The Online Library of Liberty

A project of Liberty Fund, Inc.

• Preface to the Second Edition
• One: the Doctrine of the Separation of Powers and Institutional Theory
• Two: the Foundation of the Doctrine
• Three: the Theory of the Balanced Constitution
• Four: Montesquieu
• Five: the Matchless Constitution and Its Enemies
• Six: the Doctrine In America
• Seven: the Doctrine In France
• Eight: the Rise and Fall of Parliamentary Government
• Nine: From the Third Republic to the Fifth
• Ten: Progressivism and Political Science In America
• Eleven: Political Theory, Constitutionalism, and the Behavioural Approach
• Twelve: a Model of a Theory of Constitutionalism
• Thirteen: Epilogue: the Separation of Powers and the Administrative State
• Bibliography

SEVEN

The Doctrine in France

The pattern of mixed government and the separation of powers in the seventeenth and eighteenth centuries closely reflected the institutional developments of England and America, reacting to the problems those countries faced, providing the ideological materials with which they formulated solutions to those problems. The institutional development of France, however, had followed very different lines, drawing inspiration from a different set of ideas about institutions. These conflicting English and French approaches to constitutional structure were reflected in the work of Montesquieu, who attempted to integrate them into a single theoretical framework. Although Montesquieu’s attempt to provide a general description of systems of government embracing both French and English thought was not altogether successful, his great achievement was to concentrate attention upon the means of containing arbitrary power, and, by writing in general terms, to suggest that those aspects of English government which were so admired in the early eighteenth century might be applied elsewhere, if the requisite physical and social conditions were to hand. The latter part of the eighteenth century in France presents, therefore, a strangely confused picture in relation to Montesquieu and his thought. On the one hand, he is in many ways more important in France than in England and America, for he represents a turning-point in the whole approach to the problems of government and their solution; on the other hand, the specific solutions he offers in his description of the constitution of liberty seem inappropriate to French conditions and the current of French thought. This dilemma is clearly portrayed in the difficulties his disciples faced in 1789, when the proposals they made, based upon the mixed and balanced constitution, were rejected as alien and irrelevant to the new France.

The separation of powers, however, upon which Montesquieu had placed so much stress, became an essential article of faith with the men of 1789, so that, in a different way from that of America, the process of rejecting mixed government and of turning instead to the separation of powers as the basis of a free constitution was followed in France, with a vital difference. In 1789 in France, both doctrines, the balanced constitution and the separation of powers, were rejected or accepted as theoretical principles which had had little or no institutional reality in pre-revolutionary experience, whereas in America the doctrines were regarded as the evolution of, and the reformulation of, a system of institutions that had been operated, had become obsolete, and were to be modernized. It is in this respect that the institutional changes in France were truly revolutionary, in a way that those of America were not.

Furthermore, the role played by the separation of powers was different in France and America. In France the pure doctrine was held fiercely as an explicit ideological position, whereas in America it had been more a matter of the logic of the revolutionary situation than a conviction of the necessity of the pure doctrine which had dominated events. When in America the political situation enabled a resurgence of the older ideas, these had quickly returned to modify the extremes of the doctrine of the separation of powers. In France, however, the pure doctrine of the separation of powers took hold of men’s minds with an intensity, and a durability, not paralleled in America. Part of the explanation of this lies, as will be seen, in the complexity and intractability of the political situation, but part of the explanation must be sought further back in the past, in the particular form in which the men of 1789 came to look upon the separation of powers as a result of the development of thought after 1748. This part of the explanation of the intensity and persistence of the theory must be sought in the way in which the thought of Jean-Jacques Rousseau was overlaid upon that of Montesquieu, modified it, and gave a new direction and force to the theory of the separation of powers.

It is difficult to exaggerate Rousseau’s importance in determining the particular form that the separation of powers took in France. He was himself bitterly critical of the doctrine in the form in which Montesquieu developed it, and he was supremely unconcerned with the problems most of the writers on the separation of powers had considered vitally important. Yet the form in which he cast his theory, and the vocabulary he used, when adapted to the needs of the more practical men of the late eighteenth century, combined with that part of Montesquieu’s thought which seemed to them to be relevant, produced a theory of the separation of powers very different from that of England or America. Rousseau’s central position in the Social Contract, first published in 1762, was that law can only emanate from the general will of the community; the legislative power is the exercise of the sovereign will of the people. This power cannot be alienated or delegated; any attempt to create generally applicable rules from any other source represents a usurpation of popular sovereignty and cannot result in law. Rousseau’s emphatic denial of the divisibility of sovereignty was aimed at all those political theorists who in the past had divided up the sovereign powers among different persons or branches of government. Such writers, he wrote, make of the sovereign “a fantastic creature, composed of bits and pieces.” Like a Japanese conjuror dismembering the body of a child and reassembling it by throwing it into the air, these writers dismember the social body and reassemble the pieces “without our knowing how.”1 For Rousseau the idea of the division of sovereignty was the central fallacy of political thought, from which flowed most of the obscurities in writings on the State. “Whenever we think we see sovereignty divided, we are mistaken . . . the rights which are taken for parts of sovereignty are all subordinate to it.”2 Rousseau had, then, a boundless contempt for the theories of equilibrium and balance that invested the parts of the State with an independent power to check each other. It was this belief in the indivisible sovereignty of the people, adopted in the Revolution, that made the constitutional theory of Montesquieu unacceptable, except in relation to the separation of powers in its starkest form.

Although for Rousseau sovereignty was indivisible, the functions of government were distinguishable, and it was here that he made his greatest impact upon later French thought. He developed precise, abstract definitions of the legislative and executive powers, which have an immediate appeal far greater than the confused attempts at definition made by Montesquieu. “Every free action is the effect of two concurrent causes, a moral cause, or the will which determines the act; and a physical cause, or the power which executes it.”3 In the political sphere this distinction between will and force corresponds to the difference between the legislative and executive powers. This intellectual distinction between the parts of any political act was to dominate the mind of Sieyès and of other constitution-builders for many years, as a basis for dividing the functions of government, but it did not in itself entail that the power to will and the power to execute should be placed in separate branches of government, or in separate groups of men. On the contrary, it might well be argued, as later it was in fact, that these faculties of willing and acting are inseparable. Yet Rousseau himself explicitly demanded their separation. The legislative power can belong only to the people, but by definition this power can be concerned only with generalities, whereas the power to act can be concerned only with particular cases. The executive power cannot, therefore, be placed in the hands of the sovereign, whose province is the law, and whose acts can consist only of laws. It is a mistake to identify the government with the sovereign, for which the former is merely the agent or minister.4

It is necessary to pause here, however, and look very closely at Rousseau’s usage. When he wrote of the “executive power” his meaning was clearly something very different from that of his contemporaries. His use of the term was closer to that of Marsilius than that of Locke or Montesquieu. Rousseau’s executive denoted the whole apparatus of “the government,” that part of the State which puts the law into effect. Since the law can only be made by the people, and Rousseau refused to allow the delegation of this power to a representative assembly, then any such assembly, if one exists, is part of the “executive.” Thus a pure democracy was defined by Rousseau as a State in which the legislative and executive powers are in the same hands, but an aristocracy as a State in which the executive power is aristocratically constituted, whether it be a natural, an elective, or a hereditary aristocracy.5 This becomes very clear when we examine his treatment of mixed governments. “In the distribution of executive power,” he wrote, “there is always a gradation from the greater to the lesser number”; however, “sometimes there is an equal distribution; this may occur . . . when the constituent parts are in a condition of mutual dependence, as in the government of England.”6 Thus Rousseau subsumed all of Montesquieu’s ideal types of government under the label of “the executive power.” All forms of government, however constituted, are subject to the overriding legislative power of the sovereign people. This was the vital principle. It was a matter of expediency, and not of principle, to determine just how the “executive” should be organized, and whether the branches of government should be separated or not. The problems of institutional organization which seemed vitally important to other theorists were of secondary importance to Rousseau. In certain states, he said, it might be desirable to divide “the executive power.” “When the executive power is not sufficiently dependent upon the legislative, in other words when the ratio of prince to sovereign is greater than the ratio of people to prince, this disproportion must be remedied by dividing the government; for then its several parts have no less authority over the subjects, and their division reduces their total power as against the sovereign.”7 Rousseau therefore admitted the principle of the separation of powers by the back door, but treated all the former discussions of this problem as a matter of the organization of the “executive.”

