Constitutional amendments have lost their magic (much to the relief of modern lawyers who do not care for magic and do their best to rationalize and demythologize the law). That magic resided in the quasi-religious communion of the amendment procedure with the founding moment of each political regime, the time at which authority (as opposed to mere power) expressed itself and shaped future political behaviour with the utmost legitimacy, but also with an inevitable degree of aloofness from ordinary politics. In many countries, constitutional amendments have become ordinary mechanisms used, when need be, to regulate the massive normative output of inferior institutions at the appropriate level. The aloofness has nearly gone, and the legitimacy has weakened. Since 1949 the German Basic Law has been revised more than fifty times. The French Constitution has been amended twenty-four times since 1958, including the major overhaul made by the reform of 23 July 2008. Many important constitutional amendments have been triggered, not from above, in accordance with the magical understanding of the constituent power, but from below, through a technical process of updating rather than through any re-enactment of the State’s founding. Yet, however mundane the amending power may have become, it still has a political dimension. Recourse to the constitution is made necessary by the almost instinctive belief that the constitution is a higher law, and that their entrenchment in the constitution shelters certain key political arrangements during times of upheaval.

Constitutional substance is like quicksilver: it slips from one’s grasp. Constitutional form is like crystal: it is easily broken. It was the very purpose of constitutionalism to set in stone certain rules, principles, values and institutional patterns. Judicial review has emerged as the most satisfactory means of guaranteeing constitutional substance against undesirable alterations. Constitutional legislation has evolved through time. Rationalization is a process that runs deep within constitutional lawmaking. The natural law element in constitutionalism has faded while a variety of mechanisms, normative constraints and the apparatus of judicial review have flourished. What is new is not the existence of legal limits on the amending power as such; they have been around for some time. It is rather that those limits have taken on a new profile in the era of constitutional review. In constitutions such as those of France, Italy or Germany, ‘eternity clauses’ do not point explicitly to the intervention of the constitutional courts. This raises the issue of their proper degree of involvement. Should constitutional courts meddle with the constituent power or not? In the Tribunal of Public Opinion, the burden of proof lies with those answering no. If the constitution is indeed the highest law of the land and if the courts are its staunchest guardians, one would do well to come up with fairly sound reasons as to why they should refrain from reviewing constitutional amendments. The fact that the courts concern themselves with the exercise of the amendatory power raises another issue. No one seriously believes that judicial review miraculously maintains constitutional substance in a pristine, untouched, condition. There is ample evidence that the guardian of the constitution plays a role, often a paramount one, in defining constitutional substance. When courts, as in Italy, Germany or India, declare themselves competent to review constitutional amendments, they are clearly induced to express higher constitutional principles that are the underpinnings of such review. Alongside eternity clauses, there is now a body of eternity rulings by which courts have set standards for the review of constitutional amendments. Similarities can be found in those enactments and rulings: they prompt readers to see the constitution as a complex, multidimensional, entity that is not reducible to a set of norms. Strikingly, as we shall see, the same could be said of a court that has decided to declare that constitutional amendments are not justiciable: the French Constitutional Court (Conseil constitutionnel).

This essay is divided into four parts. The first gives an account of the case law of the Conseil constitutionnel regarding the judicial review of constitutional amendments (I). In a second part, I attempt to show that the CC’s refusal to review amendments can only be understood in the light of a doctrinal background that provides its intellectual justification (II); this doctrinal literature comes up with a theory of the constitution which, in turn, helps to explain why, despite refusing to review constitutional amendments, the CC is involved in the continuing process of altering the French Constitution (III). Also, while the supporting literature insists on the absence of substantive limits to the amendment of the Constitution, and on the absence of any supra-constitutional rule in the French legal system, an analysis of the language used by the CC in these rulings offers reasons to diverge from this view. The fourth and last part of this article attempts to show that, while the CC has refused to review constitutional amendments, this has been done in a way that comes very close to the language used by those courts which state that such amendments are justiciable. Far from adhering to a mere policy of neutrality and self-restraint, the CC speaks a ‘language of eternity’ with a rich substantive content (IV).

I. THE RULINGS

When it comes to discussing the issue of the judicial review of the amending power in the French context, four cases are usually discussed. The first three occurred in a particularly dense political context. And all of them have helped shape French constitutional politics.

Before we set out to analyse those rulings, a few words should be said about constitutional review in France. Before 2008 (and the enactment of article 61(1) C) the CC was special among the world’s constitutional courts in that it could only exercise constitutional review under the form of ‘constitutional preview’. The 1958 Constitution created two distinctive procedures. Article 61(2) C allows for the review of ordinary Acts of Parliament. Under article 61(2) C:

• Acts of Parliament may be referred to the Constitutional Council, before their promulgation, by the President of the Republic, the Prime Minister, the President of the National Assembly, the President of the Senate, sixty Members of the National Assembly or sixty Senators (...).

Article 54 C has a separate purpose, namely to allow for a check on the compatibility between treaties and the Constitution before those treaties are ratified:

• If the Constitutional Council, on a referral from the President of the Republic, from the Prime Minister, from the President of one or the other Houses, or from sixty Members of the National Assembly or sixty Senators, has held that an international undertaking contains a clause contrary to the Constitution, authorization to ratify or approve the international undertaking involved may be given only after amending the Constitution.

Nowhere in the Constitution does any clause explicitly enable the CC to review constitutional amendments; but neither does any ouster clause prohibit such review.

A. The decisions and their context

1. The 1962 ruling

The first topical case is coeval with the major constitutional reform brought about by De Gaulle in 1962 regarding the way in which the Président de la République, the head of State, was appointed. Initially the President was designated by a body of about 80 000 delegates. Since 1962, he has been designated by universal suffrage. The bill introducing that reform was approved by ‘the people’ by way of a referendum based on article 11 C and not on article 89 C. Article 11 C provides for statutes to be adopted by popular referendum. The scope of subject-matters those statutes may cover is set out restrictively in article 11 C:

• The President of the Republic may, on a recommendation from the Government when Parliament is in session, or on a joint motion of the two Houses, published in the

Journal Officiel

• , submit to a referendum any Government Bill which deals with the organization of the public authorities, or with reforms relating to the economic or social policy of the Nation, and to the public services contributing thereto, or which provides for authorization to ratify a treaty which, although not contrary to the Constitution, would affect the functioning of the institutions.

Although the provisions refer to ‘the organization of the public authorities’ and ‘the functioning of the institutions’, it was generally held that this procedure was not supposed to be used for the sake of amending the Constitution. As a matter of fact, the only provision in the Constitution’s chapter 16 (‘On amending the constitution’) was article 89 C, which provided for a different procedure from article 11 C:

• The President of the Republic, on the recommendation of the Prime Minister, and Members of Parliament alike shall have the right to initiate amendments to the Constitution. A Government or a Private Member’s Bill to amend the Constitution must be considered within the time limits set down in the third paragraph of art. 42 and be passed by the two Houses in identical terms. The amendment shall take effect after approval by referendum. However, a Government Bill to amend the Constitution shall not be submitted to referendum where the President of the Republic decides to submit it to Parliament convened in Congress; the Government Bill to amend the Constitution shall then be approved only if it is passed by a three-fifths majority of the votes cast (...)

For De Gaulle to have recourse to article 11 C instead of article 89 C in order to alter the Constitution thus came as a shock to many lawyers, even, as we now know, inside the CC itself. The supreme administrative court (Conseil d’Etat) acting in its capacity as administrative counsel to the executive, maintained that article 89 C alone laid out the proper procedure for amending the Constitution. In 1962, the President of the French Senate deferred this bill to the CC. The Court ruled that it had no jurisdiction to determine whether the bill violated the Constitution or not. That decision came in the formative years of the Fifth Republic. Commentators were keen to point to the Council’s leniency or maybe self-restraint. In 1962, interpreting (or maybe standing as the guardian of) the Constitution was a task beset with specific difficulties. The author of the Constitution had not yet left the stage. The artist was still at work and the plaster not yet hardened. De Gaulle was the artifex of the Constitution in the fullest sense: not the drafter in a technical sense (that had been the role of Michel Debré and his entourage); not the legal ‘author’ (the people had adopted the Constitution by way of a referendum), but the artist himself, who had moulded the Constitution according to his own idea of constitutional power. Maybe ‘idea’ is too abstract a word: the Constitution was moulded around De Gaulle’s very person. How could a newcomer like the CC not be immensely deferent in such a context? In the 1962 ruling, the legal realist could detect the work of a newborn institution deferring to the will of a ruler at the peak of his legitimacy. Anyone steeped in the tradition of French public law might also see a reference to the general will, a concept that spanned the gap between the French Revolution and the ideas of public lawyers of the Third Republic (1875–1940). These motives were not incompatible: deference to the legitimate author of the Constitution (something very different from mere submission to political force) was expressed in terms of adherence to enduring principles of legal legitimacy. At the same time, the irony was that the creation of the CC foreshadowed the demise of the classical theory of the law as the expression of the general will.

