Separation of powers

Michel Troper


[Translator’s note: In this article, the terms pouvoir and puissance, both of which are frequently used in conjunction with the “branches” of government and their separation, are equally rendered power, there seeming to be no available equivalent for contrasting these two French terms in a political context.]

1The separation of powers, one of the fundamental principles of modern constitutionalism and of government by law, is indissociable from the name of Montesquieu. Those who write on the separatdion of powers never fail to cite the name of the Baron de la Brède or chapter six of book XI of L’Esprit des lois, “On the English constitution”, and contrariwise, when the work of Montesquieu is mentioned, it is the separation of powers that first comes to mind. Truth to tell, this association is not without its problems. The misunderstanding is not, as is sometimes said, simply a matter of Montesquieu’s failure to use that expression. It is true that he does not use it – although he writes that “there is yet no freedom if the power to judge is not separated from the legislative and executive powers” – but he could very well be the inventor of the doctrine that was later designated in this way, labelling it differently or not at all. Thus the true relationship between Montesquieu and the separation of powers can be discovered only by seeking first what is generally understood by separation of powers, then by examining whether this is indeed the doctrine exposed by Montesquieu in L’Esprit des lois. The interest of the question is not limited to an historical point in political philosophy and it is not simply a matter of whether he is indeed the inventor of this theory or whether the merit – or in the opposite case, the error – should be imputed to another.

2The discussion over Montesquieu’s paternity in reality goes to essential theoretical and practical questions. What are the functions of the state? How can they be divided up in order to guarantee freedom while preserving the efficiency of power and avoiding risks of paralysis? To these questions Montesquieu has an answer, the pertinence and specificity of which appear more clearly when we confront them it other doctrines which, under the name of separation of powers, purport to be variation on the theme developed in chapter 6 of book XI of L’Esprit des lois and which in addition present very serious flaws.

The classical conception of the separation of powers

3We will call “classical” here the conception of the separation of powers which we have been encountering, notably in works on public law, since the mid-nineteenth century. According to that conception, which moreover contains several variants, the separation of powers is a technique of constitutional engineering – the preferred term for which is a maxim of the political art – intended to guarantee freedom. Its discovery is attributed to Montesquieu and it is generally set forth with the support of quotations drawn from L’Esprit des lois, and it consists in “separation”, in other words a certain mode of distribution or repartition of the functions of the state among various authorities.

The powers

4The state excercises a great variety of functions of a social-political order – making war, rendering justice, maintaining order and security, etc. – but it does it by means of law, producing general rules and particular commands, in other words exercising different juridical functions.

5In juridical language as much as in ordinary language, the word power has a plethora of meanings. Sometimes it designates one of these juridical functions, sometimes the power necessary to exercise it, and sometimes again the authority or organism which is invested with it. Legislative power is thus either the legislative function, or else the power invested in an authority to allow it to make laws, or else that authority itself, parliament for example in modern democracies. In that case the expression “separation of powers” designates a simple distribution of the functions as well as a separation of the organisms, and one speaks of functional separation or organic separation.

6When we distinguish two functions, what is involved is the legislative function consisting in making laws, which is to say general, impersonal rules, and the executive function by which those laws are applied to concrete cases, either by material acts (the building of roads, the employment of force to assure public order), or by particular decisions.

7When it is the third function that is at issue, it concerns the judiciary or jurisdictional function. If we refuse to treat it as a third function, that is because we judge that it consists in deciding the disputes by application of the law, which makes of it merely a branch of the executive function. If we consider on the contrary that disputes cannot or should not be decided exclusively by the application of laws, but that judges disposes in fact, or should, of broad evaluative power, we conclude that we are indeed in the presence of a third function.


8On this theoretical basis, the principle of engineering declares itself in classical doctrine as the combination of two distinct rules, the rule of specialization and the rule of independence. The former is simple. There must be as many authorities or organisms of state as there are functions, and each authority must be specialized in one of these functions, in other words exercise one function and exercise it solely, remaining completely outside the exercise of the other functions. As for independence, it means that each authority must be sheltered from all influence by the others, else there would be no specialization. Independence stems above all from the absence of power of revocation of one power over another, but secondarily an authority is also independent if it does not owe its nomination to another, if its budget does not come from another and also if judicial pursuits cannot be exerted against it by one of the others. From this we see that total independence is not possible and that there is only a more or less high degree of independence. Each independent authority is thus a genuine “power”.

