A confused constitutional order. Inexpressibility and uncertainties of the French constitution

From an article published in ”Un ordre constitutionnel confus. Indicibilité et incertitudes de la Constitution française“, in Les Cinquante ans de la Constitution. 1958-2008, Études réunies par Dominique Chagnollaud, Litec-LexisNexis, 2008, p. 147-158


In France today, we are prompted to value the notion of constitution whether in political talk or in the learned discourse of jurists. The notion has in some sense been placed back at the very core of national political culture after, if not having been driven out, at least having played only a very unobtrusive role for many decades. "The Constitution" is nowadays invoked much as it is in most free countries governed by law, in what are conventionally called constitutional democracies. This reinstatement (for it does seem that is what it is, especially with respect to the beginnings of the Fifth Republic) appears to be incontrovertible, but it is extremely ambiguous even so.



In France today, we are prompted to value the notion of constitution whether in political talk or in the learned discourse of jurists. The notion has in some sense been placed back at the very core of national political culture after, if not having been driven out, at least having played only a very unobtrusive role for many decades. ‘The Constitution’ is nowadays invoked much as it is in most free countries governed by law, in what are conventionally called constitutional democracies. This reinstatement (for it does seem that is what it is, especially with respect to the beginnings of the Fifth Republic) appears to be incontrovertible, but it is extremely ambiguous even so.

It is seldom realized that there is a striking contrast between the contemporary exaltation of constitutional law—which has supposedly become ‘true law’ notably because of the development of ‘constitutional jurisdiction’ and correlatively the advent of a ‘normative’ constitution which purportedly provides its main foundation—and the haziness and uncertainties surrounding the very notion of constitution. Upon reflection, it appears that French legal thinking has curiously enough never developed a genuine general theory of the constitution[1]. Admittedly all of the textbooks on constitutional law devote some space to the concept of constitution, but those very general developments are somewhat exaggeratedly academic, abstract and very imprecise: they show scant concern for coming up with any general systematic tenets articulated on the stringent analysis of legal and political orders in general or even, more concretely, of either one of them in particular. It is only implicitly and highly approximately that a sort of French vulgate of the constitution has been outlined, which commentators apply to a greater or lesser extent to France’s contemporary political regime. But it is by no means sure that this is enough.

In the 1950s, André Mathiot[2] already evoked the relative character of the formal notion of constitution usually employed, before Georges Burdeau drove the point home in 1956 in a famous paper[3] posing the problem, that was not confined to the case of France, of how constitutional texts articulated with political and social reality. The advent of the Fifth Republic did not prompt any deeper inquiry into the subject.

At the time of the twenty-fifth anniversary of the 1958 Constitution, although Stéphane Rials had underscored some of the ‘uncertainties of the notion of constitution under the Fifth Republic’[4], he was barely heard. A short time later, Dominique Rousseau advanced to the contrary the idea of a ‘resurrection of the notion of constitution’[5], a formula that largely converged (albeit without prejudice to disagreements on other points) with the arguments of the doctrinal strand which, after a somewhat forced repudiation of political science, extolled a ‘renewal’ of constitutional law in France and was gradually becoming the mainstream[6]. This optimistic reversal was nonetheless tempered by Pierre Avril[7] who recalled that a constitution cannot be reduced to its legal nature alone and that it has a political dimension; it is first of all an instrument of government before being a guarantee of individual rights and freedoms.

Legal scholarship has not gone deeper into this issue since then. It shifts incessantly and above all without saying so from one concept of constitution to the other. Its dominant propensity to attempt to isolate (if not ‘purify’) law from politics has reduced its capacity to analyse its subject matter and it frequently and over simplistically contrasts a ‘political’ notion of the constitution and a ‘legal’ notion as if the world of politics and the world of law were alien to one another and formed separate watertight compartments. Finally, French scholarship seems to waver between a concept that seeks to be more or less normative (but when it is applied stringently, it has little or no explanatory value for the actual constitutional life of the country) and a fuzzy concept inherited from a tradition that is closer to political science as was, pretty much the equivalent of the concept of ‘political system’, but formulated in terms that are far too vague and whose legal dimension scholarship fails to explain clearly. In short, no satisfactory articulation has been found between an over-invested concept of normative constitution and the imprecise usage of the traditional concept of constitution as an institutional and political system.

