Introduction

Contemporary debates on sovereignty mostly focus on external sovereignty and the relationship between various levels of government. The growing interest of constitutional lawyers in international and transnational relations has moved the question as to who should hold the ultimate authority within the state, somewhat to the background. Yet this question remains highly relevant, not least because issues of internal and external sovereignty are often intertwined. Any position in the debate over the competence to devolve or transfer (sovereign) powers to sub-state or supranational entities requires a clear conception of domestic sovereignty. This article seeks to refocus the attention on internal sovereignty.

The starting point of our essay is Article 33 of the Belgian Constitution and more specifically this article’s first phrase: ‘Tous les pouvoirs émanent de la Nation’ (‘All powers emanate from the Nation’). This provision is often considered the keystone of the Belgian Constitution, purportedly expressing the drafters’ view on sovereignty. Yet the phrase has generated a multitude of interpretations. Today, most public law textbooks claim that this phrase gives sovereignty to the ‘nation’ (a fictitious, transgenerational entity encompassing not only Belgium’s current citizens but also all past and future generations) as opposed to the ‘people’ (the currently existing generation of citizens). In other words, the Constitution would sanction the principle of ‘national sovereignty’, rather than ‘popular sovereignty’. The aim of this article is to put into question this simplistic – but very widely held – view and to argue that we are in need of an altered interpretation of the conception of sovereignty that underpins the Belgian Constitution. This requires re-embedding the Constitution in the political-philosophical debates of its day and investigating the intellectual sources of the Belgian ‘founding fathers’. These sources are almost invariably French liberal philosophers, whose ideas were eagerly absorbed in Belgium during the years of opposition against King William I.Footnote ¹ Developing an improved interpretation of the Belgian constitutional project is not of mere historical interest. Belgium’s political system, like that of many other countries, faces increasing calls for greater citizen participation.Footnote ² Yet many of these calls – in particular, calls for referendums – are resisted by constitutional scholars and by Belgium’s Council of State as going against the spirit of the Belgian Constitution and its ingrained conception of sovereignty. But can the Constitution really be invoked to this effect? We argue that it cannot.

Although our emphasis is squarely on the Belgian Constitution and its history, we believe the relevance of this study surpasses the Belgian context for at least two reasons. Firstly, it can be mentioned that the meaning of the term sovereignty remains highly contested, especially in circles of legal and political philosophers. Martin Loughlin and Andreas Kalyvas, among others, have recently reminded us that the term, and its multiple meanings, are still poorly understood.Footnote ³ We believe that the historical meaning of sovereignty in the Belgian Constitution can help to refine and contextualise our contemporary interpretation of the term. The distinction between ‘national sovereignty’ and ‘popular sovereignty’, for instance, is familiar to French and Belgian public law scholars but is largely unknown in other constitutional traditions. Although we will reject the ‘national sovereignty’ interpretation of the Belgian Constitution, we do think that this pair of notions, taken together, points to a tension that sits at the very heart of the idea of sovereignty, namely the temporal tension between an actual population making decisions in the present and the necessity of seeing the subject of sovereignty as transcending the present and extending both into the past and into the future.Footnote ⁴ While we will not be able to delve further into this topic, it should be noted that this temporal dimension of sovereignty is very much debated in contemporary political and legal theory.Footnote ⁵

Secondly, our study might help to shed light on Belgium’s historical place in the broader trend of ‘constitutional monarchism’ in post-Napoleonic Europe.Footnote ⁶ Horst Dippel, in his oft-cited article ‘Modern Constitutionalism. An Introduction to a History in Need of Writing’, notes that the Belgian 1831 Constitution was in its time seen as modern constitutionalism’s ‘greatest triumph’.Footnote ⁷ However, while other constitutions of the era are increasingly being investigated and re-interpreted, the original Belgian Constitution remains underresearched (and most existing research is only published in Dutch and in French, making it less accessible to international audiences). We cannot offer a complete exegesis of the 1831 Constitution, but we do hope to clarify its conception of sovereignty. This is not a bad place to start, as the drafters’ understanding of sovereignty might well hold the key to a full understanding of many other pivotal elements of Belgium’s constitutional regime. More generally, it is clear that the meaning of sovereignty was shifting in the decades after the French Revolution, and continued to evolve later in the nineteenth century. The Belgian Constitution was an important stepping stone in this evolution. Thus, this article might be valuable for international researchers who are working in the field of modern constitutionalism and who have, at present, hardly any serious analyses of the Belgian Constitution at their disposal.

It should be added that this article is not intended as an exhaustive treatment of sovereignty in the Belgian Constitution, but rather as an exploratory endeavour. Our primary purpose is to show that it might be wise to revise the received interpretation of sovereignty in the Belgian Constitution, and to show a possible direction for another interpretation, one that is more informed by the views of the Constitution’s original authors. After putting forward the dominant ‘national sovereignty’ interpretation of the Belgian Constitution, we argue that this interpretation is highly implausible. We then give a brief introduction to the understanding of sovereignty developed by French liberal philosopher Benjamin Constant, who was a particularly important influence on the Belgian drafters. Basing ourselves on Constant, we then give a rough sketch of what could be an historically more sound interpretation of the conception of sovereignty that underlies the Belgian Constitution. Finally, we look into some of the potential practical consequences of this revised interpretation.

The ‘national sovereignty’ interpretation

Belgium’s Constitution, enacted in 1831, established the newly-independent state as a constitutional monarchy with bicameral representation. Like other liberal constitutions of this period, the Belgian Constitution sought to situate the source of sovereignty not in the theological or historical realm, but in society itself. The drafters’ view on sovereignty is purportedly expressed in Article 33 of the Constitution (originally Article 25),Footnote ⁸ an article that is to be read in conjunction with several other key provisions of the Constitution, such as those concerning the powers of the king and the role of the two chambers.