Rousseau, then, discussed the organization of the powers of government at two levels; at the higher level he talked of the relations between the legislative and executive powers, and at the lower level he discussed the structure of the executive as he defined it. If we are to interpret Rousseau correctly, therefore, we must be careful to remember that when the discussion is at the higher level, as it almost always is, he is not using words, or conveying the same meaning, as do other theorists who write of the legislative and executive powers. Given his particular use of words, however, Rousseau had very definite views about the division of the functions of government and their separation between different bodies. It was this which led him to reject democracy. “It is not good for the makers of laws to execute them, nor for the body of the people to turn its attention from general considerations to particular objects.”8 One of the major advantages of aristocracy, including an elective aristocracy which Rousseau described as the best form of government, was that it distinguished between the sovereign and governmental powers.9 The government has the sole duty of executing the law, and the sovereign people must not attempt to govern. “If the sovereign tries to govern, or if the magistrate tries to make laws, or if the subjects refuse to obey, order is succeeded by disorder, force and will no longer act in concert, and the state being dissolved, falls thereby either into despotism or anarchy.”10

Within Rousseau’s frame of reference the separation of powers is an essential characteristic of an ordered system of government, but it has nothing to do with the separation of the powers of government between an elected legislature and a king, president, or cabinet. The rejection of any possibility of delegating the legislative power made Rousseau’s theory inappropriate, as it stands, to a great nation like France, and his use of terms in a way which Marsilius would have understood, but which few of his contemporaries could have accepted, made it unlikely that it would be used as Rousseau intended it. However, the emphasis upon popular sovereignty suited exactly the mood of revolutionary France, and attempts were made in 1793 to adapt the structure of the State to his theory by associating the whole people with the making of law. Most people in and after 1789, however, accepted the necessity of a representative assembly wielding the effective legislative power, for such an assembly was the only practicable means of carrying on the business of government. Rousseau’s ideas and his vocabulary were taken over and adapted to a representative system. The qualities and requirements of the “legislative and executive powers” were applied to assembly and king, or assembly and directory, in a way very different from that intended in the Social Contract. Indeed, when Rousseau turned to the discussion of more practical matters in the Considerations on the Government of Poland, he himself used the more normal vocabulary and discussed the separation of powers in the usual way. This adaptation of Rousseau’s thought to the representative system had, however, serious consequences. The extremely abstract definitions of functions were applied to the activities of ordinary legislative and executive bodies; the extreme and rigid division which Rousseau insisted upon between the legislative power, that is the body of the people, and the executive power, or the whole machinery of government, was applied to the elected legislature and its corresponding executive branch. There were to be no links between these branches, whose separate functions are clear and precise; the one wills, the other acts. This is an extreme version of the separation of powers that means something very different from what Rousseau intended. He clearly meant it as a rejection of anarchy—the people cannot govern themselves, they must be governed. Applied to a representative system it means a complete, thorough, separation of the branches of government on a functional basis that is implicitly unrealizable. It is the doctrine of the separation of powers in its purest, and most unworkable, form.

Rousseau’s theory of government is a direct attack upon the ideas of the mixed and balanced constitution, and its supremacy at the time of the Revolution, when the attack upon aristocratic, and later upon monarchical powers, was at its height, meant that Montesquieu’s version of the English Constitution would be quite unacceptable. However, in the interpretation given to Rousseau’s thought there was nothing inconsistent with what Montesquieu had to say about the separation of powers as such. Thus, once again, the attributes of the mixed and balanced constitution were stripped away, and the separation of powers remained, interpreted in the light of Rousseau’s theory as that was understood. This process of merging ideas from Montesquieu and Rousseau, and adapting them to representative government, is embodied in the work of Mably and Sieyès. Mably, well before the Revolution, evolved a version of this combination of the ideas of Montesquieu and Rousseau, although it should be mentioned that much of Mably’s work predates the publication of the Social Contract. The Abbé de Mably died in 1785. Over a period of some forty years he had developed ideas that were to become popular with the Constituent Assembly, which met such a short time after his death. The theme of Mably’s work is the equality of man, and a general feeling of dissatisfaction with all hereditary privilege.11 Yet Mably accepted the mixed system of government as the best system that could be attained in practice. He believed that full equality was unattainable, and that unrestrained democracy would be a dangerous experiment. A representative Assembly was an essential part of this pattern, so that Mably would seem, so far, to be a disciple of Montesquieu. In fact, however, Mably’s ideas were very different. He rejected all the checks and balances of the balanced constitution, and all positive links between the branches of government. The royal power of prorogation or dissolution of parliament in England, or the use of a royal veto, were indefensible because they resulted in the subordination of the legislative to the executive power. The executive, said Mably, should have no part at all in the exercise of the legislative power.12 In fact he believed in the necessity for a strict division of the powers of government on functional lines, and for this reason he rejected the example of the English Constitution. There was no true equilibrium between the powers of the English government; it was really a disguised monarchy, for in the last analysis all power must result to the King. The King could do many things without parliament; parliament, on the contrary, could do nothing without the King.13

Mably, then, retained the idea of equilibrium or balance in the constitution, but it was a different balance from that of Bolingbroke or Montesquieu, for Mably emphasized the balance between the legislative and executive branches rather than a balance of estates. How did he reconcile this with mixed government? To answer this question it is necessary to look at Mably’s earlier works, in which he discussed the ancient constitutions of Greece and Rome, as he interpreted them. He described these as mixed systems in which the different orders of society exercised separate and distinct functions of government, so that none could neglect their duties or abuse their power.14 With modifications this was just the view of the Constituents of 1789. Although there was no place for an aristocracy in the new constitution, the King was to remain but only on the basis of a strict separation of the functions of government. Mably’s theory is reminiscent of that of Philip Hunton; he attempted to accommodate the theory of mixed government to new conditions, and to emphasize the functional division between a hereditary monarch and an elected Assembly, although it is true that he also criticized the Constitution of Pennsylvania of 1776, because the executive was not chosen from among the legislature, a measure, he said, necessary to the attainment of harmony between the two branches.15

Mably’s rejection of the Montesquieu version of the balanced constitution is indicative of an important aspect of French thought before and during the Revolution. The English Constitution did not lack its passionate admirers,16 but it was under constant attack from all sides of the political spectrum. The advocates of absolute monarchy, the physiocrats, and the men of the Revolution, alike rejected the balanced constitution, either as an impossible division of sovereign power, or as an outright sham. The outbreak of the American Revolution was taken as an indication that the much-vaunted balance of powers in Britain did not result in political liberty,17 and the observation of corruption in English politics confirmed the view that the system was a disguised monarchy or aristocracy rather than a truly mixed system. The sense of outrage against aristocratic power at the Revolution made a rejection of the British pattern certain, and the whole fabric of the balanced constitution was rejected along with it. In the Constituent Assembly of 1789, however, the doctrine of the separation of powers was held as an article of faith by all except the few supporters of the ancien régime. On all sides it was hailed as the essential basis of a good constitution; it was incorporated into the Declaration of the Rights of Man as Article 16. The Assembly staged the most intensive discussion of the doctrine of all time, testing every aspect of the proposed constitution against this touchstone. In a period when abstract principles were highly regarded it had become part of the law of nature. The limits to be set to each of the powers of government, said Mounier, were sacred.18

The right wing of the Assembly presented, with great ability, a version of Montesquieu’s doctrine, modified somewhat to meet the views of the Assembly concerning aristocracy. In the report of the Committee on the Constitution presented by Lally-Tollendal there was set out a system of divided powers linked by the checks and balances of the English Constitution. Lally’s presentation was not merely a regurgitation of Montesquieu, but drew also upon Blackstone and de Lolme, and appeal for support was also made to John Adams’s Defence of the Constitutions of the United States. Although the idea of a hereditary Second Chamber was rejected in favour of a Senate appointed for life by the King, Lally developed the virtues of a system of perfect equilibrium, in which the three branches of the legislature would combine all the advantages of the three simple forms of government without their disadvantages. The executive power united in the hands of the King would have a veto to defend itself against encroachments by the legislature. With the power of prorogation and dissolution granted to the King and the power of impeachment vested in the Senate the structure of a free constitution was completed.19 Thus the doctrine of the balanced constitution formed the basis of the first concrete proposal submitted to the Constituent Assembly.

