Beyond the res judicata doctrine: The nomomechanics of ICJ interpretation judgments

Abstract

Res judicata is a core belief of international law; the ICJ’s judgments are seen as final and without appeal, to doubt that is apparently equal to calling the entire international legal order into question. But the doctrine is not as absolute as the orthodoxy makes it out to be, neither as a matter of positive international law nor as a statement of legal theory. Even final judgements are not always final and appeals procedures and judicial review are not special in that they engage res judicata whereas regular legal change does not; rather, both do from a legal-theoretical vantage-point. This article makes the point by looking at ICJ interpretation judgments under Article 60; it argues that, far from leaving the original judgment’s res judicata intact, interpretation judgments actually impinge or even disrupt it. The article discusses ICJ interpretation judgments (the 2013 judgment in Preah Vihear serving as convenient example), introduces Adolf Julius Merkl’s Error Calculus theory as the theoretical framework best suited to analysing the nomomechanics and critiques the Preah Vihear interpretation judgment as change disguised as a hermeneutic exercise. It then turns the critical enterprise on its head to look at the Error Calculus theory itself to lay the groundwork for an even more audacious argument that the Error Calculus does not depend on errors in the narrow sense of the word: it is neither an ex post ratification of an imperfect norm nor a confirmation of invalidity, but the derogation of a perfectly valid norm.

1. Introduction

Nobody, but nobody, can doubt res judicata. Most statements about international law are questioned, but nobody denies that judgments of international tribunals have the force of res judicata. If there is one core belief on which we can all agree, on pain of excommunication, it is this. Article 60/1 ICJ Statute ¹ expresses the profession of our faith: the Court’s judgments are ‘final and without appeal’. Some international instruments do indeed provide for some form of appellate jurisdiction, but there must there be a final instance at some point and the legal resolution of state–state conflicts is also said to be inimical to being overturned. The ICJ’s institutional design is testament to that: one court, one instance, and one judgment which is final.

If this were merely the accurate description of the current state of international law, we would stop at this point, satisfied in the knowledge that, for once, we have found a point on which we can all agree. However, neither is this description accurate – even ‘final’ judgements are not always final – nor is the basis for our agreement an analysis of the (contingent) state of positive international regulation. As so often, it is based on quasi-positivist arguments from necessity. It can be argued that appeals procedures and judicial review are not special in the sense that they engage res judicata whereas ‘regular’ legal change does not. Both do, but in a different manner than is commonly conceptualized. While the words used to describe it may suggest that this applies only to tribunal judgments (‘judicata’), from a legal-theoretical vantage-point, the ‘nomomechanics’ (my neologism) are identical.

This point can be made by looking at ICJ interpretation judgments under Article 60/2 as a case study. In contrast to the mainstream argument that the Court’s interpretation judgments, while binding, do not engage res judicata and that they are as far removed from such an engagement as is possible to be for a post-judgment procedure, I argue that they do. Section 2 will briefly present salient points of international legal res judicata doctrine, Section 3 connects ICJ interpretation judgments to that doctrine, Section 4 introduces Adolf Julius Merkl’s theory of legal hierarchy and its solution of the problem of ‘erroneous’ legal decisions, Section 5 critiques the Preah Vihear interpretation judgment as change disguised as a hermeneutic exercise, and Section 6 in turn critiques Merkl’s theory to lay the groundwork for an even more audacious argument.

2. The res judicata doctrine in international law

International lawyers are not the only jurists to talk about res judicata. The concept has both a long history in the development of legal concepts and is probably a universal, if not identical, concept across the globe. Because Merkl’s theory, created by a German-language scholar, plays an important role in this article, the equivalent German concept of Rechtskraft (‘force of law’) needs to be introduced as well. Prima facie, the difference between res judicata and Rechtskraft is small. One could argue that while Rechtskraft places perhaps slightly greater emphasis on the immutability of the judgment in connection with a lack of means of redress (no higher instance has jurisdiction), res judicata focuses slightly more on the chose jugée aspects, i.e., the non-admissibility of a re-litigation of the case. However, there are crucial differences.

International legal doctrine on res judicata can be boiled down to a number of core arguments: (i) it causes judgments (or reasons contained therein) to be binding with finality for the parties; (ii) it is unquestionably part of positive international law; (iii) it is only applicable where persona, petitum and causa petendi are identical; ² (iv) it is also the expression of a double telos; ³ (v) a central but controversial question is to which parts of a judgment res judicata applies. Only the first two are of immediate relevance here and will be discussed.

2.1 Finality and bindingness

Res judicata, as used in international legal doctrine, is a ‘a relatively broad notion’, ⁴ as the Court pointed out in Bosnia Genocide: ‘That principle signifies that the decisions of the Court are not only binding on the parties, but are final, in the sense that they cannot be reopened by the parties as regards the issues that have been determined.’ ⁵ It concludes that ‘the applicable principle is res judicata pro veritate habetur, that is to say that the findings of a judgment are, for the purposes of the case and between the parties, to be taken as correct’. ⁶ This is a restatement of the orthodox position, but it does raise the question whether international lawyers seem to associate substantive correctness of the judgment with the notion more than they associate its formal force of law, irrespective of whether the judgment is correct.

The core of the international legal res judicata doctrine is seen as a combination of two elements found in Article 60/1 and Article 59: the Court’s judgments are ‘final and without appeal’, on the one hand, and its decisions have ‘no binding force except between the parties and in respect of that particular case’, on the other. ⁷

Res judicata, therefore, has two effects. First, that which is res judicata is definitive … the same issue may not be disputed again between the same parties … Secondly, res judicata, that is to say, what has been finally decided by a tribunal, is binding upon the parties. ⁸

The first effect of ‘finality’ or ‘definitiveness’ primarily denotes ‘excluding every ordinary means of appeal’, ⁹ but appellate jurisdiction tends to be defined narrowly. Res judicata as envisaged does not exclude any and all further procedural moves like the revision and interpretation of judgments. The other prominent aspect is certainly ne bis in idem, protecting the parties against interminable cases: ‘a successful party should not have to reargue its case’; ¹⁰ ‘an issue decided in a judgment or award may not be relitigated’. ¹¹ The third aspect is the claim that ‘the dispute is brought to an end’ ¹² or that ‘the matter is finally disposed of for good’. ¹³ Fourth, the mainstream does not recognize or accept Rechtskraft as relative immutability (relative Unveränderlichkeit) ¹⁴ of legal acts, i.e., their infinite validity ratione tempore and the absence (or inapplicability) of rules of change for this act.