Since 1831, scholars of public law have interpreted the notion of sovereignty in the Belgian Constitution in a variety of ways.Footnote ⁹ From the 1950s on, however, a consensus seems to have grown as to what the drafters of the Belgian Constitution genuinely intended when writing that all powers emanate from the nation. This consensus can be summarised as follows. The drafters wanted to break with theological or absolutist justifications of sovereignty, which they saw as potential threats to liberty. Yet they did not grant sovereignty to the ‘people’, understood as the set (or a subset) of all currently living citizens. Instead, they granted sovereignty to the ‘nation’, a fictitious, transhistorical entity that encompasses not only currently living citizens, but also earlier and future generations of citizens. The former option would be ‘popular sovereignty’ (typically associated with Rousseau), whereas the latter option is ‘national sovereignty’ (purportedly originating with Sieyès). Both types of sovereignty were supposedly propagated by the French Revolution, but the Belgian drafters – just like the French drafters of 1791 – consciously picked national sovereignty.

This choice for the nation as the subject of sovereignty has a number of ‘elitist’ consequences. i) Given that the nation does not materially exist, it can only express itself by means of representative institutions. Thus, direct democratic procedures are ruled out, as only careful deliberation by representatives can reveal the will of the nation. ii) National sovereignty provides a strong argument against universal suffrage, since voting is not a fundamental ‘right’ of the citizens but only a means (or a ‘function’) to appoint the most competent representatives of the nation. iii) In the exercise of their mandate, representatives are not bound to the opinions or interests of their voters, since they are representing the nation rather than the actual electorate.

Today, most public law textbooks insist that the Belgian drafters made a conscious choice for national sovereignty, possibly with these three ‘elitist’ implications clearly in mind. Starting with André Mast’s influential 1950 textbook,Footnote ¹⁰ this interpretation can now be found in virtually all public law handbooks.Footnote ¹¹ In contrast to this general consensus, a small number of contemporary scholars maintain that the Constitution actually sanctions the principle of popular sovereignty, in line with the ‘Rousseauist’ traditionFootnote ¹² or that the Belgian drafters simply had no deeply theoretically embedded view of sovereignty.Footnote ¹³ Yet the overwhelming consensus reads that the drafters opted for ‘national sovereignty’.

This consensual emphasis on ‘national sovereignty’ has considerable practical consequences. Most significantly, it is used as a key argument in debates over the constitutionality of forms of direct citizen participation, in particular referendums. Since 1985, the Belgian Council of State has consistently declared legislative proposals to introduce binding and non-binding referendums at the federal, regional and community levels unconstitutional, as a violation of the principle of national sovereignty enshrined in Article 33.Footnote ¹⁴ According to the Council of State, a distinction should be made between ‘souveraineté nationale’ (national sovereignty) and ‘souveraineté populaire’ (popular sovereignty), and the drafters of the Belgian Constitution deliberately opted for the former.Footnote ¹⁵ ‘The Nation’, the Council of State writes, ‘has a continuous existence; it encompasses past and future generations of citizens’.Footnote ¹⁶ With respect to direct citizen participation, the Council of State decides: ‘National sovereignty, as opposed to popular sovereignty, justifies the institutionalisation of a representative system and generally precludes direct government by the people’.Footnote ¹⁷ ‘It is impossible’, the Council concludes, ‘to give a certain number of the citizens of today the power to replace the national representation, which is constitutionally established to express the will of that abstract being which our foundational act calls “the nation”’.Footnote ¹⁸

This reading of Article 33 has often been confirmed, both with respect to ordinary legislative referendums and ‘constitutional referendums’ (e.g. referendums concerning the transfer of sovereign powers to supranational bodies). Thus, in its opinion on a proposed referendum on the adoption of a Constitution for Europe, the Council recalled that ‘the Constitution is grounded not in a system of popular sovereignty but in a system of national sovereignty, in which the Nation is represented by the established powers’, and that as a consequence ‘methods of direct democracy on the federal, the regional and the community levels’ are not allowed.Footnote ¹⁹ In a similar vein, most constitutional law scholars opposing the use of referendums refer to the concept of sovereignty in Article 33 of the Constitution as an important argument against the introduction of referendums.Footnote ²⁰

To be sure, the theory of national sovereignty is not the only legal line of reasoning against the introduction of methods of direct democracy under the current constitutional framework. The second major argument introduced by the Council of State, and supported by legal doctrine, is based on the second sentence of Article 33 of the Constitution, which states that the powers emanating from the nation shall be exercised in accordance with the procedures laid down in the Constitution.Footnote ²¹ Since the Constitution explicitly states that the legislative power is exercised by the King, the Chamber of Representatives and the Senate (Article 36), and since the Constitution does not provide for direct citizen participation, there is no room for referendums.

Today, there is a broad consensus amongst constitutional scholars that, for these two reasons, binding referendums are unconstitutional. With respect to non-binding referendums, some dissenting voices argue that these would not violate the Constitution.Footnote ²² Yet given the clear stance of the Council of State, the federal, regional and community parliaments generally refrained from organising referendums. The fact that the current constitutional framework would not allow for the use of referendums of course does not preclude a Constitutional amendment. In an attempt to overcome the constitutional hurdles, the Constitution was amended twice. Today, it provides for a non-binding referendum at the local level (Article 41, since 1999) and the regional level (Article 39bis, since 2013). Hence, the possibilities for direct democracy are still very limited.