Lally’s proposals were based upon the mid-eighteenth-century conception of English government developed by Montesquieu, but they owed little to knowledge of the contemporary operation of English government. The links between the branches were simply to operate as checks to the encroachments of one upon the other, their function was purely negative. A few voices were raised in the Assembly, however, to put a very different point of view about the articulation of the parts of the government. Thouret, in opposing the exclusion of the King’s ministers from the Assembly, argued that to do this would be to establish as a constitutional principle that there should be a constant state of antipathy and discordance between the executive and the legislature.20 It was Mirabeau, however, who put forward most strongly the argument that although the powers of government must be separated, they must also be linked together, not merely in order that they should check each other, but so that a high degree of co-operation and co-ordination between them might be achieved. He rejected the extremes of the doctrine of the separation of powers in favour of a more pragmatic approach to the structure of government. An exact analysis of the “theory of the three powers,” he argued, would perhaps demonstrate the facility with which the human mind confuses words with things, and formulae with arguments.21 It is necessary to make the King’s ministers responsible to the people through their representatives, but this could only be done effectively if ministers were in the assembly to be questioned, and could not evade responsibility. In England, said Mirabeau, the people did not consider the presence of ministers in the parliament as a danger, but as an absolute necessity if they were to be subject to control. The chief agents of the executive must be in the legislature, for they are a major source of information; laws discussed with them will be more easily shaped, more effective, and executed more faithfully. The ministers’ presence will forestall incidents between legislature and executive, and will obviate the need for troublesome legislative committees.22 In the Courier de Provence he wrote that the maintenance of a direct, daily intercourse between ministers and legislature was a necessity.23 Mirabeau argued, therefore, for the recognition of a form of parliamentary government along the lines which, a few years earlier, Burke and Fox had proposed in England.

However, the majority of the Constituent Assembly was prepared to accept neither the Montesquieu scheme proposed by Lally, nor the parliamentary system of Mirabeau. They were too afraid of royal domination, too jealous of their legislative power to wish to share it, and their main impression of English politics, which formed the basis of both proposals, was of its deep corruption. They shared rather the views of one who rejected the corrupt English system, the Abbé Emmanuel Sieyès. Sieyès’ thought was based upon a hatred of aristocracy and privilege that led him inevitably to reject the balanced constitution. He was as committed to the unity of the sovereign power as was Rousseau. In his Qu’est-ce que le Tiers Etat? published in January 1789 Sieyès had emphasized the unity and sovereignty of the nation, from which all power is derived, which establishes the constitution and determines the functions of the parts of the State. This unity precludes all privilege. The third estate is the nation; “Qu’est-ce que le Tiers?” he asked, and gave the answer “Tout.”24 A single chamber is the only form of legislature which can represent this unity. At the same time, Sieyès, in 1789 at least, saw the monarchy also as a manifestation of the unity of the nation, and supported it for this reason. Thus we are presented with the simple bipolarity of a single-chamber legislature and a royal executive. Sieyès started from the same view of popular sovereignty as Rousseau, but his acceptance of the representative principle raised a number of problems about the status of the branches of the government and their relationships to each other. He attributed to the representative legislature, and to the royal executive, the same characteristics that Rousseau, with very different concepts in mind, had attributed to the sovereign people and the government. Sieyès was able to do this by inventing a fourth “power” that was solely in the nation, the constituent power. The nation exercises this power by delegating the legislative power to the assembly and the executive power to the King. Thus each of these has a “representative” character and they enjoy a certain equality of status. Each power must be limited to its proper function and prevented from interfering with the other.

Sieyès returned, therefore, to a view of the legislative and executive branches more in line with that of Montesquieu than of Rousseau, and by giving to each of these branches a certain equality of status as a delegate of the people he introduced something of the idea of an equilibrium, rather like that of Malby. But Sieyès’ approach to the functions of these branches of the government, and to the relationships between them, had nothing in common with the theorists of the balanced constitution. He took over Rousseau’s definitions of the functions of government, attaching them now to the legislative and executive branches. The body politic, he said in the Constituent Assembly, must be endowed with a power to will, and a power to act; the former is the legislative power, the latter is the executive power, and these must be kept as distinct from each other, and as related to each other, as they are in the human body.25 This rigid definition of government functions led Sieyès to strip away all semblance of links between them, and to reject all checks and balances. A royal veto is unacceptable because the executive power has no right to enter into the making of the law.26 In England such a veto may be necessary because the English have failed to distinguish the constituent power, and the unlimited authority of parliament could therefore be used to attack the position of the monarchy if the veto were not there to protect it. In France, however, this situation could not arise, for the authority of the branches of government being drawn directly from the people, they need not fear each other, and no checks to the encroachment of one branch upon another are needed. Provided that the powers of government are divided with care, and are made independent of each other, they are then in an equally advantageous position. No veto is required, nor the power to withhold supplies, for if any part of the State should exceed its authority the people in Convention will intervene and resume the power which it has delegated.27

In the revolutionary conditions of 1789 we are back once again in the presence of that formula which Nedham had announced in 1654, the power of the community flowing in distinct channels, so that they may never meet. The conditions of the two periods are also similar in many respects: we have the attack upon privilege, the determination to limit the “single person” to the execution of the law, the attempt to pick a middle path between royal absolutism and an excess of democratic government. For at no time did Sieyès advocate an all-powerful legislature. He clearly believed that the danger of arbitrary rule threatened more from the royal power than from the elected representatives of the people, but gouvernement d’assemblée was far from his conception of constitutional government. The French Revolution was unlike the English, however, in that the attachment to the pure doctrine of the separation of powers preceded the period of convention government in France, and this government by convention, when it did come, gave way again in 1795 to another constitution of divided powers, only then to succumb to the authoritarian rule of Bonaparte. Perhaps the most important difference between the ideas of the two periods in relation to the separation of powers was the determination in France to enforce strictly the separation of the personnel of government between its branches. This aspect of the doctrine, which had recently been so important to the American constitution-makers, dominated the minds of the men of the Revolution. It led to the defeat of any attempt at a parliamentary system, and it persisted well into the period of the Convention.

The Constitution of 1791 began with a sweeping abolition of all privileges, orders of nobility, and feudal or other social distinctions. It proclaimed the indivisible, inalienable sovereignty of the people, but hurriedly added that the nation could only exercise its powers by delegation through its representatives, the National Assembly, the King, and the elected judiciary. The unicameral Assembly was a permanent body, elected every two years, over which the King had no power of dissolution. The King could not initiate legislation, but he was given a suspensive veto. The idea of ministerial responsibility was rejected in favour of a process of impeachment before a National High Court.28 Members of the National Assembly were to be incapable of appointment to ministerial office, or of accepting any place or pension in the gift of the executive, during their membership of the Assembly and for two years afterwards. Thus the legislative and executive branches were strictly divided, although ministers were allowed to speak in the Assembly and to listen to the debates. The Constitution raised “the judicial power” to a level of equality with the legislature and the executive, the result, in Duguit’s view, of the American example rather than a response to political theories.29 The Assembly and the King were expressly forbidden to exercise any judicial function, and the judges were intended to be independent of the other two branches by virtue of their popular election. However, the strict separation of powers ruled out any possibility of judicial review of legislative acts like that which resulted from the theory of checks and balances in the United States. Indeed the Constituent Assembly inserted in the Constitution a specific denial of the right of judicial review; the courts were forbidden to interfere with the exercise of the legislative power or to suspend the execution of the laws. Nor were they to entertain actions against officials in respect of their administrative activities, so that the courts were prevented from exercising authority over executive or administrative, as well as legislative, actions. Thus the Assembly laid the basis for the vitally important distinction in French law between the judicial and the administrative jurisdictions. In the Constituent Assembly this was justified by the separation of powers theory,30 but its roots went back to the practice of the ancien régime.31

The Constitution of 1791 was based, therefore, upon an extreme version of the doctrine of the separation of powers. In reality, of course, circumstances were such that there was very little likelihood of such a system of government succeeding. It assumed a degree of natural harmony between the branches of government which would be rare in the calmest of political systems, but in the turmoil of revolutionary France its chances were slender indeed. In other revolutionary situations, in England in the 1650’s and America in the 1770’s, it had proved unworkable. The circumstances of 1791–2 in France were even less propitious. The Assembly was determined to limit the power of the executive, and if this meant that it had to do more than merely pronounce general rules of behaviour, then an abstract doctrine like the separation of powers would not stop it. Had the King kept to the spirit of the Constitution the outcome might have been less disastrous, but even before the acceptance of the Constitution his flight to Varennes made it necessary for the Assembly to take over his authority, and to become for a time the sole source of governmental power.32 In spite of the potential support he enjoyed in the new National Assembly of 1791 the King antagonized the Assembly by almost immediately making use of his suspensive veto. Through the medium of its committees the Assembly began to exercise an increasingly detailed control over the executive, until the further use of the veto in May and July 1792 and the insurrection of the 10 August brought an end to any pretence of divided power. The National Convention which assembled on the 20 September 1792 exercised authority in every sphere of government action, dealing with every type of government business, itself directing the everyday affairs of the State through its committees, and through the commissars it sent into the provinces or to the armies. The régime d’assemblée was an accomplished fact.