9From the operation of these two rules a considerable benefit is expected. Each of the powers will have the means of opposing the other (or the others), were it to attempt to become despotic, in this way, thanks to this balance, freedom will be preserved. Such is the thesis traditionally attributed to Montesquieu.

10In truth, it is not always invoked in perfectly rigorous manner and there are authors who will also speak of separation of powers to designate singly either the specialization of authorities that are not independent or the mutual independence of non-specialized authorities, or yet again any balance obtained without specialization or independence. Despite all expectation, they judge that the application of just one of the two rules that make up the principle of the separation of powers does not constitute a rejection of this principle but an attenuated form or a broad construal, and in any event makes it possible to obtain the desired balance. This attitude has no more than rhetorical interest, since it allows one to pretend that, with exception of absolute despotism and direct democracy, most regimes in one way or another satisfy the principle of the separation of powers.

11But the principle is also used to realize a general classification of constitutions. Jurists thus distinguish systems of conflated powers, systems of absolute separation of powers, and finally those that apply it in attenuated fasion and realize what is called a flexible separation of powers or a collaboration of powers. Although there may be divergences on the categories, in other words on the affectation of such a such a constitution to one of them or another, today this classification is largely dominant.

12Systems of conflated powers are those in which a single authority exercises all the functions or controls their exercise. Sometimes conflation realized to the benefit of the executive – such as military dictatorships – and conflation realized to the benefit of the legislative power, also called an “regime of assembly” of which few examples exist aside from the National Convention of 1792-1795, are contrasted. These terms are obviously lacking in pertinence, since in cases of conflation there is no longer an executive nor a legislative power.

13Systems of absolute or rigid separation of powers are those in which the authorities are at once specialized and independent, and in which the specialization and independence are applied in excessively rigorous manner without any exception. French authors put the French constitutions of 1791 and Year III into this category, of which they sometimes say that it goes as far as to realize an isolation of powers. They judge in general that in the two cases the constituents have been influenced by Montesquieu, but guided by an erroneous and dogmatic interpretation of L’Esprit des lois. Certain authors add to this category the constitution of 1848 and the American constitution, but others reject these classifications. Most American authors on the other hand consider that the separation of powers organized by their constitution is not rigid.

14All other regimes would belong then to the category of regimes of flexible separation of powers or of collaboration. In reality this is a heterogeneous category, since it includes all representative systems as well as English-style parliamentary regimes which realize a specialization of mutually dependent authorities – the House of Commons can dismiss the cabinet, whereas the Prime Minister can set in motion the dissolution of the House of Commons – which the presidential regime, in which authorities are independent, since the president is not revocable by Congress and he cannot dissolve it, but not specialized.

15If Montesquieu thus appears as one of the fathers of modern constitutionalism, that is because he would have been one of the first to formulate this idea that power should be organized in such a way as to preserve freedom, that this organization should be expressed in a rule which is the constitution, and that it should institute specialized and independent powers.

16He would thus have established between constitution and separation of powers an essential connection, to the point that respect for the principle of the separation of powers was to become the touchstone of a constitution worthy of the name. That at least is how the classical doctrine interprets the formula of article 16 of the Declaration of the Rights of Man and the Citizen of 1789: “Any society in which the guarantee of rights is not assured nor the separation of powers determines, is without a constitution” (“Toute société dans laquelle la garantie des droits n’est pas assurée ni la séparation des pouvoirs déterminée, n’a point de constitution”).

17Montesquieu’s paternity is no doubt sometimes contested, as happens for all theories, and the merit of having discovered the separation of powers has been attributed to others such as Locke or Bolingbroke. To be sure, these discussions are pointless if it turns out that the separation of powers is a bad doctrine. But it is a fact that, thus presented, it has inspired numerous objections.