Note that the assumption of ‘national sovereignty’ would, mutatis mutandis, also exclude other forms of non-representative democracy, such as participation by citizens in deliberative councils that would take part in the legislative process. From the way certain constitutional scholars defend the ‘consensual’ interpretation, it can be gathered that the exact same conclusions would apply. André Alen, for instance, wrote that, given the conception of national sovereignty, the Belgian representative system operates on a very specific presumption: ‘From a constitutional perspective, there is an irrefutable presumption that the will of the representatives is the will of the people’.Footnote ²³ Given this view, it would be only logical not to discriminate between different forms of non-representative decision-making and thus to also exclude as unconstitutional the transfer of legislative tasks to participatory bodies.Footnote ²⁴

In addition to the legal considerations, there is of course a host of philosophical, political and empirical arguments for and against the use of referendums and/or forms of participatory democracy.Footnote ²⁵ It is not the purpose of this article to take a conclusive position in these debates. Let us just briefly mention one non-legal argument that is of special relevance in the Belgian context. As the Council of State observed in an obiter dictum in its 1985 Opinion, the use of referendums at the federal level might jeopardise the ‘harmonious coexistence’ of the French and the Flemish Communities, the two largest communities in Belgium, as it would bring the differing political preferences of these communities to the fore.Footnote ²⁶ It is probable that this consideration – even when not always explicitly mentioned – has played an important role in keeping the door towards more direct forms of democracy closed in Belgium, as it is usually assumed that elite compromise formation offers the best prospects for stability in this divided country.Footnote ²⁷ Thus, the context in which the debates are being conducted is a highly delicate one.

With so much at stake, one would expect the Belgian drafters’ view on sovereignty to have been the object of serious scrutiny and extensive research. Yet, surprisingly, no systematic studies of the topic are available. The ‘consensual’ interpretation seems to have been constructed, around 1950, on poorly-justified assumptions about the meaning of sovereignty in 1831 and has thenceforth been transmitted from textbook to textbook, and from textbook to courtroom.Footnote ²⁸

Dismantling the myth

The consensual understanding of sovereignty in the Belgian Constitution has many advocates, yet we believe that it amounts to little more than a myth that needs to dismantled.Footnote ²⁹ What are the arguments that are typically put forward in favour of the ‘national sovereignty’ interpretation? An important, if not the principal, argument is the supposed parallel between the 1831 ‘Belgian’ understanding of sovereignty on the one hand, and the notion of sovereignty that is enshrined in the French Revolutionary Constitution of 1791 on the other. Defenders of the consensual understanding point out that the phrasing of Article 33 (‘Tous les pouvoirs émanent de la Nation’) was inspired by Article 2 of the French Revolutionary Constitution of 1791, which contains similar wording (‘La Nation de qui seule émanent tous les pouvoirs, ...’). This is indeed rather plausible.Footnote ³⁰ Unfortunately, many authors defending the consensual interpretation conclude from there that both constitutions must share the same conception of sovereignty, namely ‘national sovereignty’ as described above. Mast, for instance, explicitly draws this conclusion. His opinion is worded rather cautiously in the first, 1950 edition of his textbook, Overzicht van het grondwettelijk recht (Overview of constitutional law),Footnote ³¹ but in subsequent editions he defends this conclusion with increasing confidence.Footnote ³²

We see at least five reasons to reject this conclusion. i) In the 1791 Constitution, Article 2 is preceded by Article 1, in which sovereignty is extensively defined (as, inter alia, ‘indivisible’ and ‘inalienable’). The Belgian Constitution contains no trace of Article 1. If the Belgian drafters intended to follow the 1791 understanding of sovereignty, it is at least remarkable that they did not copy the article on sovereignty but only bits of the next article. ii) Recent research demonstrates that the concept of national sovereignty, as specified above, was not part of the discussion among French revolutionaries and did not genuinely exist at that time.Footnote ³³ The strong conceptual opposition between national sovereignty and popular sovereignty surfaced only after 1831 and was then, wrongly, projected back onto the French Revolution, in particular by Carré de Malberg.Footnote ³⁴ iii) More generally, it is doubtful that the Belgian drafters would have stuck to sovereignty as it was understood in the early years of the French Revolution. In France, ideas on sovereignty and constitutional government had evolved considerably in these four decades. Discussions during the French Revolution tended to involve heavy, ‘Rousseauist’ questions such as how to discover the general will and how to represent sovereignty given that it should not be divided or alienated.Footnote ³⁵ These questions dovetailed with the general concern of abolishing the ancien régime and setting up a strong central authority that would empower the people as the new subject of sovereignty. In 1831, however, the debate had decidedly moved on, in particular because of the experiences of the Revolution and the first Empire. The Belgian drafters, who were perfectly tuned into Parisian political-philosophical debates (as their theoretical publications attest to)Footnote ³⁶, were certainly aware of these developments. iv) Furthermore, when the National CongressFootnote ³⁷ discusses the issue of sovereignty (which happens only very rarely, suggesting that the sovereignty question was not particularly controversial) there is no mention of the idea of ‘national sovereignty’, nor is there a trace of an opposition between defenders of ‘national sovereignty’ and defenders of ‘popular sovereignty’.Footnote ³⁸ On the contrary, certain debaters in the National Congress state that a representative, constitutional monarchy is always based on the ‘sovereignty of the people’, or even that all modern societies are based on the ‘sovereignty of the people’,Footnote ³⁹ without meeting any resistance or correction. This strongly indicates that no one in the National Congress was aware of a possible conceptual distinction between national sovereignty and popular sovereignty. This is also confirmed by the fact that debaters in the National Congress consistently use ‘people’ and ‘nation’ interchangeably, which would be confusing if the difference between the two terms was seen as decisive with regard to the source of sovereignty.Footnote ⁴⁰ Moreover, the National Congress makes all its declarations ‘au nom du peuple belge’, that is, in name of the Belgian people rather than in name in of the Belgian nation. v) Finally, it should be noted that the formal, 1831 Dutch translation of the Constitution translated ‘nation’ by ‘people’ in this article.Footnote ⁴¹ One could ascribe this to sloppiness, but if the choice of the term ‘nation’ was so momentous as later interpreters pretend, it is doubtful that the error would have gone unnoticed at the time.