Now to properly grasp their subject matter, jurists must work simultaneously with several notions of constitution, or more exactly, with several facets of the notion of constitution, which complete each other and are all essential to the analysis of it. While the subject matter of constitutional law is, in its broadest sense, to determine how a regular and legitimate power of domination is exercised within a political body, the constitution brings us back in the first place to the idea of power and government: it is from this angle that it must be analysed first (2). Its form, while not indifferent, is nonetheless secondary and even relative (3). These two dimensions might be articulated by resort to the notion of constitutional order (4).


Initially there was not much concern about a legal notion of constitution, in the sense that the legitimate worry of how the country was governed prevailed. The gist of constitutionalist literature of the early years of the Fifth Republic used the term ‘constitution’ to mean in actual fact ‘system of government’.

Its legal underpinnings were not overlooked but it essentially took on a functional dimension: it was to help to solve the fundamental political problem which was to restore the authority of the state by ‘cleaning up’ parliamentary government and establishing a stable government that was capable of acting. In other words, an instrumental conception of the constitution prevailed in 1958; this was a far cry from the constitution as a unifying symbol of the political body (in the US style). In this instrumental perspective, the formal constitution was overburdened with detailed provisions for straitjacketing parliament (especially title V); and it was kept imprecise so as to facilitate a ‘presidential’ practice (e.g. article 10 on promulgation, which was less precise than in the 1946 Constitution, or articles 29 and 30).

A constitution that was thus extended in a fairly limited way could not operate by itself: it had to be conveyed by something else, that went beyond it: this was that sort of ‘(practical) lesson in constitutional law’ propounded by General de Gaulle. The instrumental Constitution was transcended by de Gaulle’s dynamic vision, captured in his famous formula: ‘a spirit, institutions, a practice’ (press conference of 31 January 1964).

  • The spirit, the prime factor and without which the Fifth Republic would never have become what it is, was de Gaulle’s determination to break with the system of ‘government by parliamentary delegation’ and correlatively to give an autonomous foundation to the organ embodying state power. That determination was a political principle that was to impart a direction and give concrete meaning to the instrument of government that is the written constitution.
  • Institutions, they obviously included innovations (whether far-reaching technical arrangements such as dispensing with any ministerial countersignature for certain competences of the President of the Republic, or new creations like the Conseil constitutionnel), but most were part of a certain continuing historical line that must not be forgotten: especially the legal duality of the executive, a bicameral parliament, and the maintenance of the principle of the government’s answerability to parliament. In themselves, they were not highly original; only, the way they were set in motion explains that a novel political system developed after 1958.
  • A practice: it was precisely through this that, being put in a position where he could give the first main orientations, de Gaulle, with the backing of the electoral body, was able to move the institutions in a direction that was consistent with the spirit he had defined.

It was therefore a genuine conventional constitution that emerged, and that could barely be discerned from a straightforward reading of the constitutional parchment of 1958-1962. Nothing of what came about was inevitable. The emergence and above all the continuation, by and large, of this system dominated by the office of President of the Republic comes across as the particular materialization, that was not inevitable but conveyed a degree of coherence, of one of the virtualities provided by a largely indeterminate and highly flexible legal framework[8].

We shall not dwell here on the old question of how to characterize this complex system of government, and shall confine ourselves to reiterating that it seems fairer to characterize it as a ‘parliamentary system hijacked by the office of president’[9]. But it is important above all to emphasize that this conventional constitution remains precarious: its apparent entrenchment after 1969 has led to underestimation of the point that this formula which has no real equivalent elsewhere rests on a fragile construction (much more so than the UK, German or US systems): to operate in practice, it requires a subtle harmony between the President and a parliamentary majority. More than elsewhere, it is dependent upon a convergence in the behaviour of its actors. Now, while already precarious in itself, this construction was gradually subverted by the practice of presidential government without ethics (the concentration of power, the phenomenon of the ‘Sycophants’, the tendency towards political unanswerability) and phases of cohabitation which de-natured the founder’s spirit. The conserving of the written statements of the constitutional text in theory makes it possible to return to the other virtualities provided by constitutional law. Above all, French democracy struggles to live serenely with such a muddled and hap-hazard mechanism of government, the fate of which is at issue with each presidential or parliamentary election.