2. The 1992 (Maastricht) rulings

The next two relevant cases were decided in 1992, during the time when the burning issue of the ratification of the Maastricht Treaty was at the forefront of French politics.

a) On 2 September 1992, the CC was asked to review the ‘Maastricht’ Treaty on European Union. The application was based on article 54 C rather than article 61 C. This was the second time that the Treaty had been referred to the Court. In April, responding to an initial article 54 C application introduced by the President, the Council had ruled that the Treaty could only be ratified after the Constitution had been amended (‘Maastricht I’). This triggered, among other things, the enactment of article 88-3 C. On 2 September 1992 (‘Maastricht II’), the Council ruled that no clause in the Maastricht Treaty now came into conflict with the Constitution. The referral had alleged that ‘despite the adjunction [to the Constitution] of article 88-3 C [the Maastricht Treaty] remains in contradiction with the Constitution’. The claimants argued that the amending power had failed to amend article 3 of the Constitution as well as article 3 of the 1789 Declaration of Rights insofar as both clauses express, among other things, the principle of national sovereignty. The CC responded that:

• Subject to the provisions governing the periods in which the Constitution cannot be revised (Arts 7

C

• and 16

C

• and the fourth paragraph of Article 89

C

• ) and to compliance with the fifth paragraph of Art. 89

C

• (“The republican form of government shall not be the object of an amendment”) the constituent authority is sovereign; it has the power to repeal, amend or amplify constitutional provisions in such manner as it sees fit; there is accordingly no objection to insertion in the Constitution of new provisions which derogate from a constitutional rule or principle; the derogation may be express or implied.

It also ruled that:

• The constituent power is sovereign, save only for the exceptions indicated above; it has power to repeal, amend and amplify constitutional provisions in such manner as it sees fit.

…and that:

• The effect of art. 88-2 is to remove the constitutional obstacles to integration of France into the economic and monetary union established by the Treaty; it is within the constituent authority’s discretionary power to decide whether to insert a new provision in the Constitution rather than amending or amplifying Arts 3 and 24 on the powers of the representatives of the people; the argument that those articles are violated is thus devoid of substance.

b) On 23 September 1992 (‘Maastricht III’) a final attempt was made by several Members of Parliament to have the CC declare that the bill authorizing the ratification of the Maastricht Treaty had been passed in breach of the Constitution. It was an ‘article 62(1) C’ type of review as opposed to an ‘article 54 C’ one. The Act had been adopted by way of an article 11 C referendum.

The CC maintained that it had no jurisdiction in such a case as ‘the CC’s jurisdiction is strictly defined by the constitution (...)’ and ‘art. 61 only empowers the CC to establish whether the organic laws and the ordinary acts of parliament are compatible with the constitution (...)’ and does not specify whether ‘this competence extends to all legislative enactments, be they adopted by way of a referendum or be they enacted in parliament (...)’.

3. The 2003 ruling

In March 2003 several Members of Parliament referred a bill ‘regarding the decentralized organization of the Republic’ to the CC. The bill amended several clauses in the French Constitution. Along with another bill of the same nature that had not been referred to the CC, the bill had been passed on 17 March 2003 by ‘Parliament convened in Congress’, an ad hoc chamber which is competent, on the basis of article 89 of the Constitution, to approve constitutional amendments. The claimants based their application on article 61 C. They alleged that, despite this nature, the CC did have jurisdiction to review the bill. The Court took another view; in a rather laconic decision, it ruled that it had ‘no jurisdiction to decide’ the case.

The decision beautifully exemplifies the CC’s customary imperatoria brevitas. It is grounded on arguments of jurisdiction alone:

• 1. The council’s jurisdiction is defined restrictively by the constitution. It can only be expanded by way of an organic law with due consideration for the principles stated in the constitutional enactment. Claims cannot be brought before the

CC

• in cases not provided for expressly in these enactments. 2. Article 61 of the constitution grants the

CC

• power to review

lois organiques

• or, if and when they are referred, l

ois ordinaires

• . [Yet] the

CC

• does not derive from either art. 61, art. 89 or, for that matter, any other clause in the constitution, a power to review a constitutional amendment. 3. As a result, the

CC

• has no jurisdiction [in the present case].

B. The CC does not review amendments.

The cases I have just reviewed have ‘contributed to the conclusion that, in France, constitutional amendment bills are not subject to constitutional review’. The CC has repeatedly ruled that it has no jurisdiction over constitutional amendments. Only a minority of observers maintain that the matter is slightly more murky. In particular they insist on some details that lend credence to a more nuanced view of the matter.

1. This is obviously the case of the Maastricht II ruling in which the CC stated that the amending power was ‘sovereign subject to the provisions governing the periods in which the Constitution cannot be revised (Arts 7 and 16 and the fourth paragraph of Art. 89) and to compliance with the fifth paragraph of Art. 89 C (“The republican form of government shall not be the object of an amendment”) …’. It is of paramount importance to be mindful of the procedural context of this latter ruling. In the French Constitution, article 54 C and article 61(2) C create two separate procedural vehicles. On the basis of article 54 C, the CC is involved in the process of articulating international law (treaties already signed but not yet ratified) and constitutional law. The CC cannot annul the treaty or declare it void. It can only declare that, as a prerequisite to ratification, the Constitution has (or does not have) to be amended. It is then for the constituent power to decide whether such an amendment ought to be made. Such a declaration of incompatibility places the Court in a new context: as a part of the process of amending the Constitution, article 54 C puts it in a position to possibly trigger the amending process by pointing to the incompatibility between the treaty and the Constitution. The only remedy can be to amend the Constitution. Failing this the treaty cannot be ratified. As a result, article 54 C should be seen as one of the constitutional clauses regulating the amending power. Yet it is not a part of title XVI of the Constitution (‘of amendments to the constitution’) and it is not referred to in article 89 C. The Maastricht II ruling should be read in this light: the amending power had been exercised once but its use could still be required as a result of the Court’s review. At this point, the Court might be tempted to define some guidelines for the later use of its amending power by ‘Parliament convened in Congress’.

2. Be that as it may, this explanation is hardly of any use in the case of other relevant dicta regarding the amending power. In at least five decisions since 1999 the CC set out the way in which the amending power could enact rules that ‘derogated from previous constitutional rules or principles’. There were, said the Court, ‘no hindrances’ to such an exercise of the amending power, ‘under the limits laid out in articles 7 C, 16 C and 89 C of the constitution’. In both these 1999 and 2003 rulings, the procedural context is also of interest: this was not an article 61(2) review C but an article 61(1) C procedure: what was brought before the Council was an organic law, not an ordinary act of parliament or a treaty.

Here, a reader not versed in French law might ask a very simple question: the entrenched 1958 Constitution contains certain limits on the amending power. These limits are spelled out by the CC itself in the rulings that have just been quoted (‘under the limits laid out…’). How, if the CC refuses to review amendments, can such limitations be enforced? The answer is clear: de lege lata they cannot, at least in the course of constitutional review as exercised by the CC. This might appear as a blunt disregard of the blank letter of the Constitution, and indeed it might well be just that. This is also an instance of the classical distinction in legal theory between the issue of validity and the possibility of a sanction: unconstitutional amendments can remain in force, as the constitutional court has refused to review them and has thus turned down the chance to declare them invalid. To fully understand this contradiction, one has to turn to what doctrinal literature has to say on these rulings.

II. CONSTITUTIONAL SCEPTICISM

All the cases we have examined are remarkably laconic. The CC can hardly be said to give reasons, yet its decisions are not unprincipled. The decisions may not give express reasons, but they point to reasons, by using certain standards such as (in our case) ‘the spirit of the constitution’, or the ‘direct expression of national sovereignty’. One can imagine that those who decided on these cases relied on very precise and well-articulated arguments. Yet the decision merely stands as a kind of syllabus briefly spelling out these reasons. Reasons are relegated to the background as if they were better left unarticulated. In a western legal culture dominated by the values of transparency and rationality as implying (among other things) the giving of reasons, the level of rationality of the CC’s case law does not exceed a certain limit. This is ultimately bound to have a negative impact on its legitimacy. Yet this is how French courts often decide cases, and the question is not so much of approving or disapproving of this phenomenon as to inquire into its meaning. In Germany, India or the United States, opinions are given by individual judges and often reflect a lengthy and quasi-doctrinal pursuit of the meaning of the law. In France those standards are expressed in unanimous rulings that are expressive of the will of the State rather than of the opinion of a lawyer, be he a constitutional judge. Because of the way the decisions are drafted, what comes to the forefront is authority rather than what, in classical common law thinking, was termed the artificial reason of the law.

Yet these rulings can afford to be laconic. They can rely on a coherent set of doctrinal justifications which has been developed over time in the legal literature. Leading among their supporters are authors who were very close to the Court itself, either because they were former members of its judicial panel (Georges Vedel, Robert Badinter) or its Secrétaires Généraux (Bruno Genevois, Jean-Eric Schoettl). This supporting doctrinal literature tends to come to the defence of the rulings by using broadly similar arguments. I will refer here mostly to the views of Georges Vedel and Bruno Genevois. They express admirably what may be called a ‘sceptical’ theory of the constitution. In the following paragraphs, I will only attempt to sketch out the broad lines of this theory, which approaches the constitution in purely formal terms and is very reluctant to admit that there can be actual limits (other than procedural ones) to the amending power. The sceptical doctrine also rejects ‘supraconstitutionality’ as well as most legal distinctions between the initial enactment of the constitution and its amendments.

