The Idea of Discursive Constituent Power

The question addressed by this article is whether a form of constituent power exists at the EU level. It is argued that European integration has not suppressed the idea of people as constituent power. Instead, the idea of ‘people’ has been constructed through the discourses of security and rights. Ever since the early stages of European integration, the security and rights discourses have consisted in the articulation of a meta-constitutional rationale, which is here called the ‘security of the European project’, i.e. a form of political morality that is pursued by the EU as a polity over time and aims at its own survival. Security and rights discourses have contributed to constructing two ideas of ‘people-as-constituent-power’. The first idea is that of ‘mobile people’, i.e. people exercising EU free movement rights. The second idea is that of ‘peoples’ in the plural, conceived as States and citizens at the same time. Nevertheless, these discourses are characterised by a certain degree of ambiguity and have been unrolling as if the development of the EU polity was a mere technical, neutral matter. This state of affairs cannot continue: the European project has always been a political project, and, as integration reaches its more advanced stages, the time has come to disclose its political nature and address conflict openly.

formal may be opposed to informal change. Formal change consists in the amendment of one or more provisions according to prescribed procedures. In the EU this type of change takes place through the procedures established, more recently, by the Treaty of Lisbon 1 . More precisely, the ordinary revision procedure, the simplified revision procedure, the general passerelle clause and the procedure for the accession to a Treaty are all provided for by the current Treaty (Article 48 TFEU). Rather than looking at these procedures in detail, it is important to observe here that, although the European Parliament and the European Commission have the right to propose Treaty amendmentsat least in the ordinary and in the simplified revision procedure -Member States retain a considerable weight throughout the whole procedure, especially at the end, during the inevitable negotiations at the Intergovernmental Conference and the final ratification. However, this consideration needs qualifying. In fact, according to the CJEU, it is only possible to amend the Treaties through the amendment procedures laid down in the Treaties themselves 2 . While Member States can conclude reciprocal agreements, which may subsequently be incorporated into EU law (as in the case of the Schengen agreement, the European Financial Stabilisation Facility (EFSF), the European Stability Mechanism (ESM) or the Fiscal Compact), the actual provisions of the treaties can only be modified in compliance with Article 48 TFEU 3 .
Informal change takes place beyond the restrictions of a text. It consists in practices and activities that may alter the substance of a provision or customary rule. Because the meaning of norms is often involved in this type of change, the activity of interpretationincluding judicial interpretationis one of the most important examples. Informal change thus corresponds roughly to Ackerman's category of 'unconventional change', which in his approach is a form of higher law-making through collective action, which takes place outside prescribed rules for amendment (typically, Article V of the US Constitution), in five distinct phases: signalling, legitimating, proposing, triggering, ratifying and consolidating 4 .
At least theoretically, change in the EU may occur in many directions, including a progressive erosion of the values that are placed at the heart of the polity by the very Member States that have created it. Nevertheless, as will be argued in this article, the very existence of procedural and substantive limits alludes to the notion that Member States' pouvoir constituant is not absolute, but it is expressed by a form of political morality that secures EU integration.
The focus of this work is on a conceptualisation of informal change and its impact on the EU constitutional structure. While keeping in mind the remarkable differences between the USA and the EU 5 , I would like to qualify this form of change constitutional 6 .
Informal change possesses a constitutional character not in the thick sense, which normally applies within the State. It rather consists in discursive practices, including language, which spread across society at many levels (courts, media, scholarly work, etc.). Although it is not captured by the formula contained in Article 48 TFEU, informal change may sometimes be at least as profound and pervasive. The idea of discursive constituent power, suggested in this work, plays an important role in the context of informal change.
The article proceeds as follows. In Section 2 it illustrates how a unique form of constituent power has characterised the development of the EU, which is called discursive constituent power and is articulated through security and rights discourses that have emerged in the process of European integration, starting from the early years. These discourses are part of a meta-constitutional rationalethe security of European integrationwhich is a form of political morality underpinning the EU as a polity. The article shows in Section 3 the distinctive traits of discursive constituent power. This approach is then compared with similar approaches configuring constituent power at the EU level.

Discursive Constituent Power
The argument illustrated in this article is that European integration has not neglected the idea of people as constituent power. European integration is constitutional integration, insofar as it does not displace the-people-as-constituent-power, but makes an idiosyncratic use of it within both a political and legal framework. This argument needs to be clarified. It is a fact that the EU is founded on a set of Treaties. On this basis, it has historically emerged as a creature of international law. As noted by Weatherill, the EU's basic purpose is the 'management of interdependence', although it retains the ambition to transcend a system based on inter-State consensus 7 . At the same time, distinctive traits can be found, including the combined presence of institutions claiming popular representation such as the European Parliament, institutions representing the States such as the Council of the European Union and independent institutions such as the European Commission and the European Central Bank, as well as an organ, the Court of Justice of the EU (CJEU), which cooperates with national courts and carries out a self-defined process of constitutionalisation 8 . In a curious blend, the ongoing relevance of the role of Member Statesconfirmed by the restriction of the EU's powers and competences to those conferred upon the EU by the Member States themselvesmust be contrasted with the practical flexibility of the Treaties' provisions, which allow the EU institutions, within the limits mentioned above, a relatively wide room for manoeuvre. The ability and tendency of the EU to stretch the limits of its mandate is especially visible in times of crisis but can also be deduced from the interpretative activity of the CJEU. The EU legal order is also capable of conferring rights and obligations on private parties, and there exists a certain degree of control on the exercise of public power. These are all signs that EU integration cannot be easily reduced to merely international or administrative integration. Moreover, the ongoing transformation of domestic constitutional systems; the extended EU competence, e.g. in matters of social justice; and the growing penetration of EU law into national law are a peculiar phenomenon that characterises the EU. Understandablyfrom different perspectivesseveral scholars have rejected the doctrine of the autonomy of the EU, viewing the EU as based on the 'mutual recognition of the moral standing of each Member State' 9 or on a civic model of 'cosmopolitan statism' that regards the good of being an equal member of a democratic polity as valuable and worthy of respect in itself 10 . Isiksel employs a more nuanced model, called 'functional constitutionalism', i.e. a form of constitutionalism beyond the State characterised by the functionally delimited scope of the EU legal order's authority 11 . This purposively oriented public authority, cutting across national and subnational levels of governance and detached from the feature of sovereignty, would be inspired neither by the principle of the rule of law, like the Rechtstaat, nor by the principle of solidarity, like the Sozialstaat, but by principles of expediency, expertise and power-building 12 . There seems to be thus a tradeoff between democratic legitimacyjustifying the maintenance of core elements of State sovereigntyand effectiveness, justifying functional expansion. Interesting, all the approaches mentioned above highlight the importance of the EU institutional structure and consider the EU from the perspective of what it is for, rather than merely of what its nature is. This is a welcomed contribution to the debate.
