Front Page Titles (by Subject) ELEVEN: Political Theory, Constitutionalism, and the Behavioural Approach - Constitutionalism and the Separation of Powers
The Online Library of Liberty
A project of Liberty Fund, Inc.
Search this Title:
ELEVEN: Political Theory, Constitutionalism, and the Behavioural Approach - M.J.C. Vile, Constitutionalism and the Separation of Powers 
Constitutionalism and the Separation of Powers (2nd ed.) (Indianapolis, Liberty Fund 1998).
About Liberty Fund:
Liberty Fund, Inc. is a private, educational foundation established to encourage the study of the ideal of a society of free and responsible individuals.
The copyright to this edition, in both print and electronic forms, is held by Liberty Fund, Inc.
Fair use statement:
This material is put online to further the educational goals of Liberty Fund, Inc. Unless otherwise stated in the Copyright Information section above, this material may be used freely for educational and academic purposes. It may not be used in any way for profit.
Political Theory, Constitutionalism, and the Behavioural Approach
The twentieth century is an age of cynicism and scepticism as far as political theory is concerned. It is argued that political “theory” is in fact little more than the expression of opinion or prejudice, or, to put it in another way, is the expression of an ideology which is not amenable to proof or disproof. The sceptic sees the antithesis of political theory in the strictly empirical studies which have come to be known as “the behavioural approach.” At the extreme, these two approaches to political phenomena are seen as wholly unrelated and irreconcilable. Any attempt, therefore, to assess the relevance of ideas of constitutionalism in the mid twentieth century must take account of the arguments of the behaviourist, and attempt to place such ideas in a meaningful context, for the attacks which have been made upon the separation of powers over the past century have not been concerned merely with a critique of its concepts, but have been associated with the rise of the behavioural school and its philosophical forbears, whose attack has been directed towards the very foundations upon which constitutional theories of the past have been based. This is hardly surprising. A theory which has at times claimed the status of a law of nature, or which has announced universal prerequisites of good government, must expect to come under attack when, for many people, the hallmark of a meaningful proposition comes to be the extent to which it can be scientifically tested. Clearly the relevance of this criterion to theories of government like the separation of powers must be explored.
Although at times raised to higher metaphysical levels, the discussion of the separation of powers, and related theories of limited government, has usually been carried on in empirical terms. The evidence that has been summoned in their defence has been the experience of history, the assumed knowledge of human nature, or the workings of contemporary systems of government. Yet given the kind of knowledge we can have about history, human nature, and politics, exactly what would constitute “proof” or “disproof” of theories is open to considerable doubt. Propositions concerning the nature of governmental organization, as broad as those made by proponents of the separation of powers, present a rather different problem of empirical verification from that of a proposition about voting behaviour in Greenwich or Elmira at a particular point of time. Can the same criteria be applied to both types of proposition? In general, political theorists would point out that theories of constitutionalism are made up of a number of different types of proposition, not all of which could be subjected to the same rigorous treatment as studies of voting behaviour. Their antagonists argue that if this is the case, then nothing “scientific,” and, therefore, they imply, nothing of value, can result from this type of theory.
The behavioural attack upon political theory has been accompanied by a related, but separate, attack upon the idea that “constitutions” play an important role in the operation of the political system. The behaviourist’s concern with “social forces,” and his emphasis upon the real stuff of politics, seem to lead him to the view that the behaviour which he sees as the sole content of politics is not in any significant way affected by the structure of constitutional rules, but is wholly determined by economic, racial, class, and other factors. This is not simply an attack upon an outdated legalism in the study of politics. He is not concerned simply to point out that the formal structure of a “constitution” may be a very poor indicator of the actual operation of a political system. The rejection of the significance of constitutions goes far beyond this, to encompass much of what would normally be called political “institutions.” We shall examine in some detail the most distinguished work of this sort, Robert A. Dahl’s Preface to Democratic Theory, in which the definition of “constitutional” is given as the “prescribed rules influencing the legitimate distribution, types, and methods of control among government officials.” This is no mere attack upon formalism, for Dahl includes all those rules which may be prescribed “by a variety of authorities accepted as legitimate among officials: the written Constitution, if there is one; decisions of a tribunal accepted as authoritative on constitutional interpretation; respected commentaries and the like.”1
Theories of constitutionalism are attacked, therefore, both as being incapable of rigorous proof and because they deal with insignificant factors in political life. They are further subjected to attack because of their overt normative content. This objection, however, will not detain us for long. One may disagree with the recommendations of the constitutionalist on the grounds that one does not share his values, or that one does not accept the logic of his arguments, but of the importance of the views of writers on political and constitutional theory in the past, and of their role in helping to shape the patterns of history, there can surely be little doubt.2 However, it has been pointed out by J. C. Rees that political theories, which are of necessity formulated in very general terms, cannot be used to deduce unequivocal courses of action in relation to specific cases.3 Rees has in mind such general statements as that of Sir Ernest Barker that the purpose of the State is to promote “the highest possible development of all the capacities of personality in all of its members.” Although the theories of limited government, such as the separation of powers, have claimed much greater precision and much more empirical relevance than could be claimed for the above statement, nevertheless Rees’s objections do impinge upon these theories, for, as we have seen, it has been found to be virtually impossible in practice to deduce from the doctrine of the separation of powers any precise criteria for allocating the tasks of government to its various branches, even though jurists have striven to give such precise operative interpretations to these concepts.
