Using the Conference on the Future of Europe to Consider a Multilingual Continental Deliberative Model

Abstract

This article examines the linguistic and political dimensions of deliberation at a transnational level, using the Conference on the Future of Europe (CoFoE) as a case study. The CoFoE, held from 2021 to 2022, involved European citizens deliberating in 24 official languages of the European Union. This multilingual setting provides insights into the challenges and opportunities of fostering a multilingual continental democracy. While the European Parliament’s translation services were largely effective, the study reveals how linguistic diversity can both enhance and impede deliberative processes. By comparing the CoFoE with other multilingual forums such as the European Social Forum and traditional European Parliament deliberations, this paper explores the implications of multilingualism on participatory mechanisms and democratic engagement in the EU. This research employs an ethnographic methodology grounded in non-participant observations conducted during Panel 2 of the Conference on the Future of Europe, focusing on field notes, video recordings, and a live interpretation to document the dynamics of deliberative exchanges. The approach aimed to capture the diversity of interactions in plenary sessions and smaller discussion groups. The findings highlight the complex interplay between language use, political representation, and social inclusion, and suggest that true multilingual deliberation requires more than just technical translation services; it demands a commitment to linguistic equity and the accommodation of diverse voices.

1. Introduction

We propose to analyze the quality of deliberation in a transnational and highly multilingual European context through the example of the Conference on the Future of Europe (CoFE), held on the scale of the European Union in 2021–2022. This process, launched by the three major European institutions (Parliament, Commission, and Council of the EU), was developed over several stages and levels, with the setting up of a digital platform and a multitude of events at local and national levels, followed by the work of four assemblies; 200 European citizens were randomly selected to deliberate on four major themes including environment and health. The discussions in the assemblies were made possible with the help of the European Parliament’s instant interpretation services. The recommendations made by the European panels were very diverse, but the European institutions subsequently promised to integrate and follow up on these proposals, as recalled in Ursula von der Leyen’s recent 2022 State of the Union address. These deliberations have also paved the way for the creation of other new assemblies, for example on mental health, and are also the subject of research into their applications in the field of ecological transition and the European Green Pact, as in the Horizon 2020 projects PHŒNIX and EUComMeet.

It seems, then, that we are witnessing the emergence of a new model of plurilinguistic participation on an EU scale, whose long-term effects have yet to be measured and analyzed. The organization of citizens’ deliberations in the Union’s 24 official languages offers an exceptional insight into the challenges of multilingual deliberative democracy and its effects on both the content and form of exchanges. In this respect, the European Union differs from multilingual and consociationalist states such as Switzerland and Belgium or from federal states such as the USA, Brazil, or India in the number of languages it claims to treat with the principle of equality of treatment resulting in a large number of specific challenges [1,2,3]. This condition is also an indicator of the specific nature of its construction, and of how different it is from a State [4,5]. In this respect, the choice of randomly selected assemblies at a European level is particularly interesting, given the novelty and political specificity of the European Union, as explored, for example, by Kalpyso Nicolaïdis, who believes that the “demoicratic” dimension of the European Union, meaning the plurality of its “demos”, makes it an object that goes beyond what can be observed for federal states and could be represented by this kind of assemblies [6,7].

While we can assume that typical effects of multilingualism do not differ fundamentally from observations made by MEPs and European officials, already relayed by a certain number of works [8,9,10,11], the challenge of participatory mechanisms such as the CoFoE is, in distinctive terms, to be able to hear voices that are not normally present in the political arena. As such, the mechanisms inherent in multilingualism, which can either silence or amplify these voices, are of particular importance. Moreover, in recent years, political scientists have seen the use of political sortition as a potentially effective way of ensuring that the voices of people who are usually marginalized in the public arena may be heard. To this day, if the benefits and difficulties of deliberative democracy have been largely discussed in the literature [12,13,14], there are no systematic analyses on the effect of using several languages in deliberative randomly selected assemblies (also called deliberative mini-publics). The overwhelming majority of such assemblies are monolingual, or in the best of cases, have only two or three different languages [15,16] and the study of the use of sortition for political deliberation purposes is still relatively recent, since the first randomly selected assemblies at the level of an entire State exist in the contemporary era, only dating back to the late 2000s. As Lisa Verhasselt recently stated [17], while multilingualism and its effects have been studied for liberal representative democracies, there is a real gap in the academic literature concerning deliberative democracy as many democratic theorists seem to be trapped in a linguistically homogenous narrative.