1. Except for a limited number of authors, French doctrinal literature about the limits of constituent power has adhered for quite some time to a very distinctive understanding of the constitution. It has not lost its impetus with the rise of the normative account of the constitution that has followed the development of the case law of the CC. Rather, the sceptical account and the normative understanding of the constitution have coexisted rather harmoniously, despite some disagreements along the way. According to the sceptical interpretation, the concept of a constitution is purely formal. At law, there is no substantive content inherent to the constitution. George Vedel was already leaning in this direction from the beginning of the Fourth Republic. In his 1949 textbook, Vedel insisted that the concept of a constitution was defined by a specific procedure of creation and amendment, not by any specific content. He pointed out that the formal element in defining the constitution mattered more than the substantive element (what should be the content of the constitution). In fact, it was that formal element that mattered ‘essentially’ (he said) from a legal point of view. He summed up that ‘formal’ definition of the constitution (as opposed to the political definition) by saying that ‘the constitution is such a norm as can only be altered according to certain procedures’.

One striking fact about this definition is that the constitution is not directly defined by its normative force, although this aspect is obviously decisive in the context. If Vedel insists on the procedural element in defining the constitution, it is mostly because only procedural markers can provide clues to legal issues about the constitution: ‘what distinguishes the constitution from “ordinary” statutes of parliament; what is the sanction at law for a violation of the constitution’ by such an ordinary act of parliament? Another striking element is the insistence on procedure rather than origin. On some occasions, George Vedel also pointed to a second element of definition: the constitution is also defined by the authority that enacted it. His views hardened in his writings in defence of the Maastricht rulings:

• As far as law is concerned, there is no material definition of the constitution. Any enunciation emanating from the constituent power, whatever its object, is constitutional.

This is an important aspect of Vedel’s theory of constituent power. When he embarks upon the task of defining the concept of a constitution, he does not point to its author. Rather, he points to the way in which it is enacted and modified. In defining both the concept of ‘constitution’ and that of constituent power, the question of ‘how’ (is the constitution enacted and modified) has a propensity to prevail over the question of ‘who’ (should be the author of the constitution)? This is the case even when Vedel states than ‘an enactment is constitutional, whatever its content, when it emanates from the constituent power’. Here, the ambiguity of the word ‘power’ plays a role. The word may designate either an entity (the State, the people, an assembly) or a function. The ‘who’ question is obviously more political, has more substantive implications than the ‘how’ question, which appears to be more neutral. As Vedel does not go on to say who may enact constitutional norms, constituent power appears to be a procedure, an impersonal mechanism. The reader is induced to draw no distinction between different possible legal holders of the constituent power.

2. It is no surprise, then, that another central tenet of the sceptical doctrine of the constitution should be that it is largely irrelevant to reason in terms of originary constituent power or amendatory power (pouvoir constituant de révision). It matters little, it is said, whether it is a power exercised by ‘the people’ in the sense of the community of citizens that initially approved the constitution, or ‘the people’ in the sense of the entity bearing that name (in our times, universal suffrage) and to which the constitution grants certain normative powers; or even if it is not ‘the people’ at all but a special representative assembly named Congress that is empowered by article 89 C to amend the constitution. All of this appears to be absorbed into a larger, all-encompassing entity named constituent power. To this constituent power, the CC has indiscriminately ascribed the quality of being sovereign. The jurisdictional meaning of sovereignty (immunity from judicial review) prevails over all the other possible meanings, including the political one.

Both G. Vedel and B. Genevois seem to accept the view that there is a distinction between an originary constituent power and a derived or instituted one. Be that as it may, their approach tends to blur the distinction between the originary constituent power and the derived constituent power. In Vedel’s view, this involves a consequence with regard to the meaning of the term ‘sovereignty’:

• The amending power is the expression of sovereignty in all its plenitude under the proviso that it is exercised according to the procedure that identifies it as such.

Because they are one and the same thing in the eyes of the law, the amending power which is normally a creature of the constitution, is endowed with the same omnicompetence and immunity from review as the originary constituent power. While Georges Vedel refers to the ‘plenitude’ of the powers of the Congress under article 89 C, Bruno Genevois approaches all the forms of constituent power in terms of competence. Yet he defines the competence of the Congress as being unconditional. The gap between the ‘original’ constituent power and the ‘derived’ constituent power is thus bridged, and the virtue of sovereignty (in the sense of fullness of power, of non-limitation and absence of legal constraints) is extended to an authority that was supposed to have the inherently limited nature that constitutionalism attributes to all the creatures of the constitution.

As a logical consequence, the majority opinion in doctrinal literature approves of the CC’s decision to construe the original constitutional enactment and the constitutional amendment bills as exactly the same type of legal enactments. This is remarkable, especially as this seems to defeat what Raymond Carré de Malberg, arguably the most important public law scholar of the Third Republic, saw as the defining feature of statutory law in French public law: not its content (some statutes could perfectly well lack the generality of content which was generally attached to the concept of ‘loi’), but its origin: the organ that is competent to enact it. Today, what we see is a rather strange consequence of this reasoning: no matter who enacts them, all amendments to the constitution equally enjoy an identical status of ‘sovereignty’. Rules arising through different procedures, and different authors, are, for certain important concerns, treated as having an identical legal status. Enactments of the sovereign people of 1958, of an article 11 C referendum, of an article 89 C referendum, and those of the Congress are equally immune from judicial review.

Other scholars, who do not adhere to the sceptical understanding of the constitution, have been dismissive of this approach insofar as it confuses (in their view) the originary constituent power and the amending power. According to this line of reasoning, the rulings we have discussed grant the character of sovereignty to the latter while it should belong to the former only. In Olivier Beaud’s view, the amending power is a mere constitutional magistracy which has to submit to the form as well as the substance of the constitution. Indeed, the attempt to confuse amending power and originary constituent power goes against an understanding that has at least some historical credentials. It was the originary theory of Sieyès and many others during the French Revolution. Every generation of public lawyers, however, is free to break with tradition, or to interpret tradition differently. It is nowhere set in stone that the distinction should have the implications that Beaud and others have stated.

3. This victory of form over substance is also reflected in the third tenet of the sceptical doctrine: the rejection of ‘supraconstitutionalité’ (a word that does not translate well into English: I would suggest ‘supraconstitutionality’ as a neologism). The word ‘constitution’, it is affirmed, does not refer to any substantive content. Any rule can be inserted in the constitution, as long as the proper amending procedure is followed. There is no a priori definition of the substance of the constitution. If there were, it would imply that there is a set of norms that would dictate what the content of the constitution should be and that would therefore stand above the constitution itself. The rejection of a substantive definition of the constitution and this attitude towards supraconstitutionality are really two sides of the same coin. If the constitution were defined by its substantive content, this would raise the issue of what to do if such content were absent from the constitution or of whether some matters ought not to be included in a constitution by way of an amendment, for instance. This would create a paradoxical situation in which a certain substantive definition of the constitution would in itself be superior to the constitutional statute: it is as if there were a norm superior to the constitutional enactment, deciding what should or should not be included in this enactment.

More often than not French lawyers do not know what to do with such supraconstitutional norms. For very serious reasons, they are attached to the modern definition of law as a manifestation of the sovereign’s will. Outside the four corners of the normative constitution, they feel rather insecure. Where should one look for supraconstitutional principles if we cling to the belief that only popular consent legitimates modern constitutions and that only the written word of the constitutional statute matters? Legal ideas such as supraconstitutionality have often been used rather disingenuously as a vehicle for the private political agendas of their proponents. While some such theories have left their imprint on the judicial control of the amending power, they are generally seen as unsatisfactory. Lawyers, it seems, should not set foot outside the territory of constitutional norms. Any attempt to ascribe a substantive, unamendable content to the constitution is dismissed as natural law. Another aspect of the same line of reasoning consists in the assertion that there is only one level of constitutional force: there are no degrees of superiority of legal force within the sphere of constitutional norms. As a result, there can be no supraconstitutional norms in positive law in the sense of a superior tier of constitutional norms, there being also an inferior tier of such norms. There are in fact neither, but only one level: the constitution itself.

4. As a result, it is hardly surprising that the sceptical doctrine of the constitution should reach the conclusion that there are no limits to the amending power except procedural ones. As regards substantive limits (limites d’ordre matériel), a noted commentator has expressed the view that there was only one of them in the text of the Constitution: the protection of the ‘republican form of government’ in article 89(5) C. As both Vedel and Genevois see the matter, interpreting this clause is a rather straightforward exercise: it is only about forbidding a subsequent return to monarchy. That said, B. Genevois is rather diffident of the views held by those who plead for ‘going beyond the literal meaning of article 89 C’ as defending ‘a highly contentious view (...) of supraconstitutionality’. The main contention of the author seems to be that any further step would be a step too far. As a result too much leeway would be granted to courts. This justifies the CC’s reluctance to engage in the review of the amending power.

That such a restrictive view of the matter can indeed be sustained is open to debate. Article 89(4) C sets out that ‘No amendment procedure shall be commenced or continued where the integrity of national territory is placed in jeopardy’. Yet there is nothing objective or merely procedural about such a standard as ‘the integrity of the territory’. To decide whether the territory’s integrity is jeopardized obviously involves a judgement that goes beyond legal forms. Moreover, that a constraint on a revision process should be procedural does not mean that it is without any political meaning, and cannot generate political controversy. This point could be made, for instance, with regard to the clause in article 7 C which states that ‘Neither arts 49 and 50 nor art. 89 of the Constitution shall be implemented during the vacancy of the Presidency’.

