Flexibility, Argumentation and Confrontation. How Deliberative Minipublics Can Affect Policies on Controversial Issues

Over the last few decades, deliberative minipublics have been used as inclusive and innovative practices to integrate traditional policy-making. Because of their policy-oriented aim, but also owing to the usually not legally-binding nature, some scholars have recently pointed out the importance of understanding how and to what extent they actually manage to influence the decisions of public authorities, especially when they deal with highly controversial issues, such as locally unwanted land use or ethic matters. This article has the aim of offering a contribution in this direction, by presenting the findings of a comparative analysis of two deliberative processes promoted by public authorities to integrate decision-making in order to deal with highly controversial issues. The two processes had significantly different impacts on the final policies. The analysis, conducted by means of ethnographic and qualitative methods, has shown that how the minipublics were designed and how deliberation was put in practice were crucial aspects in determining the effective influence of the minipublics on policy decisions. In particular, the use of a strictly deliberative design on such contentious issues has proved to be counterproductive, because it has generated distortions and produced institutional conflicts, while the use of a hybrid path, although not free from problems, has proved to be a more effective solution.


Introduction
The literature on deliberation already counts more than twenty years of contributions and reflections, ranging from theoretical issues to empirical inquiries and normative propositions.
Although partially different theoretical conceptions have been identified 1 , the different nuances still point out a shared view on the conditions necessary for a fair democracy: dialogue as the essential foundation of any democratic process, shared knowledge that integrates diverse viewpoints as the necessary precondition for discussion, and mutual justification as the main rule for a constructive outcome of the dialogue (Mansbridge, et al. 2010;Cuppen, 2012;Steiner, 2012).
The second approach looks at deliberation as a particular "democratic innovation" (Smith, 2009), to integrate traditional decision-making through advanced forms of citizen involvement in dialogical processes, based on the previously mentioned principles of shared knowledge, dialogue and diversity of viewpoints and arguments. This perspective focuses on what is called structured deliberation (Hamlett & Cobb, 2006), which takes place in ad hoc mini-publics (Fung, 2003;Dryzek, 2010;Moore, 2012) involving lay citizens in structured discussions on a particular public decision, 2 with the support of professional facilitators who design the processes and lead the discussions (Moore, 2012;Steiner, 2012). 1 Hendriks (2006) distinguishes between "micro" deliberation, where actors tend to use primarily well argued statements, and "macro" deliberation, where discourses are articulated through several forms of communicative action. From a partially different point of view, Baechtiger, et al. (2010) refer to "type I deliberation" as an ideal process wherein actors of equal status and resources discuss rationally, sincerely justify their opinions using mutually acceptable arguments, and are willing to reach a common understanding and to be persuaded by the best arguments, while "type II deliberation" would be a more pragmatic conception that includes several forms of discussion characterized by dialectics among discourses. In this perspective, for example, not only sincere argumentations, but also some kinds of strategic uses of arguments are allowed (Mansbridge, et al. 2010), and arguments are not only the complex ones defined by Habermas (1981), but also motivations expressed through storytelling or metaphors (Mansbridge, et al. 2010;Nanz, 2010;Steiner, 2012). 2 Partially recalling Dahl's definition, Fung (2003) defines mini-publics as highly artifactual efforts to involve citizens in the public sphere through the inclusion of different voices, the provision of information material and the facilitation of structured dialogues.
The article falls into this second strand, and it has the goal of addressing the issue of the influence of structured deliberation on highly controversial policy decisions that directly affect citizens and tend to trigger social conflicts. The matter is relevant in a two-fold sense. First, since deliberative processes are currently not legally-binding for representative policy-makers, the matter of their real impact on policy decisions is fundamental, if these processes "are to be considered something more than just a human laboratory" (Font and Blanco, 2007, 580). Second, controversial policies represent one of the most interesting and at the same time challenging fields of practice for deliberative democracy. If one of the main questions concerning deliberative mini-publics is "in what kinds of policy domains are such activities appropriate" (Fischer, 2013, 104), according to Gutmann (1996, 344) "decision-making by deliberation among free and equal citizens is the most defensible justification anyone has to offer for provisionally settling controversial issues." But controversial issues seem to pertain to a policy domain that is particularly unfavorable for the application of the deliberative ideal: such issues usually raise strong social protests, the positions of the actors often seem inconcilable and the decisional processes may result in a state of stalemate or in authoritarian impositions by the political institutions. The few participatory processes in this field have mostly been designed using consensus building techniques (Susskind, et al. 1999) or referenda (Bherer & Breux, 2012).
