A theory of plural constituent power for federal systems

Abstract

Federations present difficulties for prevailing theories of constituent power, which usually attribute ultimate constitution-making authority to a singular people. This article examines how a ‘pluralized’ constituent power functions in federal systems. It argues that the operation of plural constituent power in federations reflects a distinctive model of constitutional formation according to which a ‘polity of polities’ is established and sustained through the maintenance of a tension between plurality and unity.

I. Introduction

The resurgence of interest in constituent power reflects both practical and theoretical concerns. ¹ Remarkably, over the last 30 years the majority of the world’s constitutions have either been newly adopted or significantly amended. ² Since the late 1980s, the number of nations identifying as constitutional democracies has almost doubled, with the collapse of the Soviet Bloc, the fall of Latin American dictatorships and ongoing processes of decolonization in Africa and Asia all precipitating a sharp increase in constitution-making activity. From a more theoretical perspective, the concept of constituent power offers a way of making sense of the proposition – which is core to the self-understanding of contemporary liberal-democracies – that legitimate public authority ultimately derives from citizens, or ‘the people’. This in large part explains the appeal of constituent power to normative theories of democratic political legitimacy, including those that seek to uphold a commitment to constitutionalism while opposing populist, theocratic and other avowedly illiberal political regimes. ³ Despite the prominence of constituent power in contemporary constitutional practice and theory, however, its operation within federal constitutional systems remains a relatively neglected topic. This neglect is at first sight surprising, given the prominence of federal democracies, including India, the United States, Brazil, Germany, Canada and Australia. In these democracies, it is possible to speak not only of ‘the people’, but also ‘the peoples’ of the constituent states. Our aim in this article is to redress the relative neglect by examining the implications of a plurality of peoples for the concept of constituent power in federal systems.

Constituent power is defined broadly in this article as the power to create, replace, or fundamentally amend a constitutional order. ⁴ As Lucia Rubinelli has argued, since the late eighteenth century revolutions, constituent power has provided a ‘language’ for articulating popular power. ⁵ The language of constituent power has its own logic or ‘grammar’, which distinguishes it from the broader idea of popular sovereignty, understood as a ‘supreme’ source of power. ⁶ Most fundamentally, constituent power ‘constitutes legal-political structures and … as such, is in a direct conceptual relation with the idea of a constituted power’. ⁷ The meaning and function of constituent power are determined by its relational reference to the institutional structures and competences that it constitutes. This relational aspect of constituent power is consistent with its ‘liminal’ status between politics and law, and holds even if one subscribes to an ‘extra-legal’ conception of constituent power as the pre-constitutional, political source of legal legitimacy, rather than emphasizing the need for its exercise to be ‘constitutionalized’. ⁸ The relational aspect of constituent power is, for the theory we propose, conceptual bedrock: it is presupposed both by descriptive questions regarding attribution of the power to the people, plural peoples or the ‘nation’, and by normative questions about the criteria for legitimate or authoritative constituent acts. On the juridical level, the representative character of constituent processes also allows the collective citizenry to be understood as the bearer or subject of constituent acts that they commission (and perhaps ultimately either endorse or reject). As Hans Lindahl and Martin Loughlin suggest, a constituent people is a unity that can be understood as constituted reflectively or relationally by the constituent activity carried out in its name. ⁹ On the normative level, the juridical attribution of constituent power to ‘the people’ is usually thought to entail that a constituent process will only be legitimate or truly authoritative when it is sufficiently acceptable to the majority of citizens to receive their explicit or perhaps implicit endorsement.

Some reasons for the relative neglect of plural constituent power emerge implicitly from this discussion. On a historical level, while the concept of constituent power is anticipated in Huguenot resistance theories, English Civil War parliamentary propaganda, Rousseau’s theory of popular sovereignty and elsewhere, it was paradigmatically thematized by Emmanuel-Joseph Sieyès in the context of efforts to establish and legitimate new republican forms of government during the French Revolution. Sieyès argued that the people of the entire nation is the source of a pouvoir constituant, which is delegated to representatives elected specifically for the purpose of drafting a constitution. Sieyès believed the constituent power to be vested in the Nation as a whole, opposed a division or distribution of this power among constituent units, and rejected the proposal that local or regional assemblies participate in ratification of the constitution. While Sieyès developed the idea of constituent power in opposition to the concept of sovereignty, he insisted on the unitary nature of the body politic and argued that the nation must express its will through a single legislature. ¹⁰ Carl Schmitt’s influential Weimar-period interpretation of constituent power (verfassunggebende Gewalt) also considered constituent power primarily within a monistic frame, attributing it to ‘the people as a unity capable of political action with the consciousness of its political distinctiveness and with the will to political existence.’ ¹¹ Schmitt does, as Olivier Beaud has argued, offer resources for a pluralistic account of federal constituent power. ¹² Yet Schmitt also thought federations especially vulnerable to political instability due to competing sovereignty claims. ¹³ For Schmitt, as we explore in section 4, any division or federalization involving self-governing units is likely to result in fragmentation of a nation into a multiplicity of alternative identities. ¹⁴