In the case of this paper, we focus on the effects of this type of process on deliberation in a highly multilingual context, more specifically within the European panels that took place during the Conference for the Future of Europe. Is deliberation between randomly selected citizens in a highly multilingual context fundamentally different from monolingual deliberation? And more precisely, is deliberation by sortition in a highly multilingual context affected and limited by interpretation and translation work, and if so, in which ways? Is it possible, on the contrary, that this specific context can enhance the quality of deliberation?

Thus, several hypotheses guide this analysis. First, it is hypothesized that multilingualism in deliberative settings may lead to technical restrictions on oral and written expression, potentially reducing active participation due to increased challenges in mutual understanding and effective communication. Second, multilingual contexts might enhance the quality of deliberation by bringing diverse perspectives into discussions, potentially leading to more inclusive and comprehensive debates. However, multilingual settings could also foster power dynamics that marginalize minority linguistic groups, affecting the perceived inclusivity of these panels. Furthermore, we hypothesize that contributions in multilingual discussions are not equally valued, with some languages being more dominant (usually English, and to a lesser extent, working languages such as French and German), potentially skewing the equity of the deliberative process. Finally, it is anticipated that participants in multilingual settings adopt pragmatic solutions, such as using common languages or employing improvised translation practices, to overcome linguistic barriers and facilitate effective communication. These hypotheses will be examined to better understand how multilingualism shapes democratic deliberation and impacts the inclusiveness and quality of discourse in this specific context.

2. Materials and Methodology

The methodological approach of this article is mainly based on ethnographic observations carried out in the field at CoFoE assemblies and on its online platform, relying mainly on personal field notes and video recordings of the sessions made by the event organizers. This study relied on a single observer, who is also its author. The observation was conducted in a non-participant role and through an inductive approach to analyze emerging aspects, due primarily to the particular novelty of this type of process. The number of individuals included in these observations was 15 for each of the 5 discussion groups attended, i.e., 75 people, and around 200 for the plenary sessions observed, taking into account the fact that participants in the discussion groups were also present at the plenary sessions. We were an observer for Panel 2 of the CoFoE assemblies “European democracy/Values and rights, rule of law, security” in physical presence for the meetings taking place in Strasbourg from 24 to 26 September 2021 and in Florence from 10 to 12 December 2021, and online from 12 to 14 November 2021. Parts of the deliberations mentioned here can also be found on the archived website of the Conference on the Future of Europe [18].

The aim of this ethnographic work was to witness the greatest possible diversity of exchanges during this event. We observed plenary discussions involving all participating European citizens, with 200 participants and 24 different languages, and smaller discussion groups limited to 15 people and 5 different languages. During these observations, detailed field notes were taken, focusing on participant behaviors, the use of language, and the strategies employed to manage linguistic diversity. The languages accessible to us without the need for interpreting were English, French, Spanish, and Portuguese. For the rest of the European languages used, we had to rely on live interpretation and observe potential difficulties for a non-speaker (delays in interpretation, expressions of discontent among participants, attempts to improvise translations between participants and facilitators).

The aim of ethnographic observation is not to demonstrate the frequency of an event, but rather to initially delimit and describe its existence [19,20,21]. We, therefore, hope that this observational work will subsequently open the way to more systematic analyses of the biases and influences of interpreting and translation in deliberative processes.