The article addresses the following question: when a deliberative mini-public is adopted by an institutional authority to contribute to a highly controversial policy decision, what makes the mini-public effective? In order to deal with the question, a comparative analysis has been conducted concerning two deliberative processes on land use in which high concentrated costs for the residents and dispersed benefits are involved. 3 The processes took place in Italy between 2009 and 2011. These cases had a very different impact on the final decisions, one being quite problematic and the other somewhat successful. The analysis has shown that their diverse performances stemmed in particular from their different designs: the strictly deliberative process has proved to be counterproductive, producing an uncertain efficacy and generating controversies among institutions; the hybrid format, although not free of problems, has been more effective and has not generated controversies among the institutions. Ethnographic and qualitative methods have been adopted, through in-depth participatory observation of the processes, interviews with the protagonists and actors with a privileged view on the decisional processes, and content analysis on the speech acts of the dialogic sessions. 4 The article is organized as follows. The literature on structured deliberation in policy-making is built on in the next section, in order to develop the analytical framework. Section two offers a synthetic description of the two cases. The findings of our research are explained in the third section. The main arguments in support of the analytic explanation are summarized in the conclusions.

Structured Deliberation in Policy Making
Structured deliberation in policy-making takes place through "highly artificial" processes (Fung, 2003, 338) designed and managed by skilled facilitators to create arenas in which lay citizens contribute to a public decision through informed and argument-based dialogues. The public authorities that are responsible for a decision usually make an informal promise to take into serious consideration the policy proposals provided by the participants. From a not too ambitious perspective, it can be stated that a deliberative process affects policy decisions when public authorities implement the proposals of the mini-public or when, although not actually implementing them, public authorities justify their refusal and also show they explicitly recognize the validity of some of the arguments proposed by the deliberants.
The variables that affect the efficacy of deliberative as well as participatory processes are probably numerous (Font & Smith, 2014). So far, empirical studies have suggested that at least three factors could favor the efficacy of these processes on (more or less controversial) policy decisions: a serious commitment of the political authority, the timing of the process and the active interest of civil society in the issue at stake.
In their study on 54 participatory processes on environmental issues in the Great Lakes region (Canada), Beierle and Konisky (2000) pointed out that the capacity to influence policy decisions mainly depended on the strong commitment of the institutional authorities who were responsible for the final decisions. Political commitment was also an important precondition in the case of the deliberative process on the new Charter for the city of Chelsea, in Massachusetts (Podziba, 2006), in some Spanish citizen juries (Font & Blanco, 2007) and in some participatory budgeting in Brazil (Abers, 2000;Baiocchi, 2003). This commitment may be publicly unexpressed, although in some cases public authorities make explicit promises and provide a clear and quantifiable amount of financial resources devoted to the process (Baiocchi, 2005;Nylen, 2003).
The timing of the process, that is, its start at an early policy stage in which alternative options are still available, would probably have increased the almost null efficacy of some processes. In the case of a public works decision in Oxford (UK), the participatory process did not affect the final decision of the City Council, because the institutional authority consulted citizens when the project had already been defined and significant alternatives were not really possible, making the stakes de facto non-existent (Brownill, 2009). Holzinger (2000), analyzing the reasons for the inefficacy of a participatory process concerning the policy of waste management in Neuss (Germany), stressed that the failure in part depended on the lack of a clear degree of freedom of the participatory process in the official policy-making cycle. Gauthier and Simard (2007) have emphasized this factor and highlighted that the major weakness of the citizens' involvement in environmental decisionmaking promoted by the Bureau d'Audiences Publiques sur l'Environment in Quebec was the late timing of the processes, which severely limited their possibility of changing the projects under discussion.
Finally, some studies on participatory budgeting have shown that the presence of a local civil society that is willing to actively participate in public decisions favors the adoption of citizen policy proposals by public officials (Abers, 2000;Baiocchi, 2003Baiocchi, , 2005. The activism of groups and citizen committees acts as a factor of external stimulus and pressure on the commitment of the political authorities (Nylen, 2003).
As pointed out in the aforementioned studies, the capacity of a participatory process to influence policy decisions is affected by contextual factors. However, when a process is carefully designed and facilitated, as deliberative processes are, how it is structured may also be relevant in terms of increasing or reducing its capacity to influence policy decisions (Font & Smith, 2014), especially when these decisions are controversial and likely to raise conflict and protests.
The article will show that the way in which two deliberative processes were designed affected their capacity to influence policy decisions. It will be argued how this happened, by explaining the mechanisms generated by the design choices.