However, there is a sense in which the presence of a popular element and the construction of a public sphere cannot be easily dismissed. True, the viability and appropriateness of the idea of European people seem to require further refinement. For example, in his model of 'reciprocity-based internationalism', Sangiovanni claims that obligations of social justice can apply also beyond the State, if they are understood as demands for fairness in the distribution among people of the costs and benefits associated with the joint production of collective goods 13 . Such approach places emphasis on the institutional structure of the EU, looks at the purpose of the EU as a State-enhancing, rather than State-transcending construction, and develops a notion of solidarity that embraces also workers and migrants 14 . The presence of the feature of other-regarding traits in the construction of the idea of European people can be a useful starting point for the configuration of a peculiar concept of constituent power.
While regarded with suspicion by several versions of liberalism, the-people-as-constituentpower acts as a powerful rhetorical device, if viewed from a particular angle of liberal thought. 15 The liberal-democratic idea of attributing pouvoir constituant to the people as the subject of the founding act 16 and, therefore, the author of a radical break from the pastespecially an authoritarian arrangementis intimately connected to individual empowerment and the security of a polity (in the sense of self-preservation). This happens in two ways. First, the abolition of pre-existing structures, a break of the status quo, is normally justified through a reference to individual freedom. 17 Second, a liberal legal system legitimises itself through the construction of an idea of 'people' which, however, is fictitious. A 'people' does not really exist, or does so only retroactively, post factum, in such a way that only an act of force, as it were, founds the law. 18 In a sort of 'vicious circle', as often observed, no people is authorised to constitute a polity before the actual polity is constituted. 19 This holds true at the transnational level just as it does at the national level. 11 T. Isiksel (2016). 12 Ibid. 78-83. 13 A. Sangiovanni (2013). For a similar view, see P. Eleftheriadis (2020). 14 Sangiovanni distinguishes between national solidarity, Member State solidarity and transnational solidarity. 15 For this strand of liberal thought, see, e.g. T. Paine (1995), 579 (identification of constituent power with the 'people'). In the post-absolutist era, others have located constituent power in the more or less equivalent 'nation ': E. Sieyès (1970). 16 However, conceptions of constituent power not bound to the idea of people are also possible (e.g. political parties, ruling élite, monarch). See Colón-Ríos (2014); P. de Vega (1985); C. Mortati (1961); J. Bryce (1901). 17 C. Möllers (2006), 186. 18 J. Derrida (1986), 10. 19 See, e.g. H. . For the expression 'vicious circle', see R. Carré de Malberg (1922).
Nevertheless, during the process of European integration, it is possible to trace the development of a discursive constituent power. In other words, the idea of 'people' is constructed through the discourses of security and rights. How does this happen?
The argument goes as follows. The notion of constituent power is strictly related to the notion of sovereignty as creative power, i.e. the power to constitute, to foundas opposed to sovereignty as coercive or repressive power. 20 This productive, creative dimension of sovereign power, as an original power that simultaneously grounds a constitutional order from within and generates it from the outside, has been often associated with 'the people' by some of the earlier theorists of liberal constitutionalism, as noted earlier. 21 However, constituent power does not consist merely in the exercise of a specific type of power by a people at any given moment. It also constitutes people as an entity that did not exist beforehand. 22 The consideration of the self-constituting nature of this power leads to a deeper reflection on the problem of attribution of acts to a collective subject. As Lindahl points out, 'the attribution of legislation to a collective is first and foremost an act of self-attribution, that is, an act by which the members of a community recognise legislative acts as acts of their own community'. 23 Self-constitution and self-attribution inevitably indicate that the object of our discussion is ultimately an act of self-empowerment. Yet, at this point there emerges one of the several paradoxes associated with constituent power. For, if this is really an act of self-empowerment, how can we attribute it to a fictitious subject that did not exist beforehand and whose existence is constantly debatednamely the European people? What are the limits of constituent power, if any?
In an attempt to go beyond the mere opposition between the reduction of law to mere originary fact, and the configuration of constituent power as fully inherent in the constituted order, Negri argues that constituent power poses a 'radical question' and, 'insofar as it constitutes the political from nothingness, is an expansive principle'. 24 Negri's intention is to emphasise the creative dimension of constituent power as 'first decision' by pointing to its 'original radicalness' or 'radical openness', i.e. its ability to open a number of 'grounding, innovative, linguistic and constitutional possibilities'. 25 However, in doing so he defines constituent power as a pure 'act of choice', which necessarily stands in opposition to sovereignty as an act of repression, and is linked indissolubly with an absolutistic interpretation of democracy and the image of a continuum between law, revolution and constitution. 26 In other words, 'the question is not to limit constituent power, but to make it unlimited', because, in the end, 'the process started by constituent power never stops'. 27 20 For the conceptualization of sovereignty as power of command, see J. Bodin (1992) and later versions, e.g. J. Austin (1832). 21 J. Locke (1991), 366-67;T. Paine (1995), 579;E. Sieyès (1970), 180-91. C. Mortati (1946). More recently, see also A. Negri (1999), where constituent power is defined as 'the source of production of constitutional norms that is, the power to make a constitution and therefore to dictate the fundamental norms that organize the powers of the state. In other words, it is the power to establish a new juridical arrangement, to regulate juridical relationships within a new community'. 22 M. Loughlin (2014), 229. 23 H. . 24 A. Negri (1999), 16. 25 Ibid. 20. 26 Ibid. 22-24. 27 Ibid. 24.