3. Results

3.1. Background

The European panels benefited from the European Parliament’s interpretation services throughout the process, both during their sessions in Strasbourg in the Parliament buildings, but also during online sessions and during sessions held in other European cities, as each participant carried a tablet which enabled them to connect to different live interpretation channels. In this respect, and as far as interpreting is concerned, it is fair to say that participants had access to high-quality services comparable to those available to Members of the European Parliament. The European Union spends almost 1 billion euros a year, or nearly 1% of its budget, on translation and interpreting services to ensure dialog between 24 official languages, involving a total of 552 different combinations. The European Parliament’s interpreting service (not to be confused with the translation service, which deals with writing) comprises some 275 interpreters in 24 cabins for each official language. In addition, a large number of accredited external interpreters are also mobilized and may constitute more than half of the interpreters present at plenary sessions in Strasbourg. Some specific recommendations given by the Directorate General for Conference Logistics and Interpretation on its website [22] regarding those who use their services are particularly interesting, as they already enable us to identify a number of elements that can modify the exercise of deliberation in a multilingual context, and for which we had illustrations during the CoFoE. The speakers are asked to use their native language, for logistical reasons and to ensure that native listeners are able to hear them. It is therefore not possible for a multilingual speaker to switch from one language to another during an intervention, or to speak in a language other than the one assigned to him/her for the deliberation. On a few occasions, for example, we have observed German-speaking speakers switch to English, only to be asked by the moderators to revert to their own language. Beyond this first instance of self-discipline, speakers are also asked to “speak clearly, naturally and at a moderate pace”. Interventions are therefore dependent on the interpreters’ pace of analysis. In the same vein, they are also asked to avoid using acronyms or reading text, which can make interpreting more complicated.

Furthermore, the self-discipline involved in this type of interpretation is highly space-related: if speakers wish to access live interpretation and be interpreted, they must remain seated, connected to a headset and a microphone, a position that can quickly become frustrating for some participants, who evoke the sensation of being “chained to the microphone” [23]. On the one hand, this has an impact on the informal moments that escape translation and interpretation, but it also has an effect when some speakers are on the move or come to the central area of the European Parliament in Strasbourg1 where interpretation can not be assured and where thus fluency in several languages once again becomes a significant social advantage.

3.2. Technical Restrictions on Oral and Written Expression

These instructions reflect a number of constraints on oral expression, which can be observed in the direct interaction of smaller groups and have a direct impact on the way deliberation is organized between participants. The formation of groups of fifteen randomly selected citizens also responds to specific organizational constraints: the number of languages spoken in each of these groups must not exceed five. Furthermore, setting up groups and resolving linguistic problems in interpretation or direct communication between participants implies additional organizational time and the need to find alternative systems to interpreting from time to time, using common languages or even gestures. The facilitators have to resort to a “home-made interpretation”, having to talk to each participant individually in order to resolve any technical problems2, requiring as much as thirty minutes before the actual deliberation.

Moreover, the number of language combinations (nearly 522) far exceeds the organizational capacities of the interpreting services, and since 2004 these services have been using a system of pivot languages. In this way, a minority language is first converted into a pivot language such as English, French, or German, before being converted again into other target languages. However, this method tends to slow down the interpretation process. Therefore, the position of linguistic minorities remains particularly fragile. While it is true that people from small countries and small language groups tend to be more proficient in other languages [24], other sociological factors such as age can compound this situation: for example, in one group, a 77-year-old Hungarian man found himself particularly isolated when technical problems prevented him from accessing the live interpretation3. Sometimes, he would start thinking and chatting directly with someone of another nationality, only to be called to order by a facilitator who asked him to answer on his own.

Another element that we observed is the lack of linguistic support for informal breaks, the importance of which seems to be underestimated. The intervals between deliberations are punctuated by dinners, cafés, and receptions, all of which are moments of exchange freed from the discussion frameworks imposed by the CoFoE. They can provide an opportunity for more direct discussions and exchanges of arguments, without the scrutiny of an entire group. However, our observations soon revealed a very strong recurrence of language-based groups: each group speaking French, German, Estonian, or other languages remained relatively impermeable to interactions with other language groups, except in the case of certain polyglot individuals who navigated more easily from one group to another. The potential presence of professional, visible interpreters, with the possibility of assistance on request, would probably have enabled greater interaction between language groups.

Finally, and perhaps most importantly, technical constraints made it particularly difficult to put the participants’ work into writing, combinws with the constraints of oral communication as already described. Each working group worked with an Excel document projected in the deliberation room, which the facilitator had to fill in while leading the deliberation, an overly arduous task which meant that facilitators had to be assisted later by a “note-taker” [23]. The Excel document, with its several columns, made it possible to instantly translate an entry into the group’s four other working languages via Google Translate. The use of digital translation led to major shifts in meaning and ambiguities as it passed through a multitude of translation filters: for example, a Romanian moderator would himself write in Excel the proposals he received orally translated from French participants. The Excel spreadsheet directly translated the proposals written by the Romanian moderator in his own language and then came back again in the group’s working language, including French, but in a modified and filtered way. In the case of this exchange, we can count at least two linguistic filters, with interpretation and then digital translation, not to mention possible passages through pivot languages, very likely in the case of Romanian. For example, the term “égalité” (equality) used by a French-speaking participant came back and was transformed into “émancipation” (emancipation)4.