The privileging of a unitary conception of constituent power is attributable to other factors. Most obviously, there are more unitary states than federal states. ¹⁵ Conventional state theory has assumed the ‘nation’ as its substrate, and hence focused on constitution-making through a single constituent assembly. ¹⁶ More recently, constituent power has been prominently invoked in jurisdictions – particularly within Latin America – where political actors have sought to reorganize the state unilaterally and bypass potential opposition. ¹⁷ Apex courts have also tended to adopt a conception of popular power that papers over the complexity of federal arrangements. The best known example is the United States, where there have been competing interpretations of the famously elusive constituent subject, ‘We the People of the United States’, with its ambiguity between a singular ‘people’ and a plurality of ‘states’. ¹⁸ This ambiguity gives rise to two alternative temptations: pluralized interpretations that fail to do justice to the ‘union’ presupposed and established by a constitution, and unitary interpretations that obscure the systemic effect of a pluralized constituent power at the empirical level in terms of the design, content and the functioning of federal constitutions. ¹⁹

An adequate theory of contemporary constitution-making processes must come to terms with the way plural forms of constituent power function in federations. Federal constitutions may be in the minority, but they remain numerous and influential enough for their distinctive features to warrant greater attention. There is evidence to suggest that federalism is likely to become more prevalent as a governance model, with international advisers and other actors promoting federal systems as suitable for the post-conflict societies in which most constitution-making activity is currently taking place. ²⁰ There is a risk, in this context, of interpreting federal constitutions through a ‘monist’ lens which distorts their distinctiveness. A useful analogy is ‘imposed’ constitutions. Constitutions imposed by external or international actors – while also in the minority – have received increased attention recently, and raise distinct issues of constitutional formation relative to ‘non-imposed’ constitutions. ²¹

The concept of federation is, of course, also contested in the literature, alongside related ideas such as federalism, federal system, federal compact and confederation. Ronald Watts offered an influential approach in which ‘federalism’ is a normative term, whereas ‘federal system’, ‘federation’ and ‘confederation’ are descriptive. ²² Watts argued that in a ‘confederation’, by contrast, the system of government depended not on ‘the people’, but the ‘constituent governments’. ²³ While this can be an important distinction, it passes over another possibility: that the constitution of a federation might be derived from a plurality of ‘peoples’. ²⁴ It also does not take full account of the constitutional pre-commitment inherent in any federal constitution to plural governments, whose authority may rest upon constitutionally embedded or territorially located plural sources of legitimacy. This point applies, as we explore in Section III, to both aggregative and devolutionary types of federal system.

In what follows, we employ a combination of theoretical and comparative methodologies to develop a general theory of plural constituent power for federations. The distinctive feature of the operation of constituent power in federal systems, we argue, is a constitutive and enabling tension between multiple sites of constitutional authority. Considered from a ‘genetic’ perspective, the operation of plural constituent power in federations reveals a distinctive model of constitutional formation whereby a ‘polity of polities’ is established and sustained through the maintenance of a tension between principles of plurality and unity. Federal constituent power hence establishes a structural tension between unity and plurality, such that neither the unified polity nor the plural polities assume complete predominance. The federation is a distinct form of constitutional order, our analysis suggests, which should be regarded as categorically different model of rule from that found in unitary constitutionalism. ²⁵ A federal constitutional order contains multiple sites of constitutional personhood; it is a polity comprised of a plurality of polities and sites of constituent authority.

This article has the following structure. Section II introduces the central theoretical issues through an examination of recent scholarship on mixed constituent power (pouvoir constituant mixte) in the European Union (EU). Section III sets out the conceptual foundations of our plural theory of constituent power for federal constitutions. In Section IV we then apply this model to (A) constitution-making processes, and (B) procedures for constitutional amendment. Section V outlines some of the broader implications of our analyses by exploring the relationship between the federal pluralization of sites of constituent authority and democratic pluralism.

II. Mixed constituent power (pouvoir constituant mixte) in the EU

Contemporary accounts of constituent power often reflect a monistic mindset that sees constituent power in terms of a single constitutional people. ²⁶ One notable exception to this tendency is recent scholarship on the EU’s constituent foundations. Before directly examining plural constituent power for federations, it is therefore instructive to briefly consider theoretical debates regarding mixed constituent power (pouvoir constituant mixte) in the EU.

The applicability of constituent power to the EU is not universally accepted. It has been argued that constituent power is inessential to European constitutionalism, or that it is empirically absent from the European context, or that it is best seen as nascent. ²⁷ Relatedly, some theorists contend that the EU is paradigmatic for the displacement of constituent power by rights as the main source for the legitimacy of fundamental law. ²⁸ Others continue to focus on the treaty-foundations of the EU as a decisive indicator of its ultimate legal foundations, only to be countered in turn by those who advance a ‘constitutionalism’ thesis. ²⁹ Even if one is reluctant to accept wholesale the idea of a ‘post-constituent constitutionalism’, it is important to recognize that there is a ‘reconstructive’ (in the Habermasian sense) aspect of appeals to European constituent power, and also to acknowledge the downstream implications of this for questions of democratic legitimacy. ³⁰ Despite these qualifications, debates on European constituent power offer a useful platform for reflection on pluralization of sites of constitutional authority, and there are arguably more similarities between the constitutional foundations of the EU and ‘aggregative’ federations than is sometimes acknowledged. ³¹