Objections to this doctrine

18Certain of these objections, sometimes connected with the name of Jean-Jacques Rousseau, are based on the theory of sovereignty one and indivisible. Indivisibility does not mean that it is ill-advised or forbidden to divide sovereignty, but literally that it is impossible to do so, and that any attempt would inevitably lead either to a destruction, which would be that of the state itself, in other words to paralysis of power or anarchy, or to its reconstitution by bringing all the powers together into the hands of a single person. Other authors position themselves on a practical plane and assure that if the powers can check each other, the result will be situations of institutional blockage which can only be resolved by repeated coups d’état. They invoke for example constitutions based on the separation of powers or at the least on an excessively rigid interpretation of this principle. Thus, because of the separation of powers, the regime that resulted from the French constitution of Year III led to the 18 Brumaire and the dictatorship of Bonaparte, the Second Republic to 2 December.

19However, these objections collapse if confronted with a second, decisive one directed at Montesquieu by the greatest French jurist of the twentieth century, Raymond Carré de Malberg: specialized authorities could balance themselves only on condition that their powers are equivalent. But these functions are on the contrary hierarchized, for the execution of the laws is quite evidently subordinated to the laws themselves. If the authorities are specialized, the executive authority will be strictly subordinated to the legislative function. The hierarchy of organisms follows the hierarchy of functions and never will a subordinate power be able to check a superior authority. Therefore the separation of powers does not compromise the unity of sovereignty, for it always remains a sovereign power and it is the legislative power. A constitution based on the separation of powers therefore entails no risk of blockage or coup d’état, and if the executive power has recourse to force, it is only as a subordinate rebelling against a superior power.

20The objection formulated by Carré de Malberg is irrefutable, but perhaps it should be granted that Montesquieu was terribly wrong, as were the authors of numerous constitutions after him who appealed to him? It is the merit of Charles Eisenmann to have shown that the thesis contained in chapter 6 of book XI of L’Esprit des lois is profoundly different from the separation of powers as it was classically conceived, and thus avoided objection.

Montesquieu conception

21The demonstration is rather simple: in the first place, Montesquieu is perfectly conscious of the hierarchy of functions. No doubt his classification of the state’s juridical functions – in particular his definition of the executive function – is neither precise nor coherent. He distinguishes at the beginning of the chapter between “the legislative power, the executive power over things that depend on international law, and the executive power over what depends on civil law” (“la puissance législative, la puissance exécutrice des choses qui dépendent du droit des gens, et la puissance exécutrice des choses qui dépendent du droit civil”). The third is also called “power to judge” and the second “executive power of the state”. One could thus think that the executive power bears exclusively on international relations. Yet a few lines later he writes that if “the legislative power is joined with the executive power, there is no liberty, because it can be feared that the same monarch or the same senate might pass tyrannical laws in order to execute them tyranically” (“la puissance législative est réunie à la puissance exécutrice, il n’y a point de liberté, parce qu’on peut craindre que le même monarque ou le même sénat de fasse des lois tyranniques pour les exécuter tyranniquement”). The problem then is that the executive power does not consiste solely in the execution of things that are dependent on international law, but also in the execution of concrete internal laws. In all the rest of the chapter it is in this sense that the expression “executive power” is utilized, and he even specifies that the legislative power is but “the general will of the state, and the other but the execution of that general will” (“la volonté générale de l’État et l’autre que l’exécution de cette volonté générale”). It would thus have been perfectly contradictory to try to organize a balance between an organism responsible for articulating the will of the state and another responsible for its execution.

22Such a balance is conceivable only among non-specialized and non-independent authorities, and precisely in the English constitution such as Montesquieu describes it, the authorities are neither specialized nor independent.

23They are not specialized, for the legislative power is entrusted not to one, but to three distinct authorities, an assembly composed of representatives of the people, an assembly of nobility and finally the king, who disposes of a faculty of prevention, which is to say a right of absolute veto. A law can be adopted only after obtaining the consent of these three authorities, and a single one can oppose it. This structure is indeed that of the English constitution such as Montesquieu’s predecessors and successors describe it: the legislative power belongs to Parliament, composed of the House of Commons, the House of Lords, and the king. The latter is therefore part of Parliament, as is shown by the expression “the king in Parliament”. Now each of these three authorities, far from being specialized, also exercises another function. The king exercises the executive function, the House of Lords a part of the judiciary function, and the House of Commons can exercise accusation in public affaires and control the manner in which the laws are executed.

