What Belongs in a Constitution?(1)

Credos should only be included in constitutions if they are widely shared. In principles, constitutions should specify their modes of amendment and enforcement. If a society cannot have a widely-agreed constitution, it may be better to defer its enactment until such broad agreement is possible.

1. Introduction

For this question to be intelligible at all, we must use ‘constitution’ in the sense of a formal written document, which enjoys some form of superiority over regular law-making, and some form of entrenchment. In other words, arrangements will be considered constitutional for my purposes only if they are taken away, so to speak, from the realm of ‘ordinary politics’ and from the standard discretion of regular legislatures. The level of entrenchment may differ in the same document for various provisions, and modes of entrenchment may vary among constitutions, but the question I will consider will not take these detailed features into account.2

There are three standard candidates for inclusion in a constitution: basic governmental structures and the relations between the main powers and functions of government; basic values and commitments; and human rights. Some constitutions describe language and flags and other symbols. These may either be seen as an additional group, or be seen as a part of the main commitments of the state. In addition, a constitution usually specifies the mechanisms for its own amendment and enforcement, and proposed constitutions often contain provisions about the mechanisms of their adoption.

The title question presupposes that we have made a decision that we want or need a constitution, and the question we face is what should be included in it. However, as we shall see, it is not always possible to follow this sequence. It may well be that we start out wanting a constitution, but that we can then see that it is impossible to get a good one, because too many important elements of the desired constitution cannot in fact be enacted.

Under such circumstances, it may well be rational to decide to prefer the regulation of constitutional issues by regular laws or even by convention, and let the formal constitution await better times. I will return to this question below, after I discuss the three clusters of issues which naturally ‘belong’ in constitutions.

Before I turn to each of these groups, some general comments are in place. Constitutions are required neither by logic nor by legal theory. Similarly, they are not indicated by universal human and political experience. They are not even dictated by democracy or by the idea(l)s of limited government and human rights. England is an example supporting all these claims. Nonetheless, most modern countries, and most modern democracies, have Constitutions. This in itself suggests that there is nothing logical or conceptual that makes the idea of formal, entrenched, constitutions inconsistent with democracy.3 In fact, political experience suggests that democracy and constitutionalism are a good pair, with a tendency of democracy and constitutions to reinforce each other.

It is important to notice that it is not always easy to separate the three clusters of issues. A commitment to democracy, for example, is often a central part of the credo of a state. But it has important implications for the structure of the regime. Democracies may be parliamentary or presidential or a combination thereof, but they must have regular elections for the legislature and an effective multi-party system. Similarly, democracy, even under the thinnest, most formal, characterization requires some civil and political rights such as the rights to vote and to be elected, and some freedom of speech and association. Similarly, a commitment to a ‘social’ state may require some protection of social and economic rights. But if these are merely declaratory, and do not involve effective mechanisms of enforcement, they may be a part of the credo, rather then ‘real’ affirmation of ‘rights’.

The main purpose and functions of constitutions are at least three. First, to both authorize, and to create limits on, the powers of political authorities. Second, to enhance the legitimacy and the stability of the political order. Third, to institutionalize a distinction between ‘regular politics’ and ‘the rules of the game’ and other constraints (such as human rights) within which ordinary politics must be played. The rules of the games are the shared part, which gives society some coherence and identity. This shared framework can then facilitate a robust democratic debate between different conceptions of the good life, and between different interest groups. These purposes will suggest guidelines for answering the question of exclusion and inclusion of provisions within the constitution.

Questions about constitutions can never be answered for all societies and states, for all times. Constitutions are designed to solve the problems of the societies under consideration. The need for a constitution itself, and its structure and content, should thus be discussed against the background of the social and political problems facing that society. Since these problems are different, importation of constitutional arrangements should be done, if at all, extremely cautiously. On the other hand, comparative studies are extremely important. There are aspects of the study of human societies, which are indeed ‘universal’ in important ways. These do indicate some implications in terms of immanent and persistent problems and the structural ways to deal with them. This attitude gives credence to enterprises such as universal human rights and to ‘public choice’ models of constitution. These indications, however, are often incomplete and underdetermined.4 Comparative studies, which stress both similarities and differences among societies, may thus provide a very illuminating source of learning from the experience of others.5

A major question is that of the enforcement of the supremacy of the constitution. In the framework of my concerns, the question is whether such enforcement mechanism should be explicitly included in the constitution. As we shall see below, this question relates both to the form of the institutional arrangements adopted in given societies, but also to the effectiveness of the constitutional order as a whole, especially that of protecting human rights against the legislature.