Thus, it seems safe to assume that the ‘myth’ is historically wildly inaccurate and a more sound interpretation of sovereignty in the Belgian Constitution needs to be constructed. The hypothesis we put forward holds that the conception of sovereignty that the drafters wrote into the Belgian Constitution can best be understood in light of the paradigm of French political liberalism as it developed in the aftermath of the French RevolutionFootnote ⁴² and, more specifically, in light of the work of Benjamin Constant.Footnote ⁴³ With regard to other aspects of the Belgian constitutional edifice, such as ministerial responsibility and freedom of the press, it has by now been established that they are intellectually very much indebted to Constant.Footnote ⁴⁴ We hypothesise that, with regard to the key notion of sovereignty, the Belgian Constitution is similarly – if not more strongly – rooted in French restoration liberalism, in particular in the ideas of Constant. As we will explain in this article’s final section, such a revised reading of the Constitution could have significant implications for the constitutionality of direct-democratic or participatory procedures in Belgium. Although a ‘Constantian’ view on sovereignty is probably not compatible with direct democracy across the board, we believe it can open the door to adding alternative forms of citizen participation to a representative regime. Below, we start by looking at Constant’s ideas on sovereignty.

Constant’s understanding of sovereignty

Although Benjamin Constant’s work has been the object of relatively much recent research,Footnote ⁴⁵ his view on sovereignty has never been the focus of a specific investigation.Footnote ⁴⁶ This is unfortunate, not just because his view likely holds the key to a full understanding of the ‘Belgian’ notion of sovereignty, but also because it occupies an intriguing middle ground in comparison to other positions of his time.

Looking at the different positions that were available in francophone debates during the birth years of modern constitutionalism – that is, the debates which formed the Belgian drafters’ philosophical background – one notices that they are defined in light of one core problem, namely the tension between popular sovereignty and the restraints on sovereignty.Footnote ⁴⁷ In this regard, there seems to be a continuous evolution from Rousseau to the post-revolutionary vision of the doctrinaires. Rousseau’s account of sovereignty,Footnote ⁴⁸ which rests on a ‘realist ontology’,Footnote ⁴⁹ presents sovereignty as an exercise concept that requires the constant, direct expression of the people’s monolithic will, a will that genuinely exists and is in principle unconstrained.Footnote ⁵⁰ Sieyès made this vision more practicable by distinguishing between constituent power and constituted powers.Footnote ⁵¹ Sovereignty thereby becomes the name for the source of the constitutional order. Although this introduces a strong temporal division between the foundational moment and the further life of a political regime, the constituted powers in principle remain the medium through which the current generation of active citizens expresses its united will, and new generations maintain the right to alter and improve the constitutional order. At the same time, Sieyès did seek to impose external limits on any lawmaking act, namely natural equity and the natural rights of the individual, as can for instance be gathered from his project for a constitutional jury.Footnote ⁵² All in all though, Sieyès’s main concern – just like Rousseau’s – was to ensure that the people’s will could genuinely express itself.Footnote ⁵³ In contrast, many nineteenth-century post-revolutionary theorists seem to distrust sovereignty and are primarily concerned with limiting the expression of ‘will’. This distrust is probably most lucidly formulated by Guizot, who sought to replace popular sovereignty – which he saw as an empty ‘idol’Footnote ⁵⁴ – with the sovereignty of reason and justice.Footnote ⁵⁵ Rationality thereby not only becomes the ‘transcendent’ norm for legitimate lawmaking, it also provides a justification to reserve political power – in perpetuity – for the rational few.Footnote ⁵⁶

Constant’s ideas seem to sit somewhere in between these two extremes. Uncovering Constant’s complete view on sovereignty would require a more complete investigation than is feasible in the context of this article, yet we believe that the main vectors that sustain his view are the following.

1) Popular sovereignty but no constituent power. As opposed to Guizot, Constant remained loyal to the notion of popular sovereignty and literally called the people the source from which the law ‘emanates’.Footnote ⁵⁷ At the same time, he seemed keenly aware of the theoretical puzzle that constituent power, although in principle preceding the constituted powers, only appears in constituted form.Footnote ⁵⁸ Perhaps he did not even see any use at all for the idea of a power that pre-exists the constitutional order.Footnote ⁵⁹ According to Gauchet, Constant’s search for a ‘neutral power’, the role he eventually ascribed to the monarch, reflects Constant’s awareness that society has no immediate access to itself but is only constituted through the establishment of a symbolic power above the other powers.Footnote ⁶⁰ In other words, it seems that Constant wanted to maintain – within the constitutional order – a symbolic reference to the people as the source of sovereignty, without succumbing to the illusion of a genuinely existing ‘people’ that would precede its own constitution, that is, that would precede the establishment of a constitutional order.

2) Popular sovereignty but no general will. Although Constant described the law as the expression of the people’s will, he clearly abandoned the Rousseauist illusion of a unified will, preferring to disperse the exercise of sovereignty over a plurality of opposing powers representing different social interests.Footnote ⁶¹ This Montesquieu-esque vision of sovereignty as divisible was hard to square with the Rousseauist and Jacobin tradition that dominated in France,Footnote ⁶² and some would even say that it is not compatible with the meaning of the term sovereignty at all. According to Carré de Malberg, for instance, it is not.Footnote ⁶³ Constant clearly saw no contradiction there.

3) Non-external limitations on sovereignty. Constant was emphatically concerned with the limitation of sovereignty. Most importantly, Constant believes the law can never have the authority to infringe on the rights of individuals, as the citizens from whom the law emanates do not have that authority themselves. Yet it is not obvious how Constant justifies these rights. Their justification does not stem from the fact that they have been voluntarily instituted by the citizens at a foundational moment, since these rights evidently precede the will of the citizens and limit any legitimate expression of their will. Nor are they given a natural foundation (as in the famous eighteenth-century declarations of rights). The most probable explanation seems to be that Constant, having swallowed the lessons of Montesquieu’s proto-sociological approach to law, believes that these rights are only valuable within ‘our’, modern societies (as is suggested by his Athénée-lecture).Footnote ⁶⁴ If these rights are indeed tied to the spirit of our society, than this would also imply that they might need to be altered and reformulated over time.