The second effect of ‘bindingness’ is also closely related to domestic versions of the doctrine, although international lawyers tend to emphasize that it is the parties who are bound. ¹⁵ This is probably more an attempt to distinguish international law from a true common law than a claim that the tribunal is not bound – in ICJ procedural law, the Court is indeed seen as bound by its own decisions and cannot unilaterally change them. ¹⁶ Sometimes, the range of actors bound by a judgment is extended beyond the parties and the tribunal. Some authors seek to adopt ‘issue estoppel’, which would bind the parties to case B to the decision in case A, if there was an identity of parties and issue, thus excluding causa as criterion. ¹⁷ Here, we are reminded of the limits of res judicata, but where ‘the disposition effected by a judgment is to be accepted as true in fact and in law’ ¹⁸ in an almost general sense.

This opens the way to a stronger, common-law-like precedential force for ICJ judgments. Iain Scobbie argues that at least some parts of some types of judgment must be binding on third parties: ¹⁹ he believes that the mainstream’s reading of Article 59 as a rejection of the common law tradition for international procedural law is too simplistic. It is not always possible to limit the personal sphere of application to the parties, primarily when judgments concern questions of status, such as territorial or border disputes. ‘[T]hese judgments form an exception to the relativity of res judicata, as they are opposable to all states and not simply the parties to the case[;] this is because territorial status is “une situation objective ayant effet erga omnes”.’ ²⁰ While there are only a limited number of parties in any given case, the consequences cannot be confined to the parties – a judgment delimiting territory ‘est opposable a toute revendication d’Etats tiers; elle agit in rem et non pas seulement in personam’. ²¹ Yet what could be the legal basis of such an expansion of binding force? For some authors, the Court has ‘effectively extended the res judicata of decisions to non-parties’ ²² even outside jus in rem topics; for example, the Court may even have regarded South Africa as bound by its earlier advisory opinion in its 1962 judgment on jurisdiction in South West Africa. ²³

2.2 Positive law and nomomechanics

What about the foundation of the res judicata doctrine in positive international law? For the orthodoxy, the doctrine is obviously seen as a treaty norm (Article 60/2) and also as general principle (Article 38(1)(c)), since res judicata is very much part of all major traditions of domestic law. The locus classicus for this dual source is Dioniso Anzilotti’s dissent in Chorzow 1927: ‘if there be a case in which it is legitimate to have recourse … to the “general principles of law recognized by civilized nations” … that case is assuredly the present one’. ²⁴ It may be useful to briefly dip into Merkl’s theoretical approach, because it is at this point where the difference between the legal-structural essence of res judicata (the relative immutability of all law) ²⁵ and the more or less arbitrary nature of its design in positive law comes into play. Merkl argues that immutability, not change, is the default for legal norms, irrespective of whether we are talking about statutes, the constitution, treaties, contracts, or judgments. ‘Default’ in this context means that unless there is a rule of positive law allowing a norm to be changed, it cannot be changed. This inertia or permanence (Beharrungskraft) is inherent in the nature of (legal) norms as contrafactual claims to behave in a certain way (Ought). Immutability is generally accepted by domestic and international legal orthodoxy regarding judgments, but not regarding law in general: we would ordinarily not seek a specific legal authority to amend a treaty or a parliamentary statute. Another inherent feature of norms is their bindingness – norms are their bindingness, a non-binding norm does not exist – but this theoretical insight does not tell us who is bound by a specific norm (its personal sphere of application) or whether, procedurally, it causes a ne bis in idem block. These features depend on the ‘arbitrary’ content of positive law. Whether a judicial review is possible, how a successful outcome would impact the original judgment and which part of the (text of) a judgment has res judicata status is not implied in a norm’s classification as ‘judgment’.

It turns out that both international and domestic procedural legal doctrine tend to reify this concept: both tend to aggregate and conflate a number of separate positive norms in order to allow them to create a res judicata doctrine which privileges their own predispositions over the state of positive law and over the structural rules of nomodynamics. Unsurprisingly, the international legal tradition of res judicata highlights the prohibition of reopening a court procedure and the correctness of the content of the decision whereas the finality, immutability and sphere of application of the judgment are slightly less central. In contrast, at least the German-language domestic tradition tends to highlight the procedural bar to an appeal and the immutability of the judgment as norm, particularly when it comes to ‘erroneous’ judgments: a judgment with Rechtskraft is valid law even if it is erroneous. We might summarize thus: international lawyers argue that the case is concluded with the content ‘X’ and thus ‘X’ is immutable, whereas German-speaking lawyers would say that the judgment has Rechtskraft and thus the case is concluded, however it was decided.

Merkl’s fundamental and foundation critique of nineteenth and early twentieth century German and Austrian procedural legal scholarship can therefore be applied to today’s international procedural doctrinal legal scholarship. Neither then nor today, neither in domestic nor global contexts does it seem possible to clearly delineate between (i) contingent regulations dependent on the state of positive law, (ii) the necessary ‘laws of movement’ of dynamic normative orders, and (iii) political or pragmatic arguments seeking to ensure that the judiciary is able to keep working. If ‘the authority of … international law … depends on the finality’ ²⁶ of judgments, while international treaties may be amended, even informally via subsequent practice, at any time and without specific legal authority, then pragmatic considerations trump scholarly exactness. Raising the question whether the Court had ‘violated’ its own res judicata ²⁷ is merely a symptom of international lawyers’ problematic view of nomodynamics and will be discussed in the following sections.

3. ICJ interpretation judgments

Article 60/2 empowers the Court to interpret its own judgments: ‘In the event of dispute as to the meaning or scope of the judgment, the Court shall construe it upon the request of any party.’ The wording alone is enough to raise a whole host of issues which have been discussed at length in writings and in case law. The case law of the PCIJ, the ICJ, and of a number of arbitral tribunals on interpretation judgments is significant; academic commentary on the topic is largely agreed on many issues, but bitterly divided on others. Four questions are potentially relevant from a legal-theoretical point of view: (i) Which parts of the judgment partake of res judicata? (ii) Which parts of the original judgment may be interpreted by the Court in an interpretation judgment (Section 3.1)? (iii) What, exactly, is ‘interpretation’ in connection with the Article 60/2 process (Section 3.2)? (iv) What are the legal consequences of an interpretation judgment and what does this mean for the parties and for the relationship between the original and interpretation judgment (Section 3.3)?

The PCIJ has decided two cases of interpretation, the ICJ five, of which the Court’s 2013 judgment in Preah Vihear is certainly the most comprehensive: ²⁸

• – Treaty of Neuilly, Judgment of 26 March 1925: The Court found the request inadmissible, because an answer to the question posed would go beyond the limits of the original judgment. ²⁹

• – Chorzów Factory, Judgment of 16 December 1927: The Court found the request admissible and gave an interpretation. ³⁰

• – Asylum, Judgment of 27 November 1950: The Court found the request inadmissible, mainly because the questions posed were new and an answer would go beyond the limits of the original judgment. ³¹ Its reasons for inadmissibility nonetheless contain traces of a substantive answer to the question.