Although it may be true of constitutional theories, as of political theories in general, that it is impossible to move from general propositions to unambiguous statements about particular practical issues, this does not by any means prove that such theories have no value. They can provide guiding-lines within which the solution of practical issues must be found; they help to create an intellectual atmosphere in which certain courses of action will be excluded, even though the choice between the remaining alternatives is not precisely determined; they make self-conscious the values which are to be given priority, and the choice of a means of furthering these values. Such theories have the same relation to political action as military strategy has to tactics. The strategy which has been laid down will rarely give the tactician an unequivocal course of action to be followed in a particular situation, but he does have to reconcile his tactical decisions with the ultimate aims which have been set for him by the strategist. The mere fact that there are a number of tactical courses open to him, all of which seem to be broadly consistent with the over-all strategy to be followed, does not mean that there is no relationship between strategy and tactics. Thus political theorists can set the aims, can rule out certain courses of action, can set certain limits, but like the military strategist they will be the more successful the more they keep in mind the hard facts of the terrain they survey.
The attack upon “constitutions” as significant elements in political life may be illustrated by the work of three widely separated authors—William Penn, Sir Lewis Namier, and Robert A. Dahl. It is not suggested that these men are fully representative of “the behavioural approach”; indeed, it is the extreme nature of their expression of “behaviourism” that makes them useful as a basis for discussion of the importance of constitutional structures. The traditional or institutional approach to the study of politics has at times been subjected to extreme criticism; but in recent years this has given place to a recognition that both institutional and behavioural studies are essential to a balanced approach.4 Nevertheless, these three writers provide us with the three main characteristics of the extreme approach—the emphasis upon “human nature” to the exclusion of the mechanics of government, the underlying assumption that it is the relative power of “groups” that provides the independent variables of the political system upon which all other factors depend, and the consequent belief that institutional structures can have little or no significant effect upon the outcome of political situations.
We might begin our discussion of constitutional theory and its significance by considering the views of an early “behaviourist,” William Penn, writing in 1682. Penn was himself engaged in drawing up a form of government for the province which had been granted to him the previous year by Charles II, yet he took the opportunity in the Preface to his Frame of Government of the Province of Pennsylvania to express an extreme scepticism about the value of constitutional restraints upon political action. He dispensed with any discussion of “particular frames or models” of government: “There is hardly one frame of government,” he wrote, “so ill designed by its first founders, that in good hands will not do well enough; and story tells us the best in ill ones, can do nothing that is great or good.” For Penn governments were machines that, like clocks, work according to “the motion that men give them.” Thus he concluded, “Let men be good, and the government cannot be bad; if it be ill they will cure it. But if men be bad, let the government be never so good, they will endeavour to warp and spoil it to their turn.”
It is interesting to find this point of view expressed by a seventeenth-century Englishman so soon after so much blood had been shed over constitutional principles, for it represents a popular view of the British system of government in the mid twentieth century—principles are unimportant, all that matters is the character of those who run the government. However, such a view leads to a dilemma which was characteristic of the government of Pennsylvania in the eighteenth century, and is perhaps the problem of British government in this century. That is, one either places one’s trust in the goodness of “the people,” or, if this view is considered too ingenuous, one is led to a reliance upon government by an élite, which, because of its internal moral code, can be trusted to govern justly and well. However, when the élite begins to crumble, and is replaced, the values that might have been protected by a habit of constitutional thought find no protection in a society which has been taught to despise constitutionalism.
The basic inadequacy of Penn’s view of political systems is its over-simplification. Societies are not composed solely of “good” men or of “ill” men, but are complex collections of human beings most of whom are both “good” and “bad,” if such simple terms can convey very much of the fine shades of morality to be found throughout the political life of society. The essential point about constitutions is not that they could restrain a society full of bad men, but that they may channel political behaviour in certain directions rather than others, that the ordinary citizen will not be subject to the whims of good or bad men, but will have some certainty of essential continuities of action when the personnel of government changes. The aspiration towards a government of laws and not of men is inherently incapable of being realized, but a government of men subject to the restraints of certain rules is not.