Based upon these premises, it is possible to argue that constituent power does not dissolve once it has been exercised, but 'remains alongside and above the constitution'. 28 However, unlike the voluntaristic interpretation that follows from Sieyés, I argue that the ambiguity of EU liberal constitutionalism resides in the fact that, to avoid contradicting its own premises, it must limit itself. In fact, EU liberal constitutionalism has safeguarded the interplay between change and permanence (the security of the European project) by disguising constituent powerthe truly political dimension of the process of integrationin the form of security and rights discourses. This is important, especially once we recognise the significance of constituent power as a legal power. 29 The argument above highlights the opacities of the meta-rationale of security, which operates at the same time to silence constituent power and (re-)activate it.
Invocation of change 'in the name of the people' is always already made from within the legal system. It is a rhetorical device, employing the creative force of thepeople-as-constituent-power as if this were a real entity. This is how the security of the European project is promoted. The EU liberal project thus can only survive if it does not negate constituent power, but reinstates it as discursive constituent power through the self-justifying discourses of security and fundamental rights. Consequently, I believe we should resist the temptation to characterise European integration as a process replacing constituent power with constitutional rights, on the one hand, and individual economic freedoms, on the other 30or to view it as a primarily economic process. 31 On the contrary, the political has never been really removed from the inner core of the process of European integration. 32 Although, on the face of it, political power was supposed to be confined within the national sphere, 33 in practice transnational market integration has disguised a macro-political agenda which could only operate effectively if it was kept behind the curtains. 34 The ever-present political dimension of European integration, far from being limited by security and fundamental rights discourses, is instead expressed through them. These discourses have contributed to constructing two ideas of 'people-as-constituentpower'. The first idea is that of 'mobile people', a category of people that are supposed to benefit from EU free movement rights. They circulate from one country to another mostly to satisfy professional ambitions or their aspiration for a better life or to receive a service. The second idea is that of 'peoples' in the plural: as the imagery of a unified people constantly conflicts with the reality of the European 28 C. Schmitt (2008), 125-26. Also, '[e]very genuine constitutional conflict, which involves the foundations of the comprehensive political decision itself, can … only be decided through the will of the constitution-making power itself'. 29 C. Schmitt (2003), 82-83; referring to M. Hauriou (1923), 284. 30 See, instead, M. A. Wilkinson (2017), 51-59. 31 See, e.g. K. Tuori (2015), 26-28, where the author argues that while the political and juridical constitution has enjoyed constitutive primacy, the economic constitution has enjoyed functional primacy. The distinction between constitutive and functional primacy is not entirely clear, especially if it is argued that 'European integration has been primarily an economic project, and in spite of the expansion of EU activities into new policy domains, economic integration retains a dominant position' (ibid. 26). 32 F. Scharpf (2010), 89, where the EU is defined as a polity possessing liberal credentials, but lacking republican credentials: is this not eminently a matter of political self-design, which affects the development of all sectors of EU law? (ibid. 93). 33 J. Weiler (1992). 34 M. Bartl (2015), 578. diversified landscape, the powerful depiction of 'peoples', conceived as States and citizens at the same time, is a recurring object of study for EU scholarship. 35 As has been observed recently, '[t]he presupposition of a European people, as the collective subject of the European legal order, does not exclude the continued presupposition of European peoples, in the plural, as the collective subjects of national legal orders'. 36 Discursive constituent power is thus a powerful tool for the justification of the EU liberal project through the creation of a public sphere where discourses freely develop. The position adopted here thus dissociates itself from that part of normative legal theory, which either relegates constituent power outside the legal domain, to the sphere of the political, or anyway believes it is a redundant concept. 37 Normativists sometimes label as 'negatively prescriptive political theories' those theories that deny that law's authority derives from the intrinsic qualities of legal orders and, by so doing, do not answer clearly the question whether such authority is internal or external to law. 38 Since it is argued here that this paradox is expressed by way of a fiction, taking place through discourses operating within the EU legal order, the approach followed here is not strictly decisionist either.
Scholars in the past have sought to find a middle way between normativism and decisionism, by way of the so-called relational approach. Relationalism sets out to understand the political space as a space of unresolved conflicts, and of a constant dialectic between closure and openness. 39 As a result, there emerges a tension between the sovereign exercise of constituent power and the way sovereign authority is expressed once political power is institutionalised. In other words, it is possible to identify in the moment of foundation a symbolic act, through which a multitude of people come together as a group: this act will inevitably involve the use of some degree of force. However, force and conflicts do not vanish with this symbolic transcendental act but persist during the constitutional development of a polity. The reason for this state of affairs is that the institutional arrangement designed in a given historical and political context never corresponds to the actual decisional authority of the institutions. From a relationalist perspective, power resides 'neither in "the people" nor in the constituted authorities; it exists in the relation established between constitutional imagination and governmental action'. 40 One implication of this dynamic approach is that constituent power does not emerge merely at the foundational moment, only to disappear or remain concealed in some obscure location. Rather, constituent power continues to operate within a polity, and, by doing so, it preserves the political space as an open space of contestation. This amounts to configuring polity buildingin our case, EU polity buildingas a disputed process, which does not necessarily respond to the logic of communicative rationality. 41 However, once we recognise the deeply conflictual nature of constituent power as a 'living power', we should also be prepared to admit that any claim by a part to act on behalf of a whole is fictitious. 42 No constituent claim can be universal, because there is no homogenous will. And yet, in order for this claim to be legitimate, it needs to be represented as if it were universal. It will be an all-inclusive, all-encompassing claim, but it will nevertheless always be a partial claim. The EU liberal project exemplifies these ambiguities, because, while putting forward its claim to universality, it always already draws a line between those who should be included and those who should be excluded. 43 Given the partial nature of any constituent claim, there cannot be any metaphysically unified people endowed with sovereign authority. 44 The reason is that, just like any other polity, if not more than any other polity, the EU polity is fragmented and heterogeneous. What is more, the tension between the constituted arrangement within the EU polity and the openended range of alternatives which are theoretically available for the exercise of constituent power cannot be reduced to mere disagreement. Rather, such tension can and should always constructively open the way for alternatives, which are situated beyond the purely liberal paradigm that has been followed so far. Thus, discursive constituent power contributes to a process of constant change, which could theoretically allow for possibilities of development outside the existing ones. 45 However, the practical exercise of such power is affected by the tendency of the EU liberal project to 'neutralise' the security and fundamental rights discourses. This does not mean merely to convert political disputes into technicalities, as has instead been argued. 46 The political is still present, en travesti, as noted earlier, in such a way that security and fundamental rights appear as universalistic and all-embracing, whereas they are in reality always partial, pursuing specific political agendas.