In any event, in speaking of ‘the Constitution of the Fifth Republic’, one means above all in fact a system of government the constitutional text of which forms only an approximate framework. It seems to be therefore a ‘descriptive’ notion of the constitution since it refers to the system of government as it actually operates, but, at the same time, it is not without all prescriptive scope since this system having managed to endure pretty much with some consistency for 50 years, the political actors like the citizens have come to consider it was indeed a set of constraining arrangements (however complex and precarious), in other words it is what ‘must be’.


As no one can be unaware, the development of in-depth control of the constitutionality of legal norms by the Conseil constitutionnel from the 1970s onwards has appeared to reinforce the importance of the constitution which is considered to be the supreme written law. The prejudice in favour of judges (or at least of an institutional equivalent of judges, as is the case of the Conseil constitutionnel) has reinforced the prejudice in favour of written law and led to this received wisdom that the formal constitution of France was thenceforth reinstated in its authority perhaps for the first time in the history of France since 1789, meaning that it would not be mistaken to speak of the ‘resurrection of the constitution’.

This renewal seems to have been accompanied by an at least psychological and symbolic rehabilitation of the constitution in ordinary political discourse, as revealed inter alia by F. Mitterrand’s 1986 sound bite: ‘the Constitution, nothing but the Constitution, the whole Constitution’. But this attitude is probably more apparent than real and whatever the case, the ‘resurrection’ in question here has proved in reality highly precarious and is largely trompe l’œil: the formal Constitution of the Fifth Republic is now more relative than ever.

3.1. Relativity of the formal and rigid character of the constitution

Strictly, a constitution may be characterized as formal if it is written down and contained in a single document. Now, the Constitution of the Fifth Republic meets this definition less and less.

First of all, the document promulgated on 4 October 1958 and initially comprising 92 articles (they are slightly more numerous today, although numbered from 1 to 89 since the revision of 4 August 1995, intervening articles have been added over the course of the many formal revisions, especially that of 23 July 2008) has been enhanced by an increasingly complex set of written norms whose constitutional value has been expressly enshrined. First, the Preamble, initially considered implicitly as outside the formal Constitution must, to my mind, henceforth be regarded as an integral part of it. It is this solution that was enshrined by the Conseil constitutionnel’s decision of 16 July 1971. It is also what was enshrined by the constitutional lawmaker in article 1 of constitutional law no. 2005-205 of 1 March 2005, which for the first time expressly altered the Preamble by adding the reference to the ‘Environment Charter’. Accordingly, like ‘Russian dolls’, the formal Constitution encompasses in addition to the hundred or so numbered articles, a series of texts that are both separate from it but not alien to it: the 1958 Preamble, the 1789 Declaration of Human Rights, the 1946 Preamble, the 2004 Environment Charter[10]. There are even added to it the Fundamental principles recognized by the laws of the Republic, but which are not enumerated. It is to encompass this complex set that we use the notion of ‘bloc de constitutionnalité’,  what we might call a ‘constitutional package’, which comes across as a substantial expansion of the formal constitution.

Moreover, the coherence of the formal character of the Constitution has come in for some rough treatment since 1992 by the practice of inserting references to important European (from Maastricht to Lisbon) or international treaties (of Rome on the ICC) within the articles of this same Constitution. While the reasons for being (which are probably illusory but attest to a real concern) of a written constitution are clarity and rationality, the constitutional lawmaker is clouding the issue by jumbling the meaning of the written statement, which can no longer be understood without reference to other texts.