3.3. Conceptual Restrictions

The work of translation can also reduce a whole range of elements relating to linguistic complexity and cultural differences: idiomatic expressions, historical and cultural references, and cultural humor. Interpreters and linguists are already well aware of these constraints in politics [25,26,27,28]. However, it is possible that they acquire a new dimension with the deliberation of European citizens, through, for example, the use of formulas belonging to a national common sense that may come up against the incomprehension of other participants. The political or economic concepts mobilized by participants may refer to a specific historical heritage and different everyday realities. This was observed, for example, during a session where a series of misunderstandings arose from the mobilization of the concept of the welfare state by the Swedes, grasped very differently by the Romanian participants5.

3.4. Paradoxical Effects on Deliberation

Generally speaking, the interpretation work provokes a form of slowing down and what we might call a “flattening” of exchanges between participants. By “flattening”, we mean that the interpreter’s monotone voice, which is intended to be neutral, partially covers the expression and emotions of the main speaker. The effort made by the speaker to accentuate his or her intervention, to speak loudly, or to make his or her anger or sadness resonate, as well as the use of expressions proper to his or her language or of ironic or sarcastic formulas, all contribute to this set of paralinguistic signals [29], which can be significantly lost with the exercise of interpretation. Beyond the presence of the interpreter, oral expression in a multilingual context requires a certain discipline in speaking: clear, intelligible speech with a moderate rhythm. However, it seems to us that this is not necessarily an obstacle to the deliberation and expression of the participants. Indeed, this phenomenon also provides a form of “equalization in boredom”. By flattening exchanges between participants, a live interpretation also helps to avoid the recurrence of certain forms of inequality in speaking [30,31]. This observation is reminiscent of similar findings made in other deliberation systems. For Nicole Doerr [32], the use of multilingualism within the European Social Forum was a factor in improving the deliberative quality of this meeting process between European social movements. Unlike the CoFoE, which was piloted by three major European institutions, the European Social Forum or ESF organized between 2002 and 2010 [33,34] was a bottom-up attempt to organize debates at a European level between social movements, trade unions, and NGOs, and also focused on the issue of inclusiveness and diversity of voices heard during the process. Doerr describes an important effect of multilingualism that she calls procedural slowness [32]: live interpretation time lengthens discussions and increases the effort needed to listen to each other. This effect is also described by CofoE participants, who refer to a form of “listening fatigue”.

Doerr also describes much longer preparation times for ESF meetings, in order to be sure of everyone’s precise position. This slower pace is not necessarily perceived as a bad thing by participants, who consider that it helps to produce “politically balanced” exchanges. Following her investigation, Doerr also found that a majority of participants felt that, compared with national meetings, these European meetings were more inclusive and transparent, due to the multilingual context, which meant that meetings had to be held with much more formalized and supervised decision-making procedures, with more time given over to the expression of each participant and greater respect for the floor. This context makes it easier for more marginalized or under-represented groups (women, migrants, etc.) to have their say. This procedural slowness and the strong formalization of deliberations observed for the ESF are also points that can be observed for the CoFoE. In the same way, the European Social Forum was not immune to the same difficulties encountered by the European institutions, with the prevalence of certain languages as working languages, such as English, French, Italian, and German. Nicole Doerr’s observations showed that participants’ language skills, particularly in English, remained average and that a significant number of them, around 10% for the period 2003–2006, did not speak it at all. She described the existence of a perceived minority exclusionary bias regarding language communication problems, which could be seen in the language choices of activists according to their origin, with a difference in strategy between people from Eastern Europe who tend to adapt linguistically more than Western Europeans.

