I The problem of legality

Adjudication in matters of state is constitutional adjudication and, as such, a judicial guarantee of the constitution. It is a limb of the system of legal-technical measures whose purpose it is to secure the legality of the state’s functions. The functions of the state themselves have legal character; they appear as legal acts. They are acts through which law (and this means legal norms) is created, or through which created law (legal norms already enacted) is executed. Consequently, the functions of state are traditionally divided into legislation and execution, and the application of law, as a mere reproduction, is contrasted with legislation, as the genesis, the creation, or the production of law.

The problem of the legality of execution, in the sense of its conformity with statute, and thus the problem of guarantees of this form of legality, is on the whole well-recognized. The legality of legislation, by contrast, understood as the demand for the legality of the creation of law, as well as the idea of guarantees of this form of legality, seems to run into certain theoretical difficulties. Does it not amount to a petitio principii to want to measure the creation of law by the use of a standard that is only produced together with the object to be measured? And the paradox that seems to be contained in the idea of the ‘legality of law’ becomes the greater the more – following the traditional view – one identifies legislation with the creation of law and thus statute with law as such; so that the functions that are comprised under the name of execution, namely adjudication (judicature) and administration, and in particular the latter, appear to stand outside of the law, so to speak, and not to be genuine legal acts, to represent nothing more than applications or reproductions of a law whose production is somehow already finished, whose creation has been brought to completion, in advance of acts of execution. [1486] If one takes the view that the law is exhausted by statute, the meaning of the idea of legality will reduce to conformity with statute. And in that case, an extension of the concept of legality will hardly be self-evident.

This conception of the relationship between legislation and execution, however, is false.* The two functions are not opposed to each other in the sense that there is an absolute contrast between the creation and the application of law: rather, each of them, on closer inspection, turns out to be both creation as well as application of law. The relevant contrast is a relative, not an absolute, contrast. Legislation and execution are not two co-ordinate functions of the state, but only two different levels of the process of the creation of law, a process which neither begins nor ends at the level of statute, that stand in a relationship of subordination and superiority. This process continues to the bottom via the level of the administrative decree, the level of the judicial decision, and of the individual act of administration until it arrives at the acts that put the latter two into effect (these are the acts of so-called enforcement). It continues to the top until it arrives at the level of the constitution, in order eventually to reach, transcending the area of domestic legal order, the sphere of the order of international law that stands above all particular domestic legal orders. With this succession of levels, which is relevant, for now, only insofar as it unfolds itself within the area of a particular state, we of course intend to do no more than to offer a schematic representation of the major steps of the process through which the law, while concretizing itself, regulates its own creation; and through which the state, together with the law, continuously recreates itself. Constitution, statute, decree, act of administration, judicial decision, and enforcement: these are simply the steps in the formation of the will of community that are typical, given the way in which positive law organizes the modern state. Reality may of course depart from this ideal type. For instance, it is not necessary for a decree, i.e. for a general norm issued by an administrative agency, to interpose itself between a statute and the act of enforcement; and it is therefore possible, under certain circumstances, for a decree to be issued directly on the basis of the constitution, instead of being enacted as a means for the execution of a statute, and thus to stand alongside the statute, as a general norm enacted by a representative, parliamentary body. And other modifications of the typical process of the creation of law are similarly possible. But for now, the discussion will presuppose the most common type.

Since the constitution largely determines the way in which statutes come into existence, legislation is an application of law, if seen in relation to the constitution. In relation to the decree, however, as well as in relation to other acts standing below the level of statute, legislation is creation of law. And a decree, similarly, is application of law in relation to statute, and it is creation of law in relation to the judicial decision or administrative act that applies the decree. This decision, in turn, is application in relation to the levels above, but creation of law in relation to the level below, to the matter of fact that constitutes its enforcement. The route that the law travels, on its way from the constitution to the matter of fact that enforces a judicial or administrative decision, is one of continuous concretization. While the constitution, statute, and the decree represent the general norms of the law, [1487] which are progressively more saturated with content, the judicial decision or administrative act are to be regarded as individual legal norms. A legislator, who stands only under a constitution that determines his procedure of legislation, is bound by law only to a relatively limited extent. His freedom, his opportunity to engage in creative design, is relatively large. With every further step downwards, the relation between freedom and constraint shifts towards the latter. In other words, of the two components, which determine the respective function, the application of law is strengthened, and that of free creation of law is weakened. Every level of legal order does not just represent a production of law, in relation to the level below, but also a reproduction of law, in relation to the level above. And insofar as it is application of law, reproduction of law, the idea of legality is applicable to it since legality is nothing more than the relation of conformity in which the lower level of legal order stands to the higher. Hence, the demand for legality, and for specific legal-technical guarantees of legality, exists not just with respect to the act of enforcement, in its relation to the individual norms of the administrative command, of the administrative decision, and of the judicial decision, or with respect to these acts of execution in their relation to the general norms of the decree or the statute. The demand applies as well with respect to the relationship of the decree to the statute, and with respect to the relationship of the statute to the constitution. Guarantees of the conformity of a decree with statute and of the conformity of a statute with the constitution are therefore as possible as guarantees of the legality of individual legal acts. A guarantee of the constitution, hence, is a guarantee of the legality of the levels of law that stand immediately below the constitution. That is, first and foremost, a guarantee of the constitutionality of statutes.