The last two decades have given us very rich material about processes of constitution making and amendments: It has been a time of dissolution and of birth of nations, and in many cases these transformation had significant ‘constitutional moments’. Constitutional moments are those in which there is a change in the basic framework of government, not just in the content of specific political arrangements within it. Some such changes take the form of the adoption of new constitutions, or of significant amendments of existing ones. In other cases, serious constitutional changes take place through either a radical re-interpretation of the constitution by the courts, or through apparently minor changes in an existing constitution. In addition to the former USSR and Eastern Europe, interesting constitutional developments took place in South Africa, Canada and Israel. In addition, older constitutional systems show interesting signs of reflection and modification. Notably, England has made a first step towards a Bill of Rights with semi-constitutional significance. I will draw on this rich experience in the comments that follow, but this reliance will be impressionistic rather then systematic and comprehensive.

One last general caveat is called for. No one likes to discuss unimportant questions. Hence, my discussion of the question what should be in a constitution implies that I see it as an important one. Nevertheless, I want to warn us against taking it too seriously. Constitutions affect social and political reality, and this is why they are important. Their impact on these realities is, however, indirect and limited. Often, the question of the content of an arrangement is more important than the question whether it is included in a constitution or in regular legislation or even in a convention. Always, the reality of societies is more important than the relationships and values declared in a constitution. Nonetheless, since constitutions do help in making social realities, it is important to attend to the special contribution they may make to good government.

2. Governmental Structures

Only unitary states have the luxury (or burden) of debating the question whether governmental structure should be included in the constitution. All states must have provisions regulating these subjects, but in unitary states these may evolve, and be enforced in part as constitutional conventions. This is precisely what has happened in England, and Dicey is still a powerful spokesperson for the desirability of this way of regulating the structure of government. However, federal governments cannot make this choice, since it is a central issue of such governments that the division of powers between sub-units and the central government is determined in a way that will be beyond the unilateral change of either states or the central government.6 Even in federal governments, however, major differences are apparent in the scope, and clearly in the content and arrangements, of the constitutions. In addition, unitary states do not need a constitution, but many of them do have them.

This obvious point highlights the important distinction between formal and material constitutions. The decision whether the state should be a federal state or a unitary one is itself a major constitutional decision. If the state is a federal one – the relations between central and local government must be made in a formal constitution, which cannot be changed unilaterally by the member parts. However, if the state is unitary, its (material) constitutional arrangements need not be included in a formal constitution at all.

In federal states we must have a constitution which is beyond the powers of change of the parts thereof, but many questions remain. Should the federal constitution discuss the arrangements and the powers of the sub-units, or should it be confined to the powers and structure of the central government? Who should be the arbiter between the national government and the parts? Should the constitution include provisions concerning cessation? What should these provisions be? These are normative questions, but this fact should not mislead us into thinking that the only relevant premises for their resolution are general and universal normative principles. Both the comparative study of constitutions and a closer analysis of how these questions ought to be answered in particular states suggest that the answer to this question is dictated by historical and not by theoretical considerations. The US federal constitution was a document seeking to form a union whose members were fully established. It is thus confined to the federal powers. The German basic Law, on the other hand, discusses in detail both the federal government and the Lander. Moreover, it is quite clear in the latter that the Bill of Rights part applies to all Lander, and that the constitutional court deals with all matters arising under it. It took a long development and a civil war until the federal Bill of Rights was seen as ‘incorporated’ into States’ constitutions. The necessity to incorporate governmental structures into the federal or national constitution applies only to the central government. As mentioned above, that structure must be protected against unilateral changes by one or more of the member states (or by the central government on its own). The internal constitutional arrangements of the sub-units may, in principle, be regulated by the states as they choose. In fact, it may be seen as a part of their autonomy that their own internal affairs are controlled by them. This would suggest a reason against including the details of such arrangements in the central, federal constitution. On the other hand, it is also quite clear that member states should obey some structural constraints if the federation is to be stable and viable. These general constraints on both the structure of member states and on the relationships between them should of course be included in the federal constitution.