The deputies of the National Convention found themselves exercising a supreme, unlimited power over every type of government task, yet the extraordinary importance which they attached to the idea of the separation of powers can be seen in the way they adhered to a formal separation of persons between the “executive” and the “legislature.” The Convention decreed on the 25 September 1792 that the exercise of any function of public office was incompatible with membership of the Convention.33 It rejected the argument that with the end of royal power there was no further need for a feeling of suspicion towards the executive, and that ministers might therefore now be chosen from the Convention. It even refused to allow members of the Convention to be appointed to office if they resigned their seats, because of the possibilities for corruption that this practice would open up. Lecointe-Puyraveau carried the Convention with him when he insisted that the most important argument against the choice of ministers from within the Convention was that the deputies had been sent there to make laws for the people. If they removed a man from the Convention to the executive, would not the people be able to say “I have sent this citizen to make the laws, not to execute them”?34

In March 1793 Danton took much the same position that Mirabeau had taken in 1789, asking the Convention to choose the members of the executive council from among its members. There was a great need, he said, for a more direct day-to-day co-operation, a need for greater cohesion between legislature and executive. But again he was met by the same arguments. La Revellière-Lépaux argued that if men of great ambition were chosen, one day the Convention might find itself dissolved, and “these men, invested with both the legislative power and the executive power, will then exercise the most formidable dictatorship.”35 For all this formal concern with the separation of powers, the actual situation was very different. The “ministers” became little more than subordinate officials, mere clerks who submitted every decision to the Convention and its committees. Their advice was ignored, often they were not even consulted. They became mere marionettes in the hands of the deputies.36 With the appointment of the Committee of Public Safety, first under Danton and then under Robespierre, the system bore some faint resemblance to a system of parliamentary government, but it was a system in which “ministerial responsibility” was enforced by Madame Guillotine.37

Finally, on the 1 April 1794 the Convention acknowledged reality and abolished the six ministerial posts, setting up twelve executive commissions, each consisting of three members, closely subordinated to the Committee of Public Safety. Carnot, in urging the Convention to take this step, developed a theory of revolutionary government. An executive council, he alleged, was an instrument of royal despotism, intended to maintain privilege and social distinctions; how then could it become the instrument of a representative government devoted to the principle of equality? “Government is nothing more, properly speaking, than the council of the people.”38 The people’s sovereignty must be guarded by dividing up the instruments of government and restraining them within the closest limits, to prevent the accumulation of power; at the same time the closest subordination of the active agents of government to the National Assembly must be maintained.39 This is the pure theory of gouvernement d’assemblée, the complete rejection of the separation of powers or any other theory which sets any sort of limit to the power of the legislature.

Whilst engaged upon a practical exercise of power which showed little concern for the spirit of the separation of powers, the Convention was also busy with schemes for the re-establishment of constitutional government. The year 1793 saw first the Girondin project for a new constitution introduced by Condorcet, then a large number of schemes presented by individual deputies, and finally the project of June 1793, the Montagnard scheme, which was accepted by the Convention but never put into operation. The de facto acceptance of convention government did not mean, however, that the separation of powers played no part in the thoughts of the deputies when they turned to constitution-making. Several of the projects submitted to the Convention made the doctrine the cornerstone of their proposals. Boissy d’Anglas, later to play an important role in the writing of the Constitution of 1795, rejected the idea that popular sovereignty demanded a single channel of government action as “un blasphème politique.”40 The existing structure of government, necessary in the circumstances, was itself evidence of the way in which a single all-powerful Assembly could subject the people to the oppressive acts of their own representatives. Daunou also strongly opposed the idea of a system of government in which the Assembly could become involved in all the operations of ministers, or exercise an immediate influence over matters purely executive.41 These men were as firmly opposed as Sieyès or Condorcet to the notion of a balance of powers, “the resort of a people half-enslaved”; they demanded only a strict division of the powers of government.42 It is interesting to note that the project which most closely resembled gouvernement d’assemblée was apparently submitted by an Englishman, a “Mr. J. Smith,” who argued that the executive power should reside in the legislature, to be delegated only to a legislative committee in particular circumstances and under specific and limited conditions.43

Condorcet, in presenting the Girondin plan to the Convention, emphatically denied any attempt to create an equilibrium of powers in the government. The new draft of the Declaration of the Rights of Man replaced the reference to the separation of powers in the 1789 Declaration by the statement that the limits of the “fonctions publiques” must be clearly laid down by law.44 This was plainly a further movement away from the Montesquieu position towards that of Rousseau. The rejection of the idea of the delegation of “powers” in favour of the assignment of “functions” had been proposed by Robespierre in 1791,45 and the Girondin project removed any suggestion of an equilibrium which had remained in the constitutional theory of Sieyès and the Constituent Assembly. The principle of the new constitution, said Condorcet, must be the unity of action and of principle between the legislature, the executive council, and the people. The system of balanced powers suggested the existence of divisions within the State, whereas in the French Republic there could be none.46 The project gave a wide authority to the Assembly. It had not only the function of making “laws,” but also of enacting measures of “general administration” over a wide range of subjects. The principle of unity required that the agent entrusted with the execution of the laws should be subordinated to the legislature, for the executive council might not will but only oversee (veiller).47 However, Condorcet did not wish the executive council to be the mere creatures of the legislature. They must be forced to obey, but they must also have a certain independence, and in order to accomplish this the members of the executive council were to be elected not by the legislature, but by the people whose officers they were.48

Thus, after strongly denouncing any attempt to erect a system of balanced powers, Condorcet arrived at a position where the executive council was directly elected by the people and was quite independent, in its origins, of the legislature. Indeed, as Saint-Just pointed out, the executive council would have been more directly representative of the people as a whole than the legislature, and might easily have assumed a position of greater prestige and importance in the eyes of the public.49 Nor could the executive be said to be “responsible” to the legislature in any straightforward sense. The legislature could accuse the executive, or any member of it, and demand their removal, but the final decision was to be taken by an elected “national jury” to be specially convoked to hear each case. Hardly an expeditious means of parliamentary control! Condorcet, in fact, was proposing much the same sort of structure that had been established by the Constitution of Pennsylvania in 1776, a Constitution to which he had given much thought, and he was grappling with the same fundamental problem that would have faced Bentham and other English radicals had they become the Founding Fathers of a constitution. This problem, common to all three countries at the end of the eighteenth century, was the difficult one of finding a reconciliation between the desire for legislative supremacy on the one hand, and the desire to prevent government by convention on the other. How could the power of the legislature be “unlimited” at the same time that the executive was given the independence, and the power, to prevent the legislature from committing individuals to prison, issuing commands to the army, engaging and dismissing government servants, or the thousand and one other things for which it was not fitted? This basic dilemma can be seen in the way Condorcet attempted to find a compromise between these two positions. Paul Bastid has suggested that Condorcet was making an explicit statement of the case for gouvernement d’assemblée, and the latter certainly denies to the executive council the status of a true “power.”50 But apart from the position which the Girondin project gave to the executive council by its direct election, Condorcet went out of his way to emphasize that the legislature should be given only that function in relation to the executive which was “proper” to it, that of surveillance.51 Furthermore, the creation of a national jury to judge the shortcomings of the executive removed all suspicion, he maintained, of a possible abuse of power by the legislature.52 Condorcet was in fact grappling with the problem of legislating into a constitution Rousseau’s abstract notion that one set of people can will, and another set can act, without on the one hand making the latter mere automata, or on the other enabling them to interpose their own wills in such a way as to defeat the intentions of the legislature. This problem, which Rousseau had set, Condorcet attempted to solve by the introduction of the “national jury” as did the creators of the Council of Censors in Pennsylvania earlier, and Sieyès later by his “jurie constitutionnaire.” In all cases it was hoped to avoid government by convention, to establish a rigid division of functions, and to restrict the executive exclusively to the execution of the laws. The tragedy of all these attempts was that they were based upon an unworkable and untenable definition of the functions of government, which assumed that decisions to act could be completely separated from the putting of those decisions into effect, and upon the belief that the essential discretionary powers of government had been used by kings solely to maintain their arbitrary sway.