The Features of Discursive Constituent Power in the EU
The issue of the existence of constituent power in the EU has been debated for several decades in different contexts. The most emblematic effort to provide a definition of transnational or EU constituent power was made by Jürgen Habermas, who adopts a relationalist approach, according to whichas observed in the previous sectionconstituent power simultaneously involves the exercise of power by the people and constitutes a people. In other words, he advocates a move from post-democratic executive federalism to transnational democracy, on the basis of shared sovereignty between Member States and citizens. Hence the presence of a dual constituent power, or pouvoir constituant mixte, represented co-originally by the citizens of the EU and the peoples of Member States: for these purposes it is not necessary for transnational democracy to acquire the connotations of a federal State 47 . There are interesting elements in this view. A Constitution is considered to be emerging over time through a process of bootstrapping or self-legitimation: a founding moment which wills a Constitution into existence is thus not necessary and even illegitimate 48  approach lies in the claim that citizens and elites ought to begin to self-consciously reframe European integration as constitutional integration by constructing a public conversation that leads to such outcome. Founding moments are only legitimate if they become part of a futureoriented self-correcting learning process over time, which of course also includes regressions and obstacles 49 .
One of the main difficulties deriving from Habermas' approachextremely simplified in the lines aboveis that, to remedy the so-called bootstrapping paradox or 'vicious circle' (namely, the fact that constituent power pre-supposes as already existing the very community that it purports to create), it needs an 'inspired' population 50 . How and in what sense a population can be 'inspired' to initiate a process of self-constitutionalisation is a matter of dispute. Alternative approaches have highlighted other specific aspects that should be associated with the transnational dimension of constituent power. For example, Lindseth and Fasone view the introduction of a rule of law conditionality mechanism as part of the Next Generation EU scheme as potentially producing a genuine constitutional change, because it would imply an at least partial transferral of the capacity of resource mobilisation to the EU level 51 . The resulting constitutional moment would emerge, first, from the combination of the already existing 'constraining' function of constitutional law with the 'power-enabling' function and, second, the replacement of the old rule of recognition with a new rule of recognition, which can only occur through politics and can only be durable if it retains the forms and language of the old law 52 . Others, such as Eriksen and Thornhill, although from different perspectives, point out the constituent force of fundamental rights. From a political legitimacy perspective, Eriksen criticises Habermas' 'weak supranationalism' for undermining the principle of citizens' self-rule, by blurring the distinction between the right of citizens to participate in the process of collective opinion and will formation and the rights of States 53 . The configuration of a dual constituent subject deprives the EU polity of a unifying component and is questionable, because it seems to make a we-feeling and a sense of finalité unnecessary. By way of contrast, Eriksen suggests that the individual as a right-bearing subject should be the pouvoir constituant and the only source of legitimation for the EU, because the genuine rationale of the process of integration is the protection of individual dignity 54 . From an interesting sociological perspective, Thornhill detects an inner-juridical constituent power, which has succeeded in establish a direct channel between citizens and EU institutions, especially the CJEU, precisely by constructing individuals as rights-holders 55 . Reference to international human rights norms has allowed the CJEU not only to carve out its own constituent authority through an expansive case-law, but also to promote a process through which EU public law has been recursively generated by and within the legal system itself 56 .
More explicitly, Patberg shows the emergence of 'public narratives', which call for a complete transferral of constituent authority from the Member States to the people 57 . Rather optimistically, he views signs of a democratic renewal in the setting up of the Conference on 49 J. Habermas (1996). 50 F. Michelman (1998). 51 Lindseth and Fasone (2020). 52 Ibid. Other scholars instead refer to a 'synthetic constitutional moment', which already occurred with the creation of the European Coal and Steel Community. See Fossum and Menéndez (2011). 53 E. O. Eriksen (2018). 54 Ibid. 203. 55 C. Thornhill (2018). See also C. Thornhill (2014). 56 Ibid. 23. 57 M. Patberg (2020). the Future of Europe 58 . In contrast to functionalist and retroactive notions of constituent power, which focus on governmental or judicial bodies or emphasise the relevance of ex post facto self-empowerment, Patberg reformulates constituent power as a normative, futureoriented notion, which is based on the citizens' political autonomy or self-determination, in their dual role of European and national citizens, free from the interference of the institutions. As a result, EU constituted powers should not be allowed to operate as de facto constituent powers, and the 'people' should replace the 'masters of the Treaties'. Essentially, this is a modified version of Habermas' dual constituent subject, which, instead of citizens and people, consists of citizens 'all the way down' 59 . For these purposes, Patberg resorts to a thought experiment, whereby he aims to determine under which conditions all participants in a hypothetical constitutional moment could agree to establish a supranational polity 60 . Unlike Habermas, Patberg claims that there is no need to assume that Member States will keep existing in their current form. Moreover, his strictly individualistic notion seeks to bypass collective actors and allows no stratification, because both sides of the pouvoir constituant mixte are situated at the EU level 61 . In other words, supranational constitution making requires a vertical, not only horizontal, separation of powers, which contests the idea of permanent founding, i.e. of a constitutional framework that has been created once and for all. In order to solve the problem of the 'usurpation of constituent power', he elaborates a model of 'extraordinary partisanship', i.e. a new political actor, whose sole purpose is that of preparing the ground to enable citizens to enact constitutional change 62 . Finally, Patberg proposes to turn the Convention set out in Article 48 TEU into a permanent constitutional assembly, which would act on its own initiative, would be composed of directly elected representatives and would consist of two chambers, representing the two sides of pouvoir constituant mixte.