24What then for intependence? In addition to this control of the House over the execution of the laws, which can lead as far as formal charges against ministers, we must underscore the king’s power to convoke or dissolve the Houses.

25Thus, far from prescribing specialization and independence, Montesquieu praises a system that precisely is based on a contrary principle: “Such is the fundamental constitution of the government in question. The legislative body there being composed of two parties, they will enchain each other by their mutual faculty of prevention. Both will be bound by the executive power, which itself will be bound by the legislative.” (“Voici donc la constitution fondamentale du gouvernement dont nous parlons. Le corps législatif y étant composé de deux parties, l’une enchaînera l’autre par sa faculté mutuelle d’empêcher. Toutes les deux seront liées par la puissance exécutrice, qui le sera elle-même par la législative.”) We must underscore that if the executive power can thus bind the two parts of the legislative body, in other words the two Houses, it is not really as executive power as such – insofar as it is executive power, it is subordinate – but insofar as it is itself, by right of veto, a part of the legislative power.

26Eisenmann observes moreover that none of those who, in the second half of the eighteenth century and the first half of the nineteenth, explicitly claimed Montesquieu, thought that L’Esprit des lois specified the specialization and independence of powers, neither Blackstone, nor de Lolme, nor the authors of the Federalist, nor the group of monarchists in the Constituante Assembly of 1789, nor Benjamin Constant. It can again be pointed out that the doctrine described under the name of separation of powers by twentieth-century French jurists presupposes a very different conception of freedom from that exposed by Montesquieu: for the author of L’Esprit des lois, the freedom that is to be preserved by the separation of powers is political freedom. Book XI is, moreover, entitled “On political freedom in its relation to the constitution” (“De la liberté politique dans son rapport avec la constitution”). Now political freedom is quite different from civil freedom. It is not independence, nor the enjoyment of one’s rights, but, he says, a situation in which one obeys only the laws. The relation of power to civil freedom can be conceived as a zero-sum game in which freedom is all the greater that power is more limited, and power all the stronger that freedom is restrained, but political freedom, thus defined as obedience to the laws, cannot vary in function of the extension of the sphere or intensity of power.

27On can then, after Eisenmann, remark that Montesquieu calls in reality for the application of two different principles. The first can be called “separation of powers” (“séparation des pouvoirs”), although the author of L’Esprit des lois does not use that expression. He prescribes neither specialization nor independence. In truth, he prescribes nothing, for the principle in question is purely negative, in other words it is a principle whose sole object is to indicate what one must not do. What should be avoided is very simple the conflation of powers or the gathering of powers in the hands of one man alone.

28The verb separate which he sometimes uses does not at all mean isolate. He merely uses it as the antonym of conflate or combine. When he writes: “There is also no freedom if the power to judge is not separated from the legislative and executive power” (“Il n’y a point encore de liberté si la puissance de juger n’est pas séparée de la puissance législative et de l’exécutrice”), it is to oppose this situation to one where this power “is joined to the legislative power” (“est jointe à la puissance législative”). Sometimes separate even has simply the meaning of distinguish, as in the title of a chapter in Book XXIX, “That laws must not be separated from the object for which they are made” (“Qu’il ne faut point séparer les lois de l’objet pour lequel elles sont faites”). Thus, for powers to be separated, it is enough that they not be joined together.