• – Tunisia/Libya Continental Shelf, Judgment of 10 December 1985: The Court found the request admissible and gave a lengthy interpretation. ³²

• – Cameroon/Nigeria Boundary, Judgment of 25 March 1999: The Court found the request inadmissible, inter alia, because it had already answered the questions posed in the original judgment and because to accede to the request would call into question the res judicata effect of the original judgment. ³³ Its reasons for inadmissibility nonetheless contain an implicit substantive answer to the question.

• – Avena, Judgment of 19 January 2009: The Court found the request inadmissible, because the questions posed were new and an answer would go beyond the limits of the original judgment. ³⁴ Its reasons for inadmissibility nonetheless contain an implicit substantive answer to the question.

• – Preah Vihear, Judgment of 11 November 2013: The Court found the request admissible and gave a lengthy interpretation. ³⁵

Case law can provide two kinds of information about interpretation judgments. On the one hand, judgments and attached individual opinions contain statements of the type ‘scholarship’ about the content of positive law and about legal-theoretical issues, ³⁶ endeavouring to produce knowledge about the world. This type of statement, even if proffered by the Court or one of its judges, is not endowed with an a priori greater epistemic authority than if it had been made by scholarship and should be judged solely on its consistency and rational persuasiveness. ³⁷ On the other hand, an interpretation judgment which (explicitly or implicitly) provides an interpretation constitutes an instance of tribunal practice. In the aggregate, they show us how interpretation judgments actually work and what the Court does when it engages in this modus operandi – which have the potential to conflict with its general statements as to what it should be or aspires to be doing. ³⁸ Sometimes, as mentioned above, the Court’s decision of inadmissibility hides the fact that the attendant reasoning contains an implicit interpretation of the original judgment, ³⁹ for example in Avena. ⁴⁰

In such a case, the question is whether the parties are bound by the reasoning of its interpretation in the judgment declining admissibility. On the one hand, the Court declines admissibility, but, on the other hand, the reasoning is contained in a binding judgment. A similar question arises when, in an admissible case, the Court’s statements in merito are devoid of content, as may have happened in Tunisia/Libya Continental Shelf: ‘that Judgment laid down a single precise criterion for the drawing of the line … The request for interpretation is therefore founded upon a misreading … the Court considers that there is nothing to be added’. ⁴¹

3.1 Object

The formal object of the procedure is a judgment of the ICJ – any form of judgment, even one on jurisdiction and admissibility. ⁴² An interesting doctrinal question which, to my knowledge, has not yet been raised in literature, ⁴³ is whether interpretation judgments can also be the object of a second interpretation judgment and whether a never-ending chain of interpretation judgments would be possible. The clear wording of Article 60/2 (‘the judgment’) seems to prohibit excluding the application of the procedure to interpretation judgments, because they are judgments. It is also reasonably clear that in practice the Court would find reasons to declare a request for an interpretation judgment on an interpretation judgment inadmissible, whether or not the law allows for it.

Another question is relevant from the legal-theoretical perspective as well: which part of the original judgment may the Court interpret? Ever since Anzilotti’s dissent in Chorzów, it has been assumed that the Court’s empowerment to interpret is limited to those parts of the original judgment which are endowed with the force of res judicata: the limit of the empowerment norm has been linked with the limit of res judicata. The Court in Chorzów held that ‘the second sentence of Article 60 was inserted in order, if necessary, to enable the Court to make quite clear the points which had been settled with binding force in a judgment’; ⁴⁴ Anzilotti himself is even clearer:

It is within these limits that the Court’s judgment is binding, and it is within these same limits that Article 60 provides that any Party shall have the right … to request the Court to construe it. It appears to me to be clear that a binding interpretation of a judgement can only have reference to the binding portion of the judgment construed. ⁴⁵

This statement is generally agreed; the answer to one question seems to be predicated on the answer to the other – we seem to need to find out which parts of a judgment partake of res judicata. This focus also explains the debate, ⁴⁶ inter alia in Preah Vihear, ⁴⁷ whether the dispositif alone is res judicata or whether certain parts of the reasoning, for example the ‘condition[s] essential to the Court’s decision’, ⁴⁸ also partake of it.

Therefore, a discussion of the minutiae of the ‘positive-legal form of legal norms’ ⁴⁹ seems to be necessary in order to be able to know what the scope of interpretation judgments is. Yet on a nomomechanical level, we can and should distinguish between two concepts. A small part of ‘the judgment’ contains the individual norm which, in Merkl’s terms, ‘the legal theoretical principle of immutability … refers to’. ⁵⁰ The individual norm can be distinguished from a potentially larger part which is endowed with res judicata as finality, because ‘[s]trictly speaking, it is not about the limits of the Rechtskraft, but about the limits of the act or … legal norm’. ⁵¹ The assertion of an interpretation only of those parts partaking of res judicata relies on an unacknowledged natural-law argument and may potentially and additionally also be based on a logical fallacy. In Cameroon/Nigeria Boundary, the Court held:

[Quote from Chorzów:] “[T]he second sentence of Article 60 was inserted in order, if necessary, to enable the Court to make quite clear the points which had been settled with binding force in a judgment, … a request which has not that object does not come within the terms of this provision”. In consequence, any request for interpretation must relate to the operative part of the judgment and cannot concern the reasons for the judgment except in so far as these are inseparable from the operative part. ⁵²

The same sentiment was expressed in Asylum: ‘clarification of the meaning and the scope of what the Court has decided with binding force’. ⁵³ The Court in effect held that, because the res judicata of the original judgment must not be disrupted by the interpretation judgment, the interpretation judgment, because its interpretation is itself binding, may only interpret those parts of the original judgment which partake of res judicata. Yet if, as is generally accepted, the interpretation judgment does not disrupt the res judicata of its object, then it will neither disrupt those parts of the original judgment which do partake of res judicata nor those which do not. This is so by definition, because the mainstream argues that the interpretation judgment does not create a res judicata effect, even though, like any other judgment, it is binding. But if interpretation judgments are binding, they are perfectly capable of engaging and potentially changing all parts of the original judgment; see Section 5. If interpretation judgments are not binding, then it is equally irrelevant which parts of the original judgment may be interpreted under Article 60/2.

Even on the doctrinal level, a narrowing of possibilities as argued by the mainstream is not possible; the words ‘the judgment’ and ‘it’ cannot be limited to specific parts of the original judgment. ⁵⁴ It is very easy to construct an alternative telos to the effect that the object and purpose of interpretation judgments is to eliminate all ambiguities in the original judgment by way of a second judgment whose res judicata effect displaces that of the first judgment, even if this means turning part of the reasoning of the prior judgment into part of the dispositif of the later judgment. Only on an interstitial natural law argument which interprets ne ultra petita in an excessive manner would the ICJ Statute not be able to prescribe that a second judgment could change those parts of the first judgment, even if it is ordinarily ‘final and without appeal’.