The abovementioned efforts to identify a specific nature of constituent power in the EU have been strongly criticised by those who believe that constituent authority only resides in the Member States 63 . Despite these strong objections, I believe that the attempt to single out specific features of transnational constituent power is praiseworthy. However, its distinctiveness from the original conception should be emphasised more, and the following pages represent a move in this direction.
The argument in this work is that, given the complexity and fluidity of the transnational legal setting and its fragmented nature, it is not useful to pinpoint individual elements, such as 'people', 'rights' or 'resource mobilisation' that are supposed to capture the essence of constituent power. Nor is it useful to focus on one particularreal or hypotheticalconstitutional moment, which would separate one period from the other, mostly because constitutional moments can be protracted for a very long time and it may sometimes difficult to keep analytical distance from the events.
Discursive constituent power is different from traditional constituent power, because it is not represented by a collective expression of will by a 'people' at any given time, but rather constitutes them as an entity that did not exist beforehand 64 . As a result, alongside classic formal change, informal change has been promoted by way of discourses of powervisible, more famously, through CJEU's rulings, but also EU institutions' policies, lawyering, scholarly work, the activity of other non-State actors such as social movements and private companies, etc.which contribute to constructing a certain idea of 'people'. Despite the naturally controversial definition of the idea of 'people', change is often performed or proclaimed in their name 65 . Europe is both a multiplicity of 'peoples' and a space where mobility can be exercised outside State borders with fewer restrictions than in the past.
In light of the features illustrated above, the conception of discursive constituent power, as opposed to traditional constituent power, does not aim to recall the distinction between original and derived (derivative or amending) constituent power 66 . The former is the revolutionary power, exercised prior (and leading) to the creation of a legal order. The latter is exercised within the limits set by an already existing Constitution. As well known, in this regard two approaches may be roughly distinguished, depending on whether constituent power is regarded as extra-legal and unlimited 67 or as the power to amend or replace a Constitution regulated by specific procedures 68 . These two lines of thought may be traced back to, respectively, Sieyès and Rousseau. According to the first line of thought, the amendment of a Constitution can only take place outside the legal order, via a special organ (e.g. a convention). The second view, instead, places constituent power within the legal order. However, the distinction between original and derived constituent power has been blurred by many recent theorists, who have admitted that derived constituent power can act much in the same way as original constituent power. Despite this, they have sought to find limits to constitutional change, for example, by identifying the 'fundamental aims of the State' as an absolute limit to the possibility of amendment 69 . The idea of discursive constituent power follows this second mind frame. It is based on the assumption that the distinction between original and derived constituent power relies on an either/or approach: either legal or extralegal (possibly illegal). The either/or approach does not admit of socially embedded forms of constitutional change from within the constitution, which imply various degrees of combination between constitutionalism and democracy. Going beyond that simplistic distinction points towards a form of amendment that calls for an alternative way of interpreting a norm or a set of norms, which challenges what was until then the commonly perceived sense of legality. In 64 M. Fichera (2018) Dicey (1959). 69 See, e.g. C. Mortati (1971), 121. Such 'fundamental aims' are reflected in the 'material constitution', which as opposed to the 'formal constitution' is an expression of the ideological positions of the dominant political forces in a society. other words, we may detect different nuances in the exercise of constituent power, which correspond to different combinations of constitutionalism and democracy.
Moreover, discursive constituent power is also different from the notion of secondary constituent power, which has been used to denote a power that is not simply derived from, but also subordinated to the primary constituent power 70 . The reason is that discursive constituent power is unlimited only in abstracto, because it may not contradict certain ideological premises of liberal democracy, such as the rule of law and the protection of fundamental rights, without contradicting the very preconditions of its own activity. In other words, it operates as a result of a renewal and renegotiation of the EU's commitment to a union of values. As a result, there is no need to rely on the distinction between primary and secondary constituent power, which may be useful in a domestic legal system, where the locus of sovereignty is univocal and the hierarchy of sources is well defined.
To be clear, it is assumed that, unlike US constitutionalism, in the case of European constitutionalism, neither the moment of foundation nor the process of constitutional change is characterised by a break with previous legality. Contrary to a purely Hartian perspective, followed by scholars such as Richard Kay 71innovative, including revolutionary change, does not take place 'from a point of view external to the legal system (…) as ingredients of a position based on political history and morality' 72 . The idea suggested here is exemplified by episodes of significant change in the EU constitutional framework, which modify previously existing fundamental principles or concepts but remain an expression of law. The model of discursive constituent power as theorised above presents affinities with Ackerman's model of 'unconventional change', as noted earlier. However, it refrains from placing too much emphasis on constitutional momentsunlike Ackermanbut rather focuses on the gradual process of transformation that characterises European integration. Moreover, although it is in some respects close to Lindahl's IACA-model of law, or theory of a-legality 73 , it does not cover what Lindahl defines 'unordered and unorderable' challenges to a legal order, i.e. that type of challenges that may lead to an entirely new legal order, but only 'unordered but orderable' challenges.
I would like to qualify as 'unordered but orderable' those challenges that are not meant to upset a legal order, but come from within it. Because discursive constituent power does not dissolve after the birth of a polity, but persists thereafter, it preserves the political space as an open space of contestation. We shall examine later what this function of preservation means more precisely. Suffice it to say, for the moment, that the tension between the constituted arrangement within the EU polity and the open-ended range of alternatives which are theoretically available for the exercise of constituent power cannot be reduced to mere disagreement. Rather, such tension can and should always be seen as leading constructively to renewed arrangements and, as a result, as constitutive of the EU polity. Consequently, discursive constituent power faces a number of important dilemmas. As observed in a previous work: On the one hand, security is pursued by preserving the core values of the European project, as perceived by a more or less extended section of the society. Yet, this also implies that the formulation of these values must correspond to some degree to the 70 Y. Roznai (2017), 113-134. 71 R. Kay (1987). 72 R. Kay (1997), 182. 73 H. Lindahl (2013; H. Lindahl (2018). interpretive, legislative and administrative practices which take place in that society. On the other hand, the risk is that security relies on a self-referential imagery, which does not open up a space for other realities. Essentially, constitutional regeneration across the decades cannot be truly accomplished by restricting the space for alternative visions of liberal democracy 74 .