The rejection of the Girondin project by the Assembly, and the acceptance of the Montagnard Constitution of 1793, represented a further move towards the complete subordination of the executive, and towards a constitutional system bordering on government by convention. The direct election of the executive council was rejected in favour of election by the legislature. This can be seen particularly in the fate of the “national jury” in the debate in the Convention. The draft submitted by the Committee of Public Safety had included provision for a national jury which would guarantee the citizens against oppression by the legislature or the executive. In this respect it retained some of the philosophical assumptions of the Girondin project. In the debate, however, Thuriot objected that it was ridiculous to raise up an authority superior to the legislature which the constitution endowed with sovereign power, and the Committee’s rapporteur, Hérault-Séchelles, was quick to acknowledge that the national jury introduced a germ of division into the constitutional system. The national jury disappeared from the draft.53 Although the means of bringing the executive to account were left extremely vague it is obvious that it was intended to have little prestige, and less power. Nevertheless, the Convention was seemingly not prepared to accept for a constitution a system which merely embodied gouvernement d’assemblée. In debate Barère successfully insisted that if the choice of the executive council were made at the beginning of the life of the Assembly, then the executive would simply obey in a servile fashion the men who had created it. The legislature should therefore, he proposed, renew half the executive council each year at the end of the Assembly’s period of office, and this proposal was accepted.54 Thus the most radical group in the Assembly was unable to obtain the incorporation into the constitution of an executive which had no independence of the legislature of any sort. The revolutionary system of Carnot could not be incorporated into a constitution which was to last for all time. As Boissy said, “We are in the day of chaos which precedes the creation, but the Creator must not restrict himself to the organization of chaos.”55

After the experience of the revolutionary government, and the Terror, the Convention turned again, after the fall of Robespierre, to the task of establishing a constitutional system of government. The extremes of the Constitution of 1793 were no longer acceptable. The situation was again one in which men of all points of view had, each in succession, felt the effects of unlimited power, in the hands of an absolute king or of the representatives of a faction in the legislature. Their minds turned once again towards the idea of checks to the exercise of power.56 They moved back towards the notion of an equilibrium between executive and legislature of the sort that had prevailed in 1791, but they were still quite unprepared to consider a system of checks and balances on the English pattern. The agony of the Convention, faced with the need to create a system in which no one could exercise unlimited power, yet aware of the problems of the relationships between legislature and executive in a system of divided powers, is poignantly expressed in the attempt which Sieyès made to find a solution to this dilemma.

The report of the Constitutional Committee presented by Boissy d’Anglas was squarely based upon the absolute necessity of preventing the accumulation of the powers of government in the same hands. “Whenever they are found united, whenever they are confused, liberty is at an end, there is nothing but despotism.”57 The legislature was to be prohibited from exercising executive or judicial powers, either itself or through delegates. The executive, consisting of a Directory of five members, was to be chosen by the legislature, but not from among its own members. Members of the Directory could only be removed by impeachment before the High Court of Justice. Formal checks and balances were quite absent. The attempt by one deputy, Ehrmann, to propose a veto power copied from the American Constitution brought fierce protests, and it clearly required a good deal of courage on his part even to make the proposal.58 The movement towards the establishment of checks to the exercise of legislative power did result, however, in the creation of a bicameral legislature. Thibaudeau, in defence of the Committee’s draft, argued that the division of the legislative power was all that was necessary to maintain the necessary rapport between the executive and legislative branches. The Upper Chamber, the Conseil des Anciens, would have an interest in the maintenance of an ordered system of government, and would restrain the legislature where necessary, so that there was no need for the executive to play any part in the making of law. The constitution made no provision at all for possible conflicts between the two branches of government. The same strict separation of the judicial power from the other branches of government was established as in the Constitution of 1791.

The Constitution of the year III (22 August 1795) was, therefore, another exercise in the pure doctrine of the separation of powers, with little advance over that of four years before, except in the modifications necessary to replace the King by the Directory. The most interesting aspect of this exercise in constitution-making is the attempt which Sieyès made to go beyond the constitutional pattern he had helped to create in and after 1789, and to find a solution to the problem that Condorcet and others before him had grappled with.59 Sieyès was unsuccessful in his attempt to lead the Convention away from the pure doctrine of the separation of powers, yet his ideas are interesting, because they illustrate perfectly the dilemma into which constitutional theory had fallen in France by rejecting absolutely the theory of checks and balances. Sieyès’ use of language has been described as “metaphysico-nébuleuse,” and his enormous conceit makes it difficult to give to his presentation the close thought that it deserves. He was concerned, however, to reconcile three positions. First, the absolute necessity of keeping the functions of government in separate hands. Second, the need to prevent the abuse of power by any branch of the government. Third, the need to ensure co-operation between the branches of government without either destroying the separation of powers or resorting to Montesquieu’s system of checks and balances. His rejection of the equilibrium of checks and balances was final; it was nothing more, he said, than perpetual civil war, whereas the aim of any constitution must be to create a system of “organized unity.”60 Sieyès summed up the problem in a few words: “Unity alone is despotism, division alone is anarchic.”61 It is necessary to separate those parts of the system of government that can be separated, but at the same time to reunite those parts which must co-operate if government is not to collapse altogether. This admittedly nebulous remark nevertheless sums up the whole problem of constitutional government at the end of the eighteenth century, and is perhaps even more relevant at the present time. How can the exercise of power be checked without destroying the essential harmony of the government? Sieyès’ attempt to solve this problem was, however, still dominated by Rousseau’s formulation of the nature of the functions of government. He made a distinction between the government and the executive, and then announced that “The executive power is all action, the government is all thought.”62 Thus Rousseau’s distinction between willing and acting was still, as it was in 1789, at the basis of Sieyès’ thought. The function of the “executive” was still purely mechanical, but Sieyès had a very much more complex approach to the legislative power, and the limitation of the legislature to its proper function. He wished to create, in fact, four elected representative bodies. The government was one of these, with the power of initiating proposals for legislation as well as supervising the executive. The second was the tribunat, the function of which was to criticize the government and to propose to the legislature measures that the government had failed to put forward. The third was the legislature itself, which was to have no power to initiate laws, confining itself to making decisions on the proposals submitted to it by government and tribunat. The fourth was a constitutional jury, which would consider complaints of unconstitutional action against the legislature.63 This, then, was Sieyès’ method of providing checks to the exercise of power without the use of the hated concept of “equilibrium.” The tribunat and the government would battle against each other, but this conflict would provide no dangers because the legislature would act as an impartial judge between them, without being able itself to interfere in any way with the executive, because it could not initiate action. The constitutional jury would ensure that the legislature did not exceed its powers. Thibaudeau remarked, rather unkindly, that for all his vehement rejection of the balancing of the powers of government, Sieyès had produced nothing more than a system of equilibrium under another name.64 The proposals were undoubtedly too complex and too outlandish for the Convention to give them much serious consideration; nevertheless, Sieyès was grappling here with difficulties which would in future years be met with equally strange solutions, in the form of administrative tribunals, regulatory commissions, and, in the 1960’s, the English equivalent of the tribunat, the National Economic Development Council. The Constitution of 1795, however, remained a constitution of strictly separated powers, which in a very short time became the stage for bitter conflicts between legislature and executive, soon leading to a coup d’état by a majority of the Directory, setting France on the road to the Empire of Napoleon Bonaparte.

A comparison of the circumstances surrounding the development of the pure doctrine of the separation of powers in France and in America at the end of the eighteenth century illustrates the interesting relationship between ideas and events. In America the pure doctrine emerged only briefly, as a result of the shattering impact of the break with traditional forms of government and the consequent rejection of monarchy and aristocracy. The institutional structure of the pure doctrine in America was the result of the logic of circumstances rather than of an ideological commitment to the pure theory, although in Pennsylvania there was a small-scale rehearsal of the French situation. In France the circumstances were also such that the pure doctrine fitted better than any other available constitutional theory the aspirations of the Revolution, but the influence of Rousseau and his interpreters gave to that doctrine a degree of intellectual precision and intensity it never could have achieved in America. Thus, in part at least, the particular quality of the doctrine in France must be attributable to the ideas of Rousseau, which had so much more of an impact there than they could have had in the American colonies, fed upon a diet of Montesquieu and Blackstone. In a few years the doctrine of the separation of powers was changed and modified in America by the resurgence of the theory of checks and balances, whereas in France this was not achieved until 1814, and even then the separation of powers was to burst forth again in 1848. Clearly the explanation of this is that in America the Revolution was wholly successful, monarchy and aristocracy were routed, and many of those men of wealth and status in the community who were in 1776 caught up in the onrush of democratic fervour became in a few months or years the spearhead of a movement to moderate that democracy. There was no longer anyone on the Right to fight; the only danger was from the Left, and an excess of democracy could best be combated by reintroducing a constitution of balanced powers. In France, however, the Revolution was never complete, never wholly successful, and could not be. The threat of monarchic and aristocratic privilege remained, and the theory of checks and balances must inevitably be associated with it. The doctrine of the separation of powers remained, therefore, in France, as in a very much weaker way it remained in England, a potentially revolutionary theory.