That the demand for guarantees of the constitution is still raised today – or rather, that it has not been raised before today, and has therefore only recently acquired urgency and become a subject of scientific discussion – is due not merely to the mistaken theory we characterized earlier on, which lacked a full insight into the hierarchical structure of the law or, what amounts to the same thing, into the thoroughly legal nature of the functions of the state* and of their respective relations to each other. The fact that the legal orders of modern states exhibit an abundance of institutions that ensure the legality of execution, while they do not provide for any, or only for very poor, guarantees of the constitutionality of statutes (as well as of the conformity of decrees with statutes), is to be attributed to political motives. These motives, in turn, do not remain without influence on juristic theory, which, after all, would have to take the principal initiative in enlightening the public on the possibility and necessity of such guarantees. This observation is valid in particular for those modern parliamentary democracies of Europe that grew out of constitutional monarchies. The doctrine of public law of constitutional monarchy is still of great influence today, though this form of state has largely been pushed into the background. The constitutionalist doctrine still determines the juristic theory of the state to a high degree – in part consciously, where one wants to develop the republic, after the pattern of monarchy, in the direction of strong presidential power, and in part unconsciously. Constitutional monarchy [1488] developed from absolute monarchy, and its doctrine is therefore often guided by the aim to make the reduction of power that the formerly unlimited monarch had suffered as a result of the change of the constitution appear as small or insignificant as possible, or even to veil it altogether. Even in an absolute monarchy, it is theoretically possible to distinguish between the levels of the constitution and that of statute. But this distinction is practically insignificant, for the reason that the constitution exhausts itself in the basic principle that every expression of the monarch’s will is a binding legal norm. And since the legal order therefore lacks a determinate constitutional form, that is, since there are no differentiating legal norms that regulate the enactment of statutes in a way that differs from the procedure for changing the constitution, the constitutionality of statutes is not a problem of any importance whatsoever. With the transition to a so-called constitutional monarchy, a decisive change takes place precisely at this point, which is signalled in a very characteristic way by the designation of the new system as a Verfassungs-Monarchie or ‘constitutional’ monarchy. The decisive legal shift expresses itself in the strengthened significance that is now accorded to the concept of constitution: in the legal norm that statutes may only come about in a certain way (namely with the participation of a body of representatives of the people), and in the fact (and this is the rule of the constitution) that this norm cannot be changed as easily as other general rules of law, namely the statutes, i.e. in the fact that a change of this norm necessitates observance of a special, more difficult form, the constitutional form (heightened majority, repeated decision, special constituent assembly, etc.), which differs from the ordinary form of statute. One would think that constitutional monarchy, in particular, should have been the ground on which the problem of the constitutionality of statute, and thus of the guarantees of the constitution, ought to have made itself felt with the greatest imaginable energy. The case was precisely the opposite! The constitutionalist doctrine veiled the legal shift so threatening to the position of power of the monarchies. In contradiction to the legal reality of the constitution, it presented the monarch as the unique or decisive factor of legislation, by declaring statute to be the sole will of the monarch, and by portraying the function of parliament as that of a mere ‘assent’ – more or less marginal, inessential, and inferior. To give an example of the method that was employed here: the well-known ‘monarchical principle’,* which is not deduced from the positive constitution, but imposed on it from the outside, as it were, in order to interpret it in a particular political light, or more correctly in order to reinterpret the positive law with the help of an ideology alien to it; or the famous distinction between the statutory command, which issues solely from the monarch, and the content of statute, which is agreed upon between the monarch and the representatives of the people. The fruit of this method: that it is no longer considered to be a technical imperfection, but rather seen as its deeper meaning, that a statute is to be regarded as valid once it appears in the official gazette, with the signature of monarch, and irrespective of whether the prescriptions relating to the involvement of parliament in legislative decision-taking [1489] are satisfied or not. The decisive progress from absolute to constitutional monarchy is thus, at least in theory, almost completely nullified; at any rate, the problem of the constitutionality of statutes and of its guarantees is altogether avoided. The unconstitutionality of a statute signed by the monarch, or even its nullification on grounds of unconstitutionality, cannot even penetrate into juristic consciousness as a practical legal question. What is more, the constitutionalist doctrine – based less on the letter of the constitution than on its aforementioned ideology – claims for the monarch not only the sanction of legislative decisions but also, with it and in it, the exclusive right to promulgate statutes. In signing the decision of parliament, the monarch is to confirm the constitutionality of the process by which the statute came about. According to this doctrine, then, there is at least one part of the process of legislation that is protected by a kind of guarantee. However, the function of control is claimed precisely by the one power that is itself most in need of being controlled. Admittedly, the act of the monarch is put under responsibility, by virtue of the requirement of ministerial countersignature. But ministerial responsibility,* insofar as it is directed against acts of the monarch, is without practical significance in a constitutional monarchy. And it is altogether irrelevant where defects of the process of legislation must be attributed to parliament, since it can only be enforced by parliament itself.

The view, still widely accepted today, and still defended with the greatest diversity of arguments, that any scrutiny of the constitutionality of statutes must remain off-limits to the organs tasked with the application of law, that the courts may at most claim the right to inquire whether a statute has been duly published, that the constitutionality of the genesis of a statute is sufficiently guaranteed by the right of promulgation of the head of state, as well as the realization of this legal-political view in the positive law of the constitutions of the republics of today, all this derives not least from the theory of constitutional monarchy, whose political ideas, more or less consciously, still influenced the design of modern democracies.

II The concept of constitution

If the question is to be answered whether and in what way the constitution can be guaranteed, i.e. whether and how it is possible to ensure the legality of the levels of legal order that stand directly below the constitution and are immediately related to it, then it is necessary, above all, to gain a clear concept of the constitution.* And it is precisely the insight developed here – that the structure of legal order is hierarchical – which is alone capable of fulfilling this task. We do not go too far if we claim that the immanent meaning, the meaning that was from the beginning intended by the fundamental concept of constitution, as it was already in use in the legal and political theory of antiquity, [1490] is accessible only if one starts out from the theory of legal hierarchy, since the idea of a succession of steps in the creation of law is implicit in the concept of constitution.

If, peeling off the many modifications that the concept of constitution has undergone, one isolates its firm and intangible core, what results is the idea of a highest principle that determines the whole legal and political order, a principle that is decisive for the nature of the community constituted by that order. However one defines the concept of constitution, it always appears with the claim to encompass the foundation of the state, on which the rest of its order is constructed. If one looks more closely, it becomes apparent that the concept of constitution, which, in this respect, overlaps with the concept of form of government, primarily and always refers to a foundational principle in which the distribution of political power finds its legal expression. It is the rule that determines the genesis of statutes, of the general norms whose execution constitutes the activity of the organs of state, namely of the courts and administrative agencies. This – the rule for the generation of the legal norms that primarily form the order of the state, the determination of the organs and of the procedure of legislation – is the essential, original, and narrow concept of constitution. The positing of this basic rule is the indispensable condition for the genesis of the legal norms that regulate the reciprocal behaviour of the human beings that form the community of the state, as well as for the genesis of those legal norms that determine the organs and procedures that are necessary for the application and the enforcement of those rules. The idea that the basic rule of the constitution forms the foundation of all order of the state, and that it is therefore to be as firm and unchanging as possible, leads to the view that it is necessary to differentiate between constitutional norms and statutory norms; the former are not to be as easily changeable as the latter. This gives rise to the concept of constitutional form, as distinct from the ordinary form of statute: the procedure of constitutional legislation (or constitutional amendment) that differs from the ordinary procedure of legislation, insofar as it is tied to special, inhibiting conditions. In the ideal case, this specific form is restricted to the constitution in the narrower and essential sense, with the result that – as we tend to say, though not very felicitously – the constitutional in the material sense, and only the constitution in the material sense, is also the constitution in the formal sense.*

Once positive law provides for a specific constitutional form that differs from the form of ordinary statute, nothing stands in the way of using this form also for norms that do not fall under the concept of constitution in the narrow sense; first and foremost, for norms that do not determine the mode of creation but the content of statutory norms. In this way, a concept of constitution in a wider sense comes into existence. This wider concept is in play when modern constitutions contain not merely norms that concern the organs and the procedure of legislation, but also a catalogue of basic rights and rights of freedom. The primary, though perhaps not the exclusive purpose of such a catalogue is to put up basic principles, guidelines, and limitations for the content of statutes to be enacted in the future. If the equality [1491] of citizens before the law, the freedom of the expression of opinion, the freedom of religion and of conscience, and the inviolability of property are enacted in the typical form – the form of a constitutional guarantee of the subjective rights of subjects to equality, freedom, property, etc. – this means, above all, that the constitution determines that statutes must not merely come about in the prescribed way, but, in addition, must not have a content that violates equality, freedom, property, etc. The constitution, in that case, does not merely have the character of a law of process or procedure, but also the character of material law. The unconstitutionality of a statute can consist not only in the fact that the procedure by which the law came about was defective, but also in the fact that the content of the statute contradicts the basic principles or guidelines laid down in the constitution, or that it exceeds the limitations imposed by it. If one wants to distinguish, in consideration of this point, between the material and the formal unconstitutionality of a statute, one should be mindful that this is permissible only with the qualification that any material unconstitutionality is also a formal unconstitutionality, insofar as a statute that, by virtue of its content, comes into conflict with the relevant provisions contained in the constitution will lose the defect of unconstitutionality once it is enacted as a constitutional statute.* The issue, hence, always reduces to whether a norm has been enacted in the form of an ordinary statute or in constitutional form. Of course, if there is no differentiation, in positive law, between constitutional form and statutory form, only the observance of the latter can ever be at issue. And in that case, the proclamation of basic principles, guidelines, or limitations for the content of law will be meaningless from the legal-technical point of view; it will be no more than a misleading appearance created for political ends. Such is the case with freedoms that are guaranteed in specific constitutional form as soon as the constitution, as is often the case, authorizes the ordinary legislator to restrict these freedoms.*