It is important to keep in mind that the political dimension of European integration has always been present, even when the latter was presented merely as transnational market integration. This has occurred despite the fact that security (i.e. polity self-empowerment) and rights discourses have been configured as neutral, self-justifyingas if economic measures could never have implications for the political framework of nation States 75 .
It is almost inevitable to recall in this context Ackerman's dualist democracy. As well known, his model stands, according to himself, in between 'monism', which emphasises popular sovereignty (in particular, Parliamentary sovereignty) over fundamental rights, and 'rights foundationalism', which gives priority to the commitment to fundamental rights, as expressed in a written Constitution 76 . Ackerman's 'dualism' distinguishes between a higher law, represented by the Constitution, and ordinary law: only fundamental rights enshrined in the former would need to be preserved over time, for example, through the instrument of judicial review. Judicial review presupposes constitutional, rather than judicial, supremacy, because it expresses the supremacy of 'We the People' over government officers.
Dualism represents thus an attempt to reconcile constitutionalism and democracy, which is increasingly relevant for the EU legal system.
Yet, one of the risks stemming from such endeavour is that by conflating constituent power and interpretation on the one hand, and constituent and constituted bodies on the other, we elaborate an excessively simplified account. This is why a number of important clarifications are necessary. Can the activity of the CJEU, with its features tending simultaneously towards 'creation of law' and 'preservation of law', qualify as constituent? How do we distinguish between constituent power and interpretive activity? While some scholars have labelled the activity of the CJEU as 'revolutionary' or 'transformational' 77 and have not refrained from considering the CJEU as the locus of constituent power in the EU 78 , others are loath to do so 79 . In reality, there is reason to believe that most if not all famous landmark rulings have had some genuinely innovativethus constitutivecharacter. Nevertheless, it is difficult to accept that they were part of a revolutionary movement against the Member States. After all, Member States never openly opposed them, although it is admittedly difficult to amend the Treaties as a tool to counteract judicial law-making 80 . This relative acquiescence can also be explained by the obligation of loyal cooperation and the obligation 'to abstain from any measure which could jeopardise the attainment of the objectives' of the Treaties 81 .
The role of Member States in the polity-building process cannot thus be dismissed. However, EU constituent power does not lie only with them, nor with any particular organ or body or 'elitist' group of people, but consists in the aforementioned discourses, existing 74 M. Fichera (2018), 46. 75 G. Majone (2014). 76 Ackerman (1991a, b), 6-16;32-33;35-36. 77  ; K. Alter (2000); J. Weiler (1994). 78 H. ; L. Corrias (2011) (arguing that constituent and constituted powers should be conflated). 79 Baquero , 226. 80 K. Alter (2001), 182. 81 This obligation was already contained in Article 5 of the Treaty of Rome. See now Article 4 TEU. between institutions and other social actors and historically emerging already at the beginning of European integration 82 . There are a number of reasons why this is the case. First of all, I do not believe that, even if one were to admit that a specific court may possess constituent power, the CJEU on its own would exercise it. The Luxembourg court necessarily needs the collaboration of national courts in the first place. Several important changes to the EU legal framework have been promoted thanks to a rather intense interaction between the CJEU and national constitutional and ordinary courtsthus forming the 'EU judiciary' as a system of courts that ensures 'that in the interpretation and application of the Treaties the law is observed' (Article 19 TEU). Such interactionwhich has been marked by increasing conflict in recent yearsoften takes place over an extended period and its full impact may sometimes be properly assessed only after some time.
Secondly, it bears mentioning again that the notion I am proposing is distinct from traditional constituent power á la Sieyès. As well known, Sieyès argued that, once citizens have freely agreed to commit to and obey the rules of a legal and political order, they have the right to participate in its formation and retain some form of political authority. Interestingly, he was wary of employing the notion of sovereignty 83 and preferred to elaborate the notion of constituent power as a virtually unlimited power to constitute a State, which is separate from the constituted powers and is exercised by the people. His theory therefore aimed at legitimating a legal order through mechanisms of explicit authorisation by the citizens 84 . However, in practice the notion of constituent power á la Sieyès was not as 'popular' as might appear at first sight. Unlike Rousseau's understanding of popular sovereignty, Sieyès' intention was to circumscribe the exercise of constituent power to a location and a moment that are situated extra legem. Constituent power is thus conferred upon representatives, who act on behalf of the people as a whole and as a result are not bound by specific instructions 85 . Once the legal order is formed, we are only left with constituted powers, which are supposed to operate within the boundaries set by the established constitutional framework.
As opposed to the notion above, discursive constituent power is instead a form of constituent power that goes beyond a clear-cut distinction between constituent and constituted powers 86 and relies upon a deliberative process, characterised by an intense interaction involving the EU judiciary and other organs and actorsincluding not only European institutions, but also national authorities and the civil society at large. In this configuration, constituent power operates within the framework of the polity, not outside or prior to it. It does not lie dormant after the creation of the polity but contributes to its ongoing formation. Because it stems from inclusive deliberation, it is a practice that becomes part of ordinary political life. True, in the case of the EU, not only was there never a 'people' in the first place, but there still currently is no people. This does not imply automatically that European integration has no constitutional character. As mentioned earlier, rights and security discourses were present already at the onset of European integration, despite the fact that no provision explicitly mentioned rights in the Treaties of Rome. After all, as noted perceptively by Baquero Cruz, even the principles of direct effect and primacy, which were not mentioned 82 M. Fichera (2018), 39-63. 83 'Opinion de Sieyes sur plusieurs articles des titres IV et V du projet de constitution' in M. Dorigny (1989). See also L. Rubinelli (2020), 33-74. 84 E. Sieyès (1970). 85 For a short summary, see , 2020b It has been noted that constitutional systems have achieved a degree of complexity that goes beyond a simplistic dichotomy between constituent and constituted power. See Jacobsohn and Roznai (2020), 135.
in the Treaties and have been one of the most significant innovations by the CJEU, could be considered to be a likely outcome within a range of possible interpretations of the Treaties themselves, which, inter alia, already provided for the mechanism of preliminary ruling 87 . Consequently, what has counted is the incremental process leading towards a (not yet) fullyfledged polity. In other words, our focus is here not so much on the act of 'seizing the initiative' 88 by whoever claims to speak on behalf of the people, but on the gradual colonisation of a legal and political space by way of discourses.