Lastly, the relative character of the formal Constitution was aggravated by the adoption of written provisions of formally constitutional value but without purpose or without effect: these were certain parts of constitutional laws no. 2005-204 of 1 March 2005 and no. 2008-103 of 4 February 2008. They contained provisions to modify certain articles of the Constitution (articles 88-1 to 88-5) or to add new ones (articles. 88-6 and 88-7). Such provisions have been legally validated (the constitutional law having been promulgated and published) but their effectiveness was conditional upon the entry into force of an international treaty. In the first instance, it is known that the Treaty establishing a Constitution for Europe signed on 29 October 2004 never came into force. In the other instance (the constitutional law of February 2008 having abrogated a part of that of 2005), its provisions were conditional upon the entry into force of the Lisbon Treaty of 2007. After some trial and tribulation—especially the ‘no vote’ by the Irish people in June 2008—the treaty did finally come into force (Autumn 2009) but the French method, already practised in 1992 for the Treaty of Maastricht, consisting in having the Congress adopt a revision of the formal Constitution prior to a referendum on the bill authorizing the ratification of the Treaty for France was a chance matter: it would have carried more weight to invite the people to give their opinion on the revision of the Constitution so as to avoid the fundamental law being revised pointlessly in the event of a ‘no vote’.

In any case, there survive for a time provisions of a constitutional law that are or are not integrated in the constitutional document but that cannot apply. This slapdash policy does not exactly strengthen the formal Constitution.

One can add the case of articles of the Constitution that have been altered or added by a constitutional revision the concrete application of which is conditional upon the development and entry into force of organic laws materializing them. The revision of 23 July 2008, in particular, has multiplied these hypotheses. Now, more than three years after the constitutional revision, several articles or parts of articles are still ‘paralysed’ by the absence of an organic law. This is the case, for example, of the minority initiative referendum (subparagraphs 3 to 6 of article 11 C). Likewise, the new procedure for removal of the President of the Republic (art. 68), the principle of which was posited by the constitutional law of 23 February 2007 is still not applicable.

The formal Constitution of 1958 satisfies in appearance the doctrinal requirements of the character of rigidity (which it must be recalled is conceptually separate from the formal character: there are ‘flexible’ formal constitutions, even if they have become rare since the mid-twentieth century): the constitutional text can only be explicitly altered by the adoption of a specific constitutional law adopted by a different procedure from the ordinary legislative procedure (here we leave aside the particular and problematic question of the revision of October 1962 and the 1969 referendum). Now, it must be noted that this ‘rigidity’ is in actual fact weak or even illusory. As part of a tradition that dates back at least to 1875 it is underpinned by the primacy of the essentially representative character (admittedly with some lapses) dominating French constitutional government since 1789. The referendum ratifying a revision may be avoided: the ordinary representative organs may maintain full control of the process and the requisite parliamentary majorities are ultimately not very demanding: a simple majority in each chamber, a majority of three-fifths of the votes cast (and not of the statutory number), which is few and has more than once allowed a government majority of a single stripe to effect a revision alone (the most critical case being that of November 1993 over the right of asylum, as it was disputed and the opposition did not have enough members to block the revision). In the case in point, it is known that revisions have been more numerous since 1992 (19 in 16 years) and all (save two) have been passed by means of the Congress, so that the formal Constitution no longer truly enjoys any real stability and, although deemed to be the act of the sovereign people, is at the disposal of its representatives.

Further still, constitutional lawmaking, even when done by the representatives (and not by the sovereign people itself) is not controlled by the courts (by virtue of the Conseil constitutionnel’s decision no. 2003-469 DC of 26 March 2003), which thereby practice self-restraint that one may find wise from a certain standpoint but which is not obviously so at all with respect to the reason for being of formal and rigid constitutions that are supposed to protect the highest will of a sovereign people. One rather oddly arrives at a system which, as it is practised, means a sort of de facto ‘congressional sovereignty’: the people is ousted by its representatives who may, without judicial control, instrumentalize the formal constitution.