29Nevertheless, if Montesquieu is incontestably a partisan of the separation of powers, thus understood in a very different way from the twentieth-century juridical doctrine, he is neither its inventor nor its sole defender. It is even a commonplace of political philosophy in the Enlightenment, and is even expressed in similar fashion by numerous authors whose sole point in common is hostility to despotism. We find it for example in Merchamont Nedham, a seventeenth-century English writer, who wrote: “A fifth Errour in Policy hath been this, viz. a permitting of the Legislative and Executive Powers of a State, to rest in one and the same hands and persons” (The Excellency of a Free State, 1656), or again in Locke: “And because it may be too great a temptation to human frailty, apt to grasp at power, for the same persons, who have the power of making laws, to have also in their hands the power to execute them, whereby they may exempt themselves from obedience to the laws they make, and suit the law, both in its making, and execution, to their own private advantage, and thereby come to have a distinct interest from the rest of the community, contrary to the end of society and government: therefore in well-ordered commonwealths, where the good of the whole is so considered, as it ought, the legislative power is put into the hands of divers persons, who duly assembled, have by themselves, or jointly with others, a power to make laws, which when they have done, being separated again, they are themselves subject to the laws they have made; which is a new and near tie upon them, to take care, that they make them for the public good.” (Essay on Civil Government, 1690, chapter XII, “Of the legislative, executive, and federative power of the commonwealth” [‣] ).

30Even a philosopher like Rousseau, generally represented as an adversary of the separation of powers – which would be accurate if the separation of powers aimed at guaranteeing freedom through balance of powers – recalls the negative principle: “He who makes the law knows better than anyone how it ought to be executed and interpreted. It therefore seems that in the best constitution one could have the executive power would be joined to the legislative; but that is the very thing that makes this government insufficient in certain ways, since the things that must be distinguished are not, and the prince and the sovereign being but one person form, so to speak, but one government without a government. ¶ It is not good that he who makes the laws execute them, nor that the body of the people turn its attention away from the general perspectives to give them over to individual objects.” (“Celui qui fait la loi sait mieux que personne comment elle doit être exécutée et interprétée. Il semble donc qu’on ne saurait avoir une meilleure constitution que celle où le pouvoir exécutif est joint au législatif : mais c’est cela même qui rend ce gouvernement insuffisant à certains égards, parce que les choses qui doivent être distinguées ne le sont pas, et que le prince et le souverain n’étant que la même personne ne forment, pour ainsi dire, qu’un gouvernement sans gouvernement. ¶Il n’est pas bon que celui qui fait les lois les exécute, ni que le corps du peuple détourne son attention des vues générales pour les donner aux objets particuliers”; Jean-Jacques Rousseau Du contrat social ou principes du droit politique, 1762, livre III, chap. iv.)

31This negative principle is thus a genuine commonplace in the Enlightenment. Its justification is simple and resides in a very similar conception of political freedom. If political freedom is submission to the laws – and Rousseau gives of it the same definition as Montesquieu – the negative principle is indeed the guarantee of freedom, for if he who makes the law cannot execute it and if he who executes it cannot remake it, then the subjects, by obeying the executive authority, will remain indirectly subordinate exclusively to the law. This principle cannot be conflated with specialization. Doubtless he is satisfied if the authorities are specialized, but he is also quite satisfied if an authority exercises an entire function and in addition participates in the exercise of another, for example if the executive authority participates in the formation of laws, since even in this case, it cannot alone modify the law according to its whims at the moment of execution and thus cannot combine all the powers in his hands.

32Consequently it suffices for these various competences to be distributed in some manner among several authorities. Thus understood, the separation of powers is purely and simply the constitution, since a constitution is precisely a rule the object of which is to distribute competences. Article 16 of the Declaration of the rights of man is thus not designed to fix, as is too often believed, the criteria of a good constitution, to indicate what is a constitution “worthy of the name”, but only to give the definition of a constitution. It would moreover have been surprising for the idea that a good constitution should create specialized, independent organisms to have been formulated in the Declaration of the rights of man, whereas that declaration was adopted at the moment of elaboration of a constitution that did just the opposite.

33That being said, this first, purely negative principle, accepted by all who are not partisans of absolutism, necessarily calls for a second, designed to determine what the distribution of competences shall be. In this respect, in the eighteenth century there exist two rival principles, one of which can be attributed to Montesquieu.