Markus Patberg identifies four models that predominate in the literature concerned with the application of constituent power to the EU. The first of these models is a ‘regional-cosmopolitan’ constituent power, ‘according to which the EU is a non-state political system that derives its democratic legitimacy directly from free and equal persons who appear on the scene as European citizens, rather than members of nation-states’. ³² The second is a ‘demoi-cratic’ or ‘plurality of peoples’ model, according to which the EU ‘should answer to a constituent power composed of the political communities of the member states’. ³³ The third is the ‘dual’ constituent power model, according to which constituent power in the EU is shared between a European people and the peoples of the member states. ³⁴ Jürgen Habermas’s argument for a pouvoir constituant mixte, in which ‘the constituting authority should be composed of the entire citizenry of Europe, on the one hand, and the peoples of Europe, on the other’, is the best-known example of this view. ³⁵ The fourth, and more critical, view suggests that the ‘neo-liberal hegemony’ of the EU needs to be challenged by disruptive and non-conventional political action outside the ordinary processes of European governance. ³⁶

Building on his analysis of these four models, Patberg argues for an account of the constituent power of the EU grounded in the capacity of citizens to play two roles as both (1) citizens of the ‘unifying’ polity and (2) citizens of the plural political units that comprise the ‘lower’ level. ³⁷ Patberg’s analysis of the pouvoir constituant mixte, with its focus upon the identification of the subjects of plural constituent power, has relevance beyond the EU. While Patberg’s own preferred model for the EU privileges the constituent role of citizens, this reflects a normative, ‘reconstructive’ and ‘future-oriented understanding’ of constituent power, which grants priority to the conditions under which a rationally autonomous citizen can ‘democratically determine the structure and competences of public authorities’ to which they will be subject. ³⁸ From a more juridical perspective, which is our focus here, the constituent subjects of the EU order can be regarded as (1) the member polities of the EU or (2) the EU polity. The EU, on this view, is a ‘polity of polities’ and reflects a logic of plurality-in-unity and unity-in-plurality.

While the EU has distinctive features, its status as a polity of polities mirrors at least one core structural dimension of the federal constitutional orders that we examine more closely below. ³⁹ In the case of the EU, from a historical or ‘genetic’ point of view, the polities that signed up to the EU treaties preceded the ‘polity of polities’. Yet the ‘polity of polity’ concept can be seen to apply in relation to both aggregative and devolutionary federations. From a ‘genetic’ viewpoint, there is only a uniting plurality of constituent polities at the time the federal constitution is made; it is only after making the constitution (and thus the federation, or the polity of polities) that the single polity comes into existence. At best, the single polity is nascent at the moment of federalization. In the case of devolutionary systems, the plural polities are also nascent at the moment of devolution. If, for example, the ratification vote to form an aggregative devolution resulted in a ‘no’ decision, then there would be no single people formed, just as when a devolutionary proposal is constitutionally rejected by the decision-makers, then no plurality of peoples would be formed (i.e. the outcome is always contingent on the actual result of the constitution-making or constitution-amending process).

A fully formed federation is, from a structural point of view, a polity of polities that comprises both a unified polity and plural constituent polities. Crucially, the relationship between the unified polity and the member polities can vary, potentially greatly, from one federal constitution to another. ⁴⁰ For a full picture of plural constituent power, it is necessary to recognize the ‘institutional’ dimension of constitution-making processes. On an institutional level, a polity of polities can be understood as a polity with a singular government for the whole and plural governments for the constituent parts. Further questions concern the structure and composition of the institutions empowered to act for the unified polity and for the plural polities of the federation, and the connection between these institutions and the founding authority of the federation. We now turn directly to these topics.

III. Pluralized constituent power for federal constitutions

A federal system is a polity of polities comprising both a unified polity and plural constituent polities. In abstract terms, federal systems exemplify a structural tension between unity and plurality, such that neither the unified polity nor the plural polities assume total predominance. In the next three sections, we argue that the operation of constituent power in federations is informed by this enabling tension between plural sites of constitutional authority. This section examines the conceptual foundations of this plurality-in-unity and unity-in-plurality, initially by reference to the identity and interrelationship between multiple sites of constitutional authority in aggregative and devolutionary federal models.

It is uncontentious that federal systems complicate a monist model of constituent power, understood as a power exercised by a single constituent assembly or convention and attributed to a unified people or citizenry. Federations can be created in three main ways: through ‘aggregation’ of formerly separate states; through a ‘devolution’ within a formerly unitary state; or through a ‘combination’ of these two processes. ⁴¹ It is to be expected that federations formed by aggregation will attribute their origin to a plurality of constituent states or peoples, while the unitary systems from which devolved systems are derived will attribute their origin to a singular people or state. Yet the matter is more complex, as the language used in the preambles of federal constitutions attests: some use the singular, others the plural, and yet others combinations of singular and plural language to describe the ‘people’ or ‘peoples’ from whom the constitution derives its authority. ⁴² Moreover, the constitution of a federal system formed by devolution may attribute sovereignty or constituent power to ‘the people’, but the devolutionary process may have been initiated or advocated by a plurality of nascent ‘peoples’ at a point in time following the original establishment of the relevant constitution. ⁴³