3.2. The ‘living’ Constitution and the relative character of written law

There is another order of considerations that accentuates the relative character of the formal Constitution (here more specifically in its character as written constitution): this is the growing awareness that the constitutional text is, by itself, little from which to determine the law that is effectively and concretely in force. The new effectiveness afforded to the 1958 Constitution by the control exercised (mainly) by the Conseil constitutionnel implied less a restoration of the text than a change: the transition to a largely ‘jurisprudential’ constitution, a phenomenon that was besides perceived and admitted by the doctrinal strand that celebrated the renewal of the French Constitution. That Constitution is therefore seen in a dynamic and no longer merely static light, as a living instrument (obviously one is put in mind of the theme of the ‘living constitution’ that has long been fashionable in the United States) that forces the constitution as a written instrument into the background. The success of realist theories of interpretation has reinforced this empirical observation that it was henceforth essential to refer to constitutional case law to identify the constitutional law in force. In this respect, France has thus followed along the path taken by a number of neighbouring constitutional democracies where there is substantial court control over the constitution.

However, awareness of what is in truth an inevitable development has entailed a shift and above all a substantial reduction in the view of the constitution: because the control exercised by the constitutional court largely concerns citizens’ rights and freedoms, some commentators have concluded that these matters were the most important part of constitutional law, overlooking what was the primary object of modern constitutions: the shaping of political power. Moreover, the development of a constitutional court system has tended to push into the shade whatever is not the subject of court control, as if law came down to what is sanctioned by the courts and as if the judges held unbounded jurisdiction, allowing them to intervene in all matters within the field of the constitution. Against such simplifications, the obvious point should be recalled that a large part, perhaps even the largest part of constitutional matters wholly escapes the intervention of the courts. The relation among the main institutions of government are essentially conducted through other channels; often it does not involve the application of a predetermined norm, by ensuring compliance of an act with a higher rule; it is much more a matter of a set of dynamic and complex relations, that cannot be reduced to legal norms alone.

In any event, and even if we do not wish to subscribe to a radical conception of interpretation as an act of creation, the constitutional text, the formal Constitution is still only and cannot be anything but a support for the deployment of the concrete constitutional life, in short the actual Constitution.


All things considered, there being a formal constitution does not therefore radically change the nature of the power of legitimate domination. Whether the rules governing the exercise of political power and guaranteeing citizens’ rights are written or not, whether some of those rules, among the more important ones, are endowed with greater force or not, any analysis of them is invariably complex, and marked by fluidity, by an evolving character. Without conforming in the slightest to legal stringency, there should be an awareness that it would be pointless to let oneself be mesmerized by the written constitution, in other words by what acts merely as a prop for government power.

A constitution cannot therefore be reduced to a written law, however ‘supreme’[11]. For the written word can never be self-sufficient. The provisions it contains can almost never hold without resort to notions and concepts underpinned by it, nor without a source of energy from forces outside the formal legal norms.

Just as the notion of legal order (even if it is sometimes contested) is most useful for reflecting the point that law cannot be reduced to a mere compilation of disparate norms, but is ‘an organized set of interdependent elements forming a single unit’[12], the notion of constitutional order probably captures better than the term constitution (which is misleading because of its multiple meanings which as seen reveal different aspects of an irreducibly complex phenomenon), the fact that constitutional law, the study of the government of people by means of law within a given political body, presupposes a whole that is more than the straightforward addition of the objects composing it. In particular, this notion goes beyond the sole aspect of strictly legal norms the study of which may be completely separated from political components themselves[13]. Although little used in France, the idea of constitutional order is actually implicit in many commentators. It is commoner abroad, especially in Germany where the term (verfassungsmässige Ordnung) is evoked in article 20(3) of the Basic Law and regularly used not only in scholarship but also by the constitutional court.

In terms of its content, the constitutional order rests upon institutions, those elementary bodies that precede all of the ordinary interplay of production of law (even if, outside of a custom-based constitution, they are ‘posited’ in an elementary way by a constituent act, but then develop by themselves). These institutions are possessed of competences, subjected to procedures and to structuring principles (for example, state sovereignty, the principle of compliance with law as it is posited, etc.) or material principles that limit in part at least the content of their acts, but at the same time the institutions themselves shape those rules.