34The one that comes immediately to mind as the simplest is the principle of specialization. It is the one that would be adopted by Rousseau and later by the National
Convention in 1793, precisely for reasons of hierarchy of functions. If the organisms are specialized, he who exercises the legislative power will prevail over the organism of executive power and, for partisans of specialization, it is good that it should be so, because the legislative power is nothing other than the people itself, in other words the sovereign (in Rousseau) or its representave (for the constitutionals). So it would be strange for executive power to be able to oppose the sovereign. Specialization is therefore a democratic system.

35At this point its opponents oppose it for ideological and political reasons. They fear democracy. But they can also adduce technical arguments: specialization is a self-destructive system. Indeed, if the legislative power prevails over all others, it is able to appropriate their functions to itself and exercise them alone. Then there will be no more specialization, no separation of powers moreover, understood in its negative sense.

36Montesquieu is one of those adversaries and the system he proposes is very different: it is that of the balance of powers. That system, far from self-destructing, is conceived as able to regulate itself and preserve itself automatically by virtue of its internal organization alone. According to the terms of the American revolutionaries, great readers of L’Esprit des lois, a constitution constructed according to this principle will be “a machine that would go for itself”.

37This search for an automatic principle, which is Montesquieu’s great originality, stems from his theory of the forms of government. His distinction of monarchy, despotism and republic, differs from others first by his refusal of a criterion drawn solely from the number of those who govern (one, all, a few), and by the utilization of multiple criteria: monarchy is thus the government of one, but according to laws; despotism is also the government of one, but according to the prince’s whims. Moreover, each government is defined by its nature, but is characterized by its principle and by its object. Its nature is its essence, what makes it what it is. Thus, for monarchy, its nature is precisely to be the government of one, but according to laws. The principle or mainspring is what allows it to function in conformity with its nature. Monarchy thus has as its principle honor, and despotism fear, the republic virtue. As for the object, it is that towards which the government tends, what it produces by virtue of its nature. For monarchy, it is the glory of the prince and the state; for despotism, its pleasures. There is thus a mechanical relationship between nature and principle on the one hand and the object on the other. Knowing the first two, it is a simple matter to determine the third.

38It is this relationship that Montesquieu undertakes to reverse. Let us imagine a government having political freedom as its object: what would be its nature and its principle? As it cannot be a monarchy, a despotic state or a republic, which have other objects, this government can only be a fourth, mixed form. One could invent it, but Montesquieu judges that that is not necessary because it exists. It is the government of England, such as he analyses it. “There is also a nation in the world whose constitution has as its direct object political freedom. We are going to examine the principles on which it is founded. If they are good, then freedom will show there as in a mirror” (“Il y a aussi une nation dans le monde qui a pour objet direct de sa constitution la liberté politique. Nous allons examiner les principes sur lesquels elle la fonde. S’ils sont bons, la liberté y paraîtra comme dans un miroir”, EL, XI, 5).

39Truly speaking, even there the originality is not great. First because there had long existed, at least since Charles I, interpretations of the English constitution as a mixed government. On the other hand, because Montesquieu could join a tradition going back to antiquity which vaunts the merits of mixed government by affirming that it presents the advantages of each of the simple forms, with none of its drawbacks.

40It is moreover simple to present the government of England as a mixed government since the House of Commons represents a democratic element, the House of Lords the aristocratic element, and there is a king. If the nature of this government is mixed, what is its principle? It is the conflict of interests. Thus the king will oppose the two Houses to preserve his executive power, the House of Lords will oppose the House of Commons to defend the privileges of birth or fortune, the Commons to defend the taxpayers. According to a famous formula of Montesquieu’s, “for the disposition of things, power must be checked by power” (“il faut que, par la disposition des choses, le pouvoir arrête le pouvoir” (EL, XI, 4). But the principle of the separation of powers (in the negative sense) is indeed preserved, since if these three organisms well exercise the legislative power together, they do not exercise the executive power, which is the king’s alone, nor the entire judicial power, although they collaborate in these two functions.