Aggregative federal constitutions display genetic and structural features that presuppose a pluralization of constituent power, including the formation of the constitution through institutional processes that rest on a plurality of constituent states, the representation of the people of those states in the democratic institutions of the federation, the distribution of powers between constituent states and the federation, and complex prescribed pluralistic methods of amendment. ⁴⁴ These features are also found in plurinational constitutions with a multiplicity of territorially concentrated communities that possess or aspire to self-government and are either recognized as, or claim to be, discrete sources of constituent power. ⁴⁵ In some federal constitutional systems, it is true, an originally unitary state has devolved power to subnational jurisdictions, without this necessarily pluralizing the constituent power. ⁴⁶ Yet it is equally misleading to assume that the distinction between ‘aggregative’ and ‘devolved’ federations is always clear-cut; wherever the federal model of a ‘polity of polities’ is applicable, then one must ask after the plurality of sources of constituent power in their specific application. ⁴⁷

Our analysis, to be totally clear, does not assume uniformity across federal constitutional systems. This point goes beyond the above distinction between aggregative and disaggregative systems. A federation is a polity with plural governments, which may include self-government of the constitutive polities of the federal constitution, and the sharing of central authority between constitutive polities and the dedicated central institutions of the polity. Again, this can – but need not necessarily – involve robust demotic pluralism of plural constituent peoples. As seen below, the federation, as an ideal type, rests on a structural tension between unity and plurality, but in actuality one pole usually tends to predominate.

A federation is, we have noted, a polity of polities or a political community of political communities. ⁴⁸ The significance of this core feature of federal constitutions is best seen by reflecting on the multiple subjects of constituent power. As Olivier Beaud argues, ‘the first original feature of the federative compact is the plurality of the holders of the constituent power’. ⁴⁹ A federal constitution is a constitutional order that is structured by the relationship between plural polities which have distinctive constitutional identities. The essential features of this structure are best introduced by a contrast with the unitary constitutional state model.

On a unitary model of constituent power, the collective citizenry is understood as a unified ‘people’ or ‘nation’, which serves as the bearer or subject of imputation for the creation or material amendment of a constitutional order. It is plausible that the factual reality of a shared will of the people dissolves once one moves beyond political or juridical reconstructions of the origins of legitimate power. The will of the people is usually determined by a constitutional convention or assembly, and hence requires representation. From a public law theory or juridical perspective, unitary constituent power can nonetheless be seen as follows:

‘The people’ → Exercise of constituent power → The constitution

Constituent power is exercised in a unitary fashion in this model insofar as the subject of its exercise – the people or the nation – is understood as a unity comprised of individual citizens.

For constituent power in federations, the first and most significant divergence from the unitary model is a ‘pluralization’ of ‘the people’ and of constituent acts. Schematically, there are multiple collectives of citizens – the peoples of multiple polities – that serve as plural subjects for the constituent power of a federal constitution. The founding act of constitution-making plays a significant role in defining the political bond between ‘the people’ and its new system of government. The political bond that is created (and thereby constitutionalized) in a federal constitutional foundation cannot, however, be framed in conceptually unitary terms. Such a bond must be understood as uniting territorially discrete polities – as a plurality – both to one another and to the governing system of constitutional authority at the federal level. ⁵⁰

A difficult question raised by the pluralization of constituent power in federations is whether the unified compound ‘people’ established by the member polities is a discrete bearer of constituent power. In the case of an aggregative federation, the free agreement of a plurality of polities to establish a federal constitution has temporal priority in relation to the federal polity. Yet, when the peoples of plural polities agree to federate, there is a sense in which the nascent singular people have also agreed. For when majorities of several distinct peoples vote in referendums or through conventions in favour of federation, it follows that a majority of the people of the federation as a whole have also voted in its favour. Thus, procedurally, when a plurality of peoples decides to become a ‘federal people’, there is sense in which this federal people may be said to have brought itself into existence. The relational nature of constituent power, as a power to create constitutional order attributed to a subject unified as a result of constituent activity, means the newly created unified people of a compound polity can be seen as a subject of constituent power. Rather than a mere ‘horizontal’ pluralization of the subjects of constituent power at the level of the member polities of the federation, there is here also a pluralization on the ‘vertical’ dimension, so a unified constituent subject is established comprised of the totality of citizens in the compound polity. This would allow citizens, as in Patberg’s model, to be ascribed a constituent role across two ‘levels’. At a juridical level, constituent power can also be understood as pluralized along both horizontal and vertical axes. This allows for the possibility of a ‘nationally’ aggregated citizenry that, through this aggregation, becomes a unified and in some sense unitary constituent power, albeit one mediated by citizen attachments to horizontal plural polities, and albeit that this unified ‘federal constituent power’ coexists with the discrete (and continuing) sources of constituent power that reflect the plural polities together forming the unified federal polity.