William Penn’s over-simple view of the relation between human nature and “frames of government” is not, of course, the position adopted by modern academic behaviourists, but their attitudes may be equally extreme. The emphasis upon “social forces” may be taken to the point where “government” seems to disappear altogether. Sir Lewis Namier, whose work might be seen as the first flowering of the present-day behavioural approach, has provided us with as extreme an expression of this point of view as it is possible to imagine. Live forces, Namier wrote, break through forms, and shape results to suit requirements. “Were it decided that the 615 heaviest men in the country should constitute the House of Commons, the various interests and parties could be trusted to obtain their proportionate weight in it.”5 This is so extreme that merely to state it is for it to be seen to be absurd. Namier could hardly have believed that the failure to pass the Reform Acts would have made no difference to the decisions which were actually taken in Parliament in the mid and late nineteenth century. What then was the battle over the Reform Acts about? Why should the “live forces” have bothered with such a trivial matter as the franchise? Surely the failure to reform Parliament would have led eventually to revolutions like those that took place in other European countries, yet if we accept Namier’s viewpoint revolutions would never occur, for they would serve no useful purpose. “Live forces” would inevitably attain their proportionate weights. Louis XVI would have been transformed into the President of the Republic without all that messy business of the Revolution, and a Czar of Russia would now be the guiding spirit of Russian economic planning. Yet revolutions do occur, and they occur because the nature of the political machinery, the composition of the parts of government, the accepted rules by which power is allocated, do make a very great difference to the content of government decisions, and to the outcome of political situations. To capture the machinery of government and to turn it to their own ends is the aim of revolutionaries, and it is, for them, a battle worth fighting. But Namier, and the American behaviourists, lived, or live, in a society where revolutions have long ceased to occur, and have assumed therefore that the stability provided by the political system was unimportant, whereas in reality it was the political system which was the whole context for their behavioural studies. It is unlikely that their point of view would be shared in those ages, or in those parts of the world, where the machinery of government did not allow the free interplay of the groups who form the society.
Thus the basic mistake of the Namier position is the assumption that there are absolute “proportionate weights” for the various parties and interests in society, which exist independently of the nature of the political system, or of the channels through which they must seek expression. But what are these absolutes, and how does one measure them? This is to seek a philosopher’s stone of “power” that is surely chimerical. Social forces must be seen and evaluated within a particular set of relationships, without which they cannot have existence or meaning. Perhaps only in a truly revolutionary period can there be a situation in which, through the exercise of naked force, a true “absolute” can be seen to emerge, and then only momentarily. In normal circumstances the “proportionate weight” of a party or group will depend to a very considerable degree upon the institutional patterns through which it must operate, the extent to which it has freedom of manoeuvre, and the influence it has upon the decision-making process. One can see, for example, that the “weight” of the French Communist Party in French politics since 1945 has been affected deeply by the nature of the régime, by the electoral law, and by the extent to which it was allowed to have a direct influence upon the way in which government decisions were taken. But how would one arrive at a statement of its “true” weight in these years? A discussion of its political role, except in the context of changing régimes and electoral machinery, would be meaningless. On the other hand, the history of the Third and Fourth Republics illustrates the limits upon what can be achieved if one attempts to alter the behaviour of parties and groups by tinkering with the institutional machinery. It is this complex inter-relationship of groups and structures which is the very stuff of political studies, but Namier virtually dismisses it altogether.
There is clearly a sense in which political behaviour is the sole, exclusive content of the study of politics—there is nothing else to study. Structures, institutions are patterns of behaviour. Ideas, like art, can be studied for their own sake, but the student of politics will be interested in them only in so far as they have affected, or will be likely to affect, the behaviour of men in society. Equally, material things will concern the student of politics only in so far as they are objects of political action. The stones and mortar of the Palace of Westminster or of the Capitol are of political interest only in so far as they inspire awe or revulsion in those who make political decisions in their vicinity. The pieces of paper which circulate in government offices in Paris, Washington, or London have no political importance except for the reactions of men and women to the inscriptions upon them. Money, steel, land, are not in themselves subjects of political study, but only in their effects upon political behaviour; otherwise they are the concern of the economist, the metallurgist, or the agriculturalist. Thus we are all behaviourists. Political institutions are not tangible structures of steel or wood, or even of papers held together by red tape, although these materials are incidental to these institutions. The orb and the sceptre, the seal of the United States, the flag or the ballot-box are all merely the outward symbols of the things men believe in or manipulate.
A political structure then is behaviour, but it is not random behaviour, it has a pattern. Furthermore, it is patterned behaviour of a peculiar stability and consistency, behaviour which follows certain rules whether explicit or implicit. To emphasize the importance of these rules, and the need for stability in the patterns of behaviour they regulate, is an essential aspect of constitutionalism. This is not to equate constitutionalism with conservative attitudes in politics, it is merely the recognition of the basic requirement of order in a political system. The satisfaction of the varied wants of mankind entails a society in which the future is relatively predictable, in which plans can be made and brought to fruition. Just as the fulfilment of economic wants is dependent upon a stable economic organization that will enable the seed which is planted to become bread upon the table, so all our political needs, which include the economic ones, require a stable political organization for their satisfaction. Anarchy, the absence of all rules, and therefore of predictability, is the frustration of any hope of attaining those material and spiritual satisfactions, including that of simply staying alive, which are the ultimate aims of all political action. In anarchy none but the psychopath gains satisfaction. The continued satisfaction of wants, and the expectation of future satisfactions, demands a stable, rule-governed pattern of political organization of behaviour.