Another possible difference between constitutions is illustrated by this observation. Some constitutions mainly entrench the political status quo. Constitutionalization is meant to achieve unity and the additional stability allowed by entrenchment. In these constitutions, at least for the short term, there is no need for great detail. The details will be provided by the experience of the past. Very different are the constitutions erecting a new political order, especially ones that want to make a statement of a break with tradition and the past.

In unitary states, the most important question is whether the structural details of the regime should be constitutionalized. Once an affirmative answer is given, which of the basic governmental arrangements should be included in the constitution, and which should be left to regular legislation or to constitutional conventions? The answers to all these questions depend on the functions constitutionalization is supposed to serve, and on the special conditions of the society and legal system under consideration. A general reason for entrenchment is the need for institutional stability. Clearly, a special, burdensome mechanism of change would make it harder to amend the basic institutional structure, and this will generate the stability. Some stability of basic governmental structures is a crucial condition of governability. Many decisions require long-term planning. The ability to plan long-term legislative and policy initiatives depends, among other things, on constitutional stability. Moreover, stability is very important in creating the sense of tradition and continuity, which are important elements of the legitimacy of government.

We glimpsed another reason in the explanation of the reason why federal states require an entrenched constitution: entrenchment is necessary when there is a wish to prevent a unilateral change or a change by a transitory or non-representative majority. Finally, entrenchment is crucial when the constitution is a complex package deal, and the partners to it (states or segments of the population) make their agreement to the package deal dependent on the entrenchment of a provision that is deemed crucial to them and is their incentive for going into the deal. Entrenchment in such cases is needed to give the arrangement the stability it requires, and minimizes the chance that some political players will gain the benefit of the compromise, but will undermine it to their own advantage whenever an opportunity presents itself.7 In the same vein, entrenchment is crucial when the structure of government is a complex set of checks-and-balances. A structure like this should be amended with care, since one apparently local change may frustrate the effectiveness of the system as a whole.

These comments suggest that an entrenchment is indicated strongly when the constitutional arrangement is the product of a serious compromise, and when the structures include a complex set of checks and balances, so that it is important that elements of the whole will not be changed randomly and easily.

Effective federalism also requires that there is an authoritative way to resolve controversies about relations between central government and provinces. This was one of the main purposes of creating a federal Supreme Court under Article III of the US constitution, and this function is among the ones explicitly given to the German Constitutional Court. However, even this is not totally true. In Canada, relations between central government and provinces are evolving slowly, and it is not clear whether the involvement of the Canadian Supreme Court has been either effective or desirable.8

It is important to note that, in distinction with the situation concerning rights and credos, there is no serious argument against judicial review of the ‘‘institutional’’ part of constitutions. This consensus is based on the fact that the provisions of the constitution in these matters are relatively clear, and that there is a necessity that there will be an authoritative arbiter of the disputes that do arise. Moreover, while decision of issues of these sorts may have tremendous political impact, institutional questions are usually not as emotionally charged and as controversial as matters of basic values and of human rights.

This part of constitutions always includes details of the legislative, executive and judicial powers. For these, it often includes both a specification of their powers and modes of their election or appointment. It also provides some account of their relationships. There is some variety on other elements which belong to the structural part of constitutions. I will consider just one group of such elements. In all countries, institutions such as the law enforcement agencies, the state controller, and the central bank may develop great significance because they enjoy a middle-of-the road position. On the one hand, they do form a part of the executive branch. On the other, it is recognized that the need a measure of independence from the executive to fulfill their tasks effectively. This delicate balance of accountability and independence is sometimes achieved by statutory (or even conventional) arrangements. But at times, especially in ‘younger’ constitutions, the need is felt to include these institutions within the constitution itself, so that the power of the executive, and even the legislature, to reduce their independence, will be limited. In general, this tendency to include within the constitution the structure of institutions whose independence requires special protection seems desirable. However, we should also remember that in these matters, as in all others, the realities are at least as important as the constitutional provisions. At times, the de-facto reality does not give independence despite constitutional entrenchment.9 At other times, these institutions may develop in a way that gives them too much independence, and the necessary constraints of accountability are weakened. Such phenomena may happen even if the institutions are not regulated in the constitution, or even by statute.