Yet, tensions and contradictions have always been associated with these discourses. In this perspective, it is only by looking at a gradual, incremental process of (self-) constitutionalisation that it is possible to understand the transformation of the EU from a common market into an area of freedom, security and justice in the absence of an explicit authorisation by the people 89 . The point is not, of course, whether that transformation has been accomplishedit has notbut whether the grounds for this self-reinforcing process exist. Although I would refrain from calling the constitutional change performed mainly by the EU judiciary a 'juridical coup d'état' 90 , one should take seriously Stone Sweet's argument that '[t]he ECJ can command that national judges interpret and apply [EU] law (…), but it cannot force them into following its lead', because it lacks normative authority 91 . Although such normative authority exists, it is thin, because it is not the outcome of an explicit act of authorisation. Yet, such authority exists, also because it derives from a form of political morality that is expressed emblematically inter alia by a strict cooperation between the CJEU and national courts. These courts, acting jointly, are able to alter the meaning of both the national constitutions and the EU legal framework 92 . In order for the EU judiciary's authority to be strengthened, two steps are necessary. First, a shift from a self-referential modeaccording to which EU law sets the conditions for its own existence and validityto a fully heterarchical mode, which allows for alternative options and claims to be assessed, debated and pursued, if necessary 93 . Second, a more inclusive process of deliberation in which more institutions and larger sectors of civil society are involved. This process of deliberation incorporates both practices of decision-making, collective interpretation and reflection on key issues of contemporary society. One of the implications of deliberation is the possibility to subject important decisions on what ought to be changed and what ought to be preserved to test and verification.
Thirdly, the subject of our inquiry is not merely a retrospective, but also a forward-looking constituent power. In other words, it is mostly a future-oriented perspective that lends the European project credibility and authority. The retroactive constitutionalisation of the moment of foundation would lose its credibility and its claim to innovativity and durability if it did not also project such claim towards the future. Courts are crucial actors in this context, as confirmed by Alter, according to whom much of the input given to European integration can be explained by looking at the different 'time horizons' of judges and politicians 94 . Whereas political leaders often have short-term strategies, the CJEU has tendentially pursued a long-term strategy of dialogue and rapprochement.
Fourthly, the security of the European project as a form of political morality is characterised by an interplay between constitutional change and preservation 95 . While some elements of Burke's thought may be traced here 96 , I reject his appeal to the principle of prudence as a universal, eternal and unchangeable law 97 . Moreover, preservation and continuity do not depend on a fully formed constitutional consensus, which never really exists in society. Instead, the EU constitutional framework relies on a deliberative practice 98 , which necessarily presumes contestation and contradictions, as noted earlier. In other words, constitutional fundamentals ought to be seen as part of an ongoing collective debate about the political community and the idea of a final act of closure are replaced by the idea of openness and flexibility. Hence, another distinctive trait of discursive constituent power, which sets it apart from some of the traditional versions of constituent power, is that it does not only tend to 'create', but also to 'preserve'. It is 'constituent' in the sense that it is based on an open-ended, future-oriented moral judgement on what the essential features of the EU polity should be.
Fifthly, as may have been glimpsed from the analysis above, the notion of discursive constituent power may purport, if not necessarily to replace, at least to complement the language of State or EU-level sovereignty that is sometimes employed. The idea is to show that the interaction between legal systems may be examined, both descriptively and normatively, through new conceptual tools. From the point of view of its nature and purpose, the notion suggested here bears affinities to that of the 'material constitution' developed by Mortati 99 . Just like Mortati's constituent power arises from the social forces and interests that shape society (i.e. the material, as opposed to the formal, constitution), so 'discursive constituent power' relies on discourses of power that circulate in the form of judicial rulings, official documents, scholarly writings and other speech acts that contribute to giving life to the EU constitutional framework (sometimes even conflicting with each other). Just like Mortati's constituent power is associated with the realisation of a fine politico (political goal), so 'discursive constituent power' is inspired by a political morality that expresses the need to protect the autonomy and dignity of the individual 100 . Yet, differently from Mortati, the involvement of various layers of society is more complex 101 .
To illustrate this point, I return to Patberg's version of pouvoir constituant mixte. He refers to Habermas' important consideration that we should avoid reifying constituent power as a macro-subject and conceive political autonomy as a manifestation of subjectless forms of communication circulating through forums and legislative bodies 102 . As a result, in order to ensure a genuine process of deliberative democracy, the idea of a 'sluice system' is proposed. According to this idea, the relationship between centre, namely, the political system, and periphery, namely, civil society, is legitimate only if communication flows begin at the 95 M. Fichera (2018). 96 As Burke famously observed, 'A state without the means of some change is without the means of its conservation. Without such means it might even risk the loss of that part of the constitution which it wished the most religiously to preserve'. E. Burke (1909) para. 36. 97 P. J. Stanlis (2003). 98 P. Eleftheriadis et al. (2018). 99 C. Mortati (1971). 100 Interestingly, Mortati criticised Schmitt's notion of homogeneous people as 'static', always equal to itself in time and thus impermeable to social forces. One may add that Schmitt's notion of people reflects a fixed conception of constitutional time. 101 In fact, Mortati's theory has been subject to criticism in Italy. Some scholars have recently attempted to refine it: see, e.g. A. Barbera (2010). 102 J. Habermas (1996). periphery and are channelled 'through the sluices of democratic and constitutional procedures' 103 . Sluices are sets of formal and informal rules that enable the exercise of popular sovereigntyespecially, initiatives 'from below'-within the bounds of discursive rationality. This can only happen if a representative body, analogous to domestic parliaments, allows to separate constituent and constituted powers at the supranational level, while facilitating the emergence of what Patberg calls 'higher-level constituent power' within a formal, legal framework. The creation of a permanent constitutional assembly pursues exactly the purposes mentioned above. Patberg's proposal is certainly immune from the type of criticism that is levelled at the so-called 'strong popular sovereignty model' 104 , because it invokes the emergence of a pan-European constituent power that solves the problem of institutional selfinterest by conferring on citizens a key constituent role. My objection to this argument, with which I broadly sympathise, is that it risks creating yet another representative body, thus reproducing the very entity that it seeks to displace. Patberg distinguishes between constituted actorsrepresenting citizens directlyand constituted powers: while the latter deal with daily policy making, the former are created with the specific purpose of representing citizens' constituent power. 105 The permanent constituent assembly is an example of constituent actor. Yet, even if the permanent constituent assembly does not turn into a 'sovereign dictator', which Patberg correctly aims to avoid 106 , we would still witness an additional form of institutionalization, with its own rules, its own codes of conduct and claims of autonomy, i.e. another 'we, the people' acting in name of the whole.
