The People’s Assembly: Testing the Collaborative (e)-Democracy

This paper aims to critically analyse the Estonian People’s Assembly (EPA), a crowdsourcing initiative carried out from 2013 to 2014. During the project, citizens could participate in decision-making and make proposals for laws and policies on a dedicated web-platform. Additionally, some people were invited for a traditional off-line debate. In that way, the project combined virtual communication tools with traditional discussion to apply the principles of collaborative e-democracy, in which governmental stakeholders and non-governmental stakeholders (such as local communities) join in a deliberative debate. The purpose of this paper is to observe, both, gains and problems of this crowdsourcing initiative. The analysis considered the design of the online space, if people had equal access to it, and the kind of issues proposed. It also applied critical discourse analysis (following Fairclough, 1995) and the index Quality of Understanding (Klinger & Russmann, 2015). As a conclusion, the paper suggests that virtual platforms can increase the quality of deliberative decision-making. However, they can also be seen as regulated “top-down” initiatives (Pellizzoni, 2012). In a wider perspective, the paper aims to contribute to knowledge on, both, positive and negative stances of deliberative crowdsource initiatives in a post-web society.

participants stated that the discussions had made them, at least on some issues, change their earlier opinion (Praxis Centre for Policy Research, 2014). As a conclusion, the organizers stated, that the "People's Assembly was capable of increasing civic competence and creating more interest in the general public" (Praxis Centre for Policy Research, 2014).
The EPA initiative can be considered an example of collaborative e-democracy, in other words, a political system in which (1) governmental stakeholders (such as politicians and ministers) and (2) non-governmental stakeholders (such as NGOs, local communities and business actors) collaborate with (3) ordinary citizens on the development of public laws and policies. The web forum was implemented on a governmental social networking site in which all Estonian citizens with electronic ID could participate.