Not surprisingly, the main debate on this part of constitutions relates to the content of the arrangements and the balances of powers. These fascinating debates go way beyond the realm of this paper. One feature needs a special discussion, however: the stringency of the amendment process. How flexible, or how immune from change, should the structural elements of a regime be? It is often claimed that the US federal constitution is too hard to amend. The sensitivity to particular circumstances led Sunstein and Elster to recommend that in the new post-Soviet regimes, arrangements should be flexible, and allow relatively easy amendment by the regular political branches. Recent events in Syria suggest that constitutional flexibility may indeed be an asset.10

I cannot leave this issue without reminding ourselves that the decisions incorporated in the constitutional text itself are only a part of the picture. Constitutions are living documents, and much of their practical import is the function of the way they are used, interpreted and implemented. It may well be the case that very stringent rules of amendment will give legitimacy to more creative interpretive attitudes by both courts and by the political players themselves.11

It is hard to specify a general recommendation for the balance between stability and flexibility of constitutional arrangements concerning regime structure. The fact that social realities at times do not ‘‘obey’’ constitutional mandates may mean that the practical implications of over-stringency may not be disastrous, but this does not free us from the need to try and specify considerations for the right balance. When the reasons for entrenchment are stronger – the amendment mechanism should also be stronger. I mentioned above the case of the ‘‘historic compromise’’ or the delicate system of checks and balances, which should not be upset and undermined easily. In such cases, the incentive to make changes may be great, and the consequences for the viability and stability of government may be high. A stringent mechanism of amendment may then be required.

While one principle of amendment for the whole constitution has the benefit of simplicity and elegance, it may not always be a good idea to use a uniform requirement. A more nuanced amendment mechanism may be more suitable. At least two types of differentiated amendment procedure come to mind. One is that of the German Basic Law, which specifies that the Bill of Rights part is beyond amendment, while allowing amendment of the other parts. The other is a differentiation within the structural regime part itself: the senate provision and the pro-slavery interim arrangements in the US constitution were more protected than other parts of the constitution. Clearly, the reasons behind the different treatment of constitutional arrangements in the two cases are very different. The first one is primarily symbolic and declaratory, and relates to the nature of human rights and to the historical circumstances of the enactment of the German Basic Law. Bills of Rights do not enjoy special immunity in other constitutional systems, but it is almost inconceivable that any state will derogate from the constitutional guarantee of rights such as human dignity or freedom. Changes in the actual protection will be made by legislation, adjudication or social practices. The second case, on the other hand, reflects the special need for entrenchment for ‘great compromises’. A differentiated amendment procedure may be indicated when the constitution marks a great break with the past. On the one hand, commitments to democracy and to human rights may seem to require great entrenchment. On the other hand, the political system may be very volatile, and there is a need for trial and error in some structures, calling for flexibility.12

One thing should be clear – the more stringent the amendment procedure is, the shorter the constitution should be. In some new constitutions, the arrangements are specified in very great detail. A stringent amendment mechanism may then mean that it will be very hard to adapt the details of the regime structure to changing circumstances. We saw that there may be good reasons for making such changes difficult in some contexts, and for the principles of separation of powers and their structures. On the details, flexibility should be the rule.

3. Bills of Rights

All new constitutions contain detailed Bills of Rights.13 Moreover, in many of them this section of the constitution is given pride of place, and gets to be the first chapter, preceding the description of the main organs of government and their respective powers. This attitude is quite natural when the societies making their new constitutions want to stress the difference between the old regime, which did not protect human rights, and the new one they are now initiating. Nonetheless, as we shall see, the idea that states should constitutionalize Bills of Rights is not free of controversy.

It is important to remember that the inclusion of detailed Bills of Rights does not guarantee an adequate protection of human rights. This is true for both democracies and non-democracies: the presence of glorious Bills of Rights in a country’s constitution does not, in itself, guarantee their actual protection.