These admittedly abstract formulations are best elucidated by reference to the texts of prominent federal constitutional settlements. Consider the preambles of the federal constitutions of the United States, India and Ethiopia. These read, respectively: ‘We the People of the United States’, ‘We, the People of India’ and ‘We, the Nations, Nationalities and Peoples of Ethiopia’. ⁵¹ In all three preambles, one finds a complex interplay – and indeed tension – between the singular and plural. There are subtle differences in how the constituent subject is framed in these cases. Whereas the Indian example suggests a more unitary conception of a singular national people, the Ethiopian preamble expressly refers only to multiple nations, nationalities and peoples. The US Preamble is the most ambiguous of all, as it shifts between the unity of ‘the People’ and the plurality of the different states, which are brought into a more perfect ‘Union’ by the constitutional settlement. In each case, however, the plural subject ‘We’ encapsulates the compound nature of the federal polity as a polity of polities insofar as it designates a subject that, considered as a collective unity of polities, can be viewed under the aspect of either unity or plurality. Both the plural constituent polities and the compound polity can, these examples suggest, be regarded as the constituent subjects of the federation.

This model certainly applies more neatly to aggregative than disaggregative federal compacts. In aggregative federal systems, such as Australia, Switzerland and the United States, a plurality of pre-existing polities is simultaneously preserved and transformed through their constituent activity into the form of a unified federal polity. In disaggregative federal systems, such as Belgium, Italy and Spain, a unitary state is devolved so that particular regions within the state are provided with autonomous powers of self-government. ⁵² There is clearly a difference between the two types of federal system at the level of constitutional genesis: in the aggregative case the movement is from plurality to unity, in the disaggregative case from unity to plurality. ⁵³ Ultimately, both types of federation preserve the tension between plurality-in-unity and unity-in-plurality identified above. Conceptually, it is also possible to see a disaggregative federation as establishing plural constituent subjects, even if the status of plural constituent subjects is not always reflected in the text of a disaggregative constitution, but rather is conceptualized as an ‘autonomous region’ or ‘community’ (as in Italy and Spain).

A conceptual examination of the plural subjects of constituent power can only get one so far. It is also necessary to examine processes of constitutional formation and their downstream implications for constitutional structure. This sub-section has nonetheless outlined the deep, multi-layered and complex relation of tension between plurality and unity characteristic of the operation of constituent power in federations. Importantly, this tension need not be construed negatively as involving ambiguity or confusion; it is rather best understood as a ‘constitutive’ or enabling condition for the distinctive federal constitutional system and form.

IV. Implications of plural constituent power for constitution-making and amendment

This section connects our conceptual analysis of plural constituent power in federations to its institutional operation in (1) constitution-making processes and (2) amendment procedures.

The pluralization of constituent power in federal constitution-making processes

The establishment of a federal constitution involves a process of transformation, whether as a consequence of the aggregation of a plurality of relatively independent states into a more unified constitutional relationship or the disaggregation of a unitary state into a relatively more plural constitutional arrangement. ⁵⁴ It is by considering this process of transformation, and particularly its institutional mechanisms, that it is possible to discern a clearer picture of the distinctive way plural constituent power operates in the formation of federal constitutions.

Legislatures in federations, of course, often reflect a plural mode of representation – for example, through bicameral parliaments where one house offers state representation and the other house a form of unified ‘national’ representation. ⁵⁵ On our view, constituent power is analogous to legislative power in at least this sense: just as legislative power is the capacity to enact, amend or repeal a statute, so constituent power is the capacity to make or fundamentally amend a constitution: a capacity that may be constitutionally conditioned or an act of political rupture in relation to an antecedent regime. This applies whether the power (constituent or legislative) is constitutionally absolute (e.g. in the case of the UK Parliament) or constitutionally limited (e.g. in the case of the US Congress): in both cases, it is a power to make a certain kind of law (in the one case constitutional, in the other case statutory). ⁵⁶

On closer examination, however, the founding and re-founding of federal constitutions have distinctive features that reflect a pluralism of constituent subjects. As noted in Section III, a federal founding unites a plurality of polities, and hence a plurality of territorial units, and not merely undifferentiated individuals. It is the salience of these polities to the foundational constitutional moment that gives federalism its initial – and thereafter path-determining – orientation as a distinct model of constitutionalism. A federal ‘re-founding’ compact involves a significant change to the constitutional form of the polity that enlivens the constituent power, and although such significant change may depart from an existing federal arrangement, rather than arising ex nihilo, it is still in effect a foundational constituent act.

In the case of an aggregative founding federal compact, the status of the federalizing polities changes as a consequence of their agreement to form a compound polity. Such a compact allows for the creation of a new constitutional order in which previously independent polities become ‘members’ of the federation and hence change their own constituted status. ⁵⁷ In changing their own constituted status, however, the member polities maintain their political existence and are not completely subsumed within the compound polity. This tension between sovereign independence and total subsumption can be seen to inform the way federal plural constituent power is exercised at both the institutional and the procedural levels.