The period of the Revolution in France presents in some respects a doleful view of the development of constitutional theory, for the Constitution of 1795 provided little practical advance over the Instrument of Government of 1653 or the Constitution of Pennsylvania of 1776. Like them it was a dead-end, the institutional elaboration of an unworkable theory of government. Nevertheless, during this period a few men in France realized the basic problems of a democratic representative government in a manner unparalleled in England or America. In England the energies of those who wrote about constitutional matters were occupied with the problems of the mixed and balanced constitution, but few people were really trying to relate this constitutional structure to the need for co-operation in government other than by means of corruption. Those who looked for a unified political system hardly considered the problems of co-ordination and control that it involved; they concentrated instead upon attacking the cabinet system, which was in fact to provide a solution to this problem. In America the special circumstances of a federal system of government, and an expanding society based upon a fierce individualism, concentrated attention on the means of checking the exercise of government power, to the exclusion of all consideration of the problem of obtaining co-ordinated government action. In France, however, there were a number of men who saw the problem of obtaining both control and co-operation and attempted to fashion the necessary institutions to achieve both ends. In their different ways Mirabeau, Danton, and Sieyès all had a grasp of the deep dilemma of modern constitutional systems, which was unsurpassed in the contemporary world of thought. This fact is well illustrated by the way in which, after 1814, French writers, in attempting to create an intellectual basis for a system of parliamentary government, reached a very high order of modern political theory.

In the evolution of parliamentary government in Britain, and of the Federal Constitution in America, the theory of mixed government played an essential role in providing the basis for a system of separate and balanced “powers” of government. In France before 1814, however, the idea of mixed government had played a relatively minor role. Hotman, Montesquieu, and Mably had all championed the idea, but their views, in this respect, had had little influence. The constitutionalism of the ancien régime was based upon the role of the “intermediary powers” in a “legal” monarchy, but this was a very different concept from that of mixed government. The most important manifestation of the idea had been the proposals put before the Constituent Assembly by Lally-Tollendal in 1789, but this theory of constitutionalism had then been decisively rejected, and the French nation had jumped, in a few short years, from absolutism to democracy, without going through any intermediate stages. It would, of course, be rash indeed to suggest that the successful transition from absolutism to modern democracy necessitated a period of mixed government, yet it is certainly true that France, as if she had made a false start, emerged from the débris of Napoleon’s Empire to grapple with the complexities of a system of mixed government, leaving behind, for a time, the deceptive simplicity of absolute democracy or dictatorship.

In 1814 the Charte of Louis XVIII ushered in a period of mixed government, which had started out as an attempt to recreate a more liberal version of the historic French Constitution, but which in fact provided the forum for a discussion of the nature and working of parliamentary government of a quality far higher than that of most English thought of the period. The role of the King’s ministers was clearly faced and an attempt made to fit it into the pattern of government by King, peers, and assembly. The defeat of Napoleon brought with it a surge of admiration for British government throughout Europe, but it would be wrong to suggest that the authors of the Charte found their inspiration in the English system. The Charte was granted by Louis XVIII as an act of royal grace, not as a contract by one branch of a balanced system of government.65 The Chamber of Deputies and the Chamber of Peers were seen rather as consultative assemblies which were to play a part not too dissimilar from the earlier role of the parlements,66 although the necessary safeguards to the liberty of the individual were provided by the Constitution. The Charte, wrote Chateaubriand, was a treaty of peace signed between the two parties that divided France, the ancien régime and the Revolution. However, in practice, the political discourse of the whole period 1815 to 1848 revolved around the roles of King, ministers, and Assembly in a context similar to that of the disputes in England, and one in which the experiences of English politics were used as evidence in support of one point of view or another. The most important theorists of the period were Royer-Collard, Guizot, and Benjamin Constant. Royer-Collard and Guizot developed a theory of constitutional monarchy which was suited to the period 1814–30, whilst Constant went further and formulated a theory closely related to the régime established by the Revolution of 1830.

After 1814 Royer-Collard and Guizot reverted to a mid-eighteenth-century usage; when they referred to the “three powers” of government they generally meant King, peers, and deputies, and clearly had in mind a system of mixed government. But it is very clear also that they did not see this system, as had the English eulogists of mixed government, as a system of negative checks and balances. The emphasis in their work was rather upon the unity of action among the parts of government. That problem of the concerted movement of the three branches of the legislature, which Montesquieu had brushed aside as being solved by the “necessary movement of things,” had become the central concern of the theorists of the Restored Monarchy in France. They accepted the separation of powers as a basis of government, but a basis only, upon which must be built a system of concours between the branches of the government. Royer-Collard presented a theory of the Charte which rejected parliamentary government, seeing ministers simply as secrétaires du roi.67 In a period when ministerial responsibility to parliament was demanded by the ultras who had a majority in the Chamber, and who found Louis XVIII too liberal, Royer-Collard argued that parliamentary government required a stable majority in the legislature if it was to work effectively; but in France there was no means of securing the basis of a strong party system which would be able to sustain such majorities.68 He saw clearly that a system of parliamentary responsibility was incompatible with the idea of mixed government, and that the day the government was at the discretion of a majority of the Chamber, then, to all intents and purposes a republic had been established.69 The system of the Charte was, in his view, a mixed government with a strict separation of functions between King and parliament; the King governed independently of the Chambers, and their agreement with his views was only necessary for the passage of new laws and of the budget.70

Royer-Collard, then, presented a theory of mixed government which is more reminiscent of Philip Hunton in 1643, or of the English Constitution in 1689, than of a parliamentary system at the beginning of the nineteenth century. He insisted on the “harmony” of the powers of government, but he did not suggest how this was to be achieved other than by a simple exercise of self-restraint by the legislature. Guizot, however, took the analysis deeper. In 1816, in Du gouvernement représentatif, he also rejected the idea that ministers should be dependent upon a parliamentary majority as a vicious attempt to distort the whole structure of the constitution.71 The object and tendency of a constitution must be to create unity in the system of government, for this unity is an overriding necessity comparable to that of the unity of society itself. Guizot interpreted English constitutional history in terms of a bitter conflict between the powers of government up to the Revolution of 1688; after that date unity between these powers was progressively established, by their reciprocal penetration, until they formed a single power, with internal limits to be sure, but able to exercise within these limits a plenitude of power without danger to the government or its adversaries.72 When a mixed government has reached its maturity, said Guizot, there is a unity of power, and a unity of action, in which the sovereign power, fundamentally united although divided in appearance, is subjected to certain conditions by the very nature of its internal organization, which establishes within itself limits it cannot exceed without losing its force.73 There is, therefore, in this system, “a fusion of powers,” the only possible point of equilibrium for a mixed government, because it is its necessary tendency, and because in this way “the powers, far from becoming enfeebled, or being assimilated by each other, all gain equally and are equally strengthened.”74 This was to be achieved in France, as in England, by the government’s entering into the Chambers, making them the centre of its activities, governing through them and by them. This did not mean government by the majority, but government through the majority. Legislative majorities, he wrote, are transient things which do not represent permanent interests, only ephemeral points of view. If indeed the King were to be faced with a stable majority in the Chambers that he could not alter or weaken, then either he must submit to it or abolish representative institutions. Fortunately this situation did not arise because majorities were shifting, and the means of management were available to the King. The government must obtain a majority, but it had a thousand ways of doing so, ways which, if its conduct were reasonable, and firm, would ensure its success.75

Du gouvernement représentatif is a first-class attempt to present the ministerial system as the peak of achievement of the ancient theory of mixed government. The unity of the divided powers of the State is achieved by government influence over a managed parliament. Guizot saw in the English system in 1816 a “fusion of powers,” just as Bagehot in 1867 also saw “fusion” rather than separation; but for Guizot it was a fusion which allowed the King to manage the deputies, whilst for Bagehot it was a fusion that subjected the executive to the control of parliament. Unlike Bagehot, however, Guizot believed that a separation of powers was the basis upon which the built-in limits to the exercise of arbitrary power depended, and, indeed, in 1849, when political circumstances had radically changed, he rewrote his thesis with a much greater emphasis upon the division of powers and the functions of government.76