The constitutional provisions that relate to the procedure of legislation, as well as those that put up basic principles for the content of statutory law, can only concretize themselves in statutes. Guarantees of the constitution are therefore – given this scope of the constitution – nothing but means for the prevention of unconstitutional statutes. However, as soon as the concept of constitution – through the mediation of the idea of constitutional form – is extended to objects other than the procedure of legislation and the basic determination of the content of statutory law, there is a possibility for the constitution to concretize itself in forms of law other than statute; in particular in decrees, and even in individual legal acts. The content of the constitution can make the level of statute superfluous, just as the statute may be drawn up in such a way that it does not stand in need of a decree in order to be applied through individual adjudicative or administrative acts. A constitution could determine, for example, that general legal norms may, under certain determinate conditions, be enacted not through a decision of parliament but rather through an act of government, as in the case of the so-called ‘emergency-decrees’, which stand immediately below the constitution, alongside statutes, replacing or modifying the latter, and are endowed with the same legal force. [1492] Hence, such decrees are immediately related to the constitution (by contrast to decrees that merely execute a statute), and they can therefore be directly unconstitutional, just like statutes, so that the guarantee of constitutionality has to direct itself against them as much as against unconstitutional statutes. Nothing, moreover, rules out the possibility that norms are enacted in constitutional form that are not just basic principles, guidelines, or limitations for the future content of law which can be concretized only by a corresponding statute. It is possible for norms enacted in constitutional form to regulate a subject matter so comprehensively that they can be immediately applied to concrete cases, through acts of adjudication and in particular through administrative acts. Such is the case if the constitution in this extended sense determines how certain of the highest organs of the executive, the head of state, ministers, or the members of the highest courts, etc., are to be selected, and does it in such a way that the creation of these organs can proceed without any further norms (statutes or decrees) that explicate the constitutional provisions in more detail, so that the constitution is immediately executed in the act of appointment, be it a nomination, an election, or a selection by lot. These subjects indeed appear to have been admitted into the concept of constitution typically used by legal theory. One traditionally understands by a constitution (in the material sense) not just the norms concerning the organs and the procedure of legislation, but also those which concern the position of the highest executive organs, and in addition those that determine the basic nature of the relationship of subjects to the authority of the state or, in other words, the catalogue of basic rights and rights of freedom, a catalogue that, to put the matter in juristically correct form, amounts to certain basic principles, guidelines, and limitations for the content of statutes. The practice of modern states, whose constitutional charters, as a rule, exhibit these three elements, typically conforms to this understanding of constitution. If this is the case, not only general norms, like statutes or decrees, but individual acts as well can have the character of being immediate to the constitution, and may therefore turn out to be immediately unconstitutional. The class of individual acts that are immediate to the constitution can, of course, be extended as far as we please, as long as legal norms directly applicable to the concrete case are – for one political motive or another – clothed in constitutional form; so, for instance, if the legal norms that govern the law of associations, or those that regulate the position of religious denominations, are enacted as constitutional statutes. Despite the fact that a guarantee of the legality of acts that execute such statutes formally exhibits the character of a guarantee of the constitution, it is nevertheless evident that the specific form of guarantee of the constitution whose legal-technical design is to be presented in what follows, namely constitutional adjudication, will not easily find a place here. The concept of constitution has now been extended too far beyond its original scope that we derived from the theory of legal hierarchy. The individual character of the unconstitutional act would give rise to an open competition of constitutional adjudication with the administrative courts, which form part of a system of measures that is to guarantee the conformity of execution, and in particular of administration, with statute.

In all cases discussed thus far, we dealt exclusively with acts immediate to the constitution, and therefore with cases of immediate and direct [1493] unconstitutionality. There is a clear contrast between such acts and acts that are not immediate to the constitution, and that can therefore only be mediately or indirectly unconstitutional. If the constitution explicitly lays down, in general terms, the principle of the legality of execution, and especially if it raises the demand for a conformity of decrees with statute, the legality of execution will at the same time – indirectly – constitute a form of constitutionality, and vice versa. Let me highlight in particular, because we are dealing with general norms, the decree that executes a statute. The aim to secure the legality of such a decree, for reasons to be discussed later on, can still be included in the tasks of constitutional adjudication. Apart from this, it should be noted that direct and indirect unconstitutionality cannot always be sharply separated from one another, for the reason that there may occur mixed or transitional forms between both types: for example, if the constitution immediately authorizes all or some organs of administration to enact decrees within their sphere of responsibility, in the course of the concrete implementation of the statutes that they are to apply. The administrative organs, in that case, have the authority to enact such decrees of implementation directly from the constitution. That they are at all permitted to enact decrees directly results from the constitution. However, what they have to decree, i.e. the content of their decrees, is determined by the statutes that stand between the constitution and the decrees through which they are implemented. (It probably does not have to be especially emphasized that these decrees of implementation – especially with regard to their proximity to the constitution – differ from the aforementioned decrees which replace or change statutory norms. The latter are immediate to the constitution and can only be unconstitutional, but cannot violate a statute.) A different case: if the constitution posits basic principles, guidelines, or limitations concerning the content of statutes to be enacted in the future, for example in the form of the aforementioned catalogue of basic rights and rights to freedom, then it will be possible for administrative acts to be unconstitutional in a different sense than the indirect one according to which every administrative act that violates a statute is unconstitutional. If the constitution, for instance, determines that an expropriation may only take place in return for complete compensation, and if an expropriation takes place, in a concrete case, pursuant to a statute of expropriation that conforms to the constitution in demanding full compensation, but in violation of the statute’s as well as the constitution’s determinations concerning compensation, the relevant administrative act will not merely be unconstitutional in the normal, indirect sense, namely as violating a statutory norm. The administrative act in question violates not only a statute, and thus the general constitutional principle of the legality of administration, but also a special principle explicitly enshrined in the constitution: the demand for full compensation in cases of expropriation. The act, hence, oversteps a special limit that the constitution itself imposes on legislation. It would therefore make sense to mobilize an institution that serves to guarantee the constitution as well against unlawful acts of this kind.