DELIBERATIVE POTENTIAL OF CYBER SPACE
The Merriam-Webster's Dictionary (www.merriam-webster.com) defines 'deliberation' as a "long and careful consideration of discussion done in order to make a decision". Legitimacy and transparency are conditions for an effective deliberation to take place (Burkart & Russmann, 2016). The deliberative democracy presumes participation, engagement and interactivity, and at the same time "all citizens must have an adequate opportunity to speak about public issues and citizens and their elected representatives have to engage in talk with each other" (Dahlgren, 2005, p. 149).
According to Habermas (1989), a discourse-centred or deliberative democracy has several requirements. First, all parties that might be affected must be included; second, all participants should be provided with the opportunity to interact in a free, equal and easy manner; third, there are no restrictions on topics; and fourth, the outcomes can be revised (Habermas, 1989(Habermas, , 2006. By now, political debates on the web have been in the focus of academic research for decades (e.g. Dahlgren, 2005;Davis, 1999;Lilleker et al., 2011;Norris, 2001). Academic studies approach the question of the impact of Internet on deliberative politics, either from an optimistic (equalisation) perspective (e.g. Boulianne, 2009), or from a pessimistic (normalisation) perspective (e.g. Norris, 2001).
The optimists stress that the cyber environment, for issue arena debates, has all conditions for efficient public deliberation discussions (Castells, 2012;Rheingold, 1993). The pessimists do not question the potential of the cyber space as an environment for deliberative public arena, but they critically ask if and how this deliberative potential is used by political actors Toode, 2016) and the citizens themselves Dahlgren, 2005). Supporters of the positive equalisation theory bring out, that the Internet has a positive effect on public engagement and that it boosts participate in public debates to those groups in the society which were not involved in the pre-internet era (Boulianne, 2009). The cyber optimists also suggest that the new technology helps to bring more participation on all levels, and in that way creates conditions for deliberative and direct conversation between political actors and citizens without a limiting "gatekeeping" from the side of journalists and mass media (Boulianne, 2009;Gurevitch et al., 2009;Rheingold, 1993). Moreover, some even suggested that the option to debate issues on the web in virtual issue arenas would create deliberative 'conversational democracy' (Corrado & Firestone, 1996). Therefore, from the Habermasian viewpoint (Habermas, 2006), the cyber-optimists believe in the potential of the web as a re-democratizing environment for the public sphere. They prefer to consider the web as an ideal public arena where real and free deliberation can lead to perfect and efficient debates, and in that way contribute to real participatory democracy (Boulianne, 2009;Habermas, 2006). This means, that by the optimists, the web is seen as a solution and a chance to keep the electorate more engaged, and as an environment where participatory democracy can flourish.
Others focus on the pessimistic or "realist" scenario suggesting, that the shortcomings and problems of the power relations existing in the traditional, or in other words, pre-web media environment will be transferred to the cyber space (Davis, 1999;Margolis & Resnick, 2000;Norris, 2001;Resnick, 1998). They argue that dominant stakeholders in the public debate, such as state institutions, governments, major parties or NGOs would most likely have better resources and possibilities to be present in the agenda-setting process and, thus, can influence the issue-selection for the debate . The pessimists also state that the web offers excellent opportunities for cyber propaganda and that it might not engage those groups which have been not involved so far. Sunstein (2001), for example, brings forth that people usually discuss political issues with agreeing groups or "likeminded others" and that, therefore, in web forums people are in the debate not focused on gaining different viewpoints but, instead, tend to amplify similarities (Sunstein, 2009). In this context, the "Digital Divide" might easily increase the "Democratic Divide" (Norris, 2001: 274;Witte & Mannon, 2010, p. 51;Witte et al., 2009, p. 6). In other words, the participation opportunities do not guarantee deliberative, open and pluralist participation. The rapidly developed public sphere in cyber space does not automatically bring positive impacts on democracy (Dahlgren, 2005;Blumler & Gurevitch, 2001). Therefore, some research has focused on the question if political actors really wish to create an effective dialogue with citizens via Internet or whether they rather tend to maintain a monologic top-down structure of the communication (Cogburn & Espinoza-Vasquez, 2011;Jackson & Lilleker, 2010;Lilleker et al., 2011;Toode, 2016). Margetts et al. (2016) point out that interactive social media definitely have a mobilising effect and encourage collective action and participation, but it might also result in emergence of what the authors call 'chaotic pluralism' -a phenomenon which needs to be further studied (Dahlgren, 2018;Pellizzoni, 2015). Burkart and Russmann (2016) state that online deliberation is a process of interaction in which participants should follow a particular set of normative guidelines, to ensure that the content of a discussion becomes understandable. Herewith the authors propose a model to evaluate the "quality" of deliberation. As people have to understand the message to develop their opinion, the "quality" of discourses is considered the quality of "understanding" during the debates (Burkart & Russmann, 2016, p. 4142). Brabham (2008, p. 75) describes crowdsourcing as web-based "problem-solving". In social sciences, it could be described as an open-to-all invite by any actor to other actors in issue arenas to participate in the process of problem solving (Brabham, 2008;Pellizzoni, 2015;Vos et al., 2014). If the intention is just to listen to public concerns and views, monitoring web interaction can be an alternative choice (Ruggiero & Vos, 2014;Zhang & Vos, 2015), however, when the aim is to create a dialogue among people involved, this definitely calls for deliberative crowdsourcing platforms.

CROWDSOURCING AND CRISES
According to Pellizzoni (2012), political and deliberative web-forums often have a "governed" character and are organised by power holders in a situation of institutional mistrust and severe turbulence in society. The author brings out that web-forums may show high criticism in the discussion towards other participating actors, which poses an obstacle for real deliberative practice. Pellizzoni (2012) considers web-forums also an effective action to maintain the stability and existing order in the present system. Thus, deliberative web forums organized by power holders can indeed increase the effectiveness of policy decisions. There may be regulation of participation, agenda-setting towards certain issues, and moderation of the debate, which often reflect political interests from the side of dominant agents (Fairclough, 2003;Pellizzoni, 2012;Vos et al., 2014). Professionally framed issues tend to dominate the debate and survive the competition in issue arenas . Moreover, the "knowledge gap" between more active groups or individuals and those who tend to remain more passive in cyber debates can easily increase in the web environment (Witte et al., 2009, p. 6).