Order and the political system, therefore, are synonymous. But political systems can take many forms, and the values they embody can vary widely, so that the type of future satisfactions they hold out can also differ. The arbitrary rule of a despot may be preferable to the war of all against all, but once man is protected from the unpredictable behaviour of his fellow subjects, he begins to look for further stability and predictability, that is to say in the behaviour of his governors. He demands the security of knowing how decisions will be reached, who will be consulted, the procedures which will be employed, the composition of the bodies that will decide. This is not, and cannot be, a matter of the expediency of group interaction, or the casual outcome of whatever “live forces” happen to be predominant at a particular point of time. It must be a process subject to rules, as vital as, although different in character from, the laws which govern one’s relations with one’s neighbour. This is a “constitution,” and its scope is, of course, much wider than any written constitutional law as it is usually developed by lawyers. It is the secondary layer of rules in a political system, which determine how things shall be done, how decisions are to be made.6 Just as anarchy is the antithesis of the legal system, so are despotism and totalitarianism the antithesis of the constitutional State, for they represent an “organized anarchy” at the secondary level, unpredictable and uncontrolled.
The expression of the behavioural approach in the quotation from Namier given earlier, errs, therefore, in underestimating the requirement of order in the political system, and in taking for granted the institutionalized order of constitutional States like Britain and the United States. Behaviourists concentrate upon what actually happens within the ordered system provided by the constitution, ignoring the long-term stabilities which are the context of, and the prerequisite for, such behaviour. They tend to ignore the restraints that the requirement of order places upon the political situation at any one point of time. It is this set of restraints which crystallizes institutions or structures, which sets the limits within which the “live forces” or the “social forces” must operate. Indeed, it is difficult to see how behavioural methods of investigation could handle the problems presented by these structures. The student of voting behaviour, for example, in attempting to answer the question why the electorate voted in a particular way at a particular election, has developed very refined tools of investigation to deal with this narrowly defined problem. But to answer the question of why an individual voter chose candidate X rather than candidate Y is very different from trying to answer the question of how it came to be that the choice was between X and Y in the first place, and not a choice between completely different alternatives. Such a problem can only be dealt with by using the whole range of techniques available. We should not allow our devotion to certain techniques to rule out vitally important problems of political enquiry.
Robert A. Dahl has presented us, in his Preface to Democratic Theory, with a less extreme view of the behavioural approach than Namier’s, and one which, because he uses James Madison’s thought as a starting-point for his analysis, deals explicitly with the doctrine of the separation of powers. Dahl’s main point is that “the first and crucial variables to which political scientists must direct their attention are social and not constitutional.”7 As we have seen, Dahl’s definition of “constitutional” is a fairly wide one, but he does not attempt to deal with the general structure of political institutions, except in so far as he discusses the idea of the separation of powers. Political institutions in general seem to hang in a limbo between “social” and “constitutional” factors.
Dahl takes as his starting-point a theory that he labels “Madisonian,” the central thesis of which is that there must be a constitutional separation of powers if the tyranny of the one, the few, or the many is to be averted. Dahl reassembles Madison’s statements to form a number of definitions and hypotheses, in order to present the theory in a more coherent form, and to be able to test its central hypotheses. The principal hypothesis is that, if unrestrained by external checks, any given individual or group of individuals will tyrannize over others, and therefore that the first condition for the establishment and maintenance of a non-tyrannical republic is that the accumulation of all powers, legislative, executive, and judiciary, in the same hands, whether of one, a few, or many, and whether hereditary, self-appointive, or elective, must be avoided.8 Dahl’s attack is directed to showing that this position is “demonstrably false,” and to proving that the relative importance of constitutional rules in general is trivial compared with that of social forces. Thus in discussing the extent to which minorities are protected from oppression by governments he concludes that “if constitutional factors are not entirely irrelevant, their significance is trivial as compared with the non-constitutional.”9
The problem of sustaining this point of view depends very much upon the level at which the discussion is conducted. If Dahl meant merely that “paper constitutions” are in themselves unimportant, we could easily accept this point, but clearly he means much more than this. Yet the “proof” of the demonstrably false Madisonian position which Dahl offers is a very dubious one. So far as Madison is dealing with “constitutionally prescribed authority,” the only empirical evidence which Dahl produces, in order to dismiss his hypothesis as invalid, is to argue that parliamentary systems “like that of Great Britain,” which are certainly non-tyrannical, readily prove that Madison’s first condition for avoiding tyranny is unnecessary to achieve that aim. But Dahl does not discuss the British example to prove, as he must do to sustain his argument, that in Britain all power, legislative, executive, and judicial is accumulated in one set of hands. He seems merely to assume that this is the case, but this is a complicated matter which can hardly be treated a priori. Again, it is very much a matter of levels of discussion. At a later stage it will be suggested that the practical operation of British government today does suggest doubts about undue concentration of power through the medium of the party system; but at the level of constitutionally prescribed authority, at which Dahl conducts his discussion, it is by no means clear that all power is accumulated in this way in Britain. Dahl cannot mean to equate the idea of the separation of powers with the explicit example provided by the Constitution of the United States. Certainly Madison had no such idea, and it is Madison’s formulation of the doctrine that Dahl uses as a basis of discussion. If Dahl were thinking in terms only of the precise American formulation of the separation of powers then there could be no argument, for clearly this is not essential to the prevention of tyranny. But Madison never asserted this. Indeed, he used the British Constitution of those days as an authority to which to appeal in support of his argument that the accumulation of all power in one set of hands must be avoided. Dahl, however, seems implicitly to adopt Bagehot’s view that in Britain all the powers of government are fused, and does not push the empirical approach to the point of investigating the validity of this proposition.