The constitutional principle of the legality of execution does not merely demand that every executive act must be in conformity with statute. Above all, it implies that [1494] an act of execution may only take place pursuant to a statute, only as authorized by statute, and thus never without a statutory basis. Hence, if a public authority, a court or an administrative agency, posits an act without any statutory basis, the act in question is not really in violation of a statute, as there is no statute, after all, that could be used to check the act’s conformity with statute. Rather, the act is lawless and thus immediately unconstitutional. It makes no difference here whether the lawless act in question does not even appeal to a statute or whether the appeal is evidently made in bad faith, as for instance in a case where an administrative agency expropriates an urban apartment building pursuant to a law that authorizes the expropriation of agricultural property for the purpose of a land reform. Though this case clearly differs from the case we mentioned earlier of an expropriation that, in violation of statute, takes place without full compensation, one must not overlook that the distinction between acts that are altogether lawless, and thus immediately unconstitutional, and acts that violate a statute, and are thus only mediately unconstitutional, is not, in general, a sharp one.

International treaties – in addition to statutes, certain kinds of administrative decrees, and particular acts of execution – are to be regarded, in particular, among the legal phenomena that are immediate to the constitution. As a rule, constitutions contain prescriptions concerning the creation of international treaties. They authorize the head of state to sign treaties, grant to parliament the right to consent, be it to all or only to some international treaties, prescribe the transformation of international treaties into the form of statute as a condition of domestic validity, and so forth. Since the basic principles of the constitution that determine the content of statutes also apply to the content of international treaties, or at least can be made to apply to treaties (it is also conceivable that treaties might be exempted by positive law from the demand for conformity with these determinations of the constitution, so that the latter would apply only to legislation) international treaties must be regarded as being in the same position as statutes, as far as their relationship to the constitution is concerned. They can be immediately unconstitutional both formally – with respect to the way in which they come about – and materially – with respect to their content. It does not matter here whether the treaty in question has a general or an individual character.

However, the position of the international treaty cannot be determined in a completely unambiguous way, from the point of view of the hierarchy of legal order. The conception of a treaty as a method of the will-formation of the state that is determined by the state’s constitution, and immediately subject to the latter, proceeds from a standpoint that regards the constitution as the highest level of legal hierarchy, from a standpoint that one might describe as that of the primacy of an individual state’s legal order.* If one elevates oneself above this point of view, if one assumes the validity of an international law that stands above the legal orders of the particular states or, in other words, the primacy of international legal order, the international treaty will appear as a partial legal order that stands above the contracting states, and that is created, in accordance with a legal norm of international law, by a specific organ [1495] of the international legal community, an organ which is composed of the representatives of the contracting states. To determine the mode of appointment of the partial organ (head of state, foreign minister, parliament, etc.) that functions as a member of the complete, international organ that enacts the treaty-based order, international law refers to the respective national legal orders or their constitutions. From this point of view we must conclude that the treaty takes precedence over national statute, and even over the national constitution, insofar as neither an ordinary statute nor a constitutional amendment can derogate from an international treaty, whereas the reverse is possible. An international treaty can – according to the principles of international law – lose its validity only through another international treaty or other matters of fact that are specially qualified as having that legal consequence by international law; but not through a unilateral act of one of the contracting parties, i.e. not through a national statute. If a national law, and be it a constitutional amendment, comes to contradict an international treaty, it is a legally defective law, and perhaps a legally defective constitutional law, as it fails to conform to international law. Such a law immediately violates the international treaty, and mediately the legal rule of contract, the rule of international law: pacta sunt servanda. Of course, statutes are not the only acts of state that may be in violation of international law. Other acts of state can be as well, and this not only by virtue of violating the principle of contract, mediately or immediately, but also by virtue of violating other rules of general international law. If one accepts, for instance, that it is a norm of international law that foreigners may be expropriated only in return for full compensation, any national constitution, any national statute, any national administrative act, or any national judicial decision that expropriates a foreigner without compensation is in violation of international law. We must observe, however, that international law itself does not provide for a sanction of annulment of legal acts of individual states that violate international law. International law has not yet developed a procedure through which such legally defective acts can be annulled by an international forum. If they are not quashed in a domestic legal procedure, they will retain their validity. The only available sanction under international law, in the last instance, is war against the state that refuses to remove its act in violation of international law. This, however, does not change the fact that international law – once we presuppose its primacy – is capable of providing a standard of the legality of all national legal acts, the highest, the constitution, included.

III The guarantees of legality

Now that the concept of constitution and thus the nature of constitutionality and unconstitutionality have been sufficiently clarified, we can examine the question of what guarantees are available for the protection of the constitution. These are the general guarantees that modern legal technique has developed with a view to the legality of acts of state in general. [1496] They are of a preventative or a repressive and of a personal or a material nature.

The guarantees of a preventative nature want to prevent the occurrence, from the beginning, of legally defective acts; those of a repressive nature want to react against the legally defective act that has already taken place, to prevent its future repetition, to make good the damage that it caused, to remove the legally defective act, and perhaps to replace it with a lawful act. It is possible for both of these moments to combine in one and the same protective measure. One of the possible preventative guarantees, whose number is of course very large, which is of special relevance in the present context, is the organization of the law-making public institution as a court; that is, the independence of the institution, which is guaranteed in a specific form (through protection against dismissal or transfer to another position), and which consists in the fact that the organ, in the exercise of its function, cannot be legally obligated by any individual norm (command) issued by another organ, and in particular not by the norms of any organ otherwise superior, or by those of an organ that belongs to another branch of public administration. It follows from this that the judicial organ is bound only to general norms, and, above all, only to statutes and to decrees that have a statutory basis. (The review of statutes and decrees that is granted to courts is another issue.) The idea, still frequently encountered, that it is only the legality of adjudication that can be guaranteed in this way rests on the mistaken assumption that there is some sort of essential difference between adjudication and administration, from a juristic, legal-theoretical, or legal-technical point of view. However, such a difference between adjudication and administration – or, for that matter, between execution and legislation – is not to be found, in particular with regard to the relation to the respective higher-level norm, which is decisive as far as the demand for the legality of the function is concerned. The difference between adjudication and administration consists exclusively in the organizational position of courts. The proof for this claim: the system of administrative adjudication, which either consists in the fact that acts of administration – that is, acts that would normally be performed by administrative agencies – are performed by courts, i.e. by organs that are organized as courts, or else in the fact that these acts, after they have been set by an administrative agency, are reviewed for their legality by a court, and are invalidated in case of legal defectiveness, and in some cases even reformed, i.e. replaced with a lawful act. The whole traditional opposition of jurisdiction and administration, and the whole dualism of the state’s apparatus of organs based on it, i.e. of the state’s apparatus of organs of execution, can only be explained historically, and it will – unless all symptoms mislead, symptoms that already indicate a growing similarity between judicial and administrative organs – disappear in the course of future development. And it also can only be explained historically why we regard the independence of an organ from the individual norm of another as a guarantee of the legality of the former’s function.