4) Constant as a non-constitutional constitutionalist. The main lesson that Constant took from the French Revolution was that the idea of popular sovereignty could be invoked just as well as the idea of monarchical sovereignty to justify despotism. Thus, for Constant, the principal threat was not the rule of the many, but rather the concentration of unrestrained power in the hands of the few – no matter how that power is justified. Maybe this is why Constant picked a different solution than contemporary constitutionalists to restrain the exercise of sovereignty. He put his hope not so much in a written constitution, as in the pressure of a well-informed public opinion and in the political process itself, where the existence of multiple counteracting channels of power should prevent potentially dangerous concentrations of power. Constant’s brand of liberalism is sometimes labeled ‘constitutionalist’, but maybe this label should be rejected. While he does believe that the constitutional order needs to be protected against the unrestrained expression of sovereign will, his trust in public opinion and in the political process – rather than in the constitution itself – to do so, certainly puts him apart from most other constitutionalists.

Although these points stand in need of further substantiation, they do provide us with more nuanced conceptual tools to interpret the view on sovereignty that is ingrained in the Belgian Constitution than the blunt categories of ‘popular sovereignty’ and ‘national sovereignty’.

Sovereignty in the Belgian Constitution

We strongly believe that the paradigm of French political liberalism as it developed in the aftermath of the French Revolution, and in particular the ideas of Constant, were decisive in setting the Belgian drafters’ understanding of sovereignty and thus provide the best conceptual tools to develop a coherent interpretation of the account of sovereignty that is ingrained in the 1831 Constitution. Of course, the Belgian drafters were influenced by other philosophical sources as well.Footnote ⁶⁵ And there is no reason to assume a deep, fully-fledged theoretical consensus beneath all aspects of the Belgian Constitution; like all political artefacts, it was probably as much a product of compromise as of consensus. Yet we do believe that when it comes to sovereignty, Constant’s vision comes very close to what was eventually codified into the Belgian Constitution. The fact that other pivotal elements of the new Belgian regime (e.g. freedom of the press and the idea of ministerial responsibility) were strongly inspired by Constant’s ideasFootnote ⁶⁶ and the fact that contemporary accounts describe many of the influential drafters as Constant-acolytes,Footnote ⁶⁷ provide at least circumstantial indications for this hypothesis. Moreover, the records of the extensive discussions in the 1830-1831 National Congress show that the congressmen very regularly referred to the ideas of Constant.Footnote ⁶⁸ Mention can also be made of the recent historical research by Stefaan Marteel, who has comprehensively exposed the networks and background of intellectuals and activists in the southern Netherlands in the years leading up to the Belgian revolution, proving Constant to be of particular importance.Footnote ⁶⁹ Yet rather than delving into the intellectual-historical literature, we would like to point to a number of concrete instances in the 1831 Constitution that betray a very Constant-like attitude.Footnote ⁷⁰

1) Popular sovereignty but no constituent power. The French drafters of 1791 made a logical distinction between ‘original constituent power’ (‘pouvoir constituant originaire’) and ‘instituted constituent power’ (‘pouvoir constituant institué’). The 1791 Constitution contains extensive provisions for constitutional revision by the latter, involving primarily the elected legislature. Yet this Constitution also stipulated that, despite these provisions, the nation (i.e. the citizens themselves) retained the ‘imprescriptible’ right to change the Constitution, even by non-constitutional means.Footnote ⁷¹ Thus, the appearance of the original constituent power remains a genuine and legitimate possibility. The Constitution limits itself to stating that it would be in the nation’s interest to only revise the Constitution by means of the procedures provided in the Constitution. The Belgian drafters, by contrast, wanted to have it written in the Constitution that all powers only come into being by virtue of the Constitution itself. The first formulation, proposed by Congress-member François Van Snick (‘Les pouvoirs constitutionnels n’existant que par la constitution, ils ne peuvent dans aucun cas, ni sous aucun prétexte, en suspendre l’action.’),Footnote ⁷² was literally borrowed from Benjamin ConstantFootnote ⁷³ and was precisely intended to prevent the kind of turmoil seen during the French revolutionary years.Footnote ⁷⁴ Article 195 (previously Article 131) testifies to this same intention. This article specifies the procedure for revising the Constitution, a procedure that gives a controlling role to the legislature, thereby suggesting that the ‘pouvoir constituant’ has been completely devolved to the legislature.Footnote ⁷⁵ Maybe the people had even already devolved – or simply lost – their ‘pouvoir constituant’ with the election of the National Congress in the autumn of 1830, as this Congress saw no reason to subject the newly-drafted Constitution to popular approval.Footnote ⁷⁶ Clearly, all of this increases the distance from the 1791 understanding of sovereignty, which combined a strong belief in popular sovereignty with a vision of the people as the really existing ‘pouvoir constituant originaire’ that could at any moment make a genuine appearance on the political stage – even in a non-constitutional manner – just as it had at the key moments of the Revolution. The Belgian drafters, by contrast, retained a reference to the people as the titular source of all sovereign powers,Footnote ⁷⁷ yet without having a genuine role for it in the constitutional order.