3.2 Process

Article 60/2 uses the word ‘construe’, rather than ‘interpret’; in contrast, Article 98 of the ICJ Rules of Court speaks of ‘interpretation’, rather than ‘construction’ and both writings and case law treat them as synonymous. ⁵⁵ It is generally agreed in writings and case law that the Article 60/2 process is an interpretation understood as hermeneutic function, i.e., as process of finding the meaning of a text: ‘the expression “to construe” must be understood as meaning to give a precise definition of the meaning and scope’; ⁵⁶ ‘interpretation must … genuinely relate to the determination of the meaning and scope of the decision’. ⁵⁷ It is said that interpretation judgments may only serve ‘to obtain clarification of the meaning and the scope of what the Court has decided’ ⁵⁸ – a ‘procédure de confirmation’, ⁵⁹ aiding our understanding of the original judgment. The interpretation judgment is considered to be ‘purely declaratory, since it does not add anything’ and its only use is to rid the original judgment of any ‘misunderstanding and confusion’. ⁶⁰ Finally, the Court’s interpretative process ‘involves precisely defining the terms of the operative parts of the judgment, and specifying its scope, meaning and purpose’. ⁶¹ Hence, we are faced with the orthodox international legal understanding of interpretation (Article 31(1) VCLT), even though sometimes, such as in the 2013 Preah Vihear judgment, the analogy to treaty interpretation is relativized. ⁶²

To the majority, the nature of the Article 60/2 process as interpretation – as well as the prohibition of any disruption to the res judicata of the original judgment – seems to indicate that the process should be understood in a narrow sense. Writers emphasize that the interpretation judgment must remain ‘within the “strict limits of the principal judgment”’, ⁶³ that it may not add additional content to the original judgment, may only explain but not change or question ⁶⁴ it, and that it ‘can in no way go beyond the limits of the Judgment’. ⁶⁵ Most relevant for the present topic is the assumption that this ostensibly hermeneutic process is limited by the original judgment’s res judicata status: ‘[i]nterpretation … may not change what the Court has already settled with binding force as res judicata’. ⁶⁶ Interpretation is thus reduced to a ‘confirmation of its res judicata effect’ ⁶⁷ which prohibits the Court from closing any gaps left in the original judgment: ‘the Tribunal can scarcely be explaining what has already been decided when it is deciding on a point it had earlier failed to decide’. ⁶⁸

The orthodox understanding of what ‘interpretation’ means in international law, however, goes beyond or is even independent of the hermeneutic process of text-cognition. I have called this understanding ‘interpretation_B’, ‘the applicative construction of law’s meaning … an effort to guide the concretisation of abstract general norms in individual instances, foremost in the process of rendering tribunal judgments’. ⁶⁹ This understanding is expressed well in Arnold McNair’s classic adage:

Strictly speaking, when the meaning of the treaty is clear, it is “applied”, not “interpreted”. Interpretation is a secondary process which only comes into play when it is impossible to make sense of the plain terms of the treaty, or when they are susceptible of different meanings. ⁷⁰

The texts cited above, seeking to clarify what ‘construe’ (‘interpret’) means in Article 60/2, seem to be premised on several of the classical tropes of interpretation_B. For example, it is assumed that an interpretation properly speaking is a precursor to a binding result – the outcome of an interpretation becomes part of the legal order. ⁷¹ The most basic expression of interpretation_B in the debate on interpretation judgments, the idea least accepted by modern theories of language, is that such a judgment only brings to the surface a meaning which the original judgment had held all along. ‘[T]he real meaning of the judgment was – while hidden – already contained in the original judgment’ ⁷² – as if the criminal court’s conviction of the defendant was ‘already contained’ in the penal code or an international treaty was ‘already contained’ in the pacta sunt servanda maxim.

3.3 Outcome

The ICJ Statute does not define the consequences in a legal sense; Article 100(2) of the ICJ Rules of Court, however, tells us that ‘[t]he decision … on a request for interpretation or revision of a judgment shall itself be given in the form of a judgment’. All such decisions of the PCIJ and ICJ so far have been judgments. Hence, interpretation judgments are individual norms like any other of the Court’s judgments; their personal sphere of application includes at least the parties to the case.

On a more detailed examination, there is less certainty, for example regarding the legal relationship between the two judgments and possible norm-structural changes resulting from the creation of the second judgment. If anything, writers excel at evading the tough problem that this procedure must ‘engage’ the res judicata of the original judgment. The original judgment, having the force of res judicata, is now joined by a later interpretation judgment, also having the force of res judicata. The obfuscation of that problem usually involves vastly over-privileging the res judicata of the earlier judgment and suppressing the res judicata of the later judgment, for example in this passage by Chester Brown:

It is incorrect to say that the power of interpretation violates the principle of finality or res judicata, for this position misconceives the role of the request for interpretation: it is not a method of attacking a judgment, but a way of confirming it. The power of interpretation can only be exercised to explain obscurities in a judgment, and not to replace it with a new judgment. ⁷³

‘The language and structure of Article 60 reflect the primacy of the principle of res judicata. That principle must be maintained’ ⁷⁴ and interpretation judgments succeed in this, because they ensure that the judgment is ‘correct’. ⁷⁵ It ‘leave[s] intact its res judicata’ ⁷⁶ and ‘may serve to explain but may not change what the Court has already settled with binding force as res judicata’ and it does not assume the character of an appeal. ⁷⁷

The Gordian knot which had been identified is slashed by telling us that the empowerment to create interpretation judgments is valid only if it is used ‘in a way that will not be seen as an “appeal”’. ⁷⁸ If the later judgment were to purport to change the earlier judgment, the Court would have acted ultra vires. ⁷⁹ The ‘clarification’ mentioned in Section 3.2 is thus explained as a limit which positive law sets for the empowerment in Article 60/2. This has the merit of greater coherence with legal-theoretical insights than insisting on calling it a hermeneutic process, an exercise in meaning-ascertainment. It therefore empowers a ‘clarification’ only within the frame provided by the earlier norm, which makes sense from a norm-structural point of view. What does not make sense, however, are the consequences which are sought to be drawn: ‘The interpretative judgment in itself however does not impose such obligations on the parties. It is purely declaratory, since it does not add anything to the already created res judicata, its legal force depending on the previous judgment.’ ⁸⁰ How can statements like this be reconciled with the interpretation judgment being a judgment ⁸¹ and thus having its own res judicata status? Either the interpretation judgment is a judgment and thus binding and at least potentially in (legal) conflict with the original judgment, ⁸² or it is a mere declaratory interpretation without any consequence on the legal level, although it would then be rather like an advisory opinion or a statement by an academic.