Finally, this doctrine has to overcome the very wording of article 89(5) C: ‘The republican form of government shall not be the object of any amendment’. One could make the point that the odds of a return to monarchy are rather small in contemporary France, and that this readings amounts, for all practical purposes, to neutralizing the republican clause. Eventually, this amounts to saying that there are no substantive limits at all in the Constitution. Moreover, if any attempt was made to turn the republic into some other kind of regime (a personal dictatorship, say) it would probably be done in a disguised and indirect fashion. If a constitutional amendment intended to restore personal rule in this indirect way (for instance by extending the President’s term of office to life) the CC would be bound by its own precedents to refuse to review it. Even if the ‘republican’ clause of article 89(5) C is the only substantive limit contained in the French Constitution, the CC has nevertheless renounced ever making it effective by declaring that amendments to the Constitution could not forgo judicial scrutiny. Does this not amount to a refusal to acknowledge the supremacy of the Constitution by not giving effect to one of its clauses? One could retort that this is of course a very unlikely course of events. Yet there is a kind of paradox here. In the course of the judicial review of ordinary legislation, the CC protects many principles that are relevant to maintaining the republican nature of the Constitution. This is the case, for instance, of the principle of equality before the law, or of the principles contained in article 3 C. The republican nature of the Constitution is thus protected against statutory infringements. Why then not act against a President looking to trigger an article 11 C referendum to change the constitution and become a dictator? Why not either against a parliamentary majority seeking to change the constitution in order to infringe the rights of a specific category of legal subjects? The circumstances of 1884 (the year when this ‘republican’ clause was first inserted in a French constitution) are gone, but history is contingent and unpredictable and so are the perils that a republican constitution might face.

More generally, why should this narrow reading of the republican clause in article 89 C (the prohibition of monarchy) be in any way more immune from natural law or supraconstitutional reasoning than any other understanding of it? Even if the republican clause in article 89(5) C has a very limited scope, as George Vedel himself argued since 1949 at least, it is still an eternity clause with a substantive content. Any clause defining a certain quantity (be it very small) of constitutional substance in order to protect it against amendment is no more and no less supraconstitutional than another, even when it has been interpreted as narrowly as possible.

5. Earlier in this article I said that a striking feature of the sceptical definition of the constitution lies in its insistence on procedure rather than origin. The issue of ‘who’ should be the author of the constitution was the key to the way in which Sieyès framed the concepts of what would become the orthodox account of constitutional law. The constituent power was either originary or derived. Only ‘the people’ acting as a nation could be the originary constituent power. Other authorities could only act as trustees to that originary people or nation. If the constitution had to be modified, it was only by a derived constituent power, which, being a creature of the constitution, was bound by such limitations as the said constitution had initially set out. The sceptical theory can been approached as a reaction to that classical understanding. It displays no manifest hostility to it, and sometimes pays lip service to what it has to say. Yet what it objectively achieves is to overturn that classical account of constituent power. It does so by putting forward a procedural account of the constituent power which does not insist on ‘who’ but on ‘how’ the constitution is changed. It also objectively subverts the classical theory by insisting on the essential sameness of originary and derived constituent powers.

This is not to say that the rulings we have discussed came out of the blue. They can be connected to several other long-term trends in the history of French constitutional law. More especially, they identify the nation with its representatives. Despite its name, ‘the people’ mentioned in article 11 C is not the original sovereign ‘people’ that adopted the Constitution. It is merely its constitutional representative. And so is article 89 C’s amending Congress. This is very much in the French tradition, at least the one established during the French Revolution. In a sense, this looks like a return to the pre-1958 era, when Parliament (especially during the Third Republic) could rather freely amend the Constitution without being subject to review by a constitutional court. The 1962-1992-2003 rulings are open to criticism for logical inconsistency as they seem to treat ‘the people’ that votes at referendums and the ‘Congress’ as entities endowed with the same legal characteristics as ‘the people’ that enacted the 1958 Constitution. Yet this does point to an element of continuity with the tradition of French public law. Raymond Carré de Malberg (writing in 1931) pointed to the fact that French public law…

• …is based on the notion that the people is present during the making of statutes by the legislature as well as at the time of the enactment of the constitution by a constituent assembly (...); the legislature will be competent, as it represents the people, in order to interpret the will expressed by the people in the constitution.

The unification of the different varieties of constituent power would therefore appear to stem from a long tradition of refusing to distinguish between the several entities that bear the name ‘people’ in French public law. The people of 1958 and the people of 1962 were, in this sense, one and the same. And, by extension, why not apply this definition of ‘the people’ to article 89 C’s ‘Congress’. Does not article 3 C state that the people ‘exercises’ its sovereignty ‘through its representatives’ as well as ‘by way of referendums’, thus imparting equal standing and dignity to the ‘nation’, to the originary and sovereign ‘people’, to ‘the people’ of referendums, and to all kinds of representative assemblies?

6. Be that as it may, the sceptical theory of the constitution is generally justified in democratic terms. The amending power said Vedel, is the ‘place where, at law, democracy can express itself without being shared’. This democratic justification is expressed in two famous analogies: the pointsman and the ‘bed of justice’.

The theory of the aiguilleur (pointsman, or maybe, in more modern terms, the air traffic controller) is ascribed to Kelsen, and was introduced into French doctrine by Charles Eisenmann. Vedel and Louis Favoreu frequently refer to it. According to this theory, the role of the constitutional court is not to decide on the substance of the law it reviews. It is not there to pass judgment on the legal provisions from a substantive point of view, but to say who should be competent to enact them. If the clause is in breach of the constitution, only a constitutional enactment can bring it into force. The court thus acts as a pointsman sending the train along one track (amending the constitution) rather than another (ordinary legislation). This doctrine is remarkably in line with the sceptical account. It would appear that courts, and in particular the CC, do not meddle with substance, or political choices. They only point to the right procedure.

The theory of the bed of justice (lit de justice) is often depicted as a variant of the aiguilleur theory. G. Vedel repeatedly expressed the view that only this self-restraint on the part of the court was compatible with the democratic rule: only the constituent power is able to reverse a ruling from the CC and decide that a norm should be inserted in the constitution. This decision is final. Vedel referred to this practice with an expression from the Ancien Régime: bed of justice. During a bed of justice the King came in person to one of his Parliaments (which were higher courts of justice) and could overturn its decisions. Similarly, when the CC decides a case, this theory does not see its decision as final. If the court decides that a clause is unconstitutional, the people in majesty is still at liberty to amend the constitution and force the clause into the constitution. If the CC could review constitutional amendments, this democratic check, it is said, would be neutralized.

The sceptical doctrine also offers a defence of the role of a constitutional court in a democratic regime. It involves no a priori definition of the purposes or values of the constitution. It could be related to Kelsen’s defence of democracy in sceptical terms. For that very reason, it would also appear at first glance as a doctrine of judicial self-restraint. If there is no inbuilt purpose and/or content to the constitution, courts cannot identify one, and should not attempt to impose their views on the amending power. Yet the rulings we have examined are not immune from a democratic challenge: in what way is the democratic people acknowledged as the sovereign (in the sense of being the highest authority in the State)? Does the normative process reflect the people’s supremacy? Vedel’s response was that the people can still appear in majesty and quash a decision of the CC. As a matter of fact, however, the result of the 1962-1992-2003 rulings is that it is not only the people in majesty who can change the constitution validly, but many other organs of the State ranging from the electorate to the Congress. Moreover, the people in majesty is rarely summoned to amend the constitution, if by that expression one refers to universal suffrage. Apart from 1962 (which was an article 11 C procedure), the only time that a referendum has been called to approve an amendment was in 2000. In other words, there is a certain element of circularity in the argument: only ‘the people’ can amend the constitution, but the case law of the CC has extended the legal definition of ‘the people’ that can amend the constitution way beyond ‘the people’ as understood on the basis of article 89 C. Why should anyone want to call the parliamentarians assembled in Congress the people in majesty? Why should they be able to hold a lit de justice with more legitimacy than when sitting in their respective Houses of Parliament?

Even without adhering to either the pointsman or the bed of justice justifications, the rulings we are discussing may still be defended as securities against judicial activism. If amendments were judicially reviewable, this would give the CC a free rein to create supra-constitutional rules that would bind the amending power. Yet this claim raises two questions: first, has the refusal to judicially review amendments effectively prevented the CC, as well as other courts, from creating new constitutional rules or principles? This question will be raised in the third part of this article. Secondly, has the CC avoided the pitfall of supraconstitutionality or natural law merely by declaring that constitutional amendments were non-justiciable? This will be the issue dealt with in the fourth part.

III. CONSTITUTIONAL AMENDMENTS AND CONSTITUTIONAL CHANGE IN CONTEMPORARY FRANCE

A. Beyond jurisdiction.

On the surface, the rulings of 1962, 1992 and 2003 appear as a welcome declaration of self-restraint on the part of a constitutional court ultimately expressing its deference to the will of the people. They also express the cultural recalcitrance of a majority of French academics, lawyers and judges towards natural law or supraconstitutionality. Yet there is more to these rulings than just that. One is left in the 2003 decision with a few paragraphs founded only on a competence-based approach. It does not take a very radical legal realist to understand that jurisdiction is a flexible argument that can be twisted both ways according to necessity. Guy Carcassone, for instance, has noted that the CC has twice controlled the standing orders of the Congress (Congrès: the organ empowered by article 89 C to enact constitutional amendment bills). For this, says Carcassonne, ‘there were some very good reasons for the council to declare itself competent, but it is hard to deny that this competence does not rest upon any clause in the constitution’. As everyone knows, the absence of a relevant enabling clause in the Constitution did not prevent the American Supreme Court from reviewing Acts of Congress. The silence of the constitution has sometimes, as with the Weimar Constitution, been interpreted as in fact empowering courts to engage in judicial review rather than impeding them from so doing. This is only an instance of a more general observation on the issue of jurisdiction:

• The mere absence of a grant of power does not preclude a council deciding that it has that power. Neither the German nor the Indian Constitution authorises explicitly the courts to control the constitution amending power’s exercise. Yet judges have - by their interpretation of the Constitution - claimed that power for themselves.