The Cases
The presented cases are two structured deliberation processes that were financed and promoted by local institutions in order to deal with locally unwanted land uses (LULUs). 5 The former concerned a project on a new stretch of highway in the city of Genoa, and it took place in 2009, while the latter was organized almost two years later and was about a project on a new pyro-gasifier for industrial waste disposal in a small Tuscan town. The conflictual nature of these kinds of project is well known in Italy. In 2013, 25% of 336 LULU protests concerned waste disposal plants and 10% concerned large transport infrastructures (NIMBY Forum 2014). In particular, the new stretch of highway, which could potentially produce benefits in terms of urban traffic decongestion, would also imply the abatement of several 5 The acronym LULU (Locally Unwanted Land Use) is used to identify territorial conflicts that are raised in projects characterized by diffused benefits and concentrated costs. The concept is a variant of the more well-known acronym NIMBY (Not in My Backyard). While the term NIMBY implicitly assumes the viewpoint of the promoter of a public work or project, and thus attributes a negative connotation to the phenomenon, the term LULU is used from a neutral perspective (Popper, 1987;Dear, 1992;Rabe, 1994;Hunter and Leyden, 1995;Takahashi, 1998;Schively, 2007). buildings and an increase in acoustic and atmospheric pollution for several years. The new pyro-gasifier, which was based on an innovative small-size technology that could lead to a more equitable and sustainable territorial distribution of industrial waste plants, would produce (limited) polluting emissions.
The cases are very similar, in terms of the previously mentioned exogenous factors. In both cases, the institutional authorities responsible for the final decisions had publicly expressed serious commitment toward the output of the deliberative processes, and there was a concrete possibility of influencing and even rejecting the preliminary projects. In addition, both processes were introduced in contexts with an already aware and active local civil society, since several episodes of the mobilization of environmental groups and citizen committees had already taken place in the past to protest against other similar projects. The issues were also framed in a similar way (Barisione, 2012) with much of the information phase being focused on the original project rather than on other alternative options, from different viewpoints (engineering aspects, environmental impact, health impact) and confronting different considerations (experts, stakeholders and technicians with different opinions on the public work).

A New Highway Stretch
In Genoa, the construction of a new highway stretch alongside the already existent one, which runs through the centre of the town, is a topic that has been under discussion since the eighties. A first draft, dating back to 1984, was strongly opposed by several citizen committees, and at that time the decisional process ended in stalemate. A new policy window opened in 2001, when the public work was included in a National programme of strategic public works. A second project was developed between 2003 and 2005 by Highways for Italy (HI), the public company that is in charge of the Italian highway network, but its approval was again shelved after new local opposition. In 2006, HI drew up a third project that was approved by the Region, the Province and the Municipality of Genoa, the three administrative organizations in charge of permits for the implementation of the work. The project included, amongst others, the demolition and reconstruction of a highway bridge in a highly populated area adjacent to a large factory. Again, this third project raised oppositions and protests, but this time the mayor of Genoa decided to promote a deliberative process to discuss the HI project and other alternative options with the citizens. 6 The process was in part designed according to the French model of the Debàt Public (Revel, et al. 2007). In November 2008, the City Council appointed an independent Commission to design and manage the whole process. HI soon had to draw up a dossier with various alternatives as a preliminary knowledge base for the participants. A phase of bilateral meetings between the Commission and 72 stakeholders then followed, with the task of mapping all the opinions and discourses about the topic and the various aspects that could be addressed in the public meetings. The direct participation of citizens started in 2009, and immediately raised the local mobilization of civil society leaders, associations and citizen committees.
A group of entrepreneurs started an intense communication campaign in the main local newspapers in order to promote the usefulness of the new highway. Some citizen committees instead formed a network to oppose the public work, and organized a protest march in which nearly three thousand people took part. Criticism of the Commission and the process came from both sides.
Overall, the deliberative process lasted about six months and alternated moments of information exchange and face-to-face discussion. At the same time, organizations and individuals could upload documents to address specific issues or propose alternative projects on an online platform (the "Notebook of the actors"). The public meetings took place over three phases: six meetings were focused on the alternative options projected by HI and were open to all citizens; six meetings were focused on specific issues related to the public works and again there was no selection of participants; four planning workshops were designed to deal with controversial issues that had emerged in the previous meetings, with the aim of suggesting new solutions (table 1). Overall, in the first and second phase meetings, 33 technicians and experts were involved, and 45 Notebooks of the actors were uploaded on the website by organizations and ordinary citizens.
The deliberative process ended with the Commission drawing up a report, in which the core contents of the debate were synthesized. The report stated that the process had produced new proposals for the highway stretch and a request was made to establish a local observatory composed of residents to control the implementation phase. A few days after the Commission presentation, HI presented a new project, in which one of the new options proposed by the participants was clearly taken into consideration. The new project resulted in having much less impact on the territory; in terms of the demolitions of buildings, the residents living between 25 and 60 meters from the highway decreased from 497 to 85 while the residents subjected to the demolition of their houses dropped from 503 to 122. In addition, other suggestions made at the meetings were included in the new project: the adoption of a new excavation technique to further reduce the risks associated with the extraction of asbestos rocks, the remuneration of the demolishing houses at market prices (above the minimum compensation required by the Law), and the institution of a local observatory to control the implementation phase. In 2010, the Region, the Province and the Municipality of Genoa declared they were in favor of the new project, although several political authorities, including the President of the Region, did not appreciate it and preferred the original project. After a few months, ten residents were elected to set up the local control observatory. The leader of the main local committee that was against the highway stretch was also among the elected.