3.3.1 The Preah Vihear interpretation judgment

On 15 June 1962, the Court rendered its judgment on the merits in Temple of Preah Vihear; the dispute centred around the border between Cambodia and Thailand in the area of the eponymous temple. The Court’s decision was based to a significant degree on the so-called ‘Annex I map’ (see Figure 1), drawn up in 1907 by French military cartographers detailing the border between French Indochina and the Kingdom of Siam. ⁸³ Article I of the relevant 1904 French–Siamese treaty in effect prescribed the watershed as relevant criterion; however, the 1907 cartographers seemed to have departed from this principle precisely in the area of the temple.

Figure 1. Annex I map, modified with an enlarged cut-out of the temple area.

While Thailand insisted on the terms of the 1904 treaty and on the strict application of the watershed as rule for the delimitation in this area as well, the Court famously found that its behaviour could be seen as acquiescence: ‘[T]he circumstances were such as called for some reaction, within a reasonable period, on the part of the Siamese authorities … They did not do so, either then or for many years, and thereby must be held to have acquiesced.’ ⁸⁴ Thailand was thus estopped from doubting the validity of the Annex I map:

Even if there were any doubt as to Siam’s acceptance of the map in 1908 … the Court would consider, in the light of the subsequent course of events, that Thailand is now precluded by her conduct from asserting that she did not accept it. ⁸⁵

The Court’s reasoning contains the clear holding that it ‘feels bound … to pronounce in favour of the line as mapped in the disputed area’, ⁸⁶ whereas the disposititf only contains the neutral finding that ‘the Temple of Preah Vihear is situated in territory under the sovereignty of Cambodia’. ⁸⁷ Neither do the operative paragraphs mention where, exactly, the boundary line runs in the ‘disputed area’ nor did the judgment demarcate the border itself.

The dispute flared up again beginning in 2007 and Cambodia in 2011 submitted an interpretation request, the case turning on whether the Court had in 1962 demarcated the border and which parts of the area are part of the ‘vicinity’ of the temple. Although the Court had held in the original judgment that the border line as shown on the Annex I map is binding, the interpretation judgment came to a different conclusion. In 2013, it held that the temple is situated on a promontory, and the original judgment’s holding refers to the entirety of that promontory, but not to other parts of the border area. ⁸⁸ The Court concluded its reasoning in paragraph 98:

[T]he limits of the promontory of Preah Vihear, to the south of the Annex I map line, consist of natural features. To the east, south and south-west, the promontory drops in a steep escarpment to the Cambodian plain … To the west and north-west, the land drops in a slope, less steep than the escarpment but nonetheless pronounced, into the valley which separates Preah Vihear from the neighbouring hill of Phnom Trap … the Court considers that Phnom Trap lay outside the disputed area and the 1962 Judgment did not address the question whether it was located in Thai or Cambodian territory. Accordingly, the Court considers that the promontory of Preah Vihear ends at the foot of the hill of Phnom Trap, that is to say: where the ground begins to rise from the valley. In the north, the limit of the promontory is the Annex I map line, from a point to the north-east of the Temple where that line abuts the escarpment to a point in the north-west where the ground begins to rise from the valley, at the foot of the hill of Phnom Trap. ⁸⁹

The dispositif, in turn, incorporates paragraph 98: ‘The Court … [d]eclares, by way of interpretation, that the Judgment of 15 June 1962 decided that Cambodia had sovereignty over the whole territory of the promontory of Preah Vihear, as defined in paragraph 98 of the present Judgment.’ ⁹⁰ Hence, at least in the north-western sector, while the Court did not explicitly draw a border line, it did delimit the relevant area – the vicinity of the temple, which departs from the border line as shown in the Annex I map and which was held to be binding in 1962 – in a manner that favoured Thailand (see Figure 2).

Figure 2. Further enlarged portion of the Annex I map, highlighting the border line as originally drawn, the area of the hill Phnom Trap and the approximate delimitation of the temple promontory vis-à-vis Phnom Trap by the Court in 2013.

The reactions of doctrinal legal scholarship were mixed; John Ciorciari remarks that this reasoning ‘required reaching well beyond the reasoning in the original decision’ and that the ICJ seemed ready ‘to engage in significant efforts to fill interpretive gaps’. ⁹¹ Jessica Joly Hébert argues that the Court’s interpretation ‘went beyond simply confirming or clarifying … in effect, it did considerably more than that … the Court concluded that “vicinity” applied to the whole area of the promontory on which the Temple is located, thereby rewriting, in effect, the boundary between the parties’. ⁹²

Even on a doctrinal-legal analysis, the Court’s 2013 interpretation judgment is considered to be problematic, but rather than relitigating the 2013 Preah Vihear judgment, the following sections will focus on the theoretical problem which this judgment brings to the fore.

4. Adolf Julius Merkl’s Error Calculus

At this point, a brief introduction to Merkl’s Error Calculus (Fehlerkalkül) theory is necessary to acquaint non-specialists with the theoretical framework to be applied in Sections 5 and 6. Adolf Julius Merkl, Hans Kelsen’s closest collaborator, is little known outside Austria, even among legal theorists. This is unfortunate, because Merkl developed a noteworthy theory of how legal orders are organized, in Hartian terms: a theory of secondary rules. Among other innovations, Merkl discusses solutions to the eternal problem of ‘erroneous’ decisions, recognizing that traditional scholarship’s concerns are largely misplaced. Modern law actually takes errors into account, he argues. The law provides a ‘calculus’ for how to deal with errors; the law, in other words, internalizes errors.

The story starts with the Stufenbau theory, ⁹³ which is rightly one of the most successful selling points of the Pure Theory of Law, more so even than the Grundnorm. The Pure Theory of Law in effect aims to clarify the theoretical foundations for ‘a structural analysis of positive law’, ⁹⁴ because law is formally structured and legal norms are not merely clustered, but relate to one another in specific and typical ways. The Stufenbau theory argues that it makes sense to metaphorically represent these inter-norm relationships as hierarchical orders. The name that has stuck denotes a step pyramid (Stufenbau is an arcane word for Stufenpyramide), but a tree or network structure are probably more apt metaphors. ⁹⁵ There are at least two possible types of inter-norm relationship which could provide the basis for such a structure: creation and destruction; both come into play in Merkl’s Error Calculus theory.

The more important type of relationship is that of creation, as only it is ‘a necessary corollary of the concept of law’ ⁹⁶ – destruction relationships are contingent. Law is made up of norms; norms ‘exist’ only in terms of being valid. Validity is not a property of a norm; in other words, there are no invalid norms and a norm cannot be valid without its validity being traced back to another norm. ⁹⁷ The ‘higher’ norm is the basis of the validity of the ‘lower’ norm. ⁹⁸ A special type of norm, the empowerment norm, authorizes the creation of further norms. ⁹⁹ Legal scholarship must therefore primarily conduct an analysis of the structure of that type of order. ¹⁰⁰

Hence, empowerment norms – roughly equivalent to Hartian rules of change – enable law to become dynamic in the sense that they allow for the creation of new law within the legal order. If a legal organ, authorized by the empowerment norm, performs a norm-creating act of will and fulfils the (other) conditions of the empowerment norm, a new norm of that legal order is created. That is the reason why Merkl insists that the nomomechanical core of res judicata/Rechtskraft is the relative immutability ¹⁰¹ of a norm which results from the act of creation. However, the authorization to create a norm does not imply an authorization to destroy (change) it, while the immutability is ‘relative’, because it is dependent on the content of the positive law. The mere fact that a norm has been created at a later date does not suffice to privilege it over the older norm. ¹⁰² The lex posterior maxim is not a logical necessity, but a possible content of positive law.