It is indeed remarkable that the rulings we are discussing are denying jurisdiction rather than asserting it. Jurisdiction is a key moment of the process of judicial review. It is the time at which the court’s discretionality is possibly at its highest ebb. It is also the stage at which the decisional element of judicial review is maybe most visible. By declaring that a certain matter is justiciable, courts take the risk of a potential conflict with other departments of the constitution. They set foot in political territory. For a court to declare that a matter is not justiciable, however, does not amount to a declaration of neutrality on the legal and political ground. A court involves itself no less in the political and legal system by denying jurisdiction than by affirming it. In 1962, for example, for the CC to deny jurisdiction over the amending act adopted by way of a referendum implied a certain understanding of the hierarchy of political authorities in the newborn Fifth Republic. Such a ruling went against certain political interpretations and it defeated the political strategies of their proponents. At the same time, it consolidated the position of De Gaulle and his understanding of the new regime.

Away from party politics, the rulings we are discussing have a significance that goes beyond their immediate negative consequence. What they seem to mean de prima facie is that the CC does not meddle with the amending process. It refrains from passing judgment on it. It is a declaration of non-involvement. The Constitutional Court affirms its neutrality with respect to the decisions taken by the competent amending authority. This is entirely in keeping with the sceptical theory of the constitution. As there is no inherent constitutional content, it is not for courts to get involved in the amending process. Yet this declaration of neutrality is not the end of the matter. It would be an error to interpret it as meaning that the CC has nothing to do with constitutional change in France. This is far from being the case. The CC, as we shall see, is a key actor in a larger process of managing the Constitution that goes beyond its formal amendment.

B. Managing the constitution

In 1994, Louis Favoreu insisted on the fact that ‘the issue of constitutional amendment cannot be apprehended the same way as it was in the past due to the existence of constitutional review’. As he saw it, the advent of a constitutional court in France brought about a remarkable result:

De facto

• amendments resulting from voluntary or involuntary violations of the fundamental norm, as they could be found to take place during the Third and the Fourth Republic, will not be possible anymore. Only official amendments will be admitted.

It is to be wondered whether this remarkable optimism has been warranted by the facts. Does constitutional change, even in the narrow sense adopted by Favoreu himself (the creation of constitutional norms) take only the form of formal constitutional amendments? Are we through with the dark period during which, as Favoreu has it, some sinister political interests found it ‘possible to exploit the elasticity of a constitutional text’? One is tempted to reach a more nuanced conclusion. Constitutional change takes various forms not all of which are covered by the enactment of formal constitutional amendments. Moreover, that the CC should deny jurisdiction over constitutional amendments does not even mean that this court is not involved in constitutional change. For several reasons, nothing would be further from the truth. French public law has in fact witnessed the development of new ways of changing the Constitution without formally amending it. The rulings rejecting the review of constitutional amendments are no hindrance to this phenomenon. One could almost be tempted to see them as a kind of veil behind which a larger mechanism of permanent constitutional transformation takes place seamlessly.

1. Firstly, since 1971, the CC itself has developed a voluminous body of constitutional principles that are the grounds of its review of Acts of Parliament. The CC, like a number of other courts, cannot formally amend the Constitution, but it can create constitutional norms and does so on a regular basis. This body of constitutional norms has grown way beyond the formal enactment first adopted in 1958. This has enabled the CC to develop a protection of human rights although the Constitution did not contain any bill of rights. This eases the pressure on controlling formal amendments. The Court has its own private amending procedure, which it can use whenever it sees fit. In 1971 the CC decided that the Preamble to the 1958 Constitution contained positive constitutional norms. It has used those norms as references for its judicial review of Acts of Parliament. Should an Act of Parliament come into conflict with the rules referred to by the 1958 Preamble, the CC would declare it unconstitutional. The 1958 Preamble refers to several ‘texts’ throughout French constitutional history: the Preamble to the 1946 Constitution, the 1789 Declaration of Rights, the 2005 Environment Charter. The 1946 Preamble, in turn, refers to several categories of principles: ‘principles acknowledged in republican legislation’ and ‘principles of particular importance for our times’. While the 1958 Constitution contained, as such, no bill of rights, the 1971 case paved the way for the creation of a body of jurisprudential principles that stood for such a bill of rights. The CC has discovered several such principles since 1971. The first one was freedom of association; then came, for instance, equality before the law as expressed in the 1789 Declaration, and several others. In at least one case, the Court has created an unwritten constitutional principle that was not grounded in the 1958 Preamble: the principle of the continuous functioning of public services. The importance of this new judicial policy cannot be underestimated. The Court has created new constitutional rules. As such, they stand as norms of reference for the judicial review of ordinary legislation. If an appeal is filed before the Court, these norms will trump contrary provisions in the Act of Parliament. Yet, as a matter of policy, the Court adheres to a line of conduct of self-restraint. It is reluctant to create such new principles and does so only infrequently. When it does, it is careful to fasten them to a textual peg: namely, either a provision in the 1946 Preamble or one in the 1789 Declaration of Rights. But since 1971 the Court has officially been in the business of producing new constitutional norms. This is what gives the lit de justice and aiguilleur theories an air of unreality.

2. Secondly, if by constitution one means the political institutions and their legal apparatus, the main engine of constitutional change in France has for some time been the development of the European Union and of EC law. As far as civil liberties and their protection are concerned, the driving force has, of course, been international law and especially the ECHR. Because of the constraint created in the French Constitution by article 54 C, these transformations were bound to have an impact on the formal constitutional enactment itself. Many formal amendments to the Constitution now take place as a result of a requirement emanating from the CC: this is especially the case when the CC has stated (in the course of an ‘article 54 C’ procedure) that a treaty cannot be ratified before the Constitution has been amended. External pressure, such as that of the European Union and ECHR law has generated a flow of constitutional amendments. This is no longer about ‘the people’ changing its constitution in a rather grand and Lockean fashion. Nor is it about the ‘original right’ of the people to establish its government, resulting in ‘a very great exertion’ that cannot and ought not to be ‘frequently repeated’ that Marshall referred to in his Marbury v. Madison opinion. This is a much more mundane, technical and constrained process. The triggering of this process is, as we have already pointed out, ruled by article 54 C:

• If the Constitutional Council, on a referral from the President of the Republic, from the Prime Minister, from the President of one or the other Houses, or from sixty Members of the National Assembly or sixty Senators, has held that an international undertaking contains a clause contrary to the Constitution, authorization to ratify or approve the international undertaking involved may be given only after amending the Constitution.

This has made way for a mechanism known as adjunctive amendments (révisions-adjonctions). The conflict of norms between a constitutional rule, for instance, and a treaty, was solved by inserting a clause to the effect that the legislator was permitted to authorize the ratification of the treaty. The ruling in the Maastricht II case appears as a justification of this practice of adjunctive amendments.

By creating a connexion between the inclusion of new rules of international law in the French legal system and the amendment process, article 54 C raises an issue similar in many regards to that raised by article 8 of the constitutional statute of 25 February 1875. Article 8 stated that only the Houses of Parliament could ‘declare that the constitutional laws ought to be revised’ (déclarer qu’il y a lieu de réviser les lois constitutionnelles). Carré de Malberg took this as evidence that there was no clear separation between constituent and legislative power, as Parliament dominated the amending procedure from the outset. In a similar fashion, the lesson to be learned from article 54 C is that the CC can, in quite the same way, ‘declare’ that the Constitution ought to be revised. The declaration of incompatibility of article 54 C also operates as a trigger for the amendment procedure. As the CC is not the organ that will ultimately enact the amending statute, we cannot go as far as Carré de Malberg and say that the Court dominates the whole amending procedure. Yet it is clear that it can trigger it. This power of initiative counterbalances to some degree the fact that, at the end of the chain, the Court has decided not to review the statutes adopted as a consequence of its declaration. In 1931, Carré de Malberg could come to the conclusion that the Third Republic’s Parliament ‘held sway over both the constitution and the statutes’. Could we not reach a similar conclusion about the CC under the present French Constitution?

As a result, in contemporary French constitutional practice, the CC and the constitutional lawmaker are tied into a kind of feedback loop. The amending power takes into account the case law of the CC. In return, the CC does not expressly review the amending process, but it may set out guidelines for the future uses of the amending power. More especially, the Court has reminded the amending power that, while it was at liberty to amend the Constitution, even implicitly, it was in that case bound by the ‘limitations set by articles 7, 16 and 89 C’. Such a reminder is not necessarily an anticipation of a future change in the CC’s case law about the justiciability of amendments. Rather, it shows that the Court is involved in the amending process even when it does not operate as its judicial censor. It is also notable that this reminder was repeatedly expressed in the course of reviewing organic laws, although it also appears in the context of the review of ‘ordinary’ legislation. The reminder could be (implicitly) directed at the institutions that do operate, before its enactment, a non-judicial control of the compatibility of the amendment with the Constitution. This is especially the case of the Conseil d’Etat in its advisory capacity, when the amendment is initiated by the executive. This is also the case of the parliamentary assemblies when they discuss the amending bill.