It is by no means clear that the constitutional allocation of power differs so fundamentally in Britain and the United States as Dahl suggests. Thus if we think in terms of the personnel of the branches of government, although there is not in Britain the complete separation of the United States Constitution there certainly is not a complete fusion of the personnel of the three branches, either. Indeed, in spite of recent government legislation which increased the numbers of ministers in the House of Commons, there are rigid rules preventing civil servants from being members of the legislature, and to maintain a separate judiciary except for the Lord Chancellor and the Lords of Appeal. British government would be very different if these rules did not exist. If Dahl has in mind the legislative supremacy of the “King-in-Parliament,” this concept also must be treated with care. As we have seen, legislative supremacy has never meant the dictatorship of one man or of either of the Houses of Parliament, except perhaps during Tudor times, or in the reign of the Long Parliament, neither of them very good precedents for modern British government. Constitutionally the “King-in-Parliament” is composed of a number of parts, and there has always been a distribution of power among them. Thus even in the United States a legal sovereign power exists, which, however quiescent it has been over the years, is clothed with an authority equal to that of the King-in-Parliament.10 All power, legislative, executive, and judiciary, is accumulated into the hands of this “body,” although it is certainly a great deal more difficult to organize effectively than the constituent parts of the British constitutional structure.
This, of course, is a highly formal legalistic argument, which may, or may not, have relevance to political reality; but it does point to the conclusion that a useful comparison of the British and American political systems in this respect must be based upon a close analysis of the working of political institutions and the party systems of the two countries, rather than upon superficial comparisons of their legal structure. We shall have to consider at a later stage the relationship between the legal divisions of power in these countries and the operation of their party systems, and the extent to which the one may exercise restraints upon the other, but it is certainly not possible to discuss the significance of some form of “separation of powers” in Britain without such an investigation.
Thus far, then, the main objections to Dahl’s analysis are that he does not define the separation of powers clearly enough, nor does he apply empirical rigour in his “disproof” of Madison’s hypothesis. But we must follow through Dahl’s criticisms of Madisonianism, because they reveal some of the most significant problems of constitutional theory and the behavioural approach.
Dahl represents “Madisonianism” as if it were a concern merely for an abstract constitutional doctrine of the separation of powers which would of itself obviate tyrannical tendencies in government. But Madison was in fact much too aware of the importance of “social forces” to argue in this fashion. Indeed, Madison, in the Federalist Nos. 47–51, from which Dahl draws his “Madisonian” hypotheses, was not so much concerned with asserting the importance of the separation of powers, as with insisting upon the fact that it must be modified and buttressed with checks and balances reflecting social factors. These papers from the Federalist were the great climax to the campaign against a dependence upon “parchment barriers” to the exercise of power, which had begun at the very moment when the revolutionary State constitutions were being written. Madison argued forcefully, and with a good deal of empirical evidence from the experience of Pennsylvania and Virginia, in support of the view that constitutional rules not buttressed by institutionalized structures of real power, related to social reality, would be worthless. He was in fact attacking that pure doctrine of the separation of powers which Dahl attributes to him, and attacks in his turn. But, of course, Madison does not push the argument to the same extremes as Dahl does. Constitutional rules without effective sanctions are worthless in Madison’s eyes, but when they are fashioned so as to correspond to the social basis of the political situation, they can and do make a considerable difference to the way in which governments operate, and the effect they in turn can have upon social forces.
Thus although Madison was wrong in detail about the way in which the institutions of the new government would work, he was right in the long run about the importance of the institutional structure as a means of channelling political activity in certain directions. He did not “deduce” checks and balances from the separation of powers as Dahl suggests,11 he modified the doctrine with ideas drawn from the balanced constitution of Britain, with its keen appreciation of the role that institutional structures could perform in a heterogeneous society. The application of the pure separation of powers in the revolutionary State constitutions had failed, because it depended purely upon an abstract formula for its effectiveness. Therefore, Madison argued, although each branch of government ought really to be invested with “a will of its own,” and should ideally draw its authority direct from the people, through channels having no communication whatever with one another, this in practice would provide certain difficulties, in particular with regard to the judiciary. A better scheme would be for the legislature to be divided, with different modes of election for the two Houses, and for the separately elected President to be invested with a veto power.12 This structure must be related to the division of power between the Federal and State governments, and to the breaking-up of the electorate into many “interests and classes of citizens.” It is difficult to see how it can be asserted that this constitutional system has been of little significance in the development of the American polity. In the light of American history it is a brave man indeed who can assert that it would not have mattered very much if the President had not been given a veto, or if the Supreme Court had been forbidden, as were the French courts in 1791, to invalidate legislation.
