1.1. The coercive order paradigm and the bellum justum doctrine

Kelsen's search for a typological differentiation of normative orders is predicated by the pureness of the resulting theory. In order to keep morals out of law, one has to be able to point to a difference between these two normative orders. The question is thus whether a legal norm, whether a legal normative order, is a unique type of norm or normative order: ‘Es gilt festzustellen, ob die gesellschaftlichen Phänomene, die mit diesem Wort [Recht] bezeichnet werden, gemeinsame Merkmale aufweisen, durch die sie von anderen, ihnen ähnlichen Erscheinungen unterschieden werden können.’Footnote ¹¹ Precisely what does this differentiation mean with respect to the kind of difference and differentiation? On the one hand, a mere empirical differentiation according to word use, as employed by Kelsen in Reine Rechtslehre, could produce the result ‘daß mit dem Wort “Recht” und seinen anderssprachlichen Äquivalenten so verschiedene Gegenstände bezeichnet werden, daß sie unter keinem gemeinsamen Begriff zusammengefaßt werden können’,Footnote ¹² because we would employ a descriptive approach, seeking the common denominator. On the other hand, certain criteria could be made part of the very idea of law, thus becoming necessary, a priori elements of ‘law’. However, the only necessary element for any norm is that it is a norm – that is, that it is an ‘ought’, the claim to be observed.Footnote ¹³ On that basis there can be no difference between norms.

Kelsen, however, chooses an empirical approach. The key element distinguishing positive legal orders from other kinds of positive normative orders for him is that legal orders are coercive orders (Zwangsordnungen):

Thus coercion prescribed as reaction against certain behaviour is the distinguishing feature of law – the Zwangsnormpostulat (coercive order paradigm). Every legal order contains coercive norms; they prescribe certain human behaviour by attaching a coercive act to the opposite behaviour. The typical coercive norm would make that behaviour – the unwanted behaviour – the condition for a prescription to organs to sanction the human responsible: if someone commits murder, they are to be punished by life imprisonment. This, for Kelsen, is the crucial difference between law and morals: while positive moral orders proscribe certain behaviour and may or may not attach a sanction to behaviour in a second norm, law prohibits behaviour specifically by attaching negative sanctions to the contrary behaviour.Footnote ¹⁵

In order to apply such a norm, an ‘organ’ is needed. An organ is nothing more than a bundle of norms referring to human behaviour which authorizes the application of the law and thus the creation of norms. For example, a penal procedure code authorizes a human to apply the penal code and to create individual norms stipulating sanctions. Hence if this judge authoritatively finds that A has committed murder, the individual norm thus created might prescribe that A ought to be punished by life imprisonment.Footnote ¹⁶ For Kelsen there is a development of social orders to restrict progressively the use of physical force in intra-societal relations, to establish a force monopoly of the community. Thus the use of physical coercion becomes either sanction authorized by the legal order or a delict, against which sanctions are to be directed. The exclusive legal classification of physical force as either sanction or delict is not a logical necessity; it has to be proved to be part of a given positive normative order.Footnote ¹⁷ Primitive legal orders, however – legal orders with a low grade of centralization and no division of labour – do not have special organs for that task and will most likely authorize decentralized enforcement. Such legal orders typically authorize self-help; that is, ‘the legal order leaves these functions to the individuals injured by the delict’:Footnote ¹⁸

However, there is a danger of concluding that a norm is only a member of a legal order if and when it stipulates a sanction. The Kelsen of the second edition of Reine Rechtslehre makes it plain that a norm's validity – hence both its existence and its membership in a given normative order – derives from superior norms, not from its stipulating a sanction: ‘Eine einzelne Norm ist eine Rechtsnorm, sofern sie zu einer bestimmten Rechtsordnung gehört, und sie gehört zu einer bestimmten Rechtsordnung, wenn ihre Geltung auf der Grundnorm dieser Ordnung beruht.’Footnote ²⁰ The element of coercion for Kelsen merely identifies an order as legal order and does not unmake non-coercive laws contained in a positive legal order.