The organization of the law-making organ as a court is not only the characteristic preventative guarantee of the legality of the act to be enacted, but also the first in [1497] the group of what we have called personal guarantees. The others are the responsibility under criminal or disciplinary law and the civil liability of the organ that enacted the legally defective act. The material guarantees which exhibit, at the same time, a clearly accentuated repressive character, are the nullity or the annullability of the legally defective act.*

Nullity means that an act that claims to be a legal act – i.e. an act whose subjective meaning it is to be a legal act, and in particular an act of state – objectively fails to be such, for the reason that it is legally defective because it does not conform to the conditions that a higher-level legal norm prescribes for it. An act that is null lacks any legal character from the beginning, so that there is no need for another legal act to strip it of its pretended legal quality. If such a second legal act is required, we are faced with annullability, and not with nullity. If faced with an act that is null, everyone, a public organ as much as an ordinary subject, is entitled to inquire into the act’s legality, to recognize it as legally defective, and accordingly to treat it as invalid and non-binding. Only where the positive legal order restricts the power to inquire into the legality of any act that carries the subjective meaning of a legal act, by reserving this power, under certain determinate conditions, to specific authorities, as opposed to leaving it to everyone under all circumstances, can an act that exhibits some defect of legality be regarded as merely annullable, and not as already null a priori. In the absence of such a restriction, every act that exhibits some legal defect would have to be regarded as null, i.e. as a non-legal act. As a matter of fact, positive legal orders contain very far-reaching restrictions of the competence – which originally pertains to everyone – to treat legally defective acts as null. In general, private legal acts and legal acts set by public institutions are treated differently in this respect. By and large, there is a tendency to treat a legal act that has been set by a public institution as valid and binding, even if it is legally defective, for as long as it has not been removed by another legal act of a public institution. The question whether a legal act set by a public authority is legally defective or not is not to be decided, without further ado, by the subject or by the public institution to which this act, demanding obedience, is directed. Rather, it is to be decided by the institution itself that enacted the act the legality of which is impugned, or else by another public institution whose decision can be brought about by going through some specified procedure. This principle, which is by and large accepted by the positive legal order, and that one might call the principle of the self-legitimation of the acts of public authorities, is subject to certain limits. No positive legal order can determine that absolutely every act that claims to be a legal act proceeding from a public authority is to be regarded as such as long as it has not been invalidated, due to legal defectiveness, by another act of public authority. If such an act was posited by a human being who in no way possesses the quality of a public authority, it would evidently be meaningless to have to initiate a procedure with a public authority [1498] in order to bring about the annulment of the act. On the other hand, it is equally impossible to regard every act posited by a public authority that lacks competence or that is not properly constituted, or every act enacted through a defective procedure, as a priori null. Though the problem of absolute nullity is a very difficult one from the legal-theoretical and the legal-technical point of view, it is of interest to the question of guarantees of constitutionality only insofar as it must be stated that the possibility of nullity, which can never be altogether excluded by positive law, also applies to those acts that are immediate to the constitution, and that the possibility of the nullity of these acts, in a certain sense, represents a guarantee of the constitution. Not every act that describes itself as a statute has to be regarded as a statute by subjects or by law-applying public institutions. There can, without a doubt, be acts that merely have the appearance of a statute. But if we ask for the boundary that separates the enactment of a rule that merely appears to be a statute, and which is a priori null, from a defective but valid legislative act, i.e. from a statute that fails to conform to the constitution, legal theory is incapable of answering the question with a general formula. The positive legal order alone could undertake this task, but, of course, it usually refuses to do so, or does not do so in a conscious and precise way. It usually leaves the answer to this question to the public authority that must decide in case someone has refused, as a subject or organ of state, to pay obedience to the act that is under consideration, claiming that it merely appears to be a statute but isn’t. At that point, however, the act in question already moves from the sphere of absolute nullity into the sphere of mere annullability, since the decision of the competent public authority that the act to which obedience was refused was not a legal act can only be regarded as an annulment of the act, with effectiveness ex tunc.* Things are no different in cases where the positive legal order lays down the minimum conditions that must be satisfied in order for the legal act not to be null a priori; for instance, where the constitution determines that everything that is promulgated as a statute in the official gazette must be regarded as a statute, notwithstanding other legal defects, as long as it is not annulled by the authority empowered to do so. After all, the determination of the question whether the minimal condition is satisfied or not must, at the end of the day, be made by the competent public authority, and authentically, since it would otherwise be possible for everyone to withdraw from obedience to every statute on the basis of the mere claim that the minimal condition is unfulfilled. From the point of view of the positive law, the position of someone to whom an act addresses itself, demanding obedience, is invariably this: he can, if he takes an act to be null, refuse to pay obedience, but he does so at his own risk, i.e. under the threat that he will be made responsible for disobedience, and that the public authority in front of which he is made responsible will not judge the act to be null, but will declare the minimal conditions prescribed by the positive legal order for the validity of the act to be satisfied, regardless of the possibility of later annulment. If, on the other hand, the authority in question assumes the minimal condition not to be satisfied, then its decision amounts to the annulment of the act with retroactive effect extending back to the time of its enactment. This interpretation is mandated by the fact that the [1499] decision is the outcome of a procedure, a procedure which judges the nullity of the act – a nullity at first only claimed by the party accused of disobedience – so that the nullity cannot at all be regarded as given in advance of the conclusion of the procedure, since there is still a possibility that the procedure might lead to a decision that denies the nullity of the act, and since the decision must necessarily have a constitutive character, even if, in its wording, it pronounces the nullity of the act. From the point of view of positive law – that is, from the point of view of the public authority that decides on the act that is allegedly null – the question is therefore only ever one of annullability; and only in the sense that the fact of nullity can be represented as a limiting case of annullability (annulment with retroactive effect).

The annullability of a legally defective act signifies the possibility of removing it, together with its legal effects. An annulment can exhibit several different degrees, and this with respect to its material as well as with respect to its temporal extension. With respect to the first aspect, there are the following possibilities: the annulment (cassation) of the legally defective act remains restricted to a concrete case. If the act in question is an individual act, this restriction will of course apply automatically. Things are different if we are faced with a general norm. The cassation of a general norm remains restricted to a concrete case if the legal order determines that the public authority (court or administrative institution) which is to apply the norm is entitled or obliged to refuse application in the concrete case, should it judge the norm to be legally defective, and that it is thus to decide or to decree, in the concrete case at hand, as though the general norm that it judges to be legally defective did not exist. Apart from this, however, the general norm remains valid and is to be applied in other cases by other public institutions, if these are either not authorized to investigate into and to decide on the legality of a norm they are to apply, or if they, though authorized to do so, hold the norm to be legal. Since the public authority which is called upon to apply the general norm can remove the norm’s validity for the concrete case through its judgment of legal defectiveness, it has the power to annul the general norm, as the removal of the validity of a norm and its cassation are one and the same thing. It is only that the cassation is a partial one that is restricted to the individual case. This, in effect, is the position which courts (but not administrative agencies), under many modern constitutions, hold towards administrative decrees, and in some states (for example in the United States of America) also towards statutes. However, such a far-reaching authority of the courts towards statutes is by no means the rule. In most cases, courts are not permitted to engage in a comprehensive review of the legality – that is, of the constitutionality – of statutes. As a rule, their power of review is fairly restricted. The courts are permitted only to inquire into the proper promulgation of a statute and may, accordingly, refuse to apply it to a concrete case only on the ground of a legal defect in its promulgation. [1500]

The shortcomings and the insufficiency of a cassation of legally defective norms that remains restricted to the individual case are self-evident. It is above all the lack of consistency and the lack of legal security that results from it that make themselves felt very uncomfortably, if the one court refuses to apply a decree or even a statute as legally defective, whereas another court does the opposite, while administrative institutions – if they are also tasked with the application of the same decree or statute – are in no way permitted to refuse to do so. A centralization of the competence to review the legality of general norms is surely justifiable in every respect.