Thus Dahl questions whether the separation of powers can provide effective external checks upon the tyrannical impulses of officials; loss of status, respect, prestige, or friendship would be ineffective, he argues, the monetary motive is ruled out, and coercive action against officials would hardly be in question.13 What sanctions would there be? Madison, however, saw the problem more clearly. “Ambition must be made to counteract ambition,” he wrote,14 and this must surely be seen as the key to the antagonisms between President and Congress, between Senate and House of Representatives, between Congress and the bureaucracy, and even between all of these and the Supreme Court. Ambition, whether it be simply to wield power, to serve the people, to make money, or to make history, this has been the continuing dividing force that has operated through the various channels which the Founding Fathers provided, in order that no matter how tall one man’s ambitions became, there would always be an institutional basis upon which other ambitious men could oppose him. At times one feels that Dahl believes these institutionally-supported ambitions to have been all too successful in directions of which he disapproves,15 whilst his main argument is directed towards showing that they could not have been significant. But if it is his intention to argue against the results produced by the American system of government, then this is inconsistent with his general argument that the nature of the system does not really matter.
We arrive, therefore, at the central issue between Dahl and Madison—whether a particular type of constitutional structure is, or is not, a prerequisite of a non-tyrannical republic. The definition of “tyranny” is in Dahl’s view a crucial problem of the Madisonian argument. As he points out, Madison does not give a precise and specific meaning to this concept, so that the validity of hypotheses which relate to “the prevention of tyranny” cannot be established by rigorous empirical tests. He concludes, therefore, that Madisonianism is an “ideology” rather than a “political theory,” a significant and important ideology, which has played, and will continue to play, a role in politics, but which has no contribution to make to political science.16 Yet surely the inability to make precise a definition of tyranny that will be operationally effective does not dispose of the problem with which Madison was concerned. The concept of tyranny, or the rather less dramatic idea of “the abuse of power,” may not be susceptible of very precise definition: yet of the historical fact of the existence of tyranny or abuse of power there can hardly be much doubt. The existence of tyranny cannot be denied because we are unable to say precisely where it begins and where it ends. It is a fact, of course, that there can be no absolute definition of these terms, for they are relative to the particular period and the particular culture being considered. There are inescapable value-judgements here, and we must accept that a discussion of constitutionalism can begin only by pointing to certain specific examples of societies which are asserted to be non-tyrannical, and to attempt to elucidate their major characteristics. This, in fact, was the procedure Madison followed, and, it will be contended, is the only one that can be followed.
Thus the insistence of the behavioural approach upon what its proponents consider to be strict empirical verification is taken to the point where the most important questions of the study of politics are excluded from consideration. Yet the behaviourist himself rarely appreciates the extent to which his own criteria, if strictly applied, would prohibit him from discussing even those things he considers appropriate for investigation. For the theoretical demands of the behaviourist for precision and empirical verification are not always reflected in his statements about the real world. This is simply a result of the fact that it is impossible to subject the discussion of political systems to the criteria of verifiability which the behaviourist espouses. This point cannot be illustrated more effectively than by an examination of Dahl’s work itself.
He points out that in some nations powerful minorities have not refrained from the excessive exercise of power, whereas in others they have refrained. He argues that “whether or not powerful minorities or mass-based dictatorial leaders have refrained from establishing tyranny is clearly not related to the presence or absence of constitutional separation of powers. Many variables are involved in such a situation, but the constitutional separation of powers cannot be established as one of them.” Again Dahl does not attempt to give any empirical evidence for this assertion, but a brief attempt to repair this deficiency in his argument will help to illustrate the problems of the strict behavioural approach.
We can formulate two hypotheses that the doctrinaire separation of powers theorist might adopt, and consider the problems of testing them, leaving aside for the moment the problems of definition. They are as follows:
Now these hypotheses have one characteristic of which Dahl should approve—they are strictly empirical and capable of being falsified by empirical evidence. Find any one example of a non-tyrannical republic in which there is no constitutional separation of powers and (i) is false. Find any one example of a republic with a constitutional separation of powers which lapsed into tyranny, and (ii) is false. We have already referred to Dahl’s attempt to disprove (i) by a simple reference to Great Britain. Such a disproof does not work, yet it might well be that if history were scoured such an example might be found—in the ancient world with small city-States, or in the town-meetings of New England, if local government be admitted here for discussion. In such small face-to-face societies the requirements of political organization are likely to be different from those of modern nation-States. Such a conclusion would hardly be surprising. The main point is that proposition (i) could in principle be shown to be false.