41The object will be realized in two ways. On the one hand, the principle of the separation of powers is not only respected, as we have just seen, but the system of balance of powers guarantees its preservation, as opposed to specialization, since by reason of the role played by the king through his right of veto in the exercise of the legislative function, never will the legislative power be able to subsume the executive power. The king himself will faithfully execute the laws, since he has consented to them and they express his own will. By obeying the king, one will therefore simply obey the law and political freedom will thus be established. But on the other hand, the law to which one is subjected will be moderate, for, by reason of the opposition of interests at the heart of legislative power, it can be adopted only as a result of compromise.

42This result however is not expected exclusively from the judicious interplay of checks and balances, in other words by constitutional engineering alone. It also depends on the opposition of economic and social interests, represented in the different legislative organisms. It is thus the social equilibrium that is the guarantee of the constitutional equilibrium. But the contrary is also true, for the economic and social interests can be preserved only if they are represented by one of the legislative organisms, empowered to oppose laws that might threaten them. It is in this sense that the constitution of England, such as Montesquieu describes it, is not only the constitution of a state, but also the constitution of the society, and it is precisely for this reason that, according to the formula already cited, article 16 of the Declaration of the rights of man proclaims that “Any society in which the guarantee of rights is not assured nor the separation of powers determines, is without a constitution”.

43From the standpoint of constitutional technique, the result is obtained not by specialization and independence but on the contrary by the collaboration of several organisms in each of the functions and the interdependence of authorities. It is moreover indeed thus that Montesquieu had been understood before this interpretation was forgotten and until Charles Eisenmann rediscovered it. That is how it was read in the eighteenth century; it is this doctrine of the balance of powers that inspired the constitution makers of Philadelphia, and later the Constituent Assembly of 1789 and furthermore the various monarchical constitutions of the early nineteenth century.

44If it was forgotten and if this system is losing its followers, that is because it was radically incompatible with the development of representative democracy and its triumph non only over monarchy and aristocracy, but also over mixed government. It is indeed inadmissible from the democratic point of view for representatives of the people to share the legislative power, in other words the supreme power, with non-democratic elements. Modern constitutions thus no longer expressly claim a balance of powers. How is the separation of powers to be understood today?

The separation of powers today

45The different principles which we designate by separation of powers have naturally known different destinies. The one which was the least discussed in the eighteenth century, the negative principle, was subsequently little contested except by authoritarian regimes in the name of the unity of state power. But it constitutes the fundamental dogma of representative regimes under the name of doctrine of the state of law. It is indeed, in substance, the same principle, because, according to the doctrine of the state of law, the authorities of the state can act only in conformity with law, in other words with a superior rule, that superior rule being this time not law, but the constitution. It is thus but the formulation in prescriptive mode of the idea that the juridical system is hierarchized.

46The principle in its classical sense, that is to say the absurd thesis, falsely attributed to Montesquieu, that the organisms of state must be specialized and independent in order to balance each other, is still invoked by the doctrine, but it is studded with all sorts of exceptions and nuances. Sometimes it is pointed out that specialization is not necessary and asserted that the different organisms must collaborate in the exercise of several functions. Balance is then expected from the independence of the organisms, by the absence of reciprocal means of action, as in the United States where the president cannot be recalled and has no power to dissolve Congress. Sometimes specialization is maintained and the balance has to come from the exceptions to independence, in other words from reciprocal means of action, as in the parliamentary regimes in which a legislative assembly can destitute the executive authorities, whereas these can dissolve that very assembly.

47However, even in this attenuated version, the principle of the separation of powers and in a more general way all forms of distribution of competences are distorted by the evolution of democratic systems and the development of political parties. If indeed the different authorities are composed of representatives of the same political parties, the parties or majority coalitions rapidly concentrate most of the power. This concentration can even lead to reversal of the hierarchies. It is again England that offers the best example: the House of Commons disposes of legislative power and the cabinet of executive power, the House of Commons can recall the cabinet at any time and the cabinet can request of the queen the dissolution of the House. In reality, as the Prime Minister is the head of the majority in the House, the House adopts all the laws proposed by the cabinet and never recalls the cabinet. The Prime Minister thus disposes in reality of the executive and legislative powers. What distinguishes this regime from despotism – in Montesquieu’s sense, in other words a regime in which a single person governs according to his whims – and leads to moderation is the relationship of forces within the majority, the existence of regular electioins and the possibility of an alternation. But it can be clearly seen that it is no longer a case of the distribution of the state’s juridical functions. For this reason some are tempted to say that the separation of powers is replaced either by a balance between the majority and the opposition, or by a balance between the different components of the majority. However, this balance is not really a balance, for as long as the alternation is not realized, as long as the Prime Minister has not been removed by elections or a palace revolution, he indeed exercises the totality of power.