Contrary to Patberg's alleged 'usurpation of constituent power' by the EU institutions, I believe that the institutional dimension and the societal dimension should be viewed together, as part of a continuum within which discursive constituent power, with its own contradictions and ambiguities, operates. Yet, the societal dimension should be strengthened vis-á-vis the institutions. This is not a call for a court-driven over-constitutionalisation, but rather for a more open dialogue between courts and non-State actors, especially at the local level, including NGOs, private companies, social movements, think-tanks and scholars. In other words, court rulingsat least the most important onesshould not be viewed as discrete speech acts, but as part of an intricate web of meanings, which is generated by the interaction between all social actors in the EU polity.
Habermas' idea of 'dual constituent power'exercised by individuals as simultaneously national and EU citizensultimately relies upon the role of elites and the media as catalytic agents able to 'win over' the population through a top-down momentum that 'must abandon incrementalism steered by experts' 107 . Patberg's version of 'dual constituent power', instead, advocates a full bottom-up approach, driven entirely by citizens. Neither approach is entirely satisfactory. A top-down method of integration does not fully live up to the ideal of democracy in the very moment it proclaims it: the role of elites and media as catalytic agents risks degenerating into mechanisms of manipulation and coercion if it is not transparent and open to contestation. A bottom-up technique contradicts itself, because it ultimately cannot help advocating the creation of institutions representing in some form or another citizens' constituent power. In a way, this vision betrays a partial configuration of the EU, which is not only a political system, but also a legal system. As a legal system, the EU combines two elements: the normative and symbolic features of a legal order and the institutionalised legal practices of legislation, adjudication and enforcement 108 . Both legal order and institutional practices are deeply intertwined with national practices and institutional frameworks and must therefore have a role even when the constitutional change of the nature of the EU itself is at stake. Patberg's theory seems not to account for this and treats EU law as if it were mere international law. The idea of discursive constituent power indicates thus a middle way between the approaches mentioned above, one in which both have equal value. Despite the existence of a common cultural-political ground, the considerable degree of heterogeneity and complexity of EU countries increases the likelihood of conflict whenever decision-making affects delicate areas such as human rights or economic and monetary policy, especially as European integration achieves more advanced stages of development.
In other words, political morality aims to lead toor creates the conditions fora thicker conception of democracy, one in which the premises and expectations of liberalism are put to the test throughout a collective deliberation by way of public discourses on different conceptions of the common good. A legal order acquires legitimacy not from the criteria of validity that it itself has created, but from an engagement with the demands of civil society 109 . For this to happen, the dialogue between courts is only a starting point 110 .

Conclusion
It is argued in this article that European integration has not suppressed the idea of people as constituent power. Instead, the idea of 'people' has been constructed through the discourses of security and rights. Ever since the early stages of European integration, the security and rights discourses have consisted in the articulation of a meta-constitutional rationale: that of 'securing the European project'. I interpret 'security' as a form of political morality that is pursued by the EU as a polity over time and aims at its own survival. All fundamental principles of EU law, including primacy, direct effect, autonomy and effectiveness, are an expression of such meta-constitutional rationale.
The ever-present political dimension of European integration is thus not limited by the security and fundamental rights discourses, but expressed through them. Security and rights discourses have contributed to constructing two ideas of 'people-as-constituent-power'. The first idea is that of 'mobile people', i.e. people exercising EU free movement rights. The second idea is that of 'peoples' in the plural, conceived as States and citizens at the same time. Nevertheless, these discourses are characterised by a certain degree of ambiguity and have been unrolling as if the development of the EU polity were a mere technical, neutral matter. This state of affair cannot continue: the European project has always been a political project, and, as integration reaches its more advanced stages, the time has come to disclose its political nature and address conflict openly.
However, discursive constituent power is different from traditional constituent power, because it does not consist in a collective expression of will by a 'people' at any given time, 108 K. Tuori (2014). 109 Although Böckenförde argued in favour of the distinction between constituent and constituted power, this approach bears some affinities with his idea that constituent power still resides within the Constitution and is restrained by the ethical-moral awareness that a people historically acquires of itself as a political unity. See E. W. Böckenförde (2017). 110 C. S. Nino (1998). but rather constitutes them as an entity that did not exist beforehand. As a result, alongside classic formal change, informal change has been promoted by way of discourses of powervisible, more famously, through CJEU's rulings, but also EU institutions' policies, lawyering, scholarly work, etc.
In other words, this is a form of constituent power that goes beyond a clear-cut distinction between constituent and constituted powers and relies upon a deliberative process, characterised by an intense interaction between the EU judiciary and other organs and actors including not only European institutions, but also national authorities and the civil society at large. In this configuration, constituent power operates within the framework of the polity, not outside or prior to it. It does not lie dormant after the creation of the polity, but contributes to its ongoing formation. As a result, this article advocates inclusive deliberation, as a practice that should become part of ordinary political life in the polity.
Funding Open access funding provided by University of Helsinki including Helsinki University Central Hospital.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.