Should one decide to transfer the power of review to one single public authority, it will likewise become possible to extend the cassation beyond the individual case. We would then be faced with the annulment of a general norm as a whole, i.e. for all possible cases to which the norm, according to its meaning, would have to be applied. That such a far-reaching power can only be vested in a supreme and central authority should be self-evident.

With respect to the temporal dimension, the effect of a cassation can be restricted to the future, or it may extend to the past. In other words, the annulment of the legally defective act may take place with or without retroactive effect. This differentiation, of course, is meaningful only with regard to acts that have a lasting legal effect; it is therefore relevant, above all, with reference to the cassation of general norms. Out of consideration for the ideal of legal security, one should, in general, make the cassation of a general legal norm on the ground of legal defectiveness effective only pro futuro, i.e. starting from the time of the cassation. Here, one even has to take into account the option of letting the cassation take effect at a later point in time. Just as the coming into effect of a general legal norm, such as a statute or a decree, is preceded, for good reason, by a vacatio legis, it might appear desirable, for analogous reasons, to let a general norm lose its validity only after the expiry of a certain period of time from the decision to annul. Circumstances can, nevertheless, make it necessary for the cassation of a general norm to have retroactive effect. In this context we have to think not only of the extreme case, already mentioned above, of the unrestricted retroactive effect in cases where the annulment of an act is an acknowledgement of its nullity, and where the legally defective act must be recognized, according to the free discretion of the authority empowered to annul, or according to the minimal conditions for the validity of the act prescribed by positive law, as a mere appearance of a legal act. Above all, we here have to consider the possibility of restricting the retroactive effect of an annulment that, in principle, is effective only pro futuro to certain individual cases or to a specific category of cases, i.e. to provide for a limited retroactive effect of the annulment; a question to which we will have to return in a later context. [1501]

For the legal-technical implementation of the annulment of an act it is of importance, moreover, whether the cassation can only be performed by the organ that posited the legally defective act, or whether another organ is empowered to do so. Considerations of prestige, in particular, lead to the choice of the first of these two modalities. One is concerned to avoid the danger that the authority of the public institution that enacted the legally defective norm, and that is regarded as a highest organ, or that enacted the norm under the supervision and responsibility of a highest organ (especially if the norm at hand is a general norm), will suffer, by virtue of the fact that some other institution appears to have the power to annul its acts, and thus to put itself above it, though it is supposedly to be regarded as the ‘highest’. Appeal is made not only to the ‘sovereignty’ of the organ that posited the legally defective act, but also to the dogma of the ‘separation of powers’, in order to avoid the cassation of the act of the one public authority by that of another. This argument is invoked especially in cases that concern acts of the highest administrative institutions, where the authority empowered to annul would therefore have to stand outside of the administrative organization of the state and would, with respect to its function as well as with respect to its position, have to have the character of an independent judicial authority or, in other words, of a court. Given the more than doubtful character of the distinction between jurisdiction and administration, the reference to the ‘separation of powers’ is as unsound, in this context, as the appeal to the ‘sovereignty’ of an organ. Both arguments, however, play a special role when it comes to the question of guarantees of constitutionality. Under the pretext that the ‘sovereignty’ of the organ that enacted the legally defective act or the ‘separation of powers’ are to be preserved, one leaves the annulment of the legally defective act to the discretion of the enacting organ itself and permits nothing more than a non-binding petition for annulment on the part of the affected party (the so-called ‘representation’). Or else, there is a regular procedure that is supposed to lead to the annulment of the legally defective act by its author, but the motion that initiates the procedure obliges the authority only to go through the procedure, but not to end it in a certain way, namely with the cassation of the impugned act. The cassation thus remains subject to the discretion of the organ that enacted the legally defective act, a discretion which, though bound by statute, is not controlled by any higher organ. Finally, a third case should be mentioned, though it already forms the transition to the second type of guarantee presented here: Another institution is empowered to decide on the question of the legality of the act, but the cassation of the legally defective act nevertheless remains reserved to the organ that enacted the act. The latter organ, however, may be put under a legal obligation, as a result of the finding of the former, to annul the act that has been judged to be legally defective. The fulfilment of this duty may even be subject to a deadline. That this modification is equally incapable of offering a sufficient guarantee hardly requires any further proof. A sufficient guarantee is given only if the cassation of the legally defective act is to be performed directly by an organ that is altogether distinct from and independent of the organ that enacted the legally defective act. [1502] If one holds on to the typical division of the functions of state into legislation, jurisdiction (judicature), and administration, as well as to the consequent classification of the organism of the institutions of the state into three groups of organs – one legislative, one jurisdictional (or judicative), and one administrative branch – one must distinguish whether the cassation of the legally defective act remains within the same branch of public authority; whether, for instance, an act of administration or a judicial decision will, in turn, be annulled only by another act of administration or jurisdiction, i.e. through the act of an authority that belongs to the same group of organs, a higher administrative organ in the one case, a higher court in the other, or whether the authority empowered to annul belongs to another group of organs. The guarantee of the legality of acts of state that consists in the possibility of going through successive stages of appeal within one branch of public authority belongs to the first of these types; a system of administrative justice is an instance of the second. It is characteristic for modern legal orders that the legality of the acts of courts is guaranteed, almost without exception, through means of the first type alone. It is widely believed that the so-called independence of the courts is already by itself a guarantee of the legality of the decision that is to be taken.

The cassation of a legally defective act gives rise to the question of its replacement with an act that is lawful. In this regard, we have to distinguish, from a technical point of view, between two possibilities. The public authority empowered to annul the legally defective act may also possess the competence to put a lawful act in the place of the impugned and defective act. It may have the competence, in other words, not just to annul, but also to reform. But it is also possible to leave the enactment of a lawful act to the public authority whose legally defective act was annulled. If the latter authority, in doing so, is bound by legal opinion articulated in the findings of the authority empowered to annul – for instance in the form of reasons of decision – its own independence is thereby restricted; which, in the case of the annulment of a judicial decision, is not without significance for the assessment of the independence of judges as a specific guarantee of the legality of execution.