The Environmental Impact Assessment was completed in June 2013. Currently, HI is expected to obtain a co-financing from the national government.

A Small Pyro-Gasifier
Castelfranco is a Tuscan town with 13,000 inhabitants and is part of a major industrial area that is dedicated to the treatment and processing of leather.
At the beginning of 2010, one of the private waste-treatment companies in the area, Waste Recycling, presented a project for the construction of a small pyro-gassifier to dispose of industrial waste. 7 It then received the preliminary authorizations from the Municipality, the Province of Pisa, the Region and the Azienda Sanitaria Locale (the local public health organization).
The project raised strong opposition from two citizen committees, one of which had been active throughout the territory since the eighties, promoting campaigns and protests against waste landfills and incinerators, while the other had been set up more recently and was part of the Zero Waste international movement. Aware of the tensions generated by past demonstrations against other plants, the Province of Pisa, before ending the authorization process, suggested organizing a public discussion process with the local population.
In July 2010, the Province and the Municipality submitted a request to the Tuscan Agency for Public Participation. 8 The request concerned financing for the realization of a deliberative process to discuss the project of the pyrogasifier and other alternative options to dispose of the industrial waste from the area. The proposal was approved and funded by the Agency, and it set up in the classic form of a citizen jury (Crosby, 1996;Smith & Wales, 1999). A group of facilitators was charged with conducting the whole process. Firstly, they appointed a Steering committee composed of 16 local stakeholders (the Municipality, the Province, the Azienda Sanitaria Locale, the company, some factories from the area, citizen committees and environmental associations), with the task of overseeing the implementation of the entire process. Fifty jurors were then randomly selected, by a research center, from a stratified sample of the local population, in order to create a heterogeneous group of citizens, in terms of demographic variables and in terms of opinions on the issue. Widespread skepticism about a jury composed of randomly selected citizens and about a process conducted by ambiguous actors, called "facilitators", soon emerged. The press and a local TV station ensured good media coverage, while about 10,000 copies of newsletters were distributed throughout the territory and in other surrounding municipalities.
The jurors participated in six meetings in 2011, during which they could obtain information and discuss the project and the alternatives. Four thematic meetings were alternated with plenary sessions and small group sessions, with the main aim of obtaining information from experts with different viewpoints. The third and the sixth meetings had more constructive aims, the former being oriented toward defining the lines of the following meetings and the latter being oriented toward elaborating policy recommendations (table 2).
The jury declared it was contrary to the pyro-gasifier and proposed the hypothesis of building another larger waste disposal plant, based on a different type of technology. After the publication of the jury recommendations, conflict arose among the public institutions, as the municipal council wanted to accept the jury's recommendations, while the Region preferred to insert the pyro-gasifier in the regional plan of the strategic public works (Regional Law 35/2011); Waste Recycling therefore began to construct the plant. A long legal dispute followed between the Municipality and the Region. The construction of the plant has finished, but the authorization process for its use still remains uncertain, in terms of timing and real feasibility.

Deliberative Design and the Impact on Policy Decisions
As far as the design of deliberative mini-publics is concerned, two dimensions are particularly important: the way of including different views, which is mainly linked to the participant selection method and the way of making participants discuss, which is structured through facilitation methods (Carson & Hartz-Karp, 2005;Moore, 2012). "Pure" deliberative mini-publics (Braun & Schultz, 2010), such as deliberative polls and citizen juries, which were initially imagined by academics for research purposes (Fishkin, 1991;Crosby, 1996) and only later applied to real policy-making, usually follow some procedural requirements. Participants are randomly selected from stratified samples, which are representative of the population in terms of some demographic and opinion variables (Steiner, 2012). The aim of the random selection is not only to reach ordinary citizens who are not used to engaging in political activities, but also often to create arenas of "blank slate" participants, with no clear interest or commitment toward the topic (Kleinman, et al. 2011, 234). The interactions inside these mini-publics are structured and regulated by professional facilitators, who work to make people exchange reasonable and universal arguments, reflect on values and principles, and interact respectfully (Mansbridge, et al. 2006;Steiner, 2012). The idea behind this model is that it helps to produce creative solutions that will be able to persuade not only the people who take part in the discussions but also the wider public.