DIALOGUE AND PLACE MATTER
According to Wright and Street (2007) 'deliberative democracy' involves deliberative and democratic elements which should be present also on deliberative web-forums. Deliberative elements refer to mechanisms for decision-making, such as argumentation, rationality values and impartiality. The democratic element means that all citizens who are affected by the decision can participate the discussion or at least are represented, because in the core a deliberative democracy is not only the idea that people express their preferences, but also that they discuss those preferences (Wright & Street, 2007, p. 851), taking other points of view into account. This should be enabled by the architecture of such online forums, making it easy to gain knowledge on the issues at stake and comfortable to discuss opinions via the webpage. Based on these views, the design of the EPA web-forum was analysed.

DELIBERATIVE IDEALS
Deliberative initiatives involve three main elements: (1) participation in the decision-making process by the people involved; (2) deliberation as rational discussion in which all main views are considered, and (3) equality in the sense that all segments of the population concerned with the issues have the opportunity to make their voice heard and considered. Dahlgren (2013) points out that, the more people are participating, the more financial resources and time are needed for deliberative discussions. At the same time, it is impossible for every individual to take part in a societal debate (Dahlgren, 2013). To respect equality everyone should be involved, or a representative random sample of the population should be created to participate in the discussion.
Fishkin (2011) defines deliberative democracy as a form of direct democracy, which corresponds to the requirements of deliberation and offers equality, but he does not require participation of everyone who wants to be included in the discussion. Fishkin's participatory democracy involves participation and deliberation on a certain level without total equality, because there are not enough resources to compensate overall participation to every single engaged citizen. Therefore, participation tends to involve that part of the general public with a strong interest in the issues and not the entire population (Fishkin, 2011).
Existing studies have shown that Estonian Internet users, like others, tend to be passive consumers of the Internet and do not actively participate in the deliberative politics (Toode, 2016). It is also stated that few users upload critical content to social media web pages such as YouTube and Facebook (SAAR POLL, 2008, p. 16). Pellizzoni (2012) brings out that, although many web-forums offer discussions, few studies analyse the essence of such deliberative initiatives and scrutinise if these forums can be considered deliberative practices. Pellizzoni (2015) suggests viewing deliberative forums as "additional means or arenas of political struggle". He explains that in the current political landscape conflicts between political parties or turbulences are constantly present, and that deliberative forums are often used by political actors to remain in power rather than engage citizens in real decision-making. In that way, deliberative arenas might be used to give an impression of openness while avoiding real deliberative debate (Freschi & Mete, 2009). The authors stress that deliberative forums can even be a form of control over society (Freschi & Mete, 2009), without solving the problem of participation and involvement of the citizens.
According to Freschi and Mete (2009, p. 21), an important phase in the organisation of the web-platform is the shaping of the deliberative arena, including topic selection, identification of involved actors, formation of access rules and the choice of arena type (only online or a combination of online and off-line). There, the citizens as the actual 'deliberators' in the process, are in a weak position (Freschi & Mete, 2009), considering that multi-actor issue arenas often show competing interests .