Kelsen had to ‘successfully’ apply this theory to positive international law, for had he not been able to do so, he would have been forced to deny international law as lawFootnote ²¹ or, alternatively, to abandon his formulation of the specifically legal formulation of norms.Footnote ²² He attempted to apply his theory by simply asking whether international law fulfilled the criteria he had set for law in general, namely whether international law prescribes coercive acts as sanctions:Footnote ²³

What institutions of international law could possibly be called sanctions? Kelsen's answer – as far as general international law is concernedFootnote ²⁵ – is: reprisals and war. Reprisals are a form of self-help,Footnote ²⁶ a decentralized enforcement mechanism: ‘a reaction of one state against a violation of its right by another state’Footnote ²⁷ which justifies the sanctioning state's prima facie violation of international law. That is the crux: action normally prohibited is justified as a sanction – that is, as enforcement of law. The International Law Commission has recently reaffirmed that view of the law in its Articles on State Responsibility 2001 with respect to their concept of reprisals severely restricted in scope and rechristened ‘countermeasures’:

Kelsen is adamant that (at that time, i.e. in 1960) general international law would allow for the enforcement of reprisals by physical force, if necessary,Footnote ²⁹ which would not only sound odd to any international lawyer brought up on a steady diet of the prohibition of the use of force, but – as Kelsen admits – makes reprisals difficult to discern from war. Indeed, for him the ‘difference between armed reprisal and war is only one of degree’.Footnote ³⁰ Retortion, on the other hand, is not a sanction, because the employment of physical force is not permitted,Footnote ³¹ which means that the ILC's ‘countermeasures’ are not reprisals, since these may not involve the use of force.Footnote ³²

The legal nature of war, on the other hand, at that time was a far more tricky issue.Footnote ³³ In order to interpret war as either delict or sanction within positive international law, and thus fulfil the coercive order paradigm, Kelsen had to see war as regulated by international law, not as actions ‘beyond’ or ‘outside’ the law, as had been the general view during the nineteenth century and up to the Kellogg-Briand Pact of 1928.Footnote ³⁴ In order to ‘legalize’ war, he postulated a simplified and secularized version of the bellum iustum doctrine,Footnote ³⁵ as always with Kelsen a mere formal idea without ideological or political baggage. ‘Ohne den sogenannten Grundsatz des “bellum iustum” gibt es kein Völkerrecht’.Footnote ³⁶ Yet if there were a total prohibition of the use of force without the possibility of justifying forcible actions as sanctions, international law would also lose its legal ‘label’.Footnote ³⁷

It cannot be proved by theoretical argument that these institutes form the basis of a coercive order, for if it were to be proved whether, say, war were a delict, one would have to prove that it has a sanction attached – which could only be a counter-war. That counter-war is a sanction could only be a presupposition.Footnote ³⁸ Therefore Kelsen tries to prove his scheme by seeing it manifest in positive international law. The proof offered with respect to the embodiment of the bellum justum principle in Chapter VII of the United Nations Charter not only shows Kelsen's consistency most clearly, but will also be of continued relevance for a jurist working in the twenty-first century. A second layer of arguments will explore how Kelsen's theoretical model squares with positive law, namely the provisions of the Charter – made necessary by scholars alleging that Kelsen's model fits international law only up to the beginning of the First World War.Footnote ³⁹ This is the only time in this study where ‘fit’ will be at issue, for this is an example where we do have an unquestionably authoritative text to scrutinize in the first place. However, there is no objective criterion of ‘fit’. In normative sciences, the theoretical framework determines on what conditions the object of its study is valid, hence existent. Legal theory could thus negate the validity of practice by changing the theoretical framework, where natural scientists could not do so vis-à-vis their object of study.

In keeping with the promise of his preface in The Law of the United Nations ‘to present all the interpretations which according to his opinion might be possible’,Footnote ⁴⁰ Kelsen presents two ways of construing Chapter VII: one not conforming to the bellum justum doctrineFootnote ⁴¹ and another in complete conformity.Footnote ⁴² Under the first interpretation he doubts whether the enforcement measures provided for in the Charter can be characterized as sanctions. These actions do not have to be taken exclusively against a state which violates its obligations. The key clause in Article 39 – ‘threat to the peace, breach of the peace, or act of aggression’ – is not formulated as prohibition, and the prohibition of the threat or use of force in Article 2(4) does not have the same meaning as the terms of Article 39.Footnote ⁴³ The Security Council has a very wide discretion to determine what situations fall under the three clauses. It might order measures against a state not having violated its obligations, or even against a non-member. On the other hand, it is not under an obligation to direct enforcement measures against a state which has used force.Footnote ⁴⁴ Kelsen summarizes this position thus:

The other interpretation of the Charter's provisions, Kelsen argues, would be ‘in accordance with general international law’,Footnote ⁴⁶ since to be in accordance ‘a forcible interference in the sphere of interest of a state . . . is permitted only as a reaction against a violation of law, that is to say as sanction’.Footnote ⁴⁷ Kelsen goes on to argue that under this interpretation ‘enforcement actions determined by Articles 39, 41 and 42 must be interpreted as sanctions’.Footnote ⁴⁸ In a transposition of the bellum justum doctrine from (traditional) general international law to the particular regime of the UN Charter, the term ‘war’ is transformed into ‘threat or use of force’, and ‘reprisals’ into ‘non-forcible intervention’. Consequently, measures under Article 41 are seen as playing the role of traditional reprisals while enforcement under Article 42 fulfils the function of war in the Charter regime.Footnote ⁴⁹ That transposition has weaknesses, however. In pre-Charter times reprisals could be forcible, while now countermeasures may no longer involve the use of force. Thus enforcement, strictly speaking, is no longer possible (see supra). With the notable retention of an element of self-help – that is, of decentralized enforcement in the provisions allowing for self-defence in Article 51Footnote ⁵⁰ – the enforcement of Charter law is centralized.Footnote ⁵¹

Kelsen is adamant that measures under Article 41 (seen as reprisals) can only be interpreted as sanctions, because ‘reprisals are permissible only against a violation of international law’.Footnote ⁵² In contrast, the nature of measures involving the use of force under Article 42 is called ‘disputed’, because for Kelsen the transposition of ‘war’ into ‘use of force’ has brought into the discussion the old question of the legal nature of war. The question is whether measures involving the use of force are permitted only as a reaction against a delict (presumably a prior use of force, but not necessarily so) and thus constitute a sanction,Footnote ⁵³ which again begs the question. The key argument is that ‘threats to the peace’, ‘breaches of the peace’, or ‘acts of aggression’ are indeed prohibited. Kelsen's connection – admittedly a connection of rather weak nature – would be that the term ‘in any manner inconsistent with the Purposes of the United Nations’ in Article 2(4) refers to the phrase ‘to take effective collective measures for the prevention and removal of threats to the peace, and for the suppression of acts of aggression or other breaches of the peace’ in Article 1(1), which, in turn, would refer to Article 39's ‘threat to the peace, breach of the peace, or act of aggression’.Footnote ⁵⁴ Kelsen's further conclusion is that if force is prohibited unless it constitutes a collective response, which, in turn, can only be a measure under Articles 41 and 42, then Article 39 would take on – via that flimsy connection to Articles 1(1) and 2(4) – the prohibitory character of Article 2(4) and would precisely prohibit threats to the peace, breaches of the peace, or acts of aggression by authorizing sanctions under Articles 41 and 42. Kelsen concludes,

Does the bellum justum doctrine fit international law? Since customary international law norms are very difficult to prove and since there would not be any different law required for this theoretical Überbau (except, perhaps, for the role of non-forcible reprisals, rechristened countermeasures), general international law can be seen in Kelsen's light. Does the bellum justum doctrine fit the UN Charter? Kelsen has demonstrated that while the doctrine is not a logical necessity, the Charter – given a few tight fits here and thereFootnote ⁵⁶ – can be seen in this light. Whether that is the only approach possible, whether the drafters saw their creation in this light, or whether this or any other approach is the ‘correct’ view of the United Nations Charter is quite another matter.

1.2. Unity as absence of contradictions and logical monism

Kelsen is considered to be one of the staunchest proponents of a radical monist construction of the relationship between international law and municipal law. Dualism or, as he calls it, ‘pluralism’Footnote ⁵⁷ is to him an (epistemo)logical impossibility, and he sets out to discard it on purely theoretical grounds – that is, not by comparing it with the state of the positive legal orders in question. Unlike in section 1.1, Kelsen's consistency and thorough integration of international law theories, or, in this case, his thorough ‘theorization’ of questions of international law mean that both elements of his argument, the theoretical presupposition and the application to international law, have to be integrated here.

The arguments Kelsen employs are also intimately connected with that of the legal nature of international law. That is so because he takes the ‘coercive order’ paradigm and adds to it his opinions on the relationship between norms to come to his view of a logically necessary monism. As we have seen in the preceding section, law is defined – or rather delimited vis-à-vis other kinds of normative order – as coercive order. This definition has two sides: not only is law thus different from other kinds of order, but all law is made one uniform kind of normative order.