Both the processes described above were designed and facilitated by practitioners, who clearly knew the deliberative imperatives, but designed the processes in different ways. The Tuscan mini-public was very similar to a pure model, with a rigid randomized arena and purely argumentative dialogic sessions. The mini-public of Genoa was set up in the form of a hybrid process, with a flexible arena and the discussions alternating between argumentation and confrontation. The design choices in Genoa and Castelfranco contributed to generate mechanisms that affected the impact of the processes on the final policy decisions. In particular, the hybrid design, although not free from problems from a deliberative perspective, proved to be more effective in terms of capacity to influence the contents of the final policy decision. Combining a flexible arena with a mix of argumentation-based and confrontation-based discussions, the Genoa process generated specific mechanisms that helped the contents of the process to receive more attention and more careful consideration by public institutions. On the contrary, the pure Castelfranco model was counterproductive in a two-fold sense, since it did not influence the final decision and it raised institutional conflicts. These statements are explained in more detail in the following sections.

The Arena
The Genoa mini-public arena was somewhat flexible: the first meetings were open to all the residents, and the participants in the planning workshops were recruited through the "outreach" method (Schwarz, 1994). The flexible arena was not immune to problems, but it produced three virtuous mechanisms: 1) giving space for expression to committees and associations, it favored the legitimation of the process by activists usually hostile to the deliberative approach (Young, 2000); 2) using the outreach method to include citizens during the process, it favored the access of innovative ideas and the emergence of useful "bridge-proposals" to redefine the stakes and to stimulate the formulation of constructive solutions (Sclavi & Susskind, 2011); 3) making the potential for citizen mobilization visible to the institutional authorities, it highlighted the costs (in terms of loss of consensus) of the missed consideration of the citizens' recommendations.
The Castelfranco arena was instead rigid: an arena in which 50 randomly selected citizens discussed the topic, and experts and stakeholders only intervened to present information and express their views. The rigidity of the arena proved to be counterproductive for three reasons: 1) the absence of an inner expression space for organized groups increased their skepticism toward the process and led them to put pressure on jurors outside the mini-public; 2) from the first discussion sessions, it became clear that the jurors were experiencing a group polarization process (Moscovici & Zavalloni, 1969;Myers & Lamm, 1976;Sunstein, 2002), and it was almost impossible for the facilitators to weaken this mechanism; 3) the quietness of the pure minipublic, although surrounded by high media coverage, made the contents of the process de facto silent and invisible for the external political and civil society, thus weakening the potential for persuasion and the influence of the final recommendations.
How the three mechanisms developed in Genoa and Castelfranco is now explained in detail.
The open meetings in Genoa produced unbalanced discussion arenas, which were dominated to a great extent by committees of protesting citizens. However, this phase favored a kind of socialization toward the deliberative process by some activists, which, in the end, proved to have in part legitimized the process. In fact, two episodes of high tension between the Commission and the committees did not result in the failure of the process, but only because the members of these committees in the end decided to allow the debate to continue. As one of the protagonists of the episodes mentioned: "We were going to end up in a fight [...] we wondered what to do, whether to go ahead with the process or to boycott it. [...] We looked at each other and decided not to boycott the debate, because the possible benefits would have outweighed the disadvantages (member of Committee X)." The planning workshops were instead characterized by a greater diversity and balance of opinions, and by a higher level of productivity. The outreach method essentially consisted in identifying and actively involving reflexive citizens (Bobbio, 2010), that is, people who, during the previous meetings or through the Notebook of the actors, had stood out as particularly reasonable and creative, especially in critical situations. The outreach method could be criticized as being overly manipulative, because of the power it gives to practitioners. However, thanks to the "outreached" citizens, some innovative and challenging ideas positively wrong-footed the HI experts and technicians and stimulated their curiosity. The HI technicians did not openly express appreciation of these new proposals, and remained officially skeptical, but, as noted by one of the participants, "the HI experts then came to the subsequent workshops with other technicians, who were more prepared on certain aspects, and this made it clear to everyone that they had begun to seriously consider some of the proposed ideas." In the end, these ideas became a core part of the new HI project.
Finally, the mobilization of hundreds of citizens during the open meetings (on average about 300 people per meeting) had the effect of attracting the attention of all the institutional authorities toward the process and to render them well aware of the costs (in terms of popular consensus) of ignoring the contents and the recommendations coming of the mini-public. The President of the Region, although firmly contrary to the new HI project for functional reasons, when faced with the mass participation and the public attention that the process had created, cautiously evaluated the possibility of disregarding the policy advice of the participants and in the end decided not to reject the new HI project. A HI technician recalled that "since the early crowded meetings, everybody understood that the original project and some of the other options would have triggered the opposition of the local community and the Region also understood the critical situation perfectly". At the end of the process, a regional public manager confirmed that "the President of the Region, even though convinced of the positivity of the original project, chose to stand aside, actually leaving the responsibility of the final choice to the Municipality. After seeing what happened at the meetings, with hundreds of people and the committees shouting, he was worried about causing another commotion".