48Separation of powers is also invoked with respect to federalism, insofar as it constitutes a distribution of competences between the federal state and member states. One says then that the separation in question is vertical. But there is no real balance in this case either, for the powers remain hierarchized. The member states no dout dispose of considerable autonomy, but that autonomy is granted by the federal state, whose right takes precedence to that of the states.

49It is paradoxically a variety of the balance of powers à la Montesquieu that best survives. Assuredly not as L’Esprit des lois describes it, in other words between a noble House, an elected House and a king armed with a veto, but today we know another form of it. In most countries the legislative power today is shared between parliamentary assemblies and constitutional courts. And if one proposes several justification for control of the constitutionality of laws, the most widespread and most effective is by far that which makes courts into counter-powers. Obviously Montesquieu said nothing about constitutional courts, but this justification can claim his legacy for several reasons: it allows, it is said, preservation of political freedom conceived as submission to the law, understood in a broad sense, in other words as submission to the constitution; it consists in arranging for power to check power; it allows bringing the control of constitutionality under a form of mixed government, since the will of the parliamentary majority of the moment, the democratic element, is controlled by a court composed of persons chosen for their competence, thus by an aristocratic element. The difficulty confronted by tenants of this justification is however to assume Montesquieu’s heritage entirely on two principal points: the conception of the power to judge as nil and thus unable of playing a role in the balance of powers; the conscious acceptance of mixed government and the correlative rejection of democracy.


Émile Durkheim, Montesquieu et Rousseau, précurseurs de la sociologie (1892), Paris: Marcel Rivière, 1953.

Joseph Dedieu, Montesquieu, l’homme et l’œuvre (1913), Paris: Boivin, 1943.

Raymond Carré de Malberg, Contribution à la théorie générale de l’État, Paris: Recueil Sirey, 1920; Paris, 1962.

Élie Carcassonne, Montesquieu et le problème de la constitution française au XVIII e siècle, Paris, 1927 (reprint Genève, Slatkine, 1970).

Jean Starobinski, Montesquieu par lui-même, Paris: Le Seuil, 1957.

Louis Althusser, Montesquieu, la politique et l’histoire (1959), Paris: PUF, 1964.

Robert Shackleton, Montesquieu: a critical biography, Oxford University Press, 1961.

Jean Ehrard, Politique de Montesquieu, Paris: Armand Colin, 1965.

Raymond Aron, Les Grandes Étapes de la pensée sociologique, Paris: Gallimard, 1967.

Michel Troper, La Séparation des pouvoirs et l’histoire constitutionnelle française (1973), Paris: Librairie générale de droit et de jurisprudence (LGDJ), 1980.

Simone Goyard-Fabre, La Philosophie du droit de Montesquieu, Paris: PUF, 1973.

Thomas Pangle, Montesquieu’s philosophy of liberalism, Chicago, London: University of Chicago Press, 1973.

Paul Vernière, Montesquieu et “L’Esprit des lois” ou la raison impure, Paris: SEDES, 1977.

Lando Landi, L’Inghilterra e il pensiero politico di Montesquieu, Pavia: CEDAM, 1981.

Bruce Ackerman, « The New Separation of Powers », Harvard Law Review 113, 2000/3, p. 634-729.

Charles Eisenmann (1903-1980), Écrits de théorie du droit, de droit constitutionnel et d’idées politiques, Charles Leben ed., Paris: Éditions Panthéon-Assas, 2002.

Bibliographical reference

Troper Michel , « Separation of powers », translated by Philip Stewart, dans Dictionnaire Montesquieu [online], directed by Catherine Volpilhac-Auger, ENS Lyon, September 2013. URL :