The argument starts with the assertion that all jurists wish to perceive both international law and municipal law as ‘gleichzeitig gültige Normen[systeme]’.Footnote ⁵⁸ A view that would simply deny the validity (hence existence) of one or the other legal order, Kelsen argues, would be logically consistent, yet dualists do not go that far.Footnote ⁵⁹ If one wishes to perceive both as valid, however, one has already decided in favour of unity:

This statement asserts that, but does not give any reasons why, the perception of two normative orders as simultaneously valid would necessitate a connection. Two reasons can be extracted from Kelsen's writings, one positive and one negative.

The positive aspect is that the defining and delimiting criterion of law as a special kind of normative order is that it is a coercive order, a Zwangsordnung. This delimitation creates a unity of all law and thus all law is an epistemological unit – no moreFootnote ⁶¹ and no less. This view is a direct result of Kelsen's reliance on neo-Kantian epistemological constructs;Footnote ⁶² in Kritik der reinen Vernunft Kant ‘transcendentalizes’ logical principles of classificationFootnote ⁶³ and thus transforms them into principles of epistemology.Footnote ⁶⁴ The first principle is adopted, in turn, by Kelsen: entia praeter necessitatem non esse multiplicanda (Occam's Razor).

The same ought to be cognized as the same and since law is in Kelsen's view typologically different from other normative orders all law ought to be seen as one genus and thus cognized as one. This is the unity of the object of cognition, which is to be presupposed. An oft-cited sentence of Kelsen reads, ‘Die Einheit des Erkenntnisstandpunktes fordert gebieterisch eine monistische Anschauung.’Footnote ⁶⁶

However, a student of Kant's works might add that Kant has a second principle standing in exact opposition, namely the law of specification: entium varietates non temere esse minuendas – the variety of entities is not to be reduced blindly. If the difference is sufficient, one must create a new species, ‘unter jeder Art, die uns vorkommt, Unterarten, und zu jeder Verschiedenheit kleinere Verschiedenheiten zu suchen. Denn, würde es keine niederen Begriffe geben, so gäbe es auch keine höheren’.Footnote ⁶⁷ In contradistinction to Kelsen's plan, the questions to be asked are both whether the different laws are sufficiently uniform to create an epistemological unit (genus) and whether they are sufficiently different to create an epistemological distinction (species).Footnote ⁶⁸

The negative criterion of the unity of all law is logical consistency (Widerspruchslosigkeit). The Kelsen of the relevant period – before the ‘normological turn’ – saw any given normative order as a logical unit within which the conflict of norms is excluded, because of the (indirect) applicability of the principle of excluded contradiction.Footnote ⁶⁹ Since the principle is applicable, there cannot be a ‘conflict of norms’. Only one of the norms can be valid; a conflict of norms is as senseless as a logical contradiction, because neither can exist. Since legal cognition seeks to portray its object as consistent, it seeks to eliminate conflicts of norms by way of interpreting them away.Footnote ⁷⁰ Therefore – and that is the reason for the prefix ‘negative’ – if insoluble conflicts of norms between international law and municipal law can exist, the possibility of a unity of these two normative orders is excluded.Footnote ⁷¹

In all works of relevance for this topic Kelsen seeks to demonstrate that what we might think are conflicts between international law and municipal law are not, properly speaking, insoluble conflicts.Footnote ⁷² He compares a municipal statute violating international law to the analogous case of a statute violating the constitution, where the result is not a conflict, but merely the voidability of the lower-level norm with a tacit provision that the statute is considered valid as long as it is not annulled.Footnote ⁷³ This is not the place to discuss this doctrine in detail, even though it is an important topic within Kelsen's Pure Theory of Law. Suffice it to say here that this proof once again comes directly from legal theory.