Proposition (ii) is a much easier nut to crack. There must be many States that embodied some form of separation of powers—not necessarily on the American model—which have later succumbed to tyranny, or the abuse of power, by almost any definition. The Weimar Republic, Republican Spain, the Kingdom of Italy—all had constitutions that incorporated some form of separation of functions among distinct, though not completely separate, branches of government. South America provides examples of constitutions much closer to the American pattern which have not been able to stem the advance of tyranny. There is, therefore, a wealth of evidence to dispose of (ii).
A more subtle hypothesis, however, would be as follows:
This is, in fact, the antithesis of the assertion Dahl makes that the separation of powers is “clearly not related” to the question of tyranny. Now to establish either hypothesis, (iii) or Dahl’s, presents problems of great complexity. Certainly neither can be asserted as self-evident, and it is impossible to test either of them in the strict terms which Dahl himself advocates. Such an hypothesis can be examined only by looking at the examples of non-tyrannical political systems which exist, and attempting to form a judgement about their operation. Thus we can consider the role of the separation of powers in the history of the United States, but to do so we must employ a whole range of historical and philosophical techniques, concerning ourselves with modes of thought, indeed, which Dahl rejects as merely “ideological.” Indeed Dahl’s proposition is, in his own terms, an expression of an “ideology” and not a conclusion of political science, for he is asserting that the separation of powers is not a significant variable in the United States, and this is a proposition it would be impossible to substantiate through strict empirical “tests.” To apply his own strict criteria to the proof of this statement he would have to compare the United States with a society like the United States in every respect over the past two centuries, except that in the latter all powers were accumulated in one set of hands, and to show that, as regards tyranny or the abuse of power, there were no significant differences between the two societies. The impossibility of adopting this procedure indicates the limits of the behavioural approach in politics.
This limitation upon the nature of proof in history and politics must be accepted. There is no way round it. This does not mean that strict empirical verification should not be applied wherever possible, nor that the examination of propositions, like that which Dahl contends for, should be conducted as a metaphysical enquiry, and without the closest possible reference to all known facts. It simply means that if we wish to consider the most interesting, and the most important, propositions about politics we cannot afford to restrict ourselves to techniques that will allow of the investigation only of matters of secondary importance.
That Dahl himself is really aware of this can be seen by a comparison of the later chapters of the work in question with its earlier ones. He adopts very different methods when developing his own views about American government from those he wishes to apply to Madison’s thought. Thus after a close criticism of the separation of powers doctrine, as applied by Madison, Dahl ends by acknowledging that there is a sense in which every “polyarchy” is characterized by a separation of powers. He talks in terms of the need for a “more or less representative body to legitimize basic decisions by some process of assent,” of a need for bureaucracies of permanent experts, of the need for “a specialized bureaucracy” to pass judgements upon appeals from decisions of bureaucratic officials and to adjudicate conflicts among individuals, and of the need for leaders to co-ordinate bureaucratic, judicial, and legislative decisions.17 Dahl suggests that this is a matter of the division of labour, but this, of course, is only partly true. The division of labour requires specialization, but not necessarily of this precise type. Indeed the extent to which Western society has followed the dictates of the division of labour is itself an expression of a value-judgement, for if we follow its prescription we are pursuing a value, i.e. efficiency measured in terms of output or technical effectiveness. But this type of efficiency can be, and has been, sacrificed if other values predominate. Totalitarianism and theocracy, each in its own way, sacrifice technical efficiency in order to achieve other aims. The division of labour has been emphasized in Western societies because of the values we place upon technical efficiency, and also because, happily, the requirements of the division of labour have tended to match fairly well the other requirements of these societies. Yet where “fairness” or “justice” has been considered more important than speed or the expediency of policy-makers, complex time-consuming procedures have been evolved in an attempt to give priority to the desired values.
In fact the whole section where Dahl develops this point is shot through with the values and concepts of the writers who have related democracy and the separation of powers down through the centuries. Dahl is very clearly a Madisonian. He notes indeed that all polyarchies have “strikingly similar constitutions,” but his conclusion is that this means “the constitutional variable” is even more limited than would be thought at first glance.18 Others might consider this to be a highly significant correlation.