COLLABORATIVE E-DEMOCRACY
Collaborative e-democracy is a political system in which governmental political actors, such as ministers and politicians, and non-governmental political actors, such as local communities, collaborate in the creation of public laws and policies. For this purpose, a governmental website may function as an online forum in which all can participate.
Basically, collaborative e-democracy involves three elements. The first element is the collaborative democracy, a political system in which voters and those voted for collaborate to find solutions for decisionmaking with the help of information technology to boost the participation (Collaborative Democracy, 2007). The second element is a policy-making process via a social cyber forum, open to all citizens or their representatives, so that each citizen, directly or through a representative, can make proposals, evaluate them and vote. The third element of collaborative e-democracy is proxy voting, which means that people in a collaborative e-democracy select representatives or proxies to vote for them in situations where they lack time, knowledge or interest for direct participation. It would be difficult to imagine that all people in society really can or want to vote, every time, on every single policy issue. Thus, via a web-forum or social networking site, proxy voting and representation involve the main features of, both, the direct and representative democracy.
The policy process includes periods for suggestions and ranking, during which participants can submit proposals while the highest ranked proposals appear on top of the list. This is followed by an evaluation phase in which experts evaluate the impact of the proposals for the entire population. After that, all proposals considered optimal for solving a collective problem or reaching a certain goal are voted on and, in this way, new laws or policies are introduced. In the very end, during the revision phase, it is decided whether the problem has been solved or the goal achieved and, if this is not the case, the process may start again.
For the collaborative e-democracy system, O'Reilly (2005) mentioned the following positive features: (1) it is based on the principles of direct democracy; (2) open source governance enables participation of everyone interested; (3) it facilitates, both, collaboration of likeminded people and discussion of conflicting opinions; (4) it draws on collective intelligence; (5) it shifts responsibility from politicians and government to collaborative citizens.
Although deliberative web-forums are often understood as an answer to the crisis of the representative democracy, many studies question if the Internet really fosters global democracy or that it easily may become a tool for control and manipulation of the masses by some powerful actors in governments or corporations (Bynum & Rogerson, 2004;Pellizzoni, 2012).

METHODOLOGY
The EPA project was analysed according to two perspectives: (1)  The principle for the analysis was worked out based on an argumentative literature review, which allows to collect and confront contradicting viewpoints in previous research (Dahlgren, 2018;Habermas, 2006;Pellizzoni, 2015;Wright & Street, 2007;Visser & Stolle, 2013). Next, the following criteria were considered: (1) the context of the initiative in society; (2) web-forum design; (3) the choice of topics; (4) the presence of mediators; (5) the principle for participant selection; (6) equal access to all segments of the population; (7) outcomes of the initiative. For the coding scheme the research considered, partly, the model of Wright and Street (2007) and the features of deliberative initiatives offered by Freschi and Mete (2009).
From the model Quality of Understanding (by Burkart & Russmann, 2016) the study primarily considered the distribution of the issues to potential participants who had access to the web-forum and were invited to the off-line "deliberation day". Other communication arenas such as professional media channels or social media networks were not included. Similarly, no later projects in the period 2017-2019 on the same webpage were included.

Web-forum Design
Web-forum design is considered an important point of attention. Wring and Street (2007) state, that the democratisising potential of most web-forums is limited, because the architecture and design of many such forums do not allow for deliberative discussion (Wring & Street, 2007, p. 853). The authors underline that the online forum design strongly influences how people discuss as, for example, the interface of the page can even cause polarisation.
The design of the EPA platform allowed to discuss each proposal separately. At the discussion platform participants could make proposals in five main issue-groups: (1) financing of political parties (rahastamine); (2) competition between political parties and their internal democracy (erakonnad); (3) the electoral system (valimised); (4) engagement and strengthening the role of civic society in politics between the elections (kaasatus); (5) politicisation of public offices (sundpolitiseerimine); and, lastly (6) varia was included to gather proposals which did not match with the five main areas.
The EPA website showed the colours blue, white and black, with only the entry button in red. It provided an institutional and "strict" style to the website, while the red button for entering might have given attention but also an impression of red light or limited access. Almost ¼ of the page was covered by a photo (clickable video) of the first meeting of NGOs, politicians and experts with the President, who is visible as the host. The photo might create the feeling "we are working" but also "you are out", resonating with a somewhat elitist perspective towards those not included in this meeting. The event was held in the basement of the Presidential Palace in Tallinn in a meeting room, known as the "Ice Cellar" (Jääkelder), where only few ordinary citizens have access. Considering this name, associations may be "frozenness" and "closure" or "hiding", rather than openness and participation by ordinary people.
On the top of the page there was a black stripe, showing the time frame of the different phases of the EPA initiative. Under it, next to the name of the initiative "Rahvakogu" (meaning People's Assembly) the different topical groups were presented. The slogan in grey said: "If you believe that debate is part of the solution, come and think too! Together, we make proposals for Parliament to change laws and policies." (Translation by the author of this article). Underneath the slogan, the five main topics a represented in columns. Each topical column shows also the last posts. The page was officially available in Estonian, Russian and English. The real discussion took place only on the Estonian site.
The discussion page showed on the top a blue stripe showed a short presentation of the issue. The black stripe provided small clickable sections (Esita ka!) where users could leave their proposal concerning the issue. This was not easy to find and, thus, not inviting for the users. The main part of the page was formed by the single proposals. On the right side, the users could agree (by a blue button) or disagree (by a red button) with the issue. By clicking on the title there was an opportunity to comment the post and vote. In reality all five areas were included.