But what if the process of creation goes awry? What if a legal organ, for example a tribunal, attempts to create a judgment, but makes a mistake? This is the problem of ‘erroneous’ legal decisions. For German-language jurists, one of the core functions of Rechtskraft is to ‘rectify’ or ‘ratify’ both correct and erroneous decisions as soon as Rechtskraft accrues to a judgment. In response, Merkl contends that this feat of legal magic can only be explained by a typical feature of modern, complex legal orders – the Error Calculus, the anticipation of errors by the law. ¹⁰³ The Error Calculus works by distinguishing between minimum conditions and maximum conditions: (i) If an act does not even fulfil all minimum conditions, no norm is created; (ii) if an act fulfils all minimum conditions, but not all maximum conditions – if, for example, a court makes a minor mistake – a norm is created which can and is likely to be annulled upon review/appeal, unless Rechtskraft accrues after the specified time for appeals has passed; (iii) only if the act fulfils all maximum conditions as well is it likely to survive review/appeal. ¹⁰⁴ This is a distinction which the law can, but does not have to make; it is certainly not to be supplied by legal scholarship. In Merkl’s theory, Error Calculus and Rechtskraft are connected; for him, neither one nor the other miraculously transforms non-law into law. In order for review/appeal mechanisms to work, the erroneous act first has to be considered a norm – only then can it be repealed, annulled or derogated. ¹⁰⁵

5. The nomomechanics of error in international law

Legal theory scrutinizes statements about the law made by doctrine. In our case, it assesses whether statements regarding the nature of res judicata in international law are coherent with the static and dynamic behaviour of normative orders. I call this theory of the possible inter-relationship of norms ‘nomomechanics’, because norms, in both rest and movement, follow mechanisms somewhat akin to the laws of mechanics. This section uses Merkl’s theory of legal nomomechanics as metrological tool to gauge whether and how an interpretation judgment alters the structure of the international legal order. Section 6, in contrast, turns this approach on its head: interpretation judgments can also ‘test’ Merkl’s ideas about nomomechanics.

Even on a doctrinal reading, the 2013 judgment in Preah Vihear is seen to go beyond a mere reformulation of the 1962 judgment and beyond a ‘discovery’ of a meaning hidden in that text. Even those who agree with the Court’s holdings will have to admit that the 2013 judgment is a new norm and that the Court held with binding force that Phnom Trap is at least not part of the ‘vicinity’. On a Merklian account, in contrast, ICJ interpretation judgments are not an interpretation in any meaningful sense of the word, despite the humans serving as the Court’s judges almost certainly having employed hermeneutics before the legal organ ‘International Court of Justice’ created the judgment. This is so because interpretation as text-cognition can only be performed by humans, not by legal organs. ¹⁰⁶ An ICJ interpretation judgment is an authentic interpretation ¹⁰⁷ in the strictest sense, leading to a ‘replacement of the dispositif of the impeached judgment’; ¹⁰⁸ ‘an interpretation judgment will itself amount to res judicata … notwithstanding its presumed “declaratory” quality’. ¹⁰⁹ An authentic interpretation creates a later norm of the same sort as the norm which has been authentically ‘interpreted’, such as a later treaty explaining, with binding force, how a certain term in an earlier treaty is to be understood – or a later judgment, explaining, with binding force, how the terms ‘vicinity’ or ‘promontory’ are to be understood in an earlier judgment. ¹¹⁰ Authentic interpretation only claims to interpret in the hermeneutic sense; it actually changes the sum-totality of the norms contained in a legal order.

Therefore, ICJ interpretation judgments impinge on the res judicata of the original judgment and perhaps even disrupt it. There are hints of this in the case law; for example, in Channel Arbitration, the tribunal found that the delimitation in the dispositif of the original award was incommensurable with the method developed in the reasoning of that award; this ‘contradiction … must, therefore, be resolved in favour of the findings in the reasoning’, ¹¹¹ finding that this would involve ‘some rectification’. ¹¹² This is criticized by Rosenne:

The Tribunal proceeded to do this in effect by replacing the formal wording of the impeached dispositif with what it called “rectification” … the characterization of the discrepancy as a material error which the Court was empowered to rectify per se is not convincing. ¹¹³

What, then, are the ‘limits … of the legal norm, to which res judicata and a particular degree of (im)mutability are ascribed’? ¹¹⁴ Because both judgments are norms, the res judicata effect of the original judgment is at least relativized by the interpretation judgment: a later norm (N₂) with the same personal and material sphere of application and of the same legal type (‘judgment’) joins an earlier norm (N₁). The parties are therefore bound by Article 60/1 to observe both N₁ and N₂. In addition, the issues identified by Merkl in his 1918 article on lex posterior ¹¹⁵ are relevant as well: (i) Does positive law contain an empowerment to change, rather than merely to create the law? Does the Statute contain an empowerment to derogate in addition to the empowerment to create which is enshrined in Article 60/1?; (ii) is N₂ a ‘derogation actualising norm’ (derogatorische Norm) ¹¹⁶ – a condition for the (partial) derogation of N₁?; (iii) is this a genuine case of derogation or merely an accumulation of two norms, potentially resulting in a conflict of simultaneous observance (Befolgungskonflikt)? ¹¹⁷ Accumulation is probably what happens by default and it is similar to the relationship between treaty and customary international law envisaged in Nicaragua. Joly Hébert’s argument seems to support this position: ‘a “parallel deployment” of the effects of both the original and interpretation judgments may occur … It therefore allows for the possibility of both judgments being applicable, to some extent, at the same time’. ¹¹⁸

Nomomechanically speaking, therefore, there is no difference between an appeal, a revision and an interpretation judgment – the original judgment’s res judicata is not sacrosanct in either of these means of redress – even though there is a significant difference between these three modes as a matter of positive law. The Article 60/2 process is in effect a means of appeal, even if the original judgment is not formally derogated; at the very least, a second norm joins the first which itself constitutes a command – a ‘material derogation’ in traditional terminology. The Court appears to be the final instance, because there is no possibility of appeal in the traditional sense. And in one sense, this is correct: no other ‘institution’ will sit in judgment over its pronouncements. But the claim that the Article 60/2 process is categorically different from an appeal, because the latter disrupts the original judgment’s res judicata whereas the former leaves it intact, is incorrect.