Some scholars, taking their cue from Herbert Hart's 1968 paper,Footnote ⁷⁴ suggest that there were two lines of argument in Kelsen's rejection of the pluralist construction: one arguing for a necessary connection of all normative orders and another merely arguing that a relationship is established in positive law – that is, that there is a positive delegating norm. This discussion is mentioned only briefly here, because the second argument cannot be independently made within Kelsen's theory before the mid-1960s. Afterwards it is still problematic (section 2.2). Alfred Verdross had explained why this is so in 1923: ‘Allein jeder dieser positiv rechtlichen Regelungen [Verweise] setzt schon eine bestimmte Hypothese über das Grundverhältnis der beiden Normensysteme voraus.’Footnote ⁷⁵ Without a theory that declares that delegation creates unity, all norms purporting to delegate a subordinate order would merely make the claim to create unity, and a statute's claim to create the subordinate source ‘administrative ordinances’ would have just as much a claim to unity as a madman's claim to subordinate all the world's laws. If, however, there can be no argument B without argument A, can argument A (necessary unity) exist alone? Again, it does not seem possible, because if a higher norm does not delegate norm-creation it simply does not thereby create a partial, subordinate normative order. If a municipal legal order were not to create a subordinate order ‘customary law’ among its sources of law, this – arguably ‘pre-existing’ – partial legal order would not derive its validity from that municipal legal order and the two would not be connected.Footnote ⁷⁶ Kelsen argues that this would be a criterion for disunity:

If we do not question whether theory fits positive law, but only whether doctrine fits theory and vice versa, the answer is obvious: Kelsen was consistent.

2.1. The coercive order paradigm resolved

Resolving the ‘coercive order’ paradigm requires a paradigm shift. This shift has to take place at the root of Kelsen's argument that led him to assume that all law is coercive in nature. That root lies in the choice to determine the term ‘law’ and to distinguish it from other norms by reference to a comparison of various normative orders called ‘law’.Footnote ⁷⁹ Kelsen can be seen as taking what arguably constitutes an empirical property of ‘law’ and transposing it into a necessary element of law, and that would be highly problematical. We can explain, and one can sympathize with, the desire to design a pure theory specifically of law, but in that transposition lies an admixture of Is and Ought, which would violate the basic principle of the Pure Theory.

There is another way to interpret Kelsen's incorporation of the coercive element of law. In that view, the coercive order paradigm is not at the core of the Kelsenian approach to law; it is part of it, but not nearly as important as secondary literatureFootnote ⁸⁰ wants to make us believe. Kelsen was not a ‘traditional’ positivist and by no means an Austinian. The focus lies with the norm – the norm is the a priori element. It forms an ontology of the ideal, the norm is its own validity and its existence; all other elements of the term ‘law’ are mere appendages. In this sense Kelsen argues that his inquiry into the meaning of the term ‘law’ could come to the conclusion that there is no sufficient similarity, because ‘mit dem Wort “Recht” und seinen anderssprachlichen Äquivalenten so verschiedene Gegenstände bezeichnet werden’.Footnote ⁸¹

The crucial difference (to a degree finding support in Kelsen's writings) is that the nature of the differentiation of various ‘types’ of normative system is categorically different to the delimitation of norms vis-à-vis things existing in reality. The uniqueness of legal norms is empirical in nature – that is, an a posteriori description of perceived differences. These differences are typical for the configuration of legal normative orders vis-à-vis other normative orders. The ‘normative’ element of a law is not one of its properties – law is norms and not normative (understood in an adjectival sense). Norms are an a prioriFootnote ⁸² – a dogma. Norms are their validity, they are their claim to be observed, they are their existence. The norm is (not ‘has’) a trinity of bindingness, validity, and existence. These three words express different aspects of the same thing. To declare a norm is to create; to declare a norm ‘law’ by way of its coercive nature is descriptive.

The first kind of differentiation does not ‘unmake’ a norm, it does not rob it of its existence, whereas the latter kind – the decision whether a given ‘idea’ is a norm – will determine its existence vel non. There are norms in existence which fail the test of being legal norms, but there can be no non-binding norm, no norm not claiming observance, because validity is the specific form of existence of a norm, not one of its properties, as being ‘coercive’ might be.