The institutional procedures through which plural constituent power is exercised vary considerably. ⁵⁸ At the aggregative end of the spectrum, international and supranational systems can be established by agreements among executive heads of state or heads of government (e.g. United Nations, EU), whereas federal systems can be established by decisions of the legislatures of the constituent states (e.g. Germany), by constitutional conventions established in each state (e.g. USA) or by referendums held in each state (e.g. Australia). Likewise, at the devolutionary end of the spectrum, federalized systems can be established by the legislature of an originally unitary state (e.g. United Kingdom) or through constitutional provisions or amendment procedures established within such a state (e.g. Belgium, Italy and Spain). Towards the centre of the spectrum is the formation of a federal constitution by a singular constituent assembly that is representative of the people of the nascent federation as a whole (e.g. India, Nepal) or a federal constitution ratified in a national referendum (e.g. Comoros). Moreover, among most of these examples, the constitution-making process involved several steps. In Australia, for example, the premiers of the constituent colonies first agreed to the initiation of a federating process and the legislature of each colony enacted Enabling Acts that established the exact procedures to be followed. Pursuant to these Acts, each colony was represented in a federal convention by an equal number of delegates who were either directly elected by the voters in each colony or selected by the legislature of the colony, and the federal convention debated and drafted a proposed federal constitution that was submitted to the voters in each colony for approval before the constitution was conveyed to the British Parliament for formal enactment. The federal Australian Constitution was in this sense predicated on the unanimous approval of the peoples of the six constituent states. ⁵⁹

Federal constitution-making processes can therefore involve executives, legislatures, conventions, assemblies and referendums; these institutions can be supposed to represent the constituent ‘people’, the ‘peoples’ or combinations of both; and the decision-making rules within each institution can vary from simple or absolute majorities through various levels of special majorities (two-thirds, three-quarters and so on) all the way to unanimity. As Beaud observes, in

the federal case, the relatively low number of ‘federating’ states that must ‘consent’ to the constitutional compact is very small when compared to the millions of votes in a referendum on the constitution in a unitary state. Since it is possible, therefore, to require that all states agree, the unanimity rule has practical significance so that, technically speaking, it is possible to use it to adopt the founding act of a Federation of a modern democratic state. ⁶⁰

Unanimity among the component states is an indicator and measure of the aggregative character of a federal system. Constitutive decisions taken by less than the agreement of all the states indicate a federative process relatively less aggregative. The closer the constitutive decision-making principle is to simple majority-rule, the more unitary the constitution-making process. Yet, so long as constitutive decisions are taken by institutions that represent, in some way, the constituent peoples – even if decided by a simple majority among those peoples – the location of constituent power is not yet entirely unitary and the logic of any federative process is not yet fully devolutionary. This is because a unitary state, if democratic, makes decisions with the support of representatives of a majority of its citizens, not a majority of its state polities. ⁶¹

The pluralization of constituent power in federal constitution-amendment processes

In this sub-section, we outline the significance of a pluralized constituent power for constitutional amendment processes. A pluralization of the subjects of constituent power and constitution-making processes clearly has downstream implications for the amendment procedures of federal constitutional settlements. ⁶² Most notably, the extent to which a pluralized constituent power determines the amendment procedures of a federal constitution correlates with the degree to which member polities retain their political and legal autonomy.

The predominant tendency in public law theories of the federal constitution has been to assimilate it to the unitary state model, while characterizing its distinguishing feature as rigidity. ⁶³ According to the influential analysis of AV Dicey, a federal constitution is marked by both its distribution of powers between the federal and state governments and its resistance to unilateral amendment by either of these political entities. ⁶⁴ Dicey’s account was anticipated by James Bryce, who had argued that a constitution should be regarded as rigid if the authority that is held accountable for constitutional amendment can be distinguished from the authority which is attributed with responsibility for the enactment of ordinary law. ⁶⁵

A feature of a federal constitution – most notable in, but not exclusive to, those that arise from an aggregation of previously independent states – is, however, that its structure and the procedures for its amendment will differ from the unitary model in a more fine-grained way than is captured by the general idea of rigidity. Both the unique relationship between the federation, its members and citizens, and the horizontal equality of the member polities constituting the federation will be reflected in the content of the constitution. This can be seen in four main dimensions: polity recognition; autonomous government, associational rule and processes of amendment. Firstly, a federal constitution recognizes the fundamentality of the constituent polities, whether explicitly in the title or preamble of the constitution and/or in its institutional design. Institutional design encompasses the three other dimensions. Secondly, the act of creating, guaranteeing the continuation of and empowering the autonomous institutions of government at the level of constituent polities is a categorical break from a monist conception of government with its unitary structure of government. ⁶⁶ The pluralization of sites of government within a federal system results, thirdly, in the interaction among these, and in particular in the recognition of, and involvement by, the constituent polities and their institutions in the operation of the central government of the state. This can occur at the parliamentary level (for example, through a ‘territorial’ second chamber); at the executive level (for example, through intergovernmental relations); and at the judicial level (for example, through plural jurisdictions, or constitutional recognition of the polity’s federal nature in the process of judicial appointments). ⁶⁷ Federal constitutions, finally, usually build in important roles for the constituent polities in processes of amendment, either directly through their own autonomous governmental institutions or indirectly through their engagement in associational rule. This can also manifest in forms of direct democracy that, in federal polities such as Australia and Switzerland, also take account of the polity’s territorial and jurisdictional pluralism. ⁶⁸ In both Australia and Switzerland, the constituent status of the peoples of the states and cantons is reflected in the requirement that proposed constitution amendments must secure the support of a majority of voters nationally and in a majority of voters in a majority of states. ⁶⁹ Moreover, in Australia, the representation of the states in the federal legislature and the boundaries of the states cannot be altered without the support of a majority of the voters in any state affected by the proposed change. ⁷⁰ In India, reflecting the more ‘unitary’ concentration of constitution-making power in the constituent assembly representing the Indian people as a whole, most provisions of the Constitution can be amended by the Indian Parliament by a statute passed with the support of two-thirds of the members of both houses, except for changes to particular provisions and chapters concerning the ‘federal’ elements of the Constitution, in which case the proposed changes must also be ratified by the legislatures of at least one half of the states. ⁷¹