Ironically, the interpretation procedure is not just little different from an appeal – no different, except that it is apparently the same organ which decides on its own judgments, rather than a higher court – but it is in some respects even more potent: (i) There is no time bar to lodging a request for interpretation; in Preah Vihear, the application was lodged almost 49 years after the original judgment; ¹¹⁹ (ii) the ICJ has inherent jurisdiction regarding interpretation judgments; (iii) technically, an endless number of interpretation judgments could be issued on a single original judgment – this could even extend to an endless chain of interpretation judgments on interpretation judgments – whereas appeals procedures are usually tightly limited. The fact that the law does not prescribe an ‘error’ in the original judgment as necessary condition for the issuance of an interpretation judgment does not change the fact that it can be characterized as an Error Calculus in Merkl’s sense – an Error Calculus of the supposed final instance ‘ICJ’ which is reviewed by the true final instance ‘ICJ in the interpretative mode’.

6. Error Calculus without errors

Merkl’s Error Calculus has both a constructive and a critical function. The constructive side is part of res judicata/Rechtskraft: ‘erroneous’ legal acts gain legal stability. The critical side comes to the fore where forms of redress are incorporated – certain errors are conditions for the derogation by a higher instance. In a (multi-part) article published in 1916, 1917, and 1919, Merkl develops what I call the ‘Ratification Thesis’ (Ratihabierungsthese) (see Figure 3) in order to explain what happens. He argues that when a norm is thus ratified, ‘an entity which may a priori only appear to be a law can a posteriori be cognised as proper law’. ¹²⁰

What actually happens is that the act [in question] only becomes a valid act of state in retrospect, i.e., as soon as it has achieved Rechtskraft … Hence, certain invalid acts which come close to [fulfilling the criteria for] valid acts … will under certain conditions be able to be ratified afterwards … ¹²¹

Figure 3: Merkl’s ‘Ratification Thesis’.

The moment of observation is crucial for the Ratification Thesis: at the moment of (purported) creation, an act which fulfils all minimum but not all maximum conditions can only be cognized as legal nonentity. However, the error is healed, as soon as this nonentity has achieved Rechtskraft, retroactively to the moment of creation. ¹²² By this process of retroactive ratification it becomes a norm. In later writings, Merkl discards the Ratification Thesis. In his 1923 monograph, he develops a different solution (see Figure 4); his belief in the ‘legal[] miracle’ of ‘turning illegal acts into legal acts’ ¹²³ has been abandoned:

The possibility of appeal in legal orders … does not refer to “erroneous acts” in any sense of the word, because these acts contain nothing which would necessitate a repeal or would even be capable of being repealed … Only when an act is simultaneously valid despite its deficiencies and able to be repealed on the basis of a means of redress because of its deficiencies … [will] even “erroneous” acts be able to be repealed. This happens by way of a derogation [repeal] under certain specific conditions in the same way as the derogation of properly created acts. ¹²⁴

Figure 4. Reconstructing the Error Calculus as ex post derogation.

My reading of this passage is that the minimum and maximum conditions cannot diverge in the sense that maximum conditions are conditions for the creation of a norm, i.e., are part of the empowerment norm in a narrow sense. For legal organs which serve as final instance (Grenzorgan), there are no maximum conditions (as there is no higher instance to review its decision) and all conditions are either minimum conditions or irrelevant. ¹²⁵ For organs for which the law provides an Error Calculus, the situation is different: minimum conditions are maximum conditions. If all minimum conditions are fulfilled, a norm is created. In that situation, the maximum conditions – ostensibly conditions for law creation which are within the tolerance range of the Error Calculus – serve as conditions for the derogation of the (dogmatically ‘imperfect’, but nomomechanically ‘perfect’) norm. For example, if A, B, and C are the minimum conditions for the creation of norm N₁ and D, E, and F are its maximum conditions, then the fulfilment of A and B and C suffices to create N₁. The non-fulfilment of D, E and/or F is a condition for the later derogation of N₁, even though the positive law might specify that this happens retroactively, thus minimizing the practical impact of an ‘imperfect’ norm. If all this is the case, then nomomechanically speaking, rather than legal-doctrinally (regarding the content of a particular positive law), I can see no reason why we should want to distinguish between what happens when a norm is derogated because it contains an ‘error’, on the one hand, and any other occasions and conditions for derogation, on the other hand, particularly the regular amendment of statutes and constitutions.

When we speak of creating imperfect law, erroneous quasi-law, we confound the typical shape of positive domestic law with the theoretical structure of normative orders and impose the particular shape of our (domestic) legal order on our legal-theoretical arguments. International law as less complex legal order helps us to see that the institutes of domestic law which are designed to combat erroneous law-making can neither be universalized nor fashioned into theoretical statements. The interpretation procedure before the ICJ is a good example: the Error Calculus does not enable the retroactive ratification of ‘imperfect law’, but the subsequent derogation of validly created law. Hence, on the nomomechanical level, the features of positive law which serve as Error Calculus do not really concern the correction of an error in the process of law-creation, but are rules allowing for derogation/change. ‘Errors’ are not a necessary content of the conditions for derogation or change of judgments and even the institution of ICJ interpretation judgments are an expression of the Error Calculus.

The finality of the Court’s pronouncements raises another problem (and demands another solution) ¹²⁶ which comes to the fore here as well: what if the highest court of the land makes a mistake in creating its judgments? Should we treat all errors as relevant and regard imperfect pronouncements as nullum? Should we consider only ‘significant’ errors and, if so, what is the criterion for significance and which organ should and will enforce it? Should we regard all errors of highest organs as irrelevant, because they are the final arbiter? There is no easy answer to this question. My solution adheres to the (positivist) strictum that all legal organs of a given legal order must observe all requirements for the creation of a given norm if they are to be successful in creating a norm of that legal order. That sounds very conventional, but my argument is based on three unconventional arguments: first, the positivity of a norm needs to be separate from its membership in a specific normative order; second, I seek to abandon the claim made by some Pure Theorists that legal orders are categorically different to other types of normative order and; third, we should further specify the act of will whose sense (meaning) a norm is.

The solution to the problem of decisions in the final instance (although the problem applies to all norm-creation) acknowledges the irresolvable tension between positivity and normativity. On the one hand, every act of will whose sense/meaning is a norm ¹²⁷ can be cognized as norm of some normative order – both in the immanent (subjective) sense and in the systemic (objective) sense ¹²⁸ – and is therefore always successful in creating a norm. On the other hand, a norm, in order to be part of a specific normative order, must fulfil the conditions laid down by the empowerment norm of that normative order. The tension is irresolvable, but too many solutions are caught up in a supposed uniqueness of ‘law’ vis-à-vis other types of norm – and in supposing a necessary unity and coherence of the legal order.