There is no doubt that the most remarkable achievement of this period was the work of Benjamin Constant, published in 1814 and 1815. Constant accepted the need for a King and a house of peers, and so was in some sense a theorist of mixed government, but his whole approach was quite different from that of Royer-Collard or Guizot. Well to the Left of them as he was in politics, it is not surprising that the emphasis he placed upon the separation of powers was much greater, and his whole view of the relation of ministers to the King on the one hand, and to the legislature on the other, was quite different. He argued that the responsibility of ministers to the people through a majority in the assembly was an inescapable necessity in a constitutional monarchy; but if this were so how could the principle of ministerial responsibility be reconciled with the role of the King and with the separation of powers? Constant achieved this by developing a modern theory of constitutional monarchy, and by evolving a remarkable synthesis of these potentially conflicting elements. The work of Constant represents in fact a crucial turning-point in institutional theory, a turning away from the old doctrines of mixed government to a new theory of constitutional monarchy, in which the monarch assumes a new and completely different role from that assigned to him in the balanced constitution. The checks and balances of the constitution remained, but they were applied now not as checks between classes, but as checks and balances between the legislative, executive, and judicial branches of government. It is as if the approach of James Madison to the separation of powers and checks and balances in the Federal Constitution were being systematically applied to a hereditary constitutional monarchy in a democracy, but with this difference, that Constant had as great a respect for harmony in the operation of government as all his contemporaries, and so he combined this Madisonian doctrine with the operation of the ministerial system. It is remarkable to see the ideas that Blackstone had advanced in mid-eighteenth-century England, and which were adapted by the Americans to a revolutionary government without King or Lords, being reinterpreted by a Frenchman to be applied to a system in which the King was no longer an equal and active branch of the legislature vested with the executive power, but was a constitutional monarch, above politics, and separated from legislature and executive alike.

Constant defined four powers of government—the executive power in the hands of the ministers, the legislative power in the representative assemblies, the judicial power in the courts, and le pouvoir royal in the hands of the King. Constant, therefore, made a sharp distinction between the royal power and the executive power; this distinction, which he attributed to Clermont-Tonnere during the Revolution, was, he believed, the key to all political organization.77 A constitutional monarchy has the great advantage of creating a “neutral power” in the hands of the King; this power is to be used to maintain harmony between the other three branches of government. The legislative, executive, and judicial powers need to co-operate, each taking part in its own way in the general operations of government; but if instead they are at cross purposes then the King must step in to restore harmony. The means of doing this lies in the King’s prerogatives of veto, of dissolution, of dismissal of ministers, and of pardon. These prerogatives cannot be placed in the hands of one of the potential contestants, but must remain in the hands of one who has an interest only in maintaining an equilibrium between the powers of government.78 While insisting on the need to separate the powers of government, Constant reiterated the American argument that separation itself was not enough, barriers must be placed between them, in particular to prevent the legislature from exceeding its power,79 but at the same time he stressed the necessity of linking the legislature and the executive by allowing ministers to be members of the legislature. “By reuniting individuals, whilst still distinguishing powers, a harmonious government is constituted, instead of creating two armed camps.”80

Constant’s theory of constitutional monarchy combined those elements of constitutional theory which Sieyès had attempted to reconcile: the separation of powers, effective barriers to the abuse of power, and a harmonious and unified exercise of government authority. He did this, however, by adapting that very same theory of balance and equilibrium which Sieyès had so vehemently rejected. Constant did not write of a “fusion” of powers, as Bagehot was later to do; his view of government was rather more subtle. He realized that separation was as important as co-operation, and he emphasized that although ministers might become part of the legislature it is essential that the proportion of ministers to legislators should be very small—indeed that the number of ministers seated in the Chamber should never exceed one percent of the total membership.81 If all the ministers and subordinate officials were to be deputies, the Chamber would no longer be the representative of the people, but would constitute a royal council. This subtle system of division, and co-operation, checks and links, is the forerunner of the classic theory of parliamentary government, and it is a theory which clearly shows the influence of its parentage in the theories of balanced government and the separation of powers. The system was to be a nice balance between executive and legislature, which were joined but not fused by the ministerial system. A few years later Jean-Denis Lanjuinais gave further expression to the essence of this theory. The two sovereign powers of government, legislative and executive, he wrote, must never be united in the same person or body of persons, being designed to balance and to supervise each other, but their separation must never be absolute.82 Thus the partial separation of legislative and executive powers, which Blackstone made the cornerstone of the balanced constitution when applying it to King and Parliament, and which Madison used to justify the veto power of the President of the United States, was applied to the ministers and legislature of a parliamentary democracy. The ideas and vocabulary of the theory of balanced government became the basis of a balance of legislature and executive in a system in which King and nobles found a place, but it was essentially a modern system, wherein the hereditary elements of the constitution were no longer the equal and essential powers of government that they once had been. It is important to emphasize the role of the separation of powers in this embryonic theory of parliamentary government; for, although it was recognized throughout the nineteenth century as the basis of the balance of power established in the parliamentary system, this point was to be confused and obscured when Bagehot made his indiscriminate attack upon the ideas of mixed government and the separation of powers alike.

The years 1814–48 in France were a period when the idea of mixed government shading into parliamentary government dominated the field of political thought, but, although the transition from the government of Louis XVIII and Charles X to the July Monarchy of Louis-Philippe reflected this changing pattern, it was unfortunately true that by 1848 the existing system of government was corrupt, was based upon an extremely narrow franchise, and hardly measured up, in the eyes of most Frenchmen, to the benefits that a system of “equilibrium” was supposed to bring. The demands for reform, together with the deeper social unrest that was developing, led once again to a revolutionary situation in which the trappings of monarchy, and the attributes of the system of mixed government, were rejected. In earlier revolutionary situations of this sort since the mid seventeenth century, where the mixed and balanced constitution had been under attack, the doctrine of the separation of powers surged up into a dominant position to fill the vacuum left by the rejected theory of government; and in 1848 the same pattern was repeated, except that on this occasion an attempt was made to combine the system of ministerial responsibility with a constitution which in other respects was patterned upon the pure doctrine of the separation of powers. It is hardly surprising that this last flirtation with the pure doctrine ended in the same way as others had ended in France—in absolutism.

There had been continual attacks upon the system of contrepoids, or equilibrium, since the restoration of the monarchy,83 and, during the 1830’s, in his lectures on constitutional law in Paris, Pellegrino Rossi had emphasized the importance of the separation of powers;84 so it is hardly surprising that the Constituent Assembly of the Second Republic turned to the separation of powers with almost as much faith as the Constituents of 1789. The Constitution of 1848 proclaimed that “the separation of powers was the first condition of a free government,” and it is clear from the debates in the Assembly that the majority of deputies saw it as the only possible basis of a system without King or nobles, which would yet avoid the ghastly spectre of the Convention. The Constitution as it stands on paper embodies a strict division of power between a unicameral legislature and a popularly elected President. Once again all the checks and balances of the balanced constitution were swept away. The President had no veto, no power of dissolution. On the other hand the attempt by Félix Pyat on the extreme Left to subordinate the executive to the Assembly and deny the validity of the separation of powers altogether was strongly rejected.85 The proposals of Parieu and Grévy to create a purely parliamentary executive were also refused. In an impassioned speech Lamartine supported the idea of a President directly elected by popular vote. Those who wished to allow the legislature to choose the executive, he thundered, should take their ideas to their logical conclusion: “do not merely confound in yourselves the legislative and executive powers, take also the judicial power, and then call yourselves by your correct name, the Terror.”86 The fear of the Convention was enough to ensure the defeat of Grévy’s proposal by 643 votes to 158.