Moreover, the empirical closeness of the various legal normative orders is not in fact as great as is commonly assumed. First, the coercive elements of different orders are shaped differently. Second, the proportion and importance of the coercive element vary broadly among different areas of municipal legal systems and between legal systems. Finally, the international legal order does not fit the strict criteria very well (see section 2.3). The closeness of legal systems may exist and one of the connecting properties may very well be coercion, but it may be a loose family resemblance, not according to a strict logical definition which is based on necessary elements. Wittgenstein used the term ‘game’ to describe how we can subsume such things under a general term which do not have common properties, but have properties in common only with intermediary members.Footnote ⁸³

A particular area of dissent is that it is implied in various passages of Kelsen's writings that one of the bases of the ‘coercive order’ paradigm is that there is some kind of ‘automatic’ authorization for application by all subjects in a decentralized normative order, for example in morals or international law – that is, without positive regulation to this effect:

When the law does not define an organ to apply a norm – that is, when no one is authorized to determine authoritatively when the law has been breached (i.e. to create an individual norm) – it is not the case that every subject of that normative order is suddenly authorized to apply the law for him-/her-/itself. To create rights out of thin air violates a core principle of positivism: no positive norms without positive norm creation – that is, without an act of will. Even less does each subject authorize to enforce the law without authorization. If a decentralized normative system did have norms that stipulate enforcement by self-help, then the normative order would have given each subject the right, but this would be positive regulation. When nobody is explicitly authorized, nobody is, and whether the relevant substantive prescription is breached or not (in case the norm is not formulated in the form Kelsen wishes to see in law – section 1.1) remains irrelevant. The norm can be breached and behaviour would in the abstract be unlawful, but it would not be applied, no individual norm would be created that stipulates sanctions.Footnote ⁸⁵

Kelsen acknowledges this in a more abstract sense: if a regulation proves to be practically useless, nobody is authorized to make it useful unless the norm is modified through the procedure prescribed:

Yet, unfortunately, this different view of the coercive order paradigm described in the paragraphs above is contradicted in some of Kelsen's writings.

This is to be understood in the sense described above (section 1.1): Kelsen held that the coercive element is a necessary element of legal orders, not of all legal norms. In order to be a legal norm, a norm has to belong to a legal order – that is, a normative order containing coercive orders to enforce most, all (or even any) of its norms. Maybe the reason why he conceived of the law in such a way was not merely out of piety to traditional positivism, or even because he saw it as the only possibility for keeping the Pure Theory free of close relatives such as morals (see infra), but maybe the dynamic element of the Stufenbau, of the strict hierarchical membership in a normative order, demands that law be applied and thus enforced down to the individual norm, or even to the last non-normative act of enforcement (authorized by that individual norm).

To summarize: according to the view espoused here, to try to prove any ‘inherent’ facets of a law besides the norm is a problematic endeavour. Law's classification is a matter for social-scientific methodology. Norms are obligation; whether as a legal or a moral norm, the provision of enforcement elements cannot influence law's ‘normativeness’. To make law's existence dependent on its coercive elements – which are designed to guarantee its effectivenessFootnote ⁸⁸ – in effect means making it dependent on a factual occurrence. Hence, if a normative system that contains enforcement elements is not only just called ‘law’ but is thereby also made an Ought in the first place, a factual occurrence alone would determine the existence of a prescription. Such a theory would violate Kelsen's own dichotomy of Is and Ought: ‘[Es] kann daraus, daß etwas ist oder nicht ist, nicht folgen, daß etwas sein oder nicht sein soll.’Footnote ⁸⁹ Kelsen could be read as making that mistake (i) by stipulating the ‘coercive order’ paradigm and (ii) by making the validity of a legal system dependent onFootnote ⁹⁰ (not equivalent to!)Footnote ⁹¹ its continued effectiveness. Whether Kelsen himself actually made this mistake is not as relevant as the fact that his writings are frequently interpreted in that manner. The variant of neo-Kelsenianism presented here seeks to avoid this admixture of Is and Ought.

2.2. The unity of law dissolved

Kelsen's assumption of a coercive order as the basis of all law is a major argument in favour of his theory of the unity of all law. Hence the rejection of the coercive order paradigm in section 2.1 must lead not only to a different concept of international law and, in particular, a different view of the UN Charter (section 2.3), but must also mean that normative orders called ‘law’ are not necessarily one ontological entity or epistemological unit. The interpretation of the ‘coercive order’ paradigm as an empirical–politological feature of a legal normative order means that from this connection by ‘family resemblance’ one cannot derive normative unity, because normative unity is not achieved by empirical classification, but by positive norm-making.Footnote ⁹² Because the necessary uniqueness of legal orders versus other normative orders does not exist, the positive criterion of unity of legal orders proposed by Kelsen is not fulfilled.Footnote ⁹³