3.5. Contradictory Calls for Multilingualism

All language regulations within European institutions are governed by Regulation No. 1 of the European Council of 1958, which states that each member state brings with it its own official languages. In this sense, the European Union is not supposed to differentiate between languages and must treat them all equally. In reality, ethnologists Krzyżanowski and Wodak [35] describe a process of “hegemonic multilingualism”, i.e., the use of a small set of working languages in which English is predominant, to which French and German may be added. This paradox between an official discourse making multilingualism a cardinal value of the EU and much more pragmatic working practices tending to privilege certain languages to the detriment of others has been described on numerous occasions [36,37]. One of the concerns is that this discursive divide prevents genuine democratic discussion of the issue, probably because it also reveals tension characteristics of the construction of the European Union, which is neither a nation-state nor an international organization [38]. While the CoFoE process seeks to highlight the Union’s multilingualism, a paradox between stated values and everyday practices also tends to emerge. The emphasis on multilingualism and European diversity is staged throughout the panels, particularly during the plenary sessions, with presenters repeatedly expressing wonder at the “magic of translation”6, and through explanations of how translation and interpretation work: praise for language learning and language education comes from both presenters and participants, with informal times presented as moments to “practice in a foreign language”7 or as a time to “help each other out”8. The same praise for multilingualism is to be observed among the invited experts, although this appreciation of linguistic diversity also tends to be accompanied by an emphasis on the use of English, the use of which is virtually never questioned, contributing to a naturalization of its recourse. During written assignments, English is presented as the standard of reference for checking the relevance of proposals, with one facilitator saying “if it’s good in English, then it should be satisfactory”9. When it comes to writing explanatory texts for proposals made by a group of participants, some facilitators go so far as to encourage volunteers to write directly in English10.

As mentioned above, non-native speakers will even go so far as to express themselves in English, with interpreters obliged to respond themselves to requests to speak in their native language11, but this is even more widespread among guest experts who will sometimes use presentations in English or reply directly in written English to questions asked in the virtual sessions12. Alicia Gescinska, an invited expert, justifies herself by saying that “the language of the citizens is several languages, that’s why I speak in English”13. This sentence is interesting because it reflects a fairly classic confusion between multilingualism and the use of English, the use of which tends to marginalize other languages. The call for multilingualism thus covers different linguistic regimes, depending on the interpretations of CoFoE stakeholders: it can mean using the most commonly spoken third language, in this case English, or guaranteeing equal treatment for all European speakers. However, these realities conceal contradictory political projects that include very different treatments of linguistic minorities, as well as of the most marginalized members of society.

The absolute equality of European languages may even be perceived, consciously or unconsciously, as outdated by participants, who ask experts, for example, “how to better disseminate English”14, or even propose the adoption of English as a common EU language15. In the final results of the CoFoE’s recommendations, proposal 48 of the “Culture and Exchanges” plenary assembly, while insisting on the promotion of multilingualism in measure 2, explicitly mentions only one language of which learning should also be encouraged: English [39]. In the appendix, the first European panel, covering the economy, social justice, and education, also proposes, for axis 4 on “Learning in Europe” (sub-axis 4.5 and 4.6) in proposition 38, that English should be taught as a core subject in all schools in member states, in order to “communicate effectively” and to enable “the spread of a common European language within a very short time”. It should be noted, however, that the final report, which runs to over 300 pages, also echoes concerns about language issues on the part of national panels, particularly in Belgium, where translation problems for non-English-speaking citizens are repeatedly highlighted, and in Italy, where doubts are expressed about the adoption of English as a common language following the United Kingdom’s departure from the European Union.

4. Discussion

4.1. Socially Situated Conceptions

The question of the social incorporation of linguistic norms and the use of English as a “matter of course” must not be overlooked, even among European audiences with no knowledge of this language. In this respect, a particularly interesting observation was made when we carried out a focus group for another deliberation device as part of the EUComMeet European project16. This system enabled participants from several European countries to debate environmental issues online, with the support of live interpretation, this time all automatic. One of the women interviewed, middle-aged and from a rural background, said that the deliberation could have been smoother if it had been conducted in English only. When asked about her command of English, she immediately admitted that she would have been excluded from such a system, as she would have been unable to follow the exchanges. Such a paradox reminds us that effective deliberative processes must also take into account these mechanisms of self-exclusion in participation. Unsurprisingly, the least advantaged and least educated socio-economic categories are also the least able to sustain a fluent conversation in English. Boussaguet herself [40] refers to earlier experiences, such as the RAISE project and the Move Together Consensus Conference, in which the sole use of English increased the elitist character of events initially designed to enable the diversity of European viewpoints to be heard. The risk of an elitist slide via English language proficiency was also observed for CoFoE by Alvaro Oleart [41], who describes how this inequality closely associated with that of class tends to favor the word of certain participants in the drafting of the working groups’ final recommendations. He also describes how the appointment of ambassadors to the European panels helped to favor and empower those most comfortable with English and the technocratic language of Europe, as the “official” version of the final recommendations was first drafted in English and later translated.