In some cases, the amending power is used to counteract the effect of an interpretation from the CC itself. It took for instance a formal constitutional amendment to write parity between men and women into the Constitution in 1999. Parliament had created a system of quotas with a view to increasing the number of women candidates at local elections. The referred bill provided that ‘lists of candidates may not contain more than 75% of persons of the same sex’. Yet the CC declared this mechanism unconstitutional on the ground that the constitutional principles contained in article 3 C (there is only one ‘French people’) and article 6 (equality before the law) of the 1789 Declaration of Rights contained principles that ‘preclude any division of persons entitled to vote or stand for election into separate categories; this applies to all forms of political suffrage, in particular to the election of municipal councillors’. This obstacle was overcome by the amending act of 8 July 1999.

3. While the CC does not review constitutional amendments, it can of course interpret them in the course of judicially reviewing ordinary statutes. An amendment becomes a part of the Constitution, and thus belongs to the body of norms protected through the judicial review of ordinary laws. For instance, very shortly after article 1 C was amended to include the principle that the ‘organization of the Republic is decentralized’, the CC interpreted this clause in the course of reviewing a statute on ‘local liberties and responsibilities’ (CC, DC 2004-503, e.g. § 23-27).

4. Finally, at an increasingly frequent pace, constitutional rules are not altered by way of constitutional statutes (lois constitutionnelles) but by organic law (lois organiques). In French law, organic laws are statutes of a special nature. In some cases, the Constitution empowers an organic law to develop and specify its own content. For instance, the constitutional rules about budgetary laws refer to an organic law which was adopted first in 1959 and revised in 2001. The CC itself is ruled by an important organic law of 7 November 1958, which was frequently amended between 1959 and 2010. Organic laws are subject, according to article 46 C, to the CC’s mandatory judicial review. These organic laws tend to cover an ever-increasing scope in the field of substantive constitutional law: they are constitutional by their subject-matter, if not by their degree of legal force. They now regulate many aspects of institutional life, from the normative autonomy granted to overseas territories to the new-born priority preliminary rulings on the issue of constitutionality (‘question prioritaire de constitutionnalité’).

As a matter of fact, that organic laws should be inferior in value to the Constitution does not always matter because of the way the CC has used them as norms of reference: in some cases the CC has stated that certain types of norms (lois du pays de Nouvelle Calédonie) that were adopted in compliance with an organic law were ‘pursuant to the constitution’ as far as procedure was concerned. Agnès Roblot states that this amounts to ‘assimilating’ lois organiques to the Constitution as the ‘respect of the organic law is a condition of the constitutionality’ of the inferior norm. We could go further and say that organic laws are now a major instrument of constitutional transformation. It could be suggested that they be referred to as quasi-constitutional enactments.

As there are ever more organic laws, the CC has a say in the development of constitutional legislation adopted under this form. This category of organic laws, as well as their special status with regard to judicial review, somewhat blurs the line between constitutional amendment or constitutional law and an ordinary amending act. This is especially the case as there is at least one example of an organic law abrogating a constitutional statute. Article 76 C stated that a referendum should be held in New Caledonia to validate the Noumea Settlement of 5 May 1988 (Accords de Nouméa). The same article also stated that ‘Persons satisfying the requirements laid down in article 2 of Act No. 88-1028 of 9 November 1988 shall be eligible to take part in the vote’. This amounted to entrenching (i.e. granting constitutional force to) a provision from an ordinary statute. Yet, despite having been raised to constitutional rank in this fashion, article 2 of the act of 9 November 1988 was later abrogated by a ‘loi organique’. As a result, an organic law has formally abrogated a constitutional norm without incurring the wrath of the CC in the exercise of its mandatory review power. One should note that the same organic law in fact restated the content of the abrogated article in two of its own clauses.

This is not necessarily a desirable solution, as in certain cases, organic laws have been a very long time coming (if at all) once the clause calling for them had been inserted in the Constitution. In the case of the new article 68 C, which dramatically alters the principles ruling the accountability of the French President, this amounts to neutralizing the amendment itself by not enacting the organic law necessary to its implementation. As any three-year-old could explain, an organic law that is not enacted cannot be reviewed. More disturbingly, the failure to enact it, while manifestly unconstitutional (as an instance of negative incompetence on the part of the législateur organique) cannot be sanctioned. The Constitution is treated like any other enactment: the lack of a secondary legislation to implement it (mesure d’application) prevents it from being put into application at all.

IV. THE STUFF THAT CONSTITUTIONS ARE MADE OF.

I now turn to the second challenge with which both the rulings and the sceptical doctrine backing them can be confronted: has the CC avoided backsliding into supraconstitutionality or natural law merely by declaring that constitutional amendments were non-justiciable? To make a long story short: probably not, as is shown by an analysis of the judicial language used by the CC. Far from being in keeping with the sceptical account, this analysis tends to show rather that the CC is keen to generate judicial standards that are nowhere to be found in the text of the constitutional enactment. These standards, as we shall see, are analogous to those of the courts that have decided to extend their jurisdiction to the review of constitutional amendments. In Germany or India, such broad standards as the ‘basic structure’ of the constitution or the constitutional ‘normative order of values’ are used to justify the judicial review of the amending power. In France ‘the spirit of the constitution’ or ‘the direct expression of national sovereignty’ come to the support of the opposite solution. Nevertheless, they raise the question of whether the CC’s case law is in keeping with the sceptical doctrine insofar as it rejects natural law and reasoning in terms of supraconstitutional norms.

A. The self-contained constitution

It has been noted that the sceptical theory of the constitution denied that there were such things as supraconstitutional norms. This has become a word of abuse, and one that covers a variety of approaches that do not seem to have much in common apart from their being viewed with disapproval by the sceptical school.

In fact, under this banner of supraconstitutionality, we should be careful to distinguish several approaches. a/There are indeed some authors who claim (or claimed) that there are norms (or rules, or principles) that are hierarchically superior to the constitution. This was the case of some French law professors under the Third Republic. b/A second approach is based on the notion that there are two periods in the history of any constitution: its enactment and the period during which it is in force. To its enactment corresponds an originary constituent power while, when it is in force, it can only be amended by a derived constituent power which is legally constrained by the entrenched constitution. Carré de Malberg was of the view that the originary constituent power was exclusively a matter of fact, and that lawyers did not have to concern themselves with it. But others have thought otherwise: the people had enacted a constitution, and this gave to this first expression of will an authority at law that could not be overlooked. c/Finally, there are theories according to which there are some conceptual implications in the text of the constitution that have to be taken into account as a matter of positive law. This is eminently the position held by the proponents of a modernized version of the General Theory of the State, especially Olivier Beaud and Olivier Jouanjan. Olivier Beaud, in particular, has identified substantive limits to the amending power that were based on the legal existence of the State. The State is the bearer of the Staatsgewalt (puissance publique) and its existence is a precondition to the existence of the constitution. Despite contrary appearances, this does not mean that the State should stand above the constitution, but rather that it is conceptually distinct from, and not reducible to the constitution. This results in a plea for such legal limits to the amending power as are grounded in the existence of a State: Olivier Beaud bases his own defence of material limitations on the notion that constituent power and revising power (pouvoir de révision) are two different things. The constituent power results in constitutional measures (actes constituants) that express the political decision of the people. When the revising power comes into play, it is subject to this decision as expressed in one or several constitutional measures and it also has to act in keeping with the fact that the ‘State’ pre-exists the constitution. The State is here understood as the ‘puissance publique’ (Staatsgewalt): public powers in their entirety. As a result, a theory such as Beaud’s extends constitutional reasoning to several dimensions other than norms. There are at least two interrelated dimensions in Beaud’s constitutional theory: the decisional dimension that insists that the constitution must be based on the political decision of the democratic people and the dimension that focuses on the implications of the existence of a State (‘general theory of the State’ dimension). Both these dimensions have a bearing on what the organ that is empowered to change the constitution can do.

The word ‘supraconstitutional’ has something misleading about it. It seems to mean that all three types of approaches suppose that there is a body of natural law (for lack of a better word) standing above the constitution itself in a kind of transcendent and hierarchical relationship. Yet the three approaches cannot each be defeated with the same counter-arguments. Against approach (a), it is maybe fair to say that this constitution above the constitution is nowhere to be found. But this will be of no avail against approach (b). Carré de Malberg’s argument that the originary constituent power is only a matter of fact and has nothing in it with which the lawyer should concern himself is probably the best response to that approach. Finally, it will not do to dismiss arguments belonging to the third category (c) by simply saying that their proponents presuppose higher norms above the constitution. As a matter of fact, they do nothing of the sort. The best response is probably that their reasoning amounts to a collapse into Begriffjurisprudenz. Yet it then has to be proved that legal reasoning is never dependent on analytical reasoning, which I feel is a very ambitious claim to make.

In general, it is assumed the proponents of such claims mean that some norms are superior or equal to the constitution but are not contained in the entrenched (formal) constitution. What brings together the theories that come under fire from the sceptical doctrine is not that they in fact all claim that there are some norms that are superior to the constitution (although some do) but that they claim there is constitutional material outside the constitution. The critique against Begriffjurisprudenz, for instance, is more common among legal theorists of the normative school than among members of the sceptical school. Because the sceptical school reasons in normative, and thus hierarchical terms, it reconstructs all those claims in hierarchical terms: if something bears on the constitution, it has to be hierarchically superior.