In addition to rejecting the positive criterion of necessary unity it will not be surprising that this article also affirms the negative criterion that insoluble conflicts of norms are possible. Towards the end of his life Kelsen achieved the purest form of his theory, and the normological turn is the outstanding achievement of that period. That turn led to a denial of even the indirect applicability of certain logical operations to norms, in particular of the logical deduction of norms and of the principle of excluded contradiction. With the strict consistency typical of Kelsen he sets out to assert the possibility of conflict between norms from different systems. In his groundbreaking Allgemeine Theorie der Normen (1979) he does so in explicit contradiction to his earlier views.Footnote ⁹⁴

As mentioned in section 1.2,Footnote ⁹⁵ the basic assumption throughout is that norms cannot be true or false. In the later period he asserts that we cannot even indirectly ascribe truth-values to norms by going to and from statements about the validity of norms,Footnote ⁹⁶ which can be true and false, to conduct logical operations of deduction and exclusion on the norms themselves, a view he held until publication of the paper ‘Recht und Logik’ (1965).Footnote ⁹⁷ Thus in the case of two norms being valid which contain contradicting prescriptions, the two respective statements would both be true, yet would contradict each other.Footnote ⁹⁸

Kelsen also adamantly denies a method of excluding conflict between norms of different systems he had held for a long time:Footnote ⁹⁹

Law and morals can prescribe the same behaviour. Even if there is no conflict between, say, law and morals, the non-validity of the other order cannot be denied, because of the relativity of values. More than one normative order can seen as valid for a given sphere.Footnote ¹⁰¹ In this way the argument is brought to a close, because

In this way Kelsen changes his mind on the necessary unity of all law in a posthumous publication to embrace the possibility of pluralism.Footnote ¹⁰³ Insofar as the view of the impossibility of a necessary unity espoused here is only a partial variance to Kelsen, only to the Kelsen before the normological turn of the mid-1960s.

Therefore the question of a contingent connection of two normative orders by delegation becomes acute. The focus here is not so much on the dogmatic question of whether there is a delegation by international law of municipal law or vice versa, but on the theoretical implications. First, we need to clarify what ‘delegation’ means within the Vienna School of Jurisprudence. It is the dynamic properties of normative orders that allow them to form a changeable and expandable hierarchy of norms (Stufenbau). One of these properties is that norms can authorize humans to create further norms; for example, a code of penal procedure authorizes a judge to create individual norms or a constitution authorizes a number of humans (parliamentarians) to create laws. This authorization delegates to these humans the authority to create norms. The norm-making norm is the authority for the norm created under that authority. We can speak of a hierarchical relationship, because the latter norm is dependent for its validity (hence existence) on the former norm; it is apposite to call the authorizing norm ‘higher’ and the norm thus created ‘lower’. The formal source of norms called ‘statutory law’ is thus dependent on the higher source ‘constitution’; it is a subordinate and partial normative order. In effect, it is not a normative order at all. A normative order is completed by the assumption of a Grundnorm, that is a tautological ‘as if’ assumption or fiction allowing for the cognition of a normative order which has no positive authority.

How would this apply to the relationship between international law and municipal law? One of Kelsen's answers – that of his ‘monism with primacy of international law’ – is that there is a positive norm of international law which indirectly determines the humans which are authorized to create municipal law by stipulating that the effective government's legal order is ‘the state’ for international law.Footnote ¹⁰⁴ The function of the ‘effective government’ rule is to define the organs of the state (hence of a partial legal order) for international law, just as the managing directors of a corporation as juristic person would be defined in a statute in a municipal setting.Footnote ¹⁰⁵

Yet, given the non-existence of a necessary connection, the question is whether this norm of international law which claims to create a subordinate normative order really does create a subordinate order or whether this is a fraudulent claim. Such questions throw the door open to a problem of fundamental significance: where do normative orders end?