2) Popular sovereignty but no general will. Although we think it is correct to state that the ‘people’ or the ‘nation’ – with the two terms virtually equivalent – are introduced in the Constitution as the (at least nominal) source of sovereignty, it should be noted that the Constitution nowhere uses the term ‘sovereignty’. While not completely exceptional, it is a striking difference with the constitution from which its understanding of sovereignty is derived according to so many observers, namely the French 1791 Constitution. Does this omission imply that there is actually no sovereign in the Belgian Constitution? This conclusion seems overblown. The Constitution does take care to point out the source of all powers; it is difficult not to understand this as an indication of who should be considered ‘sovereign’ (or the ‘source’ of all sovereign powers).Footnote ⁷⁸ What this omission might imply, though, is a strong suspicion towards everything the term ‘sovereignty’ connoted at the time (and according to many authors, still connotes today)Footnote ⁷⁹: a supreme, monolithic power that is by its very nature indivisible, that serves to express a unified will (à la Rousseau), and that stands by definition above any possible constraint (as exemplified in Austin’s absolute understanding of sovereignty). Such a suspicion would be perfectly in line with post-revolutionary sensitivities.Footnote ⁸⁰ The French experiences of the past decades (for which the populist moments in the Belgian revolution functioned as a vivid reminder) had raised a new fear: sovereignty of the people could be just as threatening to bourgeois liberties as ancient absolutism had been. At the same time, the Belgian drafters had every reason to be wary of excessive state power in the hands of a monarch, which they saw so recently exemplified by the rule of William I and his ministers. Thus, the Belgian drafters feared as much a despotism justified by reference to the people’s general will, as the despotism of a self-righteous autocrat. In the National Congress’s lengthy discussions about Belgium’s future regime (the existential question was: should it be a monarchy or a republic?), this double risk is constantly invoked. The drafters’ main concern was to prevent instability, and they clearly saw two potential sources of this evil: the passions of the masses and the concentration of arbitrary power in the hands of a small elite.Footnote ⁸¹ With this double fear in mind, it would not be surprising if the Belgian drafters’ interest in sovereignty was primarily negative: they wanted to restrain power – be it the power of the government or the power of the people – and not construct a strong state or empower the sovereign people. Thus, rather than be about moulding sovereignty, the Belgian Constitution would be about blocking the exercise of sovereignty and about drawing the limits which every legitimate exercise of sovereignty has to respect. This brings the Constitution close to Constant’s approach to sovereignty.Footnote ⁸² The omission of the word ‘sovereignty’ might well be a symptom of this. By talking about ‘powers’ (in the plural) instead of ‘sovereignty’, the Belgian drafters created a way of mentally picturing power that is rather welcoming to the idea of division, balance and mutual restraint (à la Montesquieu), whereas talk of ‘sovereignty’ seems to lead almost naturally to the picture of a monolithic, unrestrained will (à la Rousseau).

3) Non-external limitations on sovereignty. Another, rather straightforward symptom of this concern for restraining the exercise of sovereign will is the importance which the Belgian Constitution gives to citizens’ rights. The first section of the Constitution extensively lists these rights. Yet, significantly, the Constitution does not anchor these rights in nature, reason or any other universal source. Instead, these rights are introduced under the heading ‘Des Belges et de leurs droits’ (‘On the Belgians and their rights’), and the Constitution stipulates that the concrete modalities of these rights are to be determined ‘by law’, that is, by the legislature. Thus, there seems to be no pretence to have found supra-temporal or supra-spatial norms to which sovereignty should always and everywhere adhere. Instead, the Belgian drafters here – as elsewhere – gave proof of what has been dubbed a ‘pragmatic conservatism’.Footnote ⁸³ They rejected ‘abstract rationalism’ and instead tried to determine which norms were best adapted to ‘the character, customs and social habits of the people for whom those laws are meant’ (drafter Félix de Mérode quoted by De Dijn).Footnote ⁸⁴ Already the title of this section on rights (‘Des Belges et de leurs droits’) suggests that what follows is a local story, meant to bring out – in a Montesquieu-esque fashion – those norms that best befit the specific spirit of Belgian society.

4) A non-constitutional constitutionalism. Many other elements in the Constitution are similarly inspired by a concern to limit the expression of sovereign will. Most significant, of course, is the actual creation of a number of powers that are meant to counterbalance each other. Although nominally identifying three different powers (as in Montesquieu’s famous description of the English constitution), the Belgian founding fathers in reality established five powers. Next to the judicial power, there is the legislative power, which is distributed over the King, a chamber representing the variability of public opinion (the Chamber of Representatives) and a second chamber that is meant – as a counterweight – to embody the stability of long-term reflection (the Senate). The executive power, although in name belonging to the King, de facto falls to the King as well as to his government, the ministers of which are responsible to the legislature whereas the King himself remains inviolable. In this regard, the Belgian founding fathers evidently followed Constant. As Jeremy Jennings explains, Constant amended Montesquieu’s reading of the English constitution, pointing out that English liberties were protected, not by a triad of powers, but through a more complex set-up involving five different powers: ‘royal power, executive power, the power that represents permanence (i.e. the hereditary assembly), the power that represents opinion (i.e. the elected assembly), and judicial power’.Footnote ⁸⁵ The Belgian drafters deviated from Constant’s ideal in that the Senate was not hereditary, but this was the outcome of a long and heated debate – a debate that could easily have tipped over in the other direction and that in fact only confirmed how many participants shared Constant’s intention of counterbalancing the first chamber, feared to be too strongly under the sway of popular passion, with a second chamber.Footnote ⁸⁶ Moreover, even if the resulting decision was in favour of a non-hereditary chamber, the high censitary requirements for eligibility made this chamber all but hereditary. The drafters likely followed Constant most consciously when discussing the role of the King. The King, who stands inviolably above the political mêlée and in whose name justice is to be served, was explicitly intended to be a ‘neutral power’.Footnote ⁸⁷ It is this neutrality that, for Constant, provided the King not only ‘with the function of preserving the constitution itself but also of acting as the guarantor of all political liberties’.Footnote ⁸⁸ Thus, the King, whose position is created through the ‘fictions’ of inheritability and inviolability (as the influential drafter Jean-Baptiste Nothomb put it) was meant to be an ‘immovable center’ (as another influential drafter, Joseph Lebeau, put it) in the midst of the other ever moving powers and the fluctuations of the ‘sovereignty of the people’.Footnote ⁸⁹ Ultimately, it is the combination and the interplay of these five different powers that guarantees freedom and protects justice and reason against unrestrained eruptions of the will, be it the will of the many or the will of the few. There are further elements in the Constitution that could be analysed along similar lines. Freedom of the press and trial by jury (in particular for political crimes, i.e. crimes where the judiciary might come under the sway of the other powers) similarly served the purpose of preventing the abusive effects of possible power concentrations.Footnote ⁹⁰ These different elements not only share the same purpose, namely that of restraining the exercise of power, they also show to what extent the authors of the Belgian Constitution believed that this purpose could most reliably be reached, not just by counting on the Constitution as a safeguard in itself, but rather through a political process in which different powers, interests and actors would counteract each other.