All the theories that are condemned as instances of supraconstitutionality are in fact criticized for breaching a principle which states that everything that has constitutional value in a legal sense is to be found in the constitutional enactment. The constitution is self-contained. The CC, thus, appears to be justified in denying jurisdiction on constitutional amendments. In doing so, it is perfectly in line with the sceptical doctrine: it rejects supraconstitutionality; it adheres to a neutral understanding of the Constitution, it has no regard for limits to the amending power other than procedural limits. It sees the Constitution as a self-contained body of norms. Or does it? If one looks at the way the CC’s decisions are drafted, there is reason to doubt that it acts in keeping with the sceptical account.

B. The multi-dimensional constitution

What courts actually do in the course of dealing with constitutional amendments cannot be reduced to identifying norms and attributing different legal values to them, thus solving a problem that boils down to a conflict of norms. There is more to legal reasoning than just that. In fact, when courts extract norms from the constitution (or refer to certain norms as being constitutional) they do so by pointing to a multidimensional entity, a cloth that is made of a diversity of different threads. Constitutional review of the amending power brings to the fore at least three such categories of elements that tend to be ignored by normative thinking: (a) essentialist vocabulary; (b) institutional patterns; and (c) what I would call (for want of a better terminology) non-normative legal objects, namely entities that can be described as not lending themselves to be created or abrogated by the constituent power, or for that matter by any lawmaker.

1. Essentialist vocabulary.

Some rather bothering metaphysical assumptions come along with such terms as ‘essential conditions’ or ‘basic structure’. Yet these pitfalls seem to be inherent to constitutional reasoning. This is especially the case now that constitutional reasoning is for its larger part understood as a process of extracting norms from a legal text. The strategy of normativism consists in distinguishing the act and the norm that the act ‘contains’. This attempt to read norms into the text of the formal constitution brings about certain categories of thought that lead courts and other interpreters to think in inside/outside terms. This is visible in the very language the courts use, and emphatically in the context of their reviewing the amendatory power. The constitution, we are told, has certain core (inside) characteristics: an essence (or some essential features), a substance (or substantive elements), a basic structure, an identity. Modern lawyers tend to shun essentialist vocabulary, but no one can deny that this vocabulary is used in court. Despite their brevity, the 1962, 1992 and 2003 rulings are replete with such references to constitutional substance that are absent from the text of the constitutional enactment of 1958. The essentialist vocabulary is not restricted to these rulings regarding the amending power. The CC famously used it in reviewing (under article 54 C) the treaties amending the 1957 Rome Treaty. Its case law was based on the identification of ‘essential conditions of national sovereignty’. In France, maybe under the influence of the CC’s style, this essentialist vocabulary has spilled over into the Constitution itself. Since an amendment adopted in 2003 article 72 C and article 73 C both refer to the ‘essential conditions for the exercise of civil liberties’. Yet such references to the ‘inside’ of the constitution, its very core, has a kind of magically self-defeating virtue. It is meant to save the constitution from destruction, by identifying its core content or values. As in a Lewis Carroll kind of world, the pursuit of the interiority of the constitution directs us towards the outer world. Having reached the core of the constitutional microcosm, the traveller is suddenly redirected towards values, historical precedents, systemic properties that are features of the macrocosm within which the constitution exists. When the CC refers to the spirit of the constitution, for instance, readers have to engage in a kind of elaborate guesswork. They have to look for themselves in the history of the regime and the way in which Général de Gaulle interpreted it himself. When the CC refers to the ‘essential conditions’ of national sovereignty, there is hardly any constitutional clause that can be referred to. The essence of sovereignty has to be found outside the constitutional enactment. Such standards, as a result, depict a constitution which is neither neutral in substantive terms nor self-contained.

2. Institutional patterns.

Moreover, the language of eternity points to features of an institutional nature. It lends credence to the claim, sometimes made in legal theory, that law is not only made of norms but also has an institutional dimension. Many eternity clauses and rulings point to aspects of the political regime: France or Italy as republics, Germany or the United States as federations; the spirit of the Fifth Republic as it is made to depend upon a president elected by universal suffrage. Let us take for instance article 89 C, in fine: ‘The republican form of government shall not be the object of any amendment’. De prima facie, the normative nature of this statement is beyond doubt: an imperative is expressed and directed at the amending power. Yet what is protected against amendment (the republican form of government) is not a norm. A republic is a political form. It can be explored conceptually: the concept of republic entails certain political rules and arrangements, and may also render others impossible. The paradox here is that these institutional patterns have to be approached conceptually, through very broad terms (republic, etc.) the meaning of which has evolved over the course of time and of political changes. Limits to the amending power become dependent upon concepts that are both highly political and highly dependent on historical contingency. What ‘republic’ means in France today is vastly different from what it meant in 1791 or during the Third Republic.

This element of contingency is not merely a feature of the constitutional enactment. It is also apparent in the language that the French Constitutional Court has used. The 1962 ruling was based on ‘the spirit of the constitution’. The Court viewed itself as ‘a regulatory organ’ of the ‘activity of the public bodies’. As a result ‘the bills referred to in article 61 C of the constitution’ should be held to be only ‘those voted by parliament’ as opposed to ‘those which, being adopted by the people by way of a referendum’ are ‘a direct expression of national sovereignty’. The Maastricht III ruling stated that ‘with regard to the balance of powers (équilibre des pouvoirs) as established by the Constitution, the bills referred to in article 61 C should be only those enacted in parliament’ while article 61 C’s procedure should not extend to those enactments that are ‘voted by the French people by way of a referendum (...)’ as they ‘are a direct expression of the national will’.

Here the proponents of the ‘general theory of the State’ have a point. They draw attention to the conceptual nature of law in general and of constitutional law in particular. Eternity clauses or ‘eternity rulings’ point to concepts, or to conceptual architectures. Concepts are strange animals that tend to point to each other. A constitution involves there being a State. To have a democratic State involves there being a people endowed with certain exclusive legal capacities such as political sovereignty, etc. That kind of reasoning is frequent among the proponents of different varieties of the General Theory of the State. It is on a reasoning of this kind that Beaud and Jouanjan have based their distinction between constituent power and amending power. The reaction to this type of conceptual argument of the majority opinion among French lawyers has generally been a diffident one. Nevertheless, the CC itself, in support of the opposite solution, has had recourse to arguments of a similar nature, under the cloak of a laconic reference to ‘the spirit of the constitution’ or the nature of the political regime. That such argumentation was stamped with the authority of a court ruling does not alter its nature. The CC’s style is a very technical one: most of its decisions are based on a precise, thorough and somewhat cumbersome comparison between individual legislative clauses and individual constitutional norms. Yet there are times at which the Constitution itself, or some aspects of the constitutional system, have to be approached in a wholesale fashion, and the CC is not afraid to do so.

3. Non-normative objects.

Similarly, the limits to the amending power that tend in some constitutions (as in the German Basic Law) to protect human rights point to an object that is not of the same nature as a norm. This is eminently the case of human rights. Human rights have a legal nature, but they do not take the form of a norm. A norm, to take the most frequently received definition, is the meaning of an act of will expressed by a competent authority. In the tradition of constitutionalism, human rights are not created by an act of will and cannot be suppressed by one. Hence fundamental rights are declared and not enacted.

In the formative period of written constitutionalism, it was quite clear that human rights were not created by the sovereign’s will and, as a result, could not be abrogated by the sovereign’s will. This is the profound meaning of the American Declaration of Rights, as well as that of the French Declaration of 1789. A ‘declaration’ in this sense was a specific form that was required by the specific nature of what it declared. Its content was ‘natural’ or maybe ‘self-evident’ and as such could not be repealed. It did not require an injunction from the legislator to come to life. The very nature of rights was such that they could not be repealed.

At the same time, early modern constitutionalism allowed for the recognition of certain conceptual relationships that played a role similar to modern ‘limitations’ of the kind envisaged in article 79 of Germany’s Basic Law or France’s article 89 C. It is emphatically the case of article 16 of the French Declaration of Rights of 1789: ‘any society in which rights are not guaranteed and the separation of power not established, has no constitution’. As Michel Troper has made clear in a classic article, this is not to say that a document that fails to protect rights and/or to separate powers would be a poorly framed constitution, one of poor quality. What article 16 means is that, in the absence of such elements, this is not a constitution at all. The relationship established in the text is a conceptual one: the meaning of the word ‘constitution’ is dependent on the presence of two substantive elements: properly guaranteed human rights, and a system of separated powers. Otherwise, the word ‘constitution’ should not be used at all. The idea that the constitutional stuff was made of threads of a different nature has waned with the growth of legal positivism. During the Third Republic, Maurice Hauriou could still hold views that could be put under this banner. There were, he thought, principles of constitutional legitimacy for which ‘no text was needed, as it is in the nature of principles to exist and be valid without a text’. Hardly any contemporary French lawyer would agree with this view, which smacks of natural law and sounds hopelessly outdated. Yet, the paradox is that most of what the case law of the CC has been praised for since 1971 and the development of a body of ‘norms’ protecting human rights can be subsumed under that category. Why then dismiss the existence of unwritten principles when it comes to reviewing the amending power?