More generally, Risse [42] points out that English proficiency is strongly correlated with level of education, age, and mobility, as well as with whether or not one belongs to a small country. As Fligstein [43] points out, this type of profile is very much in line with those who are the most pro-European, which would tend to show that proficiency or lack of proficiency in English also reflects a social and linguistic divide between those who do or do not possess a high level of transnational linguistic capital, to employ Jürgen Gerhards’ concept [44]. While statistics on Europeans’ proficiency in languages, and in English in particular show major differences between countries [1,45], they also tend to demonstrate that this proficiency is dependent, on the one hand, on linguistic distance from the native language, but also on the respondent’s socio-economic profile and level of education: the higher the respondent’s professional position, the more common the use of English in everyday life tends to be [24]. This socially differentiated fluency is compounded by the fact that although 40% of Europeans believe they can hold a conversation in English, only 21% consider their level to be “very good” [45]. While the aim of random selection is to extend the range of participants in public deliberation, the exclusive use of English would have the opposite effect, favoring the social profiles of individuals who are already highly privileged and open to European issues. Such a choice would therefore appear to contradict the objectives of inclusion pursued by the use of random selection.

4.2. Inspirations for New Multilingual Deliberations

For deliberative democracy, the linguistic dimension is particularly interesting as an important bridge between, on the one hand, deliberative processes that involve only a microscopic portion of the European population, and, on the other, debates on the emergence of a European public sphere, where one of the main challenges is to successfully address the constraints of the continent’s linguistic diversity, leaving the language regime as an open question [46]. From the 2010s onwards, reflections on deliberation have emphasized the need to think about this connection and to make a “systemic” turn [47,48,49]. In other words, thinking about the multilingual environment in continental deliberation processes is not only important for improving these same processes, but also for thinking about the conditions for the emergence of a genuine European public sphere: the multilingual issue constitutes an important tactical nexus for thinking about this interconnection. Is it possible to find formulas that combine the benefits of monolingualism and multilingualism for deliberative mini-publics? Several models including specific multilingual regimes have been explored for European institutions [32,50], for example through the concept of inclusive multilingualism [51] which formalizes the use of different modes of communication (English as lingua franca, intercomprehension, use of interpreters, etc.). Regarding participatory and deliberative processes, Laurie Boussaguet [38], for example, recommends alternating monolingual and multilingual phases using the “carousel” method, based on the model of the 2006 MOM project. On a longer-term note, Nicole Doerr [33] refers to the European Social Forum as the emergence of a new intercultural “activist linguistic repertoire”, in which participating Europeans tend to mix languages as they communicate and deliberate with each other.

The emergence of a simplified form of English or a European English is also an option to be explored, as is the use of potential artificial languages such as Esperanto [52], whose simplicity of learning could enable Europe-wide debates of the kind that already exist [53]. If, according to Van Parijs [9], language is a common good, why not take hold of it on a European scale, or even think of a specific language without national branding [42,48,54] that would be as easily accessible as possible, whether it be a revised Esperanto or a simplified, “Europeanized” English, in tandem with clear protection for multilingualism [55,56]? One could also consider broadening the number of languages used for this type of process: why limit oneself to official languages? Why not include sign languages, but also the languages of linguistic minorities (regional languages or migrant languages)? While it is clear, for example, that not all Europeans can access the assistance of the European Parliament’s interpreters on a daily basis to debate with their fellow citizens, we can imagine that digital tools will be increasingly present and mobilized, whether for mini-publics or maxi-publics [57,58]. Recourse to artificial intelligence and machine translation could be valuable allies in substantially lowering interpretation costs and increasing the inclusiveness of these processes [17,59].