These cursory remarks certainly stand in need of further elaboration, but we believe they do confirm that a ‘Constantian’ reading promises to be more fruitful than the received ‘national sovereignty’ interpretation in bringing out the specificity of the conception of sovereignty that underlies the Belgian Constitution. It should be noted that such a ‘Constantian’ interpretation is not completely novel; certain older studies of the Belgian Constitution come rather close. The interpretations by Thonissen (1844) and Orban (1906), for instance, similarly give considerable weight to Constant’s influence, not just when it comes to sovereignty but with regard to many other aspects of the Constitution as well.Footnote ⁹¹

Implications of this altered interpretation with regard to more direct modes of citizen participation

In the closing section of this article, we intend to point out some of the potential implications for the constitutionality of referendums and of deliberative participation. Does our revised interpretation of sovereignty in the Belgian Constitution open the door to more direct participation by citizens in political decision-making? We will argue that it does, to a certain degree. Of course, the debate on the introduction of non-representative means of decision-making is much broader than that on the nature of sovereignty in the 1831 Constitution. One should also investigate how the sovereignty issue weighs up in relation to other constitutional arguments that can be invoked, and how Belgium’s sovereignty regime has changed through the federalisation of the country. One could even claim that the introduction of linguistic groups and territories in the Constitution has made the very notion of ‘nation’ – and a fortiori the notion of ‘national sovereignty’ – obsolete in the Belgian context. Moreover, insights from political science (about the effectiveness and the side-effects of novel types of decision-making) and more prudential considerations (for instance regarding the consequences for Belgium’s already fragile political structure) should play their proper role in this debate as well.Footnote ⁹² However, our sole focus here is on one specific constitutional argument, namely the reference to the nature of sovereignty in the original Belgian Constitution. Focusing on this specific argument, our claim is that the often-used reference to the founding fathers’ purported choice for ‘national sovereignty’ is historically highly questionable and hence cannot be invoked as an argument against direct forms of democracy. On the contrary: a proper appreciation of the 1831 account of sovereignty can actually provide reasons in favour of direct or participatory modes of decision-making. This is what we will try to show in this final section.

As a preliminary, it should be mentioned that it is not obvious that there is a direct connection between conceptions of sovereignty and modes of democratic decision making. The Council of State, as well as most public law textbooks, assumes that the Belgian Constitution rules out, not only in letter but also in spirit, direct democratic procedures. As mentioned above, there are two distinct legal arguments at play. To begin with, the text of the 1831 Constitution did not explicitly provide for direct citizen participation, and since the second sentence of Article 33 holds that the powers emanating from the nation shall be exercised in accordance with the procedures laid down in the Constitution, referendums are not to be allowed. The second argument is based on the theory of national sovereignty read into the first sentence of Article 33. National sovereignty, the Council of State has repeatedly stated, precludes direct government by the people. If the Constitution were based on a Rousseauist account of ‘popular sovereignty’, things would be different. Thus, the consensual interpretation of the Belgian Constitution clearly assumes an unambiguous, ‘one-to-one’ correspondence between conceptions of sovereignty and modes of democratic decision-making. It is doubtful, however, whether such a ‘one-to-one’ connection is really defensible. Does Rousseau’s understanding of sovereignty necessarily require direct democracy? Given that the general will only emerges when voters are physically united, could large countries not better be served with a system of representatives who – as opposed to the actual citizens – can physically meet in order to discover the general will? This would be in line with what Rousseau himself proposed for Poland.Footnote ⁹³ Inversely, one can question whether ‘national sovereignty’, as understood in the myth, excludes all forms of direct democracy. The generalised use of referendums can perhaps be construed as a usurpation by the current citizenry upon the sovereignty of the nation, defined as the set of all generations. But could one not just as well claim that the disinterested, long-term reflection required to mentally anticipate the interests of future generations, is difficult to achieve for those with a career at stake in the political game? This is exactly the accusation made by certain contemporary theorists: they claim that representative democracy tends to be biased in favour of the current generation, a tendency they call ‘presentism’Footnote ⁹⁴ or simply ‘myopia’.Footnote ⁹⁵ As a possible remedy, some authors propose to appoint citizen panels that would represent the interests of future generations.Footnote ⁹⁶ Following this train of thought, national sovereignty could well provide an argument against the exclusive reliance on elected representatives. In other words, there seems to be some argumentative connection between conceptions of sovereignty and modes of decision making. But a straightforward correspondence, as is sometimes assumed by the myth (national sovereignty equals representative democracy, and vice versa), is improbable. Below, we will conclude that it is possible to derive arguments in favour of more direct procedures (such as referendums or participatory democracy) from the drafters’ understanding of sovereignty, but we do not claim that this is an all-or-nothing issue.