I think the solution is implicit in Kelsen’s late writings, where he abandons the necessity of cognizing only one normative order at a time. ¹²⁹ On this view, it is possible for an imperative-creating act of will to succeed in creating a norm in the systemic sense, of being more than the mere pretence of a norm, because we presuppose a Grundnorm even for the sense/meaning of those acts of will which are not covered by a positive empowerment norm of a legal order. In order to understand that failed attempts at creating law do not necessarily result in nothingness, we will have to distinguish the validity of a norm from its membership in any one particular normative order. Merkl at one point asks ‘what turns an act which claims to be an act of state, what turns a norm which claims to be a valid legal norm, into a member of a particular legal … order’? ¹³⁰ Kelsen argues that ‘a norm whose creation is not determined by a higher norm at all cannot be a norm valid within the legal order and can therefore not belong to it’. ¹³¹

The solution in the classic case of a judgment of final instance which does not accord with the legal rules for its creation is twofold. On the one hand, as product of a creative act, it is clearly a norm. On the other hand, because the conditions stipulated in the relevant empowerment norm of a legal order were not fulfilled, this norm cannot be a member of that legal order, but is a member of another normative order, by default a spontaneous minimalist normative order. This solution would respect the constitutive nature of a norm-creating act of will as well as the absolute relevance of empowerment norms in the foundation of validity within a normative order.

If we conduct a thought-experiment, we can take this yet another step: what if we asked the humans whose acts of will are attributed to the organ whether they would still insist on having produced a norm if the act they had just produced did not fulfil (all) the conditions for the creation of the norm? While they will usually believe that they had fulfilled all conditions, would they still want a norm to be created if it were certain that the conditions were not fulfilled? In other words: are they possessed of a voluntas eventualis? If, for example, the International Court of Justice’s majority requirement in Article 55(1) were not fulfilled, if only five of 15 members present voted in favour, would the five judges, if they were aware of that shortfall, abandon their will to create a norm or would they insist on nonetheless creating a norm? In the former case, clearly no norm would have been created, while in the latter case, unlikely though it might be, they would succeed in creating a norm, but not of the normative order in question (international law) but of a new normative order created by the five ICJ judges.

The standard interjection at this point is that we cannot conduct such an experiment, because we cannot know objectively whether the conditions are fulfilled. After all, the interjector might continue, the Court is empowered to decide, hence it is also empowered to determine whether the conditions are fulfilled. But that conflates two levels of argument: there is no authoritative cognition ¹³² in the sense that the law authorizes someone to cognize or that the law prescribes a certain scholarly truth. The epistemic ‘assessment’ by legal scholarship is ‘authoritative’ only in the sense of academic/scholarly authority. A zoologist stating ‘five gazelles were killed by this pride of lions in Kenya in the last week’ may possess more ‘authority’ than the person on the high-street being asked about the pride’s kill ratio, but the zoologist’s statement is still only true – ‘authoritative’ in that sense – if the amount killed was, in fact, five. It is emphatically not true because the person uttering the statement is a professional zoologist. Legal scholarship is not in a significantly different position and the common misconception that it is, is one of the fundamental problems of our profession.

There is a high price to be paid for my solution, however, and it is much higher than a mere lack of practicability. This solution endangers the assumed absolute unity of the legal order, more precisely: it loosens the connection between empowerment norm and norm created by it. The Grundnorm can be posited at any place in a normative order which is thus cognized as its own unity. Vice versa, is there a norm-structural necessity to ‘connect’ a norm to any empowerment norm that ‘fits’, i.e., whose conditions fit the process that has created the norm? Imagine that I create a positive norm which retroactively empowers the creation of the original norms of customary international law, something like: ‘I herewith authorise the creation of customary international law.’ It is possible to see the creation of all customary international law as authorized by my norm and thus part of the order ‘Jörg Kammerhofer’s norms’, but should we? Do the rules of nomodynamics require us to connect the empowerment norm and its product if and as soon as they fit? Following my fiat is plainly ludicrous, of course, but is it any less ludicrous to substitute a norm of natural law? If there is no requirement as a matter of proper scholarship to connect the norms in these cases, why should the matter be any different for ‘normal’ connections ‘within’ a legal order? This is not a bug, it is a feature of the constitutive nature of all human cognition.

7. Conclusion

Hans Kelsen once wrote that legal scholarship is a province far removed from the Geist, ¹³³ that untranslatable but quintessentially German word. If legal scholarship is a province far removed from the ideal of scholarship, then clearly, international legal scholarship must be the most remote village in that province, the last village before the gates of Mordor. It is the last, the poorest, but certainly not the smallest village, whose inhabitants compulsively adhere to traditions and superstitions which others in that province have long abandoned. Ironically, at the same time, the villagers insist on following the dernier cri in scholarly trends, perhaps because in this remote place, a fashion faux-pas is less easily recognized than in the province’s capital city.

Will these uncouth villagers, pitch-forks at the ready, be able to unseat the brilliant theoretical insights of the Pure Theory of Law, all the way from glamorous Vienna, as my audience of (Viennese) legal theorists might suppose? Will cosmopolitan international lawyers show the Pure Theory for what it is, an over-complex, ivory-tower impractical sort of logicist construct, unfit for the purpose of governing the world, as some of my fellow international legal scholars might answer? Or will the Odyssean journey or Sisyphean labour (take your pick in Greek mythology) of trading arguments and almost eternally talking past each other continue, continuing the debate about the nature and inner workings of the law for the foreseeable future in the same manner it has for the past two millennia?

At this point, a cynic will argue that the outcome I have presented in this article is that the international legal mainstream position on res judicata must founder on the shoals of the Pure Theory, but only because I had foisted its premises on the discussion. They have the beginnings of a point, but not because of a surreptitious, rogue move on my part, but because doctrinal scholarship cannot itself develop (all) the foundations for its work. It must rely on legal theory for an understanding of legal epistemology and of nomodynamics. There will have to be a division of labour between doctrine and theory and if doctrine refuses the help which theory proffers, it must live with the almost inevitable inconsistencies in its arguments. Merkl’s theory is not an attack on orthodox scholarship as so many critical approaches are, but an offer of help on the basis of a largely shared understanding of law, norms and rules. It offers to help international legal doctrine, if only by allowing for a more precise understanding of the problems of the static and dynamic structure of the international legal order and by unmasking pragmatic strategies of problem-avoidance.

In another sense, however, international legal doctrine’s arguments about the content of international law are both a legal theorist’s raw-material as well as their gauge. As this article has also shown, Merkl’s theory was developed largely by those concerned with Austrian domestic law and its traditions and this shows when we switch contexts to international law and its traditions. The vast range of possibilities of structuring and changing legal orders have not yet fully taken account by this theory, regrettably turning domestic cultural prejudices into supposed theoretical necessities. I say this with deep gratitude for my intellectual forbears and teachers, but in the end we must admit that, alas, legal theory turns out to be yet another village in the province of legal scholarship.