Construction of EPA
According to Freschi and Mete (2009) the features to evaluate in the case of a deliberative initiative are: (1) issue and goal, (2) construction of the arenas; (3) profile of actual participants; (4) the deliberative device in action; (5) influence on decision making process; (6) salient profile; and (7) strategic effects. Table 1 presents these features of EPA.
In favour of EPA, as a collective deliberative e-democracy initiative, it could be said that the issue proposals, according to the chosen parameters of the Index of Quality of Understanding (Burkart & Russmann, 2016), seem transparent and accessible for all those participants with an Estonian electronic ID card. For these citizens, the project facilitated participation in the policy-making programme and, in this way, being part of the decision-making process. It offered the circa 2,000 participants an experience of deliberative policy making and, at least for those who participated in the project, it increased the public awareness of collective problems, goals or policy issues. Emphasis on ordinary citizens that possess an electronic ID card (enabling registration); participation on a voluntary basis; with a representative sample for the deliberative faceto-face meeting, prepared based on the results of the proposals 3

Profile of actual participants
Crowdsourcing for the debate on the web; first time the needed number of participants for the deliberative day was not achieved; the second call resulted in 314 effective participants of the 550 foreseen 4 Deliberative device in action Formal situation for the online and off-line initiative; fragmented discussion; of the focal topics, the electoral system received most attention (nearly 40 % of the ideas), followed by financing of political parties and the role of civic society 5 Influence on decision making process From nearly 2,000 online proposals, 18 were selected for the off-line debate, while 15 proposals were eventually sent to Parliament for further action, 3 proposals became laws 6 Salient profile Crowdsourcing of participants for the online forum by the President 7 Strategic effects Framing and legitimation of a limited number of selected issues; addressing Estonian speaking citizens, excluding those speaking Russian or without electronic ID; promoting the image of the President as an institution which stands for democracy; resulting in a calming effect on tensions in society.
The creation of the deliberative arena and the rhetoric around it helped the institutional stakeholders calm the turbulence in society. Thus, it could be considered, in part, effective as a crisis management strategy. On a general level, critics of collaborative e-democracy projects (e.g. Pellizzoni, 2015) underlined that, in most democracies, the constitution sets limits for direct democracy and governments may not be willing to give policy-making power to citizens in such a way (Toode, 2016). There are limitations also from the perspective of digital divide, because many people still do not know how to use the cyber technology for political participation. Consequently, dominant actors would have advantages in the process of issue selection and presentation . In fact, some authors bring out that citizens with the right to participate or vote might also lack a real understanding of societal issues which could make them vulnerable to manipulation and cause biased votes (Dahlgren, 2015).