5. Conclusions

Regarding our hypotheses, while the European Parliament’s interpretation services seem remarkably efficient, some specific technical constraints persist with contradictory effects on the quality of deliberations between citizens. Surprisingly, deliberation in a multilingual environment can have positive effects on the inclusion and participation of participants by ritualizing and formalizing exchanges and disciplining the process of communication. On the other hand, certain technical constraints can run counter to the European Union’s declared objectives and values in terms of promoting and showcasing linguistic plurality as a key element of European identity. Thus, while the promotion of multilingualism is also a socially and politically situated value, we observe that even in the context of the Conference on the Future of Europe, the guarantee of a multilingual exchange is never completely assured and offers outlets for the pre-eminence of certain languages. The possibilities for extending and broadening this study are considerable, in both quantitative and qualitative dimensions. In particular, the use of analytical methods to precisely quantify the number of interventions related to the languages spoken would enable us to confirm whether there are any biases in favor of certain languages. Semi-structured interviews with participants, particularly those from linguistic minorities, would enable us to observe the effects of multilingual deliberations (sense of legitimacy, self-censorship mechanisms, etc.) in greater detail. In-depth interviews with the organizers could also help uncover the existence of possible linguistic biases.

If the political success of this process remains unclear [60], the CoFoE made it possible to implement a high-level multilingual deliberation system on a continental scale, and its participants had access to high-quality interpretation. While interpreting also brings about a general transformation of the deliberation regime, it is interesting to note that it can have potentially positive effects in the inclusion and integration of the most minorized participants: the significant formalization of the deliberation framework and the procedural slowness make it possible to avoid phenomena of inequality in interventions that it would be interesting to continue to investigate and record. On the other hand, the lack of available live interpretation during the informal sessions supported collective logics of language-based gatherings, linguistic domination and isolation of monolinguals. The conclusions drawn from this study suggest that future organizers of such events should give greater thought to informal times and their linguistic accessibility, for example by providing interpreters who could intervene directly at the request of participants. The written dimension of CoFoE’s deliberative framework, while meeting multilingual constraints for the main documents produced by the conference, was much more complex in the daily exercise of deliberation as it slowed down and complicated the drafting of participants’ recommendations, and because of its shortcomings, also tended to reproduce linguistic inequalities in favor of English. The issues highlighted by this study surrounding the written production of such processes are probably the most complex to resolve here, but also the most exciting. It requires thinking about the articulation of working groups and the languages involved, as well as the physical presence of experienced human interpreters capable of differentiating between the resolution of linguistic misunderstandings and the formalization of genuine political disagreements. The development of European software and AI specialized in these issues would also be an opportunity to develop genuine European democratic expertise that could also be deployed in other transnational deliberation processes.

In both cases, the limits of translation and interpretation lead to a renewal of dominant linguistic logics under the guise of practicality: recourse to English or the major European languages in the absence of interpreters or when translations fail, valorization in deliberation of the most educated and multilingual participant profiles. This ignorance of the link between equality and multilingualism is even reflected in some of the recommendations made by participants, who suggest objectively favoring the use of English within the European Union in order to improve its functioning, thereby contradicting the values of diversity that have been promoted throughout the process. Yet there are many ways in which the values of equality and diversity could be further integrated into these initiatives. In the short term, for example, they could be deployed through the cultural and artistic promotion of the Union’s minority languages, as well as the opportunity to make participants from majority languages aware of their privileged position, through games or role-playing. In the medium term, the use of methods such as multilingual intercomprehension or presentations on linguistic inequalities in the EU could be explored. Finally, in the long term, promoting and encouraging the parallel learning of constructed, non-nationally marked auxiliary languages would also allow for the development of a form of open European identity and limit the generation of inequalities between native speakers.

These linguistic challenges are particularly important indicators of the desired framework for these deliberations. The Conference on the Future of Europe is probably one of the most successful models of multilingual deliberation to have existed to this day, it offers concrete approaches to issues that often remain highly theoretical, such as the construction of a plurinational democracy or European identity. Multilingual democracy has yet to be built. It takes original and unexpected forms. Above all, like any democratic exercise, it requires a particularly significant effort in terms of time and resources.