Federal constitutions are thus established with differing predispositions about the extent to which the polity is characterized by territorially-located societal diversity, which is reflected in their institutional arrangements. It is also the case that the institutional design can be amended, or constitutional practice and interpretation will evolve, to reflect societal change – for example, the strengthening or diminution of territorially located societal diversity. A federal polity in which this dimension was always weak and becomes more so over time (e.g. Australia and the United States) will often see a strengthening in the central government at the expense of the authority of constituent governments, while those in which territorial pluralism remains strong, or even grows over time, may often see the reverse trend (e.g. Canada and the United Kingdom).

V. Plural constituent power and democratic pluralism

The operation of constituent power in federal constitutional systems entails a pluralization of, and unresolved tension between, multiple sites of constitutional authority. Federal constitutional settlements – insofar as they rest on representative institutions for constitution-making and processes for amendment requiring the agreement of member polities – reflect, at the level of both rules of recognition and change, an enabling tension between plurality and unity. In this final section, as a way of motivating reflection on the broader implications of our theory, we consider the potential concern that an emphasis on pluralistic ‘tensions’ complicates the connection between plural constituent power in federations and democratic pluralism. ⁷²

Unitary conceptions of constituent power, as Arato has demonstrated, are vulnerable to abuse by populist movements for illiberal or authoritarian ends. ⁷³ A pluralization of constituent power may serve to impede this sort of abuse. It is equally important to recognize, however, that a multiplication of sites of constituent authority does not, of itself, necessarily entail that a pluralization of constituent power in federations is conducive to stable democratic pluralism. One might argue, in fact, that a pluralization of constituent power also has the potential to promote political instability, and that this instability may itself provide fertile ground for authoritarianism. This concern can be framed by reference to Schmitt’s well-known analysis of constituent plurality and democratic pluralism in the federation (Der Bund).

Schmitt identifies three closely related antinomies of the federal political form. The first antinomy is that the federation aims at the independent self-preservation of each member, but also involves the renunciation of the jus belli, and hence is the principal means to secure this political aim. ⁷⁴ The second antinomy is that each federation member seeks to guarantee its self-determination by entering the compact, yet the compact entails that the federation has a right to intervene in the affairs of its members. ⁷⁵ The third antinomy is that the independent political existence of the federation rests on the independent’, with the constant potential to lead to ‘many conflicts’. ⁷⁶ According to Schmitt, these antinomies can only be resolved through the homogenous character of the member polities. Every federation, Schmitt argues, ‘rests on an essential presupposition, specifically of the homogeneity of all federation members, in particular on a substantial similarity that justifies a concrete, existential agreement of member states and ensures that the extreme case of conflict does not emerge with the federation’. ⁷⁷ Schmitt’s analysis is not totally unsubstantiated by the historical record of federal founding compacts. ⁷⁸ If accepted, however, it seems to entail that the pluralization of the constituent polities of a federal political compact actually works against ‘pluralism’ in a normative sense. In an age of multinational, multiethnic and multicultural federations, this is a real challenge. ⁷⁹

Schmitt’s characterization of the tension between the independent political existence of the constituent polities of a federation and the federal polity as an ‘antinomy’ is, however, misleading, or at least one-sided. As Schmitt recognizes, it is distinctive of the federation that both the ‘collective existence of the federation and the individual existence of the members’ continue to ‘co-exist as long as a federation is to remain in place’. ⁸⁰ As a result, the ‘collective existence of the federation must not subsume the individual existence of the member states, nor can the existence of the member states subsume that of the federation’. ⁸¹ The federation hence exists in a balance and tension, resisting the impulse to subordination of the plurality to unity or vice versa. There is little reason to regard this as an antinomy in a pejorative sense. The tension between plurality and unity is in fact constitutive for federalism as a political form, as is the refusal to accept an either/or which would reduce plurality to unity or vice versa. It is revealing in this context that stronger theories of (popular) sovereignty, which privilege homogeneity and a unitary state, are particularly conducive to this type of reduction. ⁸² Schmitt posits two oppositional poles: radical autonomy within ‘federations’ against very strong central authority within a unitary ‘state’. The source of this dichotomy is Schmitt’s characterization of a ‘federation’ as a distinctive model of the state, framed as an either/or between a ‘confederation’ where the constituents are ‘sovereign’, or a federal constitution like that of the United States, wherein the constituent territories effectively surrender sovereignty and with it any sense of their distinct constituent selves. Our analysis above, by contrast, suggests it is more instructive initially to characterize federalism not in terms of a typology of states, but of constitutions. What is distinctive of federal constituent power is the multiplication of the sites of legitimation that found and sustain a coherent constitution for the state. This involves a complex conception of constituent power, which is capable of embracing a plurality of territorially located polities that remain alive in the constitution, but also combine to form a unified or a ‘national’ sense of constituent identity.

