceland, Ireland, and Spain. In all of these polities, various actors have started a number of deliberative initiatives, ranging from direct citizen involvement to new party architectures and new direct communication links between representatives and the represented.⁴⁰ We think that the positive effects of these innovations can be increased when institutional designers develop a higher sensitivity to context, asking in each setting what goals they want to realize and what conditions they confront. They should realize that there is no one-size-fits-all in organizing deliberative events. Depending on the goals of deliberation (such as epistemic advancement and accommodation), different forms of recruitment and communication are necessary. For instance, if we want to promote the interests of disadvantaged and disaffected groups, we should over-recruit members of such groups in order to enhance their public standing and adopt a cooperative communication format in order to enhance their deliberative influence. But if we want to achieve epistemic goals, then random selection of participations and more contestatory forms of engagement might be more effective.⁴¹

Overall, we should note that institutional reforms do not always work in straightforward ways. PR electoral systems and their resulting coalition arrangements may tend to enhance constructive negotiations, but in some conditions (like in Israel) this institutional arrangement does not conduce to cooperation. With regard to coalition systems, rational choice theorists have long argued that coalition settings entail mixed-motive games. On the one hand, coalition parties have reason to cooperate with their partners to pursue successful common policies. On the other hand, each party faces strong incentives to move policy in ways that appeal to their party members and to the constituencies on which the party relies for support.⁴² Coalition arrangements thus do not automatically produce high-quality deliberation. Rather, as empirical research has shown, the deliberative capacity of coalition arrangements is strongly affected by partisan strategies and motivations.⁴³ This powerful effect of partisan variables underlines that there is, at least at the moment, a clear limit to the effects of institutional design.

Another challenge to institutional design derives from institutional interactions and details. First, institutions are frequently nested in other institutions. For instance, the deliberative potential of second chambers is affected by the overall system architecture: if the larger system is dominated by partisan advocacy–as in Australia and the United States–then the deliberative potential of second chambers is limited. Second, institutional details matter. In coalition government, for example, recent research has shown that constructive political action may be strongly influenced by the coalition composition. Political scientists Stephen Fisher and Sara Hobolt have provided empirical evidence that when a coalition government is composed of two parties, the head of government's party is subject to greater punishment and reward from voters in their retrospective voting in the next election than the other coalition party. Conversely, when a coalition government has more than two parties, the effects of retrospective voting on any of the parties is substantially reduced. Accountability seems to be reduced because it is harder to know whom to blame. On the other hand, the space for deliberative engagement may be enhanced, since it is more difficult for parties to use political successes for partisan electoral advantages. Details also matter for deliberative mini-publics. Political scientist Lucio Baccaro and colleagues have shown that an institutional detail, such as asking participants to justify their positions before making a choice, can have major effects on the dynamics of deliberative process and the subsequent outcomes.⁴⁴

In short, we argue for a new contextual and functional approach to deliberation. Both electoral politics and citizen participation can become more deliberative, often without undermining other democratic values. What is needed, what will work, and what innovations make sense depend on the details and the context.