Yet there was a great deal of confusion in the minds of the deputies on the proper structure of the government. Lamartine, for all his use of the terminology of the separation of powers to defeat a proposal of which he disapproved, did not really support the ideas behind the doctrine. Elsewhere he spoke of the division of functions rather than the division of powers, and during the July Monarchy he had persistently opposed attempts to exclude placemen from the Chamber. Indeed, five months after his defence of an elected presidency he told the Assembly that there was only one power of government, and this would be enfeebled by attempts to separate it into two.87 Lamartine, whom de Tocqueville labelled the most insincere of men, no doubt modified his views to suit his ambitions, but de Tocqueville himself contributed to the confusion over the structure of the government. As one of the principal authors of the Constitution, he supported, if in a very lukewarm way, the direct election of the President, but he wished also to introduce a system of ministerial responsibility to the Assembly that would have made the President a mere constitutional figure-head. He defended this strange combination, which in a somewhat different guise was to re-emerge in the course of the life of the Fifth Republic, as a completely new system of government. Till 1848, said Tocqueville, there had been two distinct systems of constitutional government; in the one the King could do no wrong, but his ministers were responsible, in the other, as in the United States, the chief of the executive was himself directly responsible, but could “act freely.” In the new Constitution a responsible President had beside him a council of ministers equally responsible without whom he could do nothing, and who could reduce him to the impotence of a constitutional monarch.88 It seems rather unlikely that the National Assembly accepted this remarkable scheme, which one member described as a monster with two heads, in the spirit de Tocqueville intended, although it is true that after the election of Louis-Napoleon to the presidency the Assembly refused to exclude ministers from membership in the legislature because the Constitution, in Billault’s words, was intended to obtain all the advantages of a republic and a constitutional monarchy.89 The majority of republicans in the new Assembly of 1849, overawed by the popular victory of Louis-Napoleon, and out of deference to the separation of powers, tolerated a minority Ministry, and when the minority turned into a majority the President showed his independence by dismissing his ministers, although they had not been defeated in the Assembly.90 The Bonapartist historian, de Cassagnac, describing the ending of the system of ministerial responsibility, said, “The Constitution wished to make the President responsible; France wanted him to govern.”91

The Revolution of 1848 brought an end to a period of development in political theory which had extended over two hundred years from the outbreak of the English Civil War. This period was one in which the two theories of constitutionalism had played out a complicated and difficult game. The proponents of both theories had rejected the absolutism of extreme Right and extreme Left, the exercise of arbitrary power by either a monarch or an all-powerful assembly. Yet they became bitterly divided whenever the attack upon hereditary privilege reached a point where it exploded into revolution, for in such an extreme situation only the separation of powers could claim to be an adequate theoretical basis for a free constitution. For two hundred years the theories of balanced government and the separation of powers were the only serious contenders in the field of rival theories of constitutionalism, but the nineteenth century was to see a fundamental change in this situation. In England and France the theory of the balanced constitution was to be transformed into a theory of parliamentary government in which most of the elements of the older theory were retained, but in which the monarchy and aristocracy of the old theory became in England merely “the dignified parts” of the constitution, or in France were wholly transformed into a republican garb. In the United States the theory of the balanced constitution had been transformed into something uniquely American, which was to follow its own rules of development. The element of mixed government in the theory of the balanced constitution was, therefore, largely suppressed, but the separation of powers continued to be an essential ingredient in the structure of Western government. The pure doctrine had finally been rejected as unworkable in favour of more complex systems, which borrowed heavily also from the theory of balanced government.

The explanation of the rejection of the pure doctrine of the separation of powers is not to be found solely, however, in this development of new institutional theories of government. The Revolution of 1848 in France had revealed new social and political movements that were to change drastically the environment in which constitutional theory must operate. The year before the Revolution the publication by Marx and Engels of the Communist Manifesto indicated the course that political thought would now be taking, with less and less emphasis upon legal and constitutional forms, and more and more attention given to the social and economic factors in political life. In the two hundred years before 1848 the concept of political liberty which had dominated political thought was essentially a negative one, a desire for freedom from restraint, and particularly from the arbitrary restraints of government action. “Liberty is a right of doing whatever the laws permit” said Montesquieu, and the desire to gain control of the making of those laws, and to ensure that no one, including each of the elements of the government itself, could exceed the rules laid down by law, was the driving force behind the whole movement that produced the theory of the separation of powers in all its varieties during this period. But the Revolution of 1848 ushered in a new view of political liberty of a more positive kind; the right to work, Louis Blanc’s conception of a fundamental duty of the State to provide its citizens with the means of livelihood, which was taken up so fiercely by the people of Paris in 1848, gave a completely new twist to the notion of “a free constitution.” Freedom from restraint alone was no longer enough. The idea that the State should concern itself with creating the environment in which its people would be free to live and develop a full life (whatever that might mean) came to dominate more and more the thought of the nineteenth and twentieth centuries. Such a philosophy of freedom had, however, little in common with the motives of the pure doctrine of the separation of powers, the whole concern of which was essentially negative in its conception of a constitutional provision for liberty. The pure doctrine was, in conception, more laissez-faire even than its rival theory of the balanced constitution, and it was this fact that ensured that, in the long run, the pure doctrine would be rejected in favour of constitutional theories which contained strong elements of the theory of the balanced constitution, elements which could be turned to provide co-operative links between the parts of government rather than mere checks.

After 1848, then, the history of institutional thought was to become more and more concerned with the problem of creating a government machine for an age of collectivism, and the theory of the separation of powers came under fierce attack, as a barrier to the development of an efficient machine of this sort. The next step in the history of the doctrine is therefore to examine the criticisms and attacks to which it, and the institutional structures that incorporated it, were to be subjected in the nineteenth and twentieth centuries, before we may turn to a consideration of the role the theory still plays today in Western thought and institutions.

[1. ]Social Contract, II, ii, p. 27. Quotations are from Rousseau: Political Writings, ed. by F. Watkins, London, 1953.

[2. ]Ibid.

[3. ]Ibid., III, i, p. 59.

[4. ]Ibid., III, i, p. 60.

[5. ]Ibid., III, iv, p. 71; III, v, pp. 73–74.

[6. ]Ibid., III, vii, p. 83.

[7. ]Ibid.

[8. ]Ibid., III, iv, p. 71.

[9. ]Ibid., III, v, p. 74.

[10. ]Ibid., III, i, p. 61.

[11. ]Paul de Mellis, Le principe de la séparation des pouvoirs d’après l’Abbé de Mably, Toulouse, 1907, pp. 25 and 34.

[12. ]Droits et devoirs du citoyen, 1758, Oeuvres, Vol. 11, p. 474.

[13. ]De l’étude de l’histoire, 1778, Oeuvres, Vol. 12, p. 232.

[14. ]Observations sur les Romains, 1751, Oeuvres, Vol. 4, pp. 280–1.

[15. ]Mellis, op. cit., pp. 113–14.

[16. ]See E. Carcassonne, Montesquieu et le problème de la constitution française au XVIIIe siècle, Paris, 1927, pp. x–xii.

[17. ]G. Bonno, La constitution britannique devant l’opinion française de Montesquieu à Bonaparte, Paris, 1932, pp. 142 ff.

[18. ]Archives parlementaires, 1st Series, Vol. 8, p. 243.

[19. ]Ibid., Vol. 8, pp. 514–22.

[20. ]Ibid., Vol. 29, pp. 399–400.

[21. ]Ibid., Vol. 8, p. 243.

[22. ]Ibid., Vol. 9, p. 70.

[23. ]No. 41, Sept. 1789, p. 2.

[24. ]Op. cit., p. 14.

[25. ]Archives parlementaires, 1st Series, Vol. 8, p. 259.

[26. ]Ibid., Vol. 8, p. 595.

[27. ]Ibid., Vol. 8, p. 596.

[28. ]See the discussion by L. Duguit, “La séparation des pouvoirs et l’Assemblée nationale de 1789,” Revue d’Economie Politique, Vol. 7, 1893, pp. 364–72.

[29. ]Ibid., pp. 571–3.

[30. ]Ibid., pp. 601–3.

[31. ]See A. de Tocqueville, L’ancien régime, Oxford, 1904, pp. 60–62.

[32. ]On this period see P. Bastid, Le gouvernement d’assemblée, Paris, 1956, pp. 135 ff.

[33. ]Archives parlementaires, 1st Series, Vol. 52, p. 128.

[34. ]Ibid., Vol. 52, p. 225.

[35. ]Ibid., Vol. 60, p. 91.

[36. ]Bastid, Le gouvernement d’assemblée, pp. 153, 158.

[37. ]See the discussion of the Convention under the Committee of Public Safety, as a system of parliamentary government, by B. Mirkine-Guetzévitch: Le gouvernement parlementaire sous la Convention, in Cahiers de la Révolution française, No. VI, 1937, pp. 66 ff. However, as Robert Villers has pointed out, neither Danton nor Robespierre exercised the power of dissolution which characterizes the classic system of parliamentary government, and gives to the parliamentary executive some degree of autonomy. “La Convention pratiqua-t-elle le gouvernement parlementaire?” Revue du droit publique, April–June 1951, p. 386.

[38. ]Le Moniteur Universel, No. 194, reprinted Paris, 1841, Vol. 20, p. 114.

[39. ]Ibid., pp. 115–16.

[40. ]Archives parlementaires, 1st Series, Vol. 62, p. 288.

[41. ]Ibid., Vol. 62, p. 358.

[42. ]Ibid., Vol. 62, p. 359.

[43. ]Ibid., Vol. 62, p. 573.

[44. ]Art. 29.

[