Nonetheless, as a matter of general principle, there are good structural and expressive reasons to give Bills of Rights constitutional status. The high visibility and solemn nature of most constitutions help in making the commitment to human rights a part of ‘‘civil religion’’ and civil-shared culture.14 A deep acceptance of this ethos is required if protection of human rights is to be effective. Moreover, there is an obvious reason for taking the protection of human rights away from the realm of regular politics. We believe one’s entitlement to basic human rights does not depend on the arrangements adopted by one’s society. Human rights are those that persons have because of their humanity. Their states are obligated to protect them, but they do not have the power to deny them. Today, these commitments are stemming from both moral theory and from international law.15 It is thus sensible to make the protection of such rights be superior to the acts of government and even legislatures. This is especially the case if the society in question is rifted, and if it contains ‘chronic’ minorities, who cannot defend their interests through the regular political channels.16 Even if Bills of Rights are neither necessary nor sufficient for adequate protection of rights, there are obvious reasons for including such Bills in constitutions, or even of entrenching Bills of Rights independently.17 Even if Bills of Rights and judicial review have not guaranteed adequate protection, on the whole the existence of such documents and institutions permitted processes and developments that increased protection of human rights. Bills of Rights very rarely have been invoked to frustrate progressive legislation, and they have often been invoked to invalidate oppressive laws.18 It is thus not surprising that some scholars have argued that the best institutional defense of human rights is judicial review.

Against this background, it is interesting to note that people, whose commitment to human rights is unquestionable, have also argued forcefully against constitutional Bills of Rights, and especially against judicial review of legislation which allegedly violates them.19

The objections to Bills of Rights are very rarely phrased as objection to the idea of rights. Critics often support the educational force of declarations of rights. What they object to is the discourse of rights, which, according to them, destabilizes the fabric of life in society, which is built on a combination of rights and duties. In addition the critics say that Bills of Rights tend to underestimate the fact that in many cases rights conflict with other rights and with other interests, so that often the mere fact that an action or a law infringe a right is not very significant. We need to know if the infringement is justified, and to do that we need more than rights discourse. Often, the critics claim, a rights-discourse leads to an underestimation of conflicting interests, and may thus lead to either wrong decisions, or to frustrated expectations, or both. It is better, the critics say, to specify the social arrangements in more detail, and not to derive them in an unmediated way from general declarations of rights.

Most critics would probably have accepted that the educational benefits of declarations of rights easily overcome these drawbacks had it not been for the institutional implications of constitutional rights: the fact that they often transfer the power to make ultimate decisions on matters concerning rights away from the political process. Judicial review and invalidation of laws allegedly violating human rights is a frequent way of giving rights the ‘teeth’ they require to effectively protect individuals. On the other hand, such review raises the counter-majoritarian difficulty, and may create serious backlashes when the court is perceived to declare laws unconstitutional when they deviate from their own partisan ideology or conception of the good life.

In states with constitutional Bills of Rights and judicial review, the debates usually center around the desirable scope of judicial review. Naturally, these countries do not consider the possibility of abolishing their Bills of Rights or of judicial review. In countries where these institutions do not exist at present, I believe these questions should be raised and discussed before a constitutional decision is made. The mere fact that most countries live well with constitutionalized bills of rights and with judicial review does not mean that all countries should have them. I have indicated above that the main problem is not the inclusion of a Bill of Rights in a constitution, or even entrenching it. It is the nature of the enforcement agency and its powers. I shall return to this question below.

4. Basic Values and Commitments

The third cluster of candidates for inclusion in constitutions is a declaration of basic values and commitments. Frequent commitments declared in such contexts are to a republic (or monarchy), to social justice, to happiness, democracy, freedom or a national character of the state. Some states declare their religious (or secular) identity. Many believe that these elements are the least important parts of the constitutional arrangement. Often, they are just expressive, and it is hard to find any practical implication that follows from them. Yet the inclusion of these elements may be extremely significant in highlighting one of the main features of constitutions, suggesting a source of either strength or weakness, as the case may be. It indicates clearly the degree to which the society governed by the constitution has an inner cohesion and civic identity that is in fact accepted by the large majority of citizens. And, in some cases, these declaratory parts are used, by courts or other state organs, to derive important practical implications as well.