Here, the hierarchical language disguises the fact that what is pointed to is a legal entity that lies out of the reach of the constituent power. This is maybe what the German Constitutional Court (BVG) meant in the Southwest State Case, one of its very first cases, when it expressed the view that the Constitution reflects an ‘objective order of values’. The most important of these basic values are the principles of human dignity and democracy (others including the separation of powers and popular sovereignty). Article 79.3 protects these most important basic values by making articles 1 and 20 unamendable. Some of these ‘values’ describe an institutional pattern: this is the case of the separation of powers, for instance. Others are probably of a non-normative nature: this is obviously the case of human rights. Very plausibly, for reasons I cannot expatiate on in this article, this also appears to be the case of the concept of sovereignty. In both cases (human rights and sovereignty) the essentialist vocabulary is used in order to protect a certain constitutional substance against alterations. This is obviously what the CC did when it developed its case law on the ‘essential conditions of national sovereignty’.

In the case of human rights, the change in constitutional culture has created an awkward situation. Human rights were supposed to elude abrogation by their very nature, rather than by some normative mechanism such as an eternity clause. As this proved to be somewhat mythical, any lawmaker being in a position to curtail rights or even suppress them altogether, eternity clauses and judicial review were developed (amongst other things) to afford better guarantees. Yet, where a court such as the French CC refuses to review constitutional amendments, human rights are worse off than they were initially. The normativist legal culture is such that they are not understood as being exempt from abrogation, while the court acknowledges that a constitutional amendment can curtail or suppress them. This is an obvious drawback of the 1962-1992-2003 rulings. A (partly) satisfactory practical response seems to be that, while constitutional amendments cannot be judicially reviewed by national constitutional courts, other mechanisms ensure they are upheld. While such a response may be valid in practical terms, it goes some way towards undermining constitutional law. The constitution is shown as being unable to protect liberties.

C. Speaking the language of eternity

That such a vocabulary should be used in order to justify the non-justiciability of the amending power, does not detract from the fact that the CC has disregarded some items of constitutional substance (such as article 89(5) C’s republican clause) and created new ones. In so doing it has used legal categories and modes of reasoning that are strikingly close to those used by those ‘activist’ courts that accept to review amendments. This may very well be fully justified in terms of judicial policy (and politics). My purpose here is by no means to pass judgment on this choice. What matters more is that, whether or not courts declare that constitutional amendments are justiciable, their rulings contain standards that point to similar features. These features also appear in certain constitutional provisions. I would submit that it is to these classes of objects that most references to supraconstitutionality (rather misleadingly) point.

In such cases, the essence of the constitution, its substance, the nature of its basic structure, is more often than not left unarticulated. It is a feature of such jurisdictional standards that their content never needs to be fully developed or expounded. To use military language, readers are informed of the content of such concepts on a ‘need to know basis’ only. Yet I see no reason not to take such language seriously. It could be read as a manifestation 1/of certain specific legal categories; 2/of certain basic modes of judicial reasoning. These categories and these modes of reasoning are not specific to the judicial review of the amending power. That they should pop up like mushrooms in this context, however, is no accident. This is a moment at which there is a risk that the integrity of the constitution might be impaired. A court is supposed to say whether an apparently minor change is not in fact going to destroy the whole constitutional edifice. The broadness and open-endedness of the courts’ standards fit the generality of the question: if this amendment is upheld, are we still going to live under the same constitution? What the court accomplishes in such a situation has nothing to do with solving a conflict of laws in the ordinary sense of the word. The court cannot compare an inferior norm with a superior one. The French cases show that the same judicial methods are used when a court decides to deny review. In 1962, the ‘spirit of the constitution’ belonged to that category of standards. Unsurprisingly, the CC has recourse to very similar categories when it mandatorily reviews organic laws. For instance, when a constitutional amendment changed the President’s term of office from 7 to 5 years, the CC did not of course review the amendment itself but it did review the organic law and upheld it. It was not unconstitutional for the lawmaker (the législateur organique) to extend the term of office of ‘Deputies’ in order for the presidential election to take place before their own election as such a postponement was in keeping with ‘the position held by the election of the President (...) in the functioning of the institutions of the Fifth Republic’. These words are not eternity rulings properly so to speak but they make use of a very similar judicial language.

Why, then, is this language of eternity common to courts that review amendments and to courts that do not? I can only hope to point to a possible explanation. Our legal culture is very much based on a certain understanding of norms as manifestations of sovereign (or authorized) will. We understand the will of a legal authority as being free of all restraints except those created by other (superior) manifestations of will: it is a matter of ‘sollen’ (ought) and has no roots in ‘sein’ (is). Reality is not binding on the lawmaker or any authority that can create law. To say the contrary would involve an authoritative definition of what reality is, a dive into realism which most of us do not want to take because it would imply that there is an outer world that we can know objectively. Hence the widespread scepticism of contemporary lawyers. We refer to this kind of reasoning as natural law, supraconstitutionality or any other suitable word of abuse. Yet, when we make use of norms, we do not act without some given conception of an objective reality. The reality that is presupposed by our normative activity is a reality ‘within’ our legal reasoning: there is absolutely no need to presuppose something objective out there, in this ‘real world’ where we do not see any trace of law at all. But, as Jellinek says, law has to prove its own existence. It does so in many different ways. Notably, law has to thrust its roots into something which law itself calls reality. This legal reality is not the reality modern science has accustomed us to. It is not within the reach of our senses. It confuses facts and values. The mark of this legal reality is that it does not depend on volition: institutional patterns (the ‘spirit of the French constitution’) that only descriptive propositions can grasp, legal characteristics that are not created by acts of will (sovereignty, human rights, unwritten legal principles) and entities to which those characters are ascribed (the sovereign, the human being as a right-holder, the State maybe). When courts review constitutional amendments, they have a propensity to refer to such a pre-existing, pre-conditioning reality: the substance, the essence, or the identity of the constitution. Yet when the French CC decides not to review these amendments, it justifies its self-restraint with exactly the same kind of language, and refers to the same classes of objects. The case law of the CC describes a pre-existing reality made of institutional patterns, of extra-normative entities (national sovereignty for instance, and maybe beyond it, as Olivier Beaud has suggested, the State envisaged as Staatsgewalt), and of superior values. All of this is a matter of description. What is described, though, lies not at the surface of the legal landscape: it is essential because it is more important and less visible than the rest. It is the business of courts, among other things, to make that kind of reality visible.

CONCLUSION

In France, constitutional review has grown as an offshoot of the judicial review of administrative action. Its intellectual equipment is very much that of administrative law rather than the founding principles of constitutionalism or a British or American common law that takes a different attitude to the concept of State. At the end of the day, and despite many affirmations to the contrary, constitutional review in France approaches Parliament as a subordinate branch of the administrative State, the work of which should be submitted to in-depth scrutiny by the Constitutional Court. Lip service is still being paid to the principle that Parliament expresses the general will. Yet this statement is hardly compatible with the way in which the executive and the top courts, amongst which the CC, carefully and aptly manage the process of legal change. The rulings that this article has tried to scrutinize are a part of this larger process. For better or for worse (my purpose here is not to indulge in political romanticism) the idea of an original constituent power has, for all practical purposes, been altered in a very significant way. This cannot ultimately be without very significant political consequences. Today, the republic is not in peril, and neither is representative democracy. Constitutional lawmaking is subject to the same transformations as parliamentary lawmaking or the executive’s regulatory power. The executive and its administrative arm are the effective lawmakers, while Parliament, article 89 C’s Congress, or even, for that matter, ‘the people’ (i.e. the electorate) are merely called upon to give a stamp of legitimacy to the decisions taken at another level. There is nothing abnormal about this in a representative democracy. The same phenomenon has taken place, under different forms, in most liberal democracies. This, however, may create some tensions with the democratic element in the constitution. This democratic element is of paramount importance in French politics. It has a tendency to express itself in rare but destructive bouts of political upheaval, at times when the oligarchical (or technocratic) stratum has lost its hold over political change. The 1962 ruling has a different meaning in the present state of the Fifth Republic from in the heroic era at which it was made. In 1962 it expressed deference for the overarching legitimacy of De Gaulle and his understanding of the Constitution. Today, it justifies what has become a process of technical management of the Constitution. Legal expertise, political and administrative convenience, have been given precedence over political legitimacy. The name of the people, that involves democratic legitimacy, has been stamped on a wider, complex, and efficient process of technical adaptation of the Constitution. Adjunctive amendments are remarkable illustrations of what happens when a premium is awarded to convenience and legal expertise over those political values that should (also) matter in the life of the law. While the 1962-1992-2003 rulings seem to express a respect for the integrity of the Constitution, the process of constitutional mutation that they make possible has in fact too often transformed the constitutional enactment into a palinody. Certain principles or rules are expressed, yet directly contradicted by adjunctive amendments authorizing Parliament and the treaty-making power to derogate from more general constitutional principles. The Constitution has become a cobweb of contradictions. This is because it is minutely modified according to the needs of the day, under the constraint of other spheres (EU law, international law, administrative expediency, political arrangements, economic concerns, etc.). Yet the magic is not entirely lost. The metaphysics of the courts has superseded the metaphysics of constitutional lawmakers. Judicial metaphysics now rules the day. Nowhere is this more obvious than in France, where the legal limitations as set out by the written Constitution have, for all practical purposes, been set aside while the Constitutional Court has had ample recourse to a ‘language of eternity’.

Denis Baranger is Professor of public law at the University Panthéon-Assas (Paris II) and fellow of the Institut Universitaire de France. He notably published : Écrire la constitution non écrite : Une introduction au droit politique britannique, PUF, Léviathan, 2008 and Parlementarisme des origines, PUF, Léviathan, 1999.