The fundamental error of all three of our “behaviourists,” William Penn, Sir Lewis Namier, and Robert A. Dahl, is that they draw a false dichotomy between “constitutions” and “social forces.” This suggestion, that social forces and constitutional structures are quite distinct entities, is a result of, and a reaction against, the legalism and formalism which once dominated the study of politics. But it leaves the whole question of the nature of “political institutions” in an awkward limbo. The Congress of the United States is surely an important subject of study for political scientists, but is it “constitution” or is it “social forces”? Merely to pose the question is to reveal its absurdity. Of course, it is neither, but it involves both. The problem arises from the attempt to reduce the material of politics to one set of “independent,” and one set of “dependent,” variables, in a way that will give a specious mathematical neatness to what is in fact an enormously complex situation. Yet the inter-relationships of this material can be seen once again by taking an illustration from Dahl’s own work. In his discussion of “polyarchy,” Dahl hypothesizes that polyarchy is a function of the consensus on the norms which are its definitional characteristics, and further that the extent of this consensus is dependent upon the extent of social training in these norms. The concluding hypothesis is: “Polyarchy is a function of the total social training in all the norms.” Now Dahl conceives of “social training” as carried on by the family, schools, churches, clubs, literature, newspapers, “and the like.” Yet social training in most societies, and not least in America, may also be carried on by governments, and all the above-mentioned media of social training may to a considerable extent consciously or unconsciously direct their activities towards training in the constitutional rules, and in the operation of the political institutions, of the country concerned. “Social training” is not an “independent variable” which gives rise to a particular constitutional structure; it is itself the result of the historical evolution of the political system and knowledge of its operation. Dahl acknowledges this “hen–egg” relationship in regard to “consensus” and “social-training,” but still doggedly removes constitutional factors from the realm of “significant variables.”19
The role of “constitutions” and of the study of “constitutions” is, then, much more complex and much more important than the behaviourists suggest. But do not let us seem to be claiming too much. We must accept the view that much of the Madisonian approach is inadequate to an understanding of the role of constitutions and political institutions in the twentieth century, although not for the reasons Dahl gives. The concepts of the eighteenth-century theorists were too crude to cope with the complexity of modern government.
Furthermore we have accepted that what was labelled (ii) above was an untenable hypothesis, and that (i) was unlikely to be true if applied to all societies at all times. There clearly are limits to what constitutional rules can achieve, however broadly we conceive of constitutionalism. In 1892 Jowett criticized Plato and Aristotle for having entertained a dream of a mixed State which would escape the evils and secure the advantages of both aristocracy and democracy. Such a creation he believed was beyond the legislator’s art. “No system of checks and balances, such as Plato has devised in the Laws, could have given equipoise and stability to an ancient state, any more than the skill of the legislator could have withstood the tide of democracy in England or France during the last hundred years, or have given life to India or China.”20 The primacy of “social forces” in this sense of the great movements of social life can hardly be questioned. Constitutions, or political systems, are not iron-clad structures which can withstand the transformation of the social assumptions on which they are based, resulting from great technological or ideological developments. But this is not to say that even in such situations constitutions are insignificant or unimportant. The Constitution of the United States is today a very different structure from that of 1789, but it is still a Constitution, possessing a recognizable continuity with that of the earlier age, and it would be very difficult to prove that it had had a negligible part in creating the present political structure of the United States.
Jowett’s objection was an aversion to constitution-building, rather than to constitutions, and it illustrates the difference between an abstract constitution on paper, and the operation of a set of political institutions which has been evolving slowly over centuries. Even so, we do live in an age that provides a veritable storehouse of political experience with new constitutions, offering an opportunity to study the circumstances in which they can and cannot have a significant effect. Jowett’s reference in 1892 to India is of great interest, for seventy years later we do find a constitutional democracy in existence in India, although it is one the stability of which is still not very certain. The life that has been given to India must surely in part be due to the history of the development of constitutionalism in that country during the present century. The study of politics must, therefore, very largely consist of the examination of the ways in which constitutional and political institutions, and the social forces and movements in a particular society, interact with each other; of the limits upon the extent to which stable constitutional modes of behaviour can be developed and maintained; and of the effects they can have in moulding behaviour. To do this we shall have to adopt all the techniques of study which are relevant to the solution of such complex problems, but however much we learn about politics in this way we shall never, and can never, have final and completely “verified” answers to any of these questions.
[1. ]Robert A. Dahl, A Preface to Democratic Theory, Chicago, 1956, p. 135.
[2. ]See J. Plamenatz, “The Use of Political Theory,” Political Studies, 1960.
[3. ]J. C. Rees, “The Limitations of Political Theory,” Political Studies, 1954.
[4. ]See the essay by E. M. Kirkpatrick, in A. Ranney (ed.), Essays on the Behavioral Study of Politics, Urbana, 1962.
[5. ]England in the Age of the American Revolution, 2nd edn., London, 1961, p. 3.
[6. ]This is an extension, to the political system as a whole, of the terms which Professor H. L. A. Hart has used in regard to the legal system. See The Concept of Law, Oxford, 1961.
[7. ]Op. cit., p. 83.
[8. ]Ibid., pp. 4–11; Federalist No. 47.
[9. ]A Preface to Democratic Theory, p. 135.
[10. ]Two-thirds of both Houses of Congress, together with three-quarters of the legislatures of the States, in all matters other than equal representation in the Senate, in which case the States affected by any alteration must consent.
[11. ]Preface to Democratic Theory, p. 14.
[12. ]Federalist, No. 51.
[13. ]Preface to Democratic Theory, p. 20.
[14. ]Federalist, No. 51.
[15. ]See for example his remarks p. 81, op. cit., about the role of checks and balances in depriving the unpropertied masses of political equality.
[16. ]Op. cit., pp. 30–31.
[17. ]Op. cit., p. 136.
[18. ]Ibid., pp. 135–6.
[19. ]Ibid., pp. 76–78; 83.
[20. ]The Dialogues of Plato, 3rd edn., Oxford, 1892, Vol. V, pp. ccxvii–ccxviii.