The citizen committees and local associations in Castelfranco could not participate as jurors: they only had access as "witnesses", who were allowed to express their views to the jurors through presentations. They were deprived of a space in which to express their activism inside the process. This induced them to surreptitiously influence the final recommendations by putting pressure on jurors from the outside. As a facilitator explained, "Some of the jurors came to some of the sessions with prepared documents containing comments and requests formulated outside of the mini-public, and everyone knew that the documents came from parallel meetings with the local committees." The rigidity of the arena, which was composed exclusively of citizens who had been recruited before the process, also limited the degree of freedom of the facilitators, who had to deal with a group polarization process without any chance of intervening through the recruitment of people with other viewpoints or opinions during the process. In spite of the random selection, in the end the composition of the arena was not balanced, because the members were strongly biased against the idea of any waste burning plant in the territory, 9 in this sense confirming the criticism of Kleinman, et al. (2011) toward random selection. From the first sessions, the jurors who were potentially favorable toward the idea of a new plant to dispose of the industrial waste of the local factories fell into a spiral of silence, that the facilitators were not able to tackle simply through dialogic techniques. Over time, the process of group polarization developed and all attempts by the facilitators to deal with it failed. One of the facilitators recalled the last day of the jury with these words: "in a climate of growing tensions, in the end we were forced to do something that a facilitator should never do: we started the day with the jurors voting on a verdict on the pyro-gasification, which was not exactly the mandate of the process, but we had no alternative. The verdict was a unanimous rejection of the original project, and we tried to make the jurors at least focus on the reasons for their rejection, because no other serious reflection on alternative solutions had emerged in the previous discussion sessions. This was a failure for us. Deliberative recommendations should be the consequence of the arguments, and the arguments should not be built to justify a previous decision".
The jury recommendations did not contain any viable alternative to the original project or credible general arguments and suggestions about the problem of the industrial waste disposal from the territory. The alternative proposal, "a larger cool treatment plant to serve a wider area," was not accompanied by arguments about its feasibility or sustainability, while the more general suggestion -"to wait for a Province planning of waste management before starting any other waste disposal works" -was actually a dilatory argument. The power of the Province was very limited, since the management of industrial waste is mainly ruled by national legislation, which promotes a free market in the field. Although the Municipality ratified the verdict, the recommendations were easily criticized by the company and by the Region, which accused the document of being simply a more sophisticated form of a NIMBY refusal.
When a group polarization process takes place, it is difficult to use facilitation techniques to combat it. A flexible arena does not guarantee that the minipublic is immune to such a vicious circle, but it provides facilitators with more degrees of freedom, and gives them the opportunity of intervening on the balance of the viewpoints and on the characteristics of the participants. In the absence of truly constructive solutions from an almost invisible mini-public, closed within the rigid boundaries of a group of a few dozen citizens, the company felt free to disregard the results of the process, and the President of the Region did not hesitate to declare that only a small group of citizens refused to accept the project: "If a company has to make a settlement or build a plant for waste recovery [...], it cannot be blocked by administrative inertia or populism. [...] Some mayors expressed critical statements about the pyrogasifier because a small group of citizens opposed it. We cannot lapse into populism" (Interview with the President of the Tuscany Region, Il Tirreno, November 14, 2011).

The Dialogic Interactions
Human beings are not used to arguing and exchanging conflicting opinions with unknown people, especially in public contexts and when problems directly affect personal identity (Johnston Conover, et al. 2002). This is one of the main reasons why deliberative mini-publics are designed to reduce the spontaneity of common talks and why they are "assisted" by facilitators (Moore, 2012). However, when the issues are highly controversial -as in the case of land use policies -the pressures of the deliberative setting can collide with common cognitive dynamics. When a conflict is developing, the actors tend to accentuate the common mechanism of categorization, attributing negative prejudices, such as incompetence or opportunism, to the people with opposed opinions (Tajfel, 1978(Tajfel, , 1981. In purely deliberative processes such prejudices tend to be silenced because of the setting: individuals feel forced to quietly argue, using universalistic justifications and not partisan interests, and to interact respectfully instead of overwhelming the opponents (Duchesne and Haegel, 2006). This is what Elster (1994) wisely called the "civilizing force of hypocrisy" and Warren (2008) labelled "deliberative diplomacy." which does not exclude the underlying persistence of individualistic orientations, strategic behavior or prejudices, but "forces or induces speakers to hide them" (Elster, 1998, 111). Although this hypocrisy seems to help the recommendation to be anchored to public reason, it risks rendering the process too artificial and incapable of building really constructive solutions (Adler, 2008;Lenard, 2008).