Schmitt’s account, along with many others, insufficiently appreciates that a federal constitution is not an unhappy outlier from the somehow ‘natural’ unitary order of modern constitutional rule, which is essentially monist in disposition and unitary in constitutional design. ⁸³ Reflection on the multiple sources of constituent authority within federal constitutions helps to dispel this prejudice. Federalism is a discrete and self-conscious idea of constitutional rule for the modern state, and the relationship between autonomy and associational government, which ultimately rests on a pluralization of constituent polities, is inherent in its very purpose. Autonomy and associational rule are not antonyms, and they are also not necessarily antagonistic. While federal constitutions may include a ‘competitive’ dimension, they must also encourage sufficient cooperation and harmony for the overall federal project to prosper.

This critique of Schmitt’s views on unified state sovereignty and federalism motivates the following conclusions regarding the relationship between pluralism as a constitutional principle and normative claims about democracy and pluralism. A central purpose of federal constitutionalism, our analysis suggests, is to establish a constitutional union that gives foundational recognition and accommodation of constituent pluralism. ⁸⁴ A democratic order should reflect the composition of the society it serves, and the multiplication of sites of authority in a federal constitution responds to this requirement. It is important not to overstate this point. There is no strictly necessary connection between the federal constitutional form and democratic pluralism. A federalized conception of constituent and constituted power is certainly, moreover, no panacea for the deep disagreements that beset many contemporary polities. There are nonetheless structural features of the federal constitutional form that are uniquely conducive to the fulfilment of the ‘democratic promise’ of modern constitutions. Federalism can potentially promote or revitalize democracy not only by sensitivity to territorial difference, but by fostering political processes that are reflective of, or conducive to, societal pluralism. ⁸⁵ In a federally constructed bicameral legislature, most obviously, both ‘the people’ and ‘the peoples’ can be represented. ⁸⁶ This representative structure is an institutional acknowledgement of the subsistence of plural sites of constituent authority in a federation and the need to reconcile principles of autonomy and harmonious association. For polities with a complex history and territorially pluralized societies, federal constitutions, which reflect multiple sites of constituent authority, respond to concerns of democratic legitimacy better than monist conceptions that assume an undifferentiated popular source of authority.

These points regarding the multiplication of sites of constituent authority and democracy are only strengthened by the fact that many federal constitutions do not simply operate on a dual conception. Federalism can go ‘all the way down’ and ‘all the way up’, as evidenced by federal constitutions that reflect principles of local autonomy or subsidiarity. ⁸⁷ Such constitutional orders can be interpreted as grounded in an account of human sociality and political organization that begins conceptually with the most local and builds progressively outward towards the most universal in a nested association of communities within communities. ⁸⁸ Although it is true that most federal systems focus on only two orders of government, a third tier of local government is also almost always an intrinsic part of the political system as a whole, and many nation-states, including those that are federal, are parts of larger regional systems of states formed to secure, for example, freedom of trade or maintain human rights standards.

The plural model of constituent power we have developed here is consistent with this critique. In a federal compact, the representatives of the member polities within a federation, and by extension the peoples of the polities, do not contract together only in the name of the people as one demotic entity, but as representatives of the member polities. ⁸⁹ This impacts the ‘recognition’ question with respect to constitutional legitimacy and representation. It is possible that the creation of a federal compact might ‘subsume’ the member peoples within a unified people. This is not necessary, however, and as a matter of normative theory it is cogent to assume that the federation rests on the ongoing or ‘retained’ salience of the constitutional authority of the subjects (peoples/citizens) of the constituent polities. This is once again a way in which the tension between plurality and unity in a federation can be seen as ‘constructive’.

A plural mode of constituent power, as found in federal constitutions, presents a conceptual and empirical challenge to some common patterns of constitutional thinking. On a plural model, there are multiple sites of constituent authority, in contrast with a unified national citizenry, understood as undifferentiated by territorial boundaries. At the same time, federal constitutions with plural constituent polities can, like unitary constitutions, formalize processes of state-building, establishing a matrix of polities that constitutes and supports a national state. The pluralization of the sources of constituent power, and thereafter of ongoing constitutional authority, is not a pathology as presented by Schmitt and others, but a reflection of the empirical diversity of federal constitutions and their complex sources of popular authorship and ongoing legitimacy. Whereas the ongoing constitutional authority of a unitary constitution, moreover, rests upon a representation of ‘the people’ as an inter-temporal abstraction, in a federal constitution the constituent peoples who legitimized the foundational moment remain active constitutional actors within the constitution, strengthening the reflexive relationship between foundational constituent power and constitutional authority.

VI. Conclusion

Federalism is, and always has been, a constitutional form that challenges the homogenizing assumptions of state centralization. Our analysis suggests that the potential democratic nexus between popular legitimacy and constitutional authority that attends federal polities warrants far greater attention than it has received. It is at least plausible, on the account of plural constituent power developed here, that federal constituent power, with its mediation between constituent polities and citizens in multiple roles, could motivate a more dynamic and flexible relationship between authority and democratic legitimacy in complex societies. In any case, and allowing for the fact that, at the level of specific institutional forms, the operation of pluralized constituent power does vary between federal constitutions, reflection on the multiple sites of constituent authority within constitutions of federations speaks to the inter-woven and over-lapping patterns of identity and loyalty found in the modern federal polity.