We said above that constitutions seek to give special visibility, legitimacy and stability to the shared framework of political life. The importance of this declaration is often a matter of historical contingency. Homogeneous nation states, not involved in a conflict or a challenge of their distinctness, do not ordinarily need this function of constitutions. They gain their cohesiveness from the homogeneity in fact of their population. But modernity means, among other things, that this form of nation-state becomes less likely. Even in non-democratic states, effective government includes the need to deal with the fact that there are serious conflicts of interests among parts of the population, as well as between the state and other groups and individuals who are their citizens or members. In democracies, and especially in divided democracies, the constitutional framework facilitates the inevitable negotiations and compromises of effective government.20 It is extremely helpful when these negotiations and campaigns, which can at times be extremely divisive, are conducted against the background of a secure and wide acceptance of a shared framework, affirming and declaring the shared commitment and allegiance of all to the welfare of this shared entity. If the shared framework is successful, the debates are debates about different conceptions of the public good of the whole, and on the meaning of distributive justice within it, not a challenge to the political and social order as a whole.

Often, these declarations are so broad that they do not have much practical import. This may be the status of the French commitment to ‘‘Freedom, Equality and Fraternity’’. Yet even declarations of this sort have been interpreted in some manner. Thus, the commitment in the US constitution to liberty and happiness, and the absence of a commitment to equality from it, were used by some to argue that the ideal of substantive equality is foreign to the American national ideals.

At times, the ideals specified in the solemn and festive openings of constitutions are universal ideas. When this is the case, they function in ways similar to human rights. The declaration is an affirmation of an ideal, which is supposed to be shared by all. The declaration itself does not add a particularistic characterization to the society in question, and so does not raise debates about the essence of that society.21 At other times, these declarations simply reflect the types of regime choices articulated in the part of governmental arrangements (e.g. the declaration that the state is a democracy). The most interesting – and controversial – ideals are those that are neither. They reflect a particular choice, which is not universal, and is not a choice of a regime type. Obvious examples are declarations that the state is, for example, ‘socialist’. Should statements like these go into a constitution? We saw that for the two previous candidates for inclusion, broad agreement is important and desirable, but that it may not be absolutely necessary. Inclusion in the constitution may be a way of affirming a normative commitment to a shared framework or to human rights even if in fact we know that some parts of the population find these commitments difficult. Part of the reason for constitutionalization is precisely the wish to give these arrangements a security that cannot be supported, in each particular case, by the political system. We may have a broad agreement on this need for stability and entrenchment even if we do not have agreement on the merits of the particular arrangements that we have adopted. The same reasons for entrenchment and superiority do not obtain for inclusion in the constitution of a value commitment endorsed by the majority, but rejected and challenged by a significant minority. The elected majority can enact its preferences, but why should it impose them on everyone as a part of the shared constitution? The inclusion would make people who disagree with the ideal, or feel they cannot identify with it, not only political opponents of the majority, but also ‘enemies’ of, or at least strangers to, the constitution. It seems to follow that commitments like these should be included in a constitution only if they do reflect a general sentiment.

The same applies to self-definition of the state in terms of a particular religious or national character. A close-to-home example is the fact that Israel is defined in its 1992 Basic Laws concerning human rights as a ‘Jewish and democratic’ state. ‘Democratic’ is neutral and innocent, but the ‘Jewish’ in the formula has already raised a heated debate. Many constitutions contain an affirmation of a particular identity. But usually, the reference is ambiguous, and it can be interpreted as an affirmation of the nation whose state it is, and the nation can be credibly identified as the body of citizens of that state. In many other cases, states are defined as ‘nation-states’ in ethnic terms. This does create a problem, especially when in that country there is a sizeable minority of people not belonging to that ethnic group.