In this sense, when deliberation concerns controversial issues, the problem is not only to favor the participation of citizens who tend to remain silent, but also to favor the construction of creative solutions based on deeper consciousness of the adversary viewpoints. The open meetings during the first two phases in Genoa were mostly arenas of positional confrontation, although structured by rules and conducted by facilitators. In sessions of about three hours, the confrontation took place following a speech-demand-answer format: HI technicians and experts presented the project, then citizens put written questions in an urn, the President of the Commission drew some questions out of the urn and asked the authors to explain them more clearly, and then the word passed to the HI technicians and experts who answered the questions and provided clarifications. However, the questions that were not extracted during the meetings received a response from the technicians in the following days, on the process website. In this structured setting, the facilitators allowed the participants to vent their accusations, claims and fears, and most of them were expressed in a truly non deliberative way, with insults and shouting against the experts and the Commission members. More deliberative type meetings followed this confrontational phase. This combination of confrontational and argumentative sessions had two effects. On the one hand, it introduced a distorting element into the overall process, since it gave space to a really confrontational mode of interaction (which also risked leading the process to failure in two critical moments, as already mentioned). On the other hand, it allowed some prejudices to be addressed openly and partially deconstructed, thus favoring a better understanding of the reciprocal interests and above all of the different viewpoints on the issue. As mentioned by the President of the Commission, "latent conflicts were made manifest, allowing the lid to be taken off a boiling pot, which sooner or later would have exploded. In this case, it exploded inside the process, which was probably better than having it explode outside and later on." From the analysis of the speech acts, it emerged that, in the first confrontational meetings, 36 of 284 contained explicit negative prejudices and many other statements alluded to them. 10 The main objection of citizens was that the original HI project was formulated on mere profit motivations. The HI technicians also emerged as being prejudiced about the residents' motivations.
In the original project, many of the buildings located under the highway bridge would have had to be destroyed. These buildings had a low monetary value because of their undesirable location. This is why the HI managers believed the residents would have been happy to move to a more attractive neighborhood and consequently thought that any opposition would simply have been instrumental to obtaining higher monetary compensation. A regional civil servant recalled that "the neighborhood was a squalid suburb… there were only ugly houses, with railway and the highway bridges above them, thus blocking the sky. We all believed that the opposition to the expropriations was a strategy to obtain a higher price for the houses." Thanks to the clarification of these prejudices, their at least partial deconstruction was possible. On the one hand, the HI managers and technicians came to understand that, behind the opposition to the building abatements, there were also relational motivations, since even a squalid suburb had developed social capital over the years. In reaction to this unexpected reading of the social impact of the public works, the HI technicians began to think more seriously about alternative projects that could reduce this kind of impact. On the other hand, although some organized groups maintained an attitude that was clearly oriented toward protest and opposition, some members of the citizen committees showed a greater willingness to listen to the HI proposals and partially recognized their reasons. As a member of a citizen committee recalled: "We shouted and protested, even loudly, but after these heated meetings we felt that the approaches were changing, both in the HI technicians and in us… we became more aware of what they would have been able to redefine and what they could not redefine, while the HI managers changed their attitude toward us and proved to have developed a more mature understanding of our preoccupations." An HI technician confirmed this reading of the usefulness of the confrontation phase: "In the end, I would say that the first meetings, although very hard, were also useful, because they allowed us to highlight that the belief of some committees had no foundation and, I must admit, that such a project in the city could generate impacts that we had not considered in the formulation of the project." In Castelfranco, the absence of a confrontational phase between the citizens and the Waste Recycling managers prevented the clarification of explicit prejudices: only nine of the 306 speech acts contained explicit prejudices against the company and no deconstruction mechanism was developed. The interviews after the process revealed that the respectful climate of discussion was highly unnatural, and that it concealed a reciprocal lack of confidence. The experts who expressed positive evaluations on the pyro-gasifier were not blamed openly, but subtly put under pressure by the jurors with questions that hid hostility and showed a lack of consideration of their arguments. Some jurors admitted that "many of us posed questions to the experts in order to somehow discredit them and strengthen our position," because "most of us had a preconceived idea, even though we could not state it, that the project only existed because of the financial interests of the company." The prejudice of the Waste Recycling managers and technicians about the NIMBY attitude of the jurors was not publicly expressed in the quiet jury meetings. At the end of the process, the head of Waste Recycling recalled this experience with bitterness: "the jurors listened to the experts and analyzed all the documents in an atmosphere of respect and calm, but the orientations were already clear, although no one revealed them openly, and we did not have the possibility of criticizing them. [...] The NIMBY syndrome was too strong. No constructive process is possible, where something is burning." Overall, it is possible to conclude that a confrontation phase, while weakening the integrity of the process in Genoa, allowed the foundations to be laid for a more sincere and conscious discussion in the deliberative phase, thus increasing the level of confidence and the productivity of the process. This finding in some ways empirically justifies the advice of Dryzek (2001) and Steiner (2012), that is, to give space to contestation in deliberative processes.