There is one further disclaimer. For our conclusion to make sense, we need to assume that the original, 1831 understanding of Article 33 (and its ingrained conception of sovereignty) remains relevant to its contemporary application. The Council of State, as well as most Belgian constitutional scholars, seems to take this for granted. Their rejection of referendums is at least partly based on their belief that Belgium’s founding fathers deliberately opted for national sovereignty as opposed to popular sovereignty, and that this determines how we should continue to apply Article 33 today. A minority of constitutional law scholars, by contrast, argues that the historical meaning of the text should not necessarily be controlling today.Footnote ⁹⁷ This disagreement can be explained by a difference in opinion as to what method of constitutional interpretation is most suitable. The use of the historical method (‘originalism’ in Anglo-American parlance) would arguably lead to another result than the use of a more dynamic method. While there is an extensive body of international literature in this domain,Footnote ⁹⁸ this issue has largely been neglected in Belgian constitutional law. We do not have room here for an exhaustive discussion concerning what method of interpretation is most appropriate in the Belgian context. But we do want to confess our partiality for a certain ‘middle road’ in constitutional interpretation. Clearly, on a literal reading the Constitution leaves no room for direct democracy. And it is equally clear that the Belgian drafters did not expect nor hope that ‘their’ constitution would be interpreted – or actively altered – to that effect.Footnote ⁹⁹ But if we are willing to read the Constitution in a more principled or ‘moral’ manner,Footnote ¹⁰⁰ not just looking at its literal prescriptions (which can and often have been altered by new generations), but by actively searching for these commanding principles that the Belgian drafters saw as essential for the continuation of the Belgian constitutional project, the results might be surprising. Such a reading of the Constitution would steer clear of the Scylla of originalism and the Charybdis of ‘living constitutionalism’. It would commit us to a certain fidelity to the normative commitments beneath the original Constitution, while allowing us to take liberties with the explicit prescriptions of that constitution.

Given this reading, the original meaning of Article 33 is at least to some extent relevant to its application today. And given this reading, we believe the Belgian Constitution need not be as inimical to forms of direct or participatory democracy, as is usually assumed. For instance, the drafters were strongly concerned with setting up separate channels of representation for current public opinion on the one hand and for ‘longer term’ reflection on the other hand, thus creating a productive opposition between variability and stability. This was the basis of their argument for establishing not one, but two chambers, whereby the senate (with its high censitary requirements) would function as a counterweight for the fluctuations and passions of current public opinion. This concern remains valuable and inspiring, even (or especially) today. But does it require maintaining a Senate, especially as the censitary requirement – which, according to the drafters, was crucial to the senate’s mission – has been dropped a long time ago and cannot be reinstated as it conflicts with many of our more basic normative commitments?Footnote ¹⁰¹ Maybe this concern could today better be recast as an argument for citizen panels that represent the interests of future generations. Such proposals are sometimes made by those who believe that the interests of future generations cannot be mentally internalised by regular political representatives, but instead require separate institutional representation.Footnote ¹⁰² These citizen panels are often proposed with environmental issues in mind, but we believe they would fit very well in the Belgian constitutional framework, where they could find broader usage and cover many different policy areas. As such, they would be a novel embodiment of a very old commitment to a type of long-term reflection that needs to be shielded from short-term interests and from electoral pressure.

Similarly, the drafters’ ‘Constantian’ concern with restraining power through a deeply entrenched pluralism of powers, could well lead to an argument for referendums and/or other forms of direct participation. The drafters (just like Constant) did not favour direct-democratic mechanisms. Yet if it can be shown that such mechanisms form an appropriate counterweight against power concentrations which could not have been foreseen at the beginning of the nineteenth century (we are thinking in particular of Belgium’s so-called particracy),Footnote ¹⁰³ can they not be depicted as adequate expressions of this concern in contemporary circumstances? Constant himself, for instance, was well aware of the fact that representatives can become a class apart and can therefore be a greater danger to liberty than the people themselves.Footnote ¹⁰⁴ The members of Belgium’s National Congress, as explained above, feared not only the passions of the masses but also the concentration of power in the hands of a small ruling elite. And just like Constant, they believed the most reliable way to prevent this danger is by setting up a variety of political channels so that power is dispersed and counterbalances itself. In light of this, forms of direct participation by citizens appear as a natural and appropriate addition to Belgium’s decision-making architecture.Footnote ¹⁰⁵ If all institutions that are occupied by professional politicians are susceptible to falling under the sway of political parties (thereby concentrating disproportionate amounts of power in the hands of party leaders, as many claim is the case in contemporary Belgium), bodies populated by non-professionals or direct decisions by the electorate might well be the best way to oppose such power concentrations.Footnote ¹⁰⁶ Again, this is an outcome that nor Constant, nor Belgium’s founding fathers could have foreseen or would have deemed desirable. But we do claim that it is the most coherent continuation of the story that they started to write, a story in which the prevention of power concentrations through the dispersion of power was definitely meant to be a leitmotif. Thus, fidelity to their key normative concerns might well require us to go against their own expectations.

In Belgium, several proposals for referendums and deliberative councils have recently gained traction.Footnote ¹⁰⁷ Although such calls can be heard everywhere, they find a particular resonance in Belgium, where many citizens, local governments, as well as various political parties feel that more direct participation might offer a remedy for the fatigue of existent political structures. If, as we conjecture, the founding fathers did not opt for ‘national sovereignty’ (and its presumed pro-representative implications), but were rather concerned with the dual representation of current public opinion and permanent long-term reflection, as well as with preventing power concentrations through a multiplication of institutionalised powers, direct-democratic mechanisms need not necessarily go against the grain of the Constitution.

Of course, these brief musings on the drafters’ views on sovereignty and on the distribution of powers should not be taken as decisive arguments in the debate on the introduction of more direct or participatory forms of democracy. As mentioned above, many other types of arguments, not only constitutional ones, but also more empirical arguments (as provided by political scientists) as well as prudential considerations should be taken into account. But we are clearing one important hurdle – the ‘national sovereignty’ interpretation of the Constitution – that has so far functioned as a lock on the introduction of direct or participatory forms of democracy in Belgium. At the very least, this provides a strong argument to reconsider the settled jurisprudence of Belgium’s Council of State regarding the unconstitutionality of referendums.