In terms of form, the constitutional order is structured in part by legal instruments, especially written statements (inscribed at the highest level of the formal Constitution but also in other instruments such as organic laws, assembly rules and regulations, etc.) that must be materialized by the actors of constitutional life. The constitutional order is also structured by unwritten rules, that are also produced by the action of institutions, and that may have a different standing and variable normative value: interpretations that are more or less binding depending on whichever institution issued them, conventions or more simply political practices. As in the other branches of law, case law here has a growing although not exclusive part to play in materializing the various statements and formulating rules.

All of that interacts in an eminently complex manner that is anything but mechanical. Above all, when taken together these elements form a system: the constitutional order is the resulting force of the relatively coordinated interplay between institutions and the rules that they contribute to formulating and to which at the same time they are subjected. It is the relatively stable product of the interaction between what must be (or what may be), that is, the constricting arrangements of the normative constitution, and what actually is, the real or effective constitution.

This system, which is necessarily impressionistic, and especially confused in the French case, wavers between a certain static component and an inevitable dynamic component. The static component is supplied notably by the permanence of the main constitutional organs and the body of the comparatively stable main legal rules governing relations among them. From then onwards, a constitutional order is erected, in its general organization, for the most part in the early years and is often marked by a degree of constancy. Such is the case for the Fifth Republic, in that the system of government remains similar to what was constructed in its first decade. But the constitutional order is also wrought by an incessant dynamic component: a large proportion of the norms in force may change (either by formal channels or by the channels of interpretation), vanish, or new norms appear. Sometimes the organs themselves may undergo changes in status (e.g. the head of state’s five-year term of office) but more often than not such changes pertain to some relatively minor aspect of the institution (e.g. the president’s refusal to sign ordinances). Such modifications may be variable in scope and it is rare that they completely call into question the overall identity of the constitutional order itself (in the case in point, the growing role of the Conseil constitutionnel has managed to graft itself to De Gaulle’s constitution in what in the end is a fairly harmonious way). In this sense, despite all of the changes that have affected it over fifty years, the Fifth Republic remains. But that permanence is a poor mask for the confused character that, from the outset, it has had in its formal and material aspects alike. Perhaps impressionism in constitutional matters does not have the beauty it has in painting.

[1] This absence goes way back; thinking about  it, it is even one of the most gaping holes in the fine treatises of the great  scholars of old like Esmein, Duguit, Hauriou and even Carré de Malberg.

[2] ‘Agonie de quelques vieux principes’ (1950), reprinted in Pages de doctrine (LGDJ, 1980), t. 2, p. 81.

[3] ‘Une survivance : la notion de constitution ’, Mélanges A. Mestre (LGDJ, 1956), p. 53.

[4] RDP, 1984, p. 587.

[5] RDP, 1990, p. 5.

[6] Particularly Louis Favoreu, ‘Le droit constitutionnel, droit de la constitution et constitution du droit’, RFDC, no 1, 1990, p. 71.

[7] ‘La Constitution, Lazare ou Janus ?’, RDP, 1990, p. 949.

[8] Cf. the present  author’s developments:  ‘Le Prince inapprivoisé. De l’indétermination structurelle de la Présidence de la Ve  République’, Droits, no  44, 2007, p. 101.

[9] ‘Un régime parlementaire à captation présidentielle’, ‘Le Prince inapprivoisé’, p. 120 ff.

[10] One might add the Noumea Agreement of 5 May 1998 included in article 76 of the Constitution by constitutional law no. 98-610 of 20 July 1998.

[11] Strictly it would be more accurate  to call it a collection of norms rather  than  a norm as such.

[12] Charles Leben, Droits, no33, 2001, p 20.

[13] Which is why we shall not run with the definition of constitutional order suggested  by J.-M. Blanquer (Mélanges J. Robert, 1998, p. 235) as ‘everything that  must be complied with by virtue of the Constitution’, because  it appears overly normativist.