CONCLUSION AND DISCUSSION: CAN (E-)DEMOCRACY WORK?
The effort of political actors to involve citizens in political debate has a decisive role in the success or failure of deliberative democracy Habermas, 2006). Herewith, two main questions remain. First, are power holders willing to open true deliberative discussions in cyber space, even though they may lack sophistication and experience in doing so? Or would political actors, to maintain the current power system, rather avoid a change from a one-way monologic pre-web communication model to a dialogic twoway communication model in the post-web era?
Previous research has shown that the web could contribute to deliberative discussion in society, but often no conditions are offered for a participatory equal debate between decision makers and voters (Wright & Street, 2007;Strandberg & Grölund, 2012;Toode, 2016). This is often the case when citizens are considered to be policy consumers rather than responsible citizens ). In other words, responsibility lies with the existing political elite to create conditions for involvement of the wider public. Participation remains a core element of deliberative dialogic politics. Habermas (1989, p. 66) notes that public opinion can only be formed if a public that engages in rational discussion exists.
Moreover, Pellizzoni (2012) notes that in the case of web-forums the posts and criticism can be strong but may not bring a true deliberative democracy. Therefore, the basic question is not whether deliberation works, but how deliberative democracy fits into and effects the existing broader policy processes. The problem is that deliberative web-platforms may also bring regulation of participation and some sort of political interests, directly or indirectly. Thus, one should take note of how the participants in deliberative discussions are chosen and if all citizens have equal access to these spaces (Pellizzoni, 2012).
Another widely discussed problem is the electronic divide. A study carried out by Trechsel et al. (2010) on e-voting in Estonia demonstrated that e-voting, in 2009, still predominantly was a property of voters aged 18 -49 years, involving less of the older electorate. As is also the case in the traditional electorate, e-voters with high education and income participated more in the elections, while the retired or unemployed population is clearly underrepresented (Trechsel et al., 2010). Similar results are likely to still apply for e-democracy using web platforms.
Furthermore, one could question how the essence of government is perceived in the new cyber era. Some authors argue that the present forms of government are dominated by neoliberal rationality, which applied on, both, the individual and the collective level, the logic of the market and strategic planning leaving little space for social justice and political equality (Pellizzoni & Ylönen, 2012). This implies, that when people are passive and do not "demand" participation, the current parties tend to maintain their power and discoursesetting role.
In the case of the EPA project, the aim was "to get open government in Estonia to the next, more advanced level" (according to the organisers, published on the platform) and, thus, to involve people in the decisionmaking process. Some challenges could be noted: -Proposals related only to specific topics.
-On the website, the proposals and debate were in Estonian only, even though one third of the Estonian population speaks Russian (their proposals could be sent in and were translated, but they could not follow the debate in their language).
-The interest of the citizens in participation was not as high as expected; 70 % of randomly chosen representatives in the first round did not want to participate.
-Finally, when a new sample was formed of 501 representatives who accepted the call to the Deliberation Day, only 314 came to Tallinn to debate and vote.
-The representatives were divided in 8-10 person groups, the debate was mediated -controlling the discourse? Wright (2011, p. 249) stresses the importance of the question whether an internet-based deliberative initiative actually achieves what was expected. There has been some debate in Estonia on who gained from this initiative, power holders or citizens, for example, questioning the self-renewal capabilities of the political parties and if the expectations of the EPA project were met (e.g. Alas, 2013;Hõbemägi, 2013). Definitely, the EPA project was an innovative initiative and an interesting experiment of deliberative democracy. Foremost, it was a good case to study, not only for its virtues but also the challenges of such an initiative. A wider research concerning more deliberative web-platforms would enable a better understanding of the democratic capacity of such initiatives in current society. One of the most important outcomes of EPA is, that the required number of members to establish a political party fell from 1,000 to 500 which clearly provided grass-root movements with more opportunities to take part in policy making.
Interestingly, a study of Jonsson (2015, p. 11) showed that two-third of the participants in crowdsourcing, in the case of EPA, were higher-educated male professionals with right-leaning views. A total of 96 % of participants were of Estonian ethnicity (Jonsson, 2015, p. 11), which is high considering that approximately one-third of the Estonian population is made up by a large Russian-speaking minority (Eesti Statistikaamet 2019).