I have dealt elsewhere with the substantive problems arising from the definition of Israel as both Jewish and democratic (Gavison 1999a, 1999b). Here, my concern is only with the question whether such descriptions should be included in a state’s constitution. For the Founding Fathers of a state like Israel, writing a constitution, which will not include an affirmation of the fact that it is a Jewish state, would be almost unthinkable. Even the UN decision supporting the establishing of the state described it as a Jewish state. This description is emphatically affirmed in Israel’s Declaration of the Founding of the State of 1948. However, declarations like this would make non-Jewish citizens of Israel feel less-than-full citizens of the state. This consideration was raised in the discussion of Israel’s constitution in 1950. It fitted well with the decision adopted then – that Israel will not enact a constitution, and will instead seek to develop legislation of constitutional materials in a series of ‘basic laws’ (Gavison 1985). In the same spirit, Israel has decided not to entrench its Law of return, adopted the same year. Many see this law as the essence of the Jewishness of the state. Yet when it was proposed that this law would contain a provision making changes thereof impossible, Israel’s first PM, David Ben-Gurion, declined. He conceded, of course, that no Israeli legislature will change the law, but preferred to leave the matter to political realities, without a formal entrenchment.

No constitution can change the realities of states. Israel is, in many senses, a Jewish state even if it does not define itself in this way in its constitution. However, because of the role constitutions play, incorporating such a declaration into Israel’s constitution in an explicit way may not be an easy and a simple decision. I would suggest that such complications might well justify not including any commitments to general values in the constitution, so that if issues of this nature exist, it may well be to exclude the entire third cluster from the constitution. At the very least, the part of the credos should indeed contain only commitments which are in fact widely shared. The story of Israel teaches us another lesson. At times it is wiser to decide not to include a contested commitment in a constitution or a basic law. It is much easier to agree not to mention a feature that is a part of social reality, then it is to remove an already existing affirmation.22

5. Enforcement Mechanisms

Some constitutions explicitly specify ways in which their provisions may be enforced. Frequently, when this is the case, special constitutional courts are established to deal with questions of constitutionality. Some constitutions explicitly specify that the constitutional provisions will be self-enforced by the political branches. Yet other constitutions grant superiority and entrenchment to the constitution, and do not specify an enforcement mechanism. In two such cases – notably the US and Canada – the regular courts, and especially the Supreme Court, assumed the power of judicial review of the constitutionality of primary legislation.

This variety of constitutional schemes shows that the choices are contingent and do not follow from the very notion of supreme constitutions. Nonetheless, I believe that explicit constitutional regulation of the enforcement mechanism is preferable to the US model. First, explicit regulation enhances certainty in this important area, and thus increases the legitimacy of whatever enforcement mechanism has been chosen. Secondly, it permits the reasoned decision to combine constitutional superiority and entrenchment without judicial review. Thirdly, and most important, it stresses the distinction between judicial review of legislation and judicial review of actions by the government and the administration. Fourthly, it allows thinking explicitly about the institution most suitable for judicial review in constitutional matters, and designs it in a way that will reflect the special nature of reviewing laws passed by the majority of representatives. When judicial review is a judge-made declaration derived from the ‘meaning’ of the superiority of the constitution, as Marshall reasoned in Marbury v. Madison, it is inevitable that regular courts will deal with constitutional challenges. However, these courts may not be well suited to the constitutional task.

Let me elaborate a bit on this last point. Judicial independence is a primary feature of the rule of law. We expect judges to be removed from political pressures, and to have the structured autonomy and the professional integrity to apply the law to the facts of the case without prejudice or fear and without wishing to please the powers that be. One way of encouraging this independence is the emphasis on the professional integrity of judges, and a structural autonomy from political fiat. Most often, this is achieved by granting judges life tenure in their job. Another feature aiding judicial independence is the emphasis on the professional ethos of judges. Often, this professionalism is seen as incompatible with clear and explicit involvement in public life, especially in partisan politics. Clearly, sitting judges are expected to be non-partisan. Consequently, there is a structured attempt in adjudication to refrain from addressing questions of ideology and policy, and to derive the judicial conclusion from exclusive reliance on the law. However, all analytical theories of jurisprudence agree that the law cannot always dictate the bottom line of judicial decisions. The explicit attempt to ignore non-legal premises of judicial action only leads to confusion, and weakens the belief in judicial autonomy. Legal theory usually concedes that in adjudication there is a mixture of application of pre-existing law and of creative discretion, and that the legitimacy of adjudication is based on the fact that application of the law is central and paradigmatic, whereas creative discretion is parasitic and incidental.