The answer in Kelsenian theory can be said to be the Grundnorm (or the highest positive norm of a normative order, but we shall leave this thought aside here). The membership of a norm is ascertained by its ‘derivation’ from higher levels of norms by way of the creation of norms (section 1.1). As mentioned above, the highest norm and ‘capstone’ of any normative order is the Grundnorm. This is not a positive norm, but an assumption in order to be able to cognize the normative order. It establishes the unity of a given normative order by allowing the perception of that order and thus also postulating the reason for validity of all norms of that order.Footnote ¹⁰⁶

As a fiction or hypothesis,Footnote ¹⁰⁷ however, its content is determined by the positive normative order, not vice versa.Footnote ¹⁰⁸ When a normative order claims to delegate a subordinate source – as a moral order may claim to authorize the creation of a municipal legal order – the problem becomes obvious. In contrast to Hart, the question of ‘fraudulent’ or ‘real’ delegation cannot be decided by recognition by the supposedly lower order,Footnote ¹⁰⁹ for if the relationship were to exist, the lower order's ‘recognition’ would be irrelevant.Footnote ¹¹⁰ In contrast, if the relationship were not to exist, the ‘recognition’ would constitute a simplified act of norm creation by the lower order and the supposedly higher order would be norms of the lower order.Footnote ¹¹¹

If the Grundnorm is an assumption by anyone cognizing a normative order, it cannot be the objective end to a normative order. In any given normative order there can only be one Grundnorm and if a partial legal order is ‘given’ a Grundnorm, the unity is destroyed. Thus if the constitution of a given state is given a Grundnorm – for example, in the case of post-1945 Austria, ‘The constitution claimed to be reactivated by the leading political parties ought to be observed’Footnote ¹¹² – the delegation by international law is no longer a ‘real’ normative link and the Austrian municipal legal order is detached from the ‘unity of all law’.

This is the fundamental problem: there is no objective criterion to cognize the coherence of a normative order. This is a problem of Kelsenian theory, a problem merely hidden behind the veil of the coercive order paradigm, but which further purification has brought to the fore. The partial legal order ‘Administrative orders of the Minister of Defence under the Military Exclusion Zone Statute’,Footnote ¹¹³ for example, is part of the legal order ‘Austrian law’, but Austrian law, in turn, is not a subordinate order to some moral order, or subordinate to the whim of, say, Mr Smith of Birmingham, but we cannot prove whether there is or is not a normative connection. Both the claim made by the Military Exclusion Zone Statute and by Mr Smith's command are mere claims to be observed – as all norms are.

This is an incredibly destructive force, and the very possibility of normative systems could be denied. On the other hand, it could have an incredibly constructive effect, gluing together the most unlikely norms and normative systems into one whole. According to this view, Kelsen's unity is established, yet it is a unity of all norms, under the general Grundnorm repeated in every particular Grundnorm: norms are to be observed, or, in other words, norms are norms.

The theory of norms does not have to be that dramatic, however. We must realize what Kelsen had taken on from Kant, namely that one's epistemological position influences the world we perceive, that the world is only what we perceive.Footnote ¹¹⁴ Ultimately it is predicated on the presumption by the person cognizing law whether he or she wants to see it as one or not and where he or she assumes the Grundnormen. For the monism–pluralism debate we can say that while both constructions are possible, neither construction is necessary.Footnote ¹¹⁵ It is both a normative scientific assumption (where does one assume one's Grundnorm?) and predicated upon positive regulation (does international law claim super-ordination or even some municipal law?). Or perhaps – in the light of the relativeness of the claim to subordination – the positive regulation is no different whether they are seen as one or not; it is merely a matter for epistemology.

Behind Kelsen's problematization of the question of which normative system is valid for a given sphere, and his answer that there can ‘really’ only be one ‘true’ law, lies piety more than consistency. It is the absolutization of law, the traditional opinion that there can only be one ‘real’, ‘true’, or ‘correct’ law, that, independently from natural law doctrines, there cannot be concurrent legal orders in one state; one of them has to be the legal order, only one legal order can be subject of international law. Piety is this motive's origin, piety to the absolute, incommensurable with a consistently relativistic Weltanschauung and normative theory. The late Kelsen moved in the direction espoused here. His scepticism towards certain logical operations and his acceptance of the conflict of normative systems in one sphere are expressions of a relativistic theory. The super-elevation of law was traditional positivism's greatest vice, ‘Rechtspositivismus, welcher der Kundgebung rein menschlicher Gesetze eine trügerische Majestät verleih[t]’, as Pope Pius XII wrote in 1942.Footnote ¹¹⁶

2.3. A concept of international law

What would a theory of international law look like if it were based on a neo-Kelsenian view? To answer that question exhaustively would require more space, effort, and wisdom than is available here, hence this last subsection will remain a cursory glance. Taking the basic theoretical assumptions derived from sections 2.1 and 2.2 and combining them with earlier writings, the following is a selection of topics from the vast field of international law.