IV The guarantees of constitutionality

Of the legal-technical measures that serve the purpose of ensuring the legality of the functions of the state which have been presented in the preceding pages, one above all is to be regarded as the most effective guarantee of the constitution: the annulment of the unconstitutional act. This is not to say that other means should not be used as well in order to secure the legality of acts standing under the constitution. Of course, the preventative, personal guarantee, namely to organize the organ enacting the act as a court, is out of the question from the start. The power of legislation, which is the most relevant here, cannot be vested in a court; not so much because of the difference in the respective functions of legislation and [1503] adjudication, but rather for the reason that the organization of the legislative organ is inevitably determined by considerations other than that of the constitutionality of its function. Here, the fundamental opposition between democracy and autocracy alone is decisive. The repressive guarantees of the responsibility of the organ that enacts the legally defective act, under criminal or civil law, however, may usefully be taken into consideration. Of course, insofar as we are dealing with the process of legislation, this kind of guarantee cannot be applied to the parliament as such, or to all those of its members who participated in the decision. There are several different reasons why a collegial organ is not a suitable subject of criminal or civil legal responsibility. But it is possible to make the individual organs that participate in the process of legislation, such as the head of state or the ministers, responsible for the constitutionality of a statute, the more so if the constitution determines that the head of state, or the ministers only, assume responsibility for the constitutionality of the legislative process, with their countersignature or their promulgation of the statute. It is indeed the case that the institution of ministerial responsibility, peculiar to modern constitutions, also stands in the service of the constitutionality of statute. It is self-evident that the personal responsibility of the organ can also be used in order to provide a guarantee of the legality of decrees, and in particular of the legality of individual acts that stand immediately under the constitution. In this last respect, the liability for damage caused by the legally defective act is particularly relevant. However, ministerial responsibility – as can easily be established from constitutional history – is in and of itself not a very effective means; all personal guarantees, moreover, are insufficient insofar as they leave the continuing validity of the legally defective act, and in particular also that of the unconstitutional statute, untouched. Where a legal state of affairs of this kind persists, one cannot, strictly speaking, say that the constitution is guaranteed. That is the case only once there is the possibility of annulling the unconstitutional act.

V The juristic and political significance of constitutional adjudication

As long as a constitution lacks the guarantee, presented in the foregoing, of the annullability of unconstitutional acts it also lacks the character of full legal bindingness in the technical sense. A constitution according to which unconstitutional acts, and in particular unconstitutional statutes, must remain valid because they cannot be annulled on the ground of their unconstitutionality amounts to little more, from a legal-technical point of view, than a non-binding wish; though one is in general unaware of this fact, for the reason that a politically motivated jurisprudential theory prevents the growth of that awareness. Any statute whatsoever, any simple decree – yes, even any general legal transaction of private parties – surpasses such a constitution in legal force, surpasses it though the constitution stands above them all, though all lower levels of legal order draw their validity from it. Legal order, after all, takes care that every act which puts itself in contradiction with any norm of a level lower than the constitutional can be annulled.

And this lesser degree of real legal force stands in a stark disproportion to the appearance of rigidity, bordering on inflexibility, that one confers on the constitution through the enactment of conditions of amendment that are very difficult to meet. Why take such precautions if the norms of the constitution, though more or less un-amendable, are almost non-obligatory? To be sure, even a constitution that does not provide for a constitutional court or a comparable institution for the annulment of unconstitutional acts is not altogether irrelevant legally. Its [1525] violation can, at least where the institution of ministerial responsibility exists, lead to some kind of reaction against certain organs who were involved in the enactment of unconstitutional acts, on the assumption that the behaviour of these organs was culpable. But apart from the fact that this guarantee – as has already been stressed – is in itself not very effective, since it leaves the validity of the unconstitutional statute untouched, we will not be in a position to assume, in the absence of proper guarantees of constitutionality, that it is the meaning of the norms that determine the procedure of legislation and the permissible content of statutes or, in other words, that it is the meaning of the constitution to indicate a unique procedure of legislation as the only possible way to legislate and to provide a real direction to the content of legislation. Admittedly, the constitution says, given its wording and given its subjective meaning, that statutes are to be enacted in such and such a way, and only in such and such a way, and that they may or may not have this or that content. But its objective meaning is: statutes are to be regarded as valid even if they come into being in some other way and even if their content violates the directives of the constitution. One is forced to interpret the constitution in this way if unconstitutional statutes, despite their legal defects, are to be regarded as valid; since even such statutes must be able – as valid statutes – to base themselves on some constitution, they must draw their validity from somewhere, and hence from the constitution; they must somehow, since they are valid, also be constitutional. But this means that the procedure of legislation explicitly laid down in the constitution, and the guidelines contained in it, do not, contrary to appearances, constitute an unambiguous determination. They must be understood, rather, in the sense of an alternative:* either in this way or, if not in this way, then in other ways, almost without limit. That constitutions which lack the guarantee of the annullability of unconstitutional acts are not understood in this way is, ironically, the result of the theory – it has already been mentioned here several times – that conceals the true state of affairs for political reasons, reasons which are at odds with the political interests that are expressed in the constitutions in question. A constitution, whose prescriptions regarding legislation may be violated, without a consequent annulment of the unconstitutional statutes, has no more legal force, vis-à-vis the lower levels of the legal order internal to the state, than international law vis-à-vis the legal order of the individual state. If the latter puts itself in contradiction with international law, with any one of its acts, from the constitution down to the last administrative act, the validity of the acts in question will remain unaffected. Of course, the state that is affected by the violation of international law may, as a last resort, go to war against the state that violated international law. This, however, is only a penal sanction, not an annulment of the legally defective act. In just the same way, a constitution that lacks constitutional adjudication can react to its violation only with a penal sanction made available by the institution of ministerial responsibility. It is this reduced legal force of international law that leads some authors – though mistakenly – to deny its legal character altogether.* And the interests that oppose the institution of an international court endowed with a competence to annul, as well as the legal-technical strengthening of international law that could be brought about in this way, are very similar to those [1526] that work against the increase in the legal force of the constitution that goes along with the function of a constitutional court.

One needs to keep all this in mind, in order to appreciate the significance of the question of constitutional adjudication that we are discussing here.

Apart from this general significance that it has for every constitution, the question may also enjoy a particular significance, depending on the specific structure of the constitution. This applies, above all, to a democratic republic. Institutions of control belong to the conditions of existence of the latter. The best way to defend this form of state against the many criticisms, some of them justified, to which it has been subjected in recent times is to put in place all the guarantees that can be given for the legality of the functions of the state. The control of the functions of the state must be strengthened to the same degree as democratization progresses. This is the point of view from which constitutional adjudication is to be evaluated here. Insofar as it makes sure that statutes come into existence in conformity with the constitution, and in particular also that their content is constitutional, constitutional adjudication serves the function of an effective protection of the minority against assaults on the part of the majority, whose rule becomes tolerable only by virtue of the fact that it is exercised in legal form. The specific form of constitution, which typically consists in the fact that a constitutional amendment is tied to the requirement of a heightened majority, ensures that certain fundamental questions can only be resolved with the participation of the minority. The simple majority does not have the right – at least when it comes to certain issues – to impose its will on the minority. It is only through a statute that is unconstitutional, because it has been enacted by a simple majority, that the majority can interfere with the minority’s constitutionally protected sphere of interest against the latter’s will. The constitutionality of statutes is therefore a pre-eminent interest of the minority; regardless of what type of minority it is, be it class-based, national, or religious, whose interests are protected by the constitution in some way.