All this is true for all judges. Interpretation of statutes often requires invocation of values and ideologies not explicitly reflected in the written law. The democratic ‘corrective’ to this aspect of adjudication is the ability of the legislature to amend the law, so as to indicate that the judicial interpretation did not reflect its judgment. Clearly, this corrective cannot work as easily when the court pronounces on the constitutionality of statutes. One of the main purposes of a constitution, as we saw, was precisely to remove certain topics from the simple discretion of regular legislatures. Many feel that this feature of judicial review of statutes requires a profile quite different from that of regular judges. First, it is thought that no person should have the power to review primary legislation for life.23 Secondly, while explicit prior involvement with partisan politics may indeed be a problem for a regular judge, raising fears that he may be distrusted by people not sharing his political commitments, such involvement may in fact be an asset for a term on the Constitutional court. Finally, while judges for regular courts should be elected and promoted on the basis of their professional competence alone, judges for the constitutional court should be of the highest professional standards, but should also be sensitive to the public and political complexities of constitutional adjudication. This requirement may be met by a special mode of appointments, which will guarantee that the judges of the constitutional court will enjoy the respect and the trust of the public and of its representatives. All of these aspects of the constitutional court get lost if we treat the constitutional court as an extension of the regular courts of the system.

The debate about judicial review and the organ most suitable to pass it goes far beyond my topic. In these days of global growth in the power of courts, based in part on our disillusionment with representative democracy and its institutions, it may seem completely out-of-tune to suggest caution when this question is raised. It may be proper to remind ourselves, therefore, that the debate is still raging.24 As we saw, the main issue is not the entrenchment and superiority of Bills of Rights, but the question whether un-elected judges should have the power to undermine the considered judgments of policy approved by a significant majority in the legislature.

The debate about the desirable scope of judicial review on such matters is connected to the fact that rights discourse may be expanded in many ways. Mostly, we do want our courts, Supreme or Constitutional, to help us save our society from moments of rage or indiscretion leading to gross violation of the human rights of minorities. In many cases, Bills of Rights and constitutional courts are established precisely to indicate a break from a past regime in which this protection was wanting. We hope these institutions will minimize the chance that our government may torture or execute people without trial, or discriminate against them for reasons of religion or race. Courts cannot always deliver on these hopes, but they can clearly help. The problem is that courts then move on to decide on the legality of laws prohibiting (or permitting) abortion; or on matters of social and economic policy; or on sensitive issues of the relationship between state and religion. On these issues, reflecting deep moral controversy within society, it is not at all clear that the courts have a relative advantage over the legislature, or that they enjoy the legitimacy required to make authoritative decisions for society at large.

For these reasons, the decision concerning the scope of judicial review over primary legislation, and the identity of the organ performing such review, should be discussed explicitly and decided on the merits.

6. Summing Up

I can now return to the question with which I started. In most cases, a new state will do the right thing if it seeks to express its credo and its political structure and major commitments in the form of a constitution. For one thing, it will be easier for it to join the family of nations, among which such constitutions are the rule.

Regime structure is something every country must have, and there are good reasons to include the principles of the regime structure in a constitution. However, strong entrenchment may be a danger in fluid situations. Situations of transitions may thus provide both reasons for a constitution (entrenchment of break with the past) and reasons against it (unclear what the political situation is, and what the desirable arrangements are).

Detailed Bills of Rights and credos may make the constitution more meaningful and powerful. But they can also make it divisive rather then the cohesive, unifying civic force it is supposed to be. The latter functions should be preferred to the former, because a divisive constitution is unlikely to be able to perform its political tasks. It will weaken, rather then strengthen, the legitimacy of government.

This is why it may be wise, under some circumstances, to defer the enactment of a constitution (or to limit its contents to widely shared commitments). Such deferment may permit processes such as round tables and thoughtful negotiations, which may make the resulting document more acceptable and suitable than the idealized draft suggested initially.

All these points are not directly related to the question what should be in a constitution. But they show that this question must be seen as a part of a broader analysis of the role and the functions the constitution is supposed to play in the society which is considering its adoption. In some countries, when we start looking at the question what should be in a constitution, and who should be the body designed to enforce it, we may well reach the conclusion that a good supreme and entrenched constitution is impossible. If this is the case, supporters of a constitution should s