Conclusion
The present study has analyzed two deliberative mini-publics organized to integrate decision-making on highly contentious issues. The aim of the comparative analysis has been to understand why one of the processes had a significant effect on policy decisions while the other did not, although both were similar in terms of the contextual and institutional aspects identified in the literature as important factors of political efficacy of participatory practices. Since such kinds of mini-publics do not usually have the formal power to make the final decisions of the political authorities binding, understanding how they could improve their capacity to influence policy decisions becomes a crucial matter.
The empirical studies that have so far addressed this issue have shown that some factors are relevant in specific cases of participatory and deliberative processes: the existence of an active civil society interested in the topic, a clear commitment by the institutional authorities who are responsible for the final decisions, and a proper timing of the mini-public, when several options are still available in the decisional process.
The present analysis has brought to light how the design of the mini-publics can contribute to increasing or weakening their impact on the policy decisions.
The deliberative design of the Genoa mini-public was structured as a hybrid path, in which the arena of participants was flexible; a purely deliberative phase was mixed with a phase of weakly argued and sometimes even disruptive confrontational interactions. This hybridation produced some distortions, but it also generated some mechanisms that positively affected the influence of the mini-public on policy decisions. First, a confrontation phase within the deliberative path favored the legitimation of the process by groups who would usually have opposed the deliberative mode of public action (Young, 2000). It also allowed the fears and prejudices, which do not usually emerge in purely deliberative mini-publics, to be expressed. Their unveiling and their at least partial deconstruction favored a better mutual understanding, which resulted to be an important precondition for a less artificial climate in the subsequent deliberative phase. At the same time, the constant outreaching of the facilitators also allowed particularly reflective citizens to be included, citizens who were distinguished during the process as being capable of challenging the topic by formulating innovative and constructive ideas. All this, combined with the public visibility that the confrontational phase gave the process, produced an important impact on the institutional authorities, in terms of level of attention, and careful consideration of the mini-public recommendations.
In the Tuscan case, the process was designed as a pure deliberative minipublic: a previously randomly selected arena called on to work on the issue according to the rules of mutual respect and rational argumentation. Dealing with a highly contentious issue, the mini-public immediately collided with the real dynamics and the pressures of the context. Despite the use of random selection, the research centre that recruited the participants did not manage to build a balanced group of citizens, in terms of opinions toward the issue, since it concerned a land use policy, which usually raises opposition by local populations. At the same time, the one-shot random selection, which excluded organized groups from the process, induced these groups to frequently attempt to influence the jurors from the outside. The discussions of the jurors, facilitated in a purely deliberative way, were somehow "caged" within rules of polite and quiet interaction, focusing on rational arguments and avoiding space for confrontation and disrespectful statements. This produced a deliberative climate, but also a generalized feeling of hidden mutual distrust. All this favored the development of a group polarization process that prevented the formulation of constructive recommendations on sustainable solutions, as alternatives to the original project. The consequent low credibility of the minipublic recommendations, together with the low visibility of the deliberative process, held exclusively with a small preselected group of citizens working behind closed doors, made it easier for the Region to disregard the contents of the deliberative process.
From these findings, it can be suggested that, on highly controversial issues (or at least on LULU matters), a pure deliberative format can induce mechanisms that significantly undermine the deliberativeness of the minipublic and weaken its influence on policy decisions. Our conclusions are partially consistent with those of Karpowitz and Mansbridge (2005). On the basis of an analysis of a deliberative process in Princeton, they argued that a purely deliberative process failed to discover the underlying conflicts and thus "seemed to have produced not consensus but a backlash of anger and frustration" (p. 353). If "the larger question is the degree to which deeply opposed conflicts in interest, when discovered, can be handled within the deliberative process itself," (p. 349) without promoting or rejecting a particular type of instrument or technique (public debate vs. citizen jury), our conclusions suggest the hypothesis that the use of hybrid paths could be more effective in countering the distortions that necessarily emerge in real contexts, and in enhancing the political efficacy of deliberative mini-publics.
The kind of hybridization identified in this article is not the only possible one, nor necessarily the best. Other design aspects, such as the forms of interaction between experts and participants or the timing of the different phases, could also be important. This is why an interesting line of research could be to test this hypothesis through other case studies and through specific experiments, aimed at comparing different forms of hybridization, in order to identify their specific effects.