This applies in particular to the case of a shift in the proportion between majority and minority, when a majority turns into a minority, but still remains strong enough to prevent the qualified decision that is necessary to bring about a lawful change of the constitution. If one does not take the essence of democracy to consist in unfettered majority rule, but rather in the continuing compromise between the different parts of the people that are represented in parliament by the majority and the minority, then one should acknowledge that constitutional adjudication is a particularly suitable means to realize that idea. The mere threat of making appeal to the constitutional court may well turn out to be a sufficient instrument in the hands of the minority to prevent unconstitutional violations of its interests on the part of the majority, and thus, in effect, to prevent a dictatorship of the majority that is no less dangerous to social peace than the dictatorship of a minority.

Constitutional adjudication attains its greatest importance, however, in a federal state.* One does not go too far in making the claim that this political idea, [1527] from a legal point of view, cannot reach its completion without the institution of a constitutional court. The essence of a federal state – assuming one regards the issue not as a problem of the metaphysics of the state but rather, from a realistic perspective, as a problem of organizational technique – consists in the fact that the legislative and executive activity of a legal community that is regarded as a state are divided between one central organ, an organ referred to as ‘union’ or ‘Reich’, etc., with competence for the state as a whole and all of its territory, and several local organs with competence only for parts of the territory or for some limbs of the state, which are referred to as ‘constituent states’, ‘Länder’, ‘cantons’, etc., but in such a way that elected representatives of the constituent parts, who may be appointed indirectly (through election by local parliaments or designation by local governments) or directly (by the people of a constituent state), participate in central legislation, and perhaps also in central administration. The federal state represents a specific case of decentralization. The legal regulation of this decentralization forms the essential content of the constitution of the whole. The latter will determine, above all, which matters are to be regulated by statutes of the union and which are to be governed by local statutes of the constituent states; and it will also divide the responsibility for executing the law between the central and the constituent states. This division of competences is the political core of the idea of a federal state. From a legal-technical point of view, the constitution does not only, as in the case of a unitary state, determine the procedure of legislation and provide certain guidelines for the content of statutes. It also delimits the material sphere of the validity of Reich as well as of Länder statutes. Every violation of these boundaries, as drawn by the constitution, is a violation of the fundamental law of the federal state, and the preservation of the boundaries of competence that the constitution draws between the union and the constituent states is a political question of life and death. Of course, it is also perceived as such in a federal state, a state that is invariably a stage of the most passionate in-fighting about competence. If it exists anywhere, then the need for an objective authority that can mediate these fights in a peaceful way, for a forum in which these quarrels can be raised as questions of law, and be decided as such, exists here. The institution required here, needless to say, is none other than a constitutional court. After all, every violation of the competence of the union by a constituent state, or of that of a constituent state by the union, is a violation of the constitution; a violation of the common constitution which integrates the union and the Länder, the Reich and the constituent states, into one whole. This common constitution, whose most essential element is the division of competences, must not be confused with the special constitution of the union (Reich), which stands below it, and which, just like the constitutions of the constituent states (Länder), is only the constitution of a part of the whole community, even where one and the same organ is competent to change the constitution of the whole and to change the constitution of union (Reich).

When it comes to judicial or administrative acts that violate the distribution of competences, the possibility of making appeal within the judicial or administrative hierarchy of the union or the constituent states provides an initial opportunity to annul such acts on the ground of their failure to conform to statute. Whether this guarantee suffices effectively to prevent acts [1528] of union administration interfering with the competences of the constituent states, or acts of the administration of the latter interfering with the competences of the union, must remain open to doubt, especially if the union and its constituent states lack a common supreme administrative court. Such a court would, if it had to control acts for their conformity with the distribution of competences, i.e. for their constitutionality, already function as a constitutional court – at least indirectly. The question of the competences that are to be allocated to the constitutional court must here be answered in a slightly different way from that of a centralist and unitary state, due to the opposition of the interests of the central and the constituent states that is a characteristic of federal states, and due to the strong need that prevails here for an objective institution that can, as an organ of the overarching community, play the role of a referee between the fundamentally co-ordinate legal communities of the union and the constituent states. It appears, at the least, to be debatable whether a federal constitutional court should also be tasked with the legal control of individual administrative acts – but only with regard to their conformity to the distribution of competences. The demand, at any rate, should be uncontroversial that statutes and decrees of union as well as of the constituent states be open to challenge in a constitutional court that can guarantee sufficient objectivity due to its composition, which should give equal representation to both levels of government, and that is competent – as an organ of the constitution of the whole, and not as a one-sided organ of the union or of the constituent states – to annul such statutes and decrees for the reason that they violate the constitution of the whole, and in particular, hence, for the reason that they disturb the constitutional distribution of competences.

It is one of the paradoxes of the theory of the federal state that it advocates the principle: the law of the Reich takes precedence over the law of the Land* as a principle that allegedly conforms to the essence of a federal state. With this alone, it has already obscured the necessity of a constitutional court in a federal state. It is easy to show that nothing can be as much in conflict with the idea of a federal state as this principle, which makes the political and legal existence of the constituent states depend on the discretion of the union, and thus on the discretion of a mere part of the federal state as a whole, by allowing the union to interfere with the competences of the constituent states by means of ordinary statute, or even through simple decrees, in a way that conflicts with the constitution of the whole, and to usurp the competences of the constituent states in an unconstitutional manner. If the idea of a federal state, as it has found expression in the constitution of the whole, is to be preserved, the law of the Reich must not be allowed to violate the law of constituent states any more than the law of constituent states must be allowed to violate the law of the Reich. Rather, one as much as the other is, both are, to be evaluated in the same way, in their respective relationship to one another, according to the constitution of the whole which delimits their spheres of validity against one another. A legal act of the union that steps over the boundary drawn for it by the constitution of the whole, and that penetrates into the sphere of competence of the constituent states, has, from a legal point of view, no more right to exist than the legal act of a constituent state that interferes with the competence of the union. This principle alone comports with the essence of the federal state, and it cannot be realized in any other form than by way of a constitutional court. It would, finally, have to belong to the natural competence of the constitutional court, according to the idea of a federal state, [1529] to decide on all violations of duty of which both the constituent states as well as the union can make themselves guilty if their relevant organs, in the pursuit of their office, violate the constitution of the whole. What is usually called the federal execution of the constitution,* and poses such a difficult problem for the theory and practice of the federal state, should be permissible – irrespective of whether it takes place through the primitive form of a strict and collective liability of the community as such, or in the technically more advanced form of an individual criminal liability of the responsible organ – only as the implementation of a judgment issued by the constitutional court, a judgment in which the court has determined some behaviour, be it on the part of the union or of a constituent state, to have been unconstitutional.

The tasks that are to be discharged by a constitutional court in the framework of a federal state put into stark relief the affinity between constitutional adjudication and an international adjudication that serves the purpose of the preservation of international law, not least with respect to the mutual proximity of the levels of legal order that are to be guaranteed. And just as the one aims to make war between peoples superfluous, the other proves its worth – and serves its most fundamental purpose – as a guarantee of political peace within the individual state.