Of Planets and The Universe: Self-Contained Regimes in International Law

The European Journal of International Law Vol. 17 no.
3 © EJIL 2006; all rights reserved
...........................................................................................
Of Planets and the Universe:

Self-contained Regimes
in International Law
Bruno Simma* and Dirk Pulkowski**
Abstract
Contemporary legal practice requires the allocation of authority within a complex system of
legal prescriptions. As international law has extended to areas as diverse as trade,
environmental regulation and human rights, the consequences of breach of international
legal obligations become more difficult to assess. The authors probe the role of the lex
specialis maxim as a tool for the effective placing of special secondary rules within the
general international law of state responsibility. The central question is: Are the general
rules on state responsibility to apply residually? The authors answer in the affirmative.
‘Conceptual’ arguments for so-called self-contained regimes are unconvincing. Scholars who
perceive international law as a unified legal order might be led to apply a presumption in
favour of the applicability of the general international law of state responsibility. Scholars
who regard international law as no more than the sum total of loosely interrelated
subsystems tend to advocate a presumption in favour of the normative closure of a particular
regime. In the authors’ view, neither presumption is helpful, since both tend to obfuscate the
value judgments that legal decision-making inevitably involves. Instead, the authors propose
that a fallback on general international law, including resort to countermeasures, may be
justified on normative grounds. A closer analysis of four subsystems that have often been
associated with the notion of self-contained regimes – diplomatic law, European Community
law, the WTO and human rights – concludes the discussion.
* Judge, International Court of Justice (The Hague); Professor of International Law, retired, Ludwig-
Maximilians-Universität (München); Member of the Affiliate Overseas Faculty, University of Michigan
Law School (Ann Arbor). Email: simma@icj-cij.org.
** Doctoral candidate in international law, Ludwig-Maximilians-Universität (München); LL.M., Yale Law
School (New Haven); Ass. jur. (München). Email: dirk.pulkowski@aya.yale.edu.
The authors wish to thank Andreas L. Paulus for his well-directed and constructive critique of an early
draft of this article, which proved invaluable for the final text. Unless specified otherwise, translations of
French texts by Euan Macdonald.
........................................................................................................................................................
EJIL (2006), Vol. 17 No. 3, 483–529 doi: 10.1093/ejil/chl015
484 EJIL 17 (2006), 483–529
1 Self-contained Regimes: A Classic Revisited

The needs of an increasingly interconnected and functionally diversified society have
radically transformed the international legal order. Early 20th-century accounts con-
ceived of the international system as a hierarchical pyramid structure comprising rel-
atively few norms, in which states, perceived as opaque and unitary actors (‘billiard
balls’1), interacted in a largely unconstrained manner. Contemporary international
law, by contrast, resembles a dense web of overlapping and detailed prescriptions in
subject areas as diverse as environmental protection, human rights and international
trade. In light of such transformations, political and legal theorists have introduced
the concept of the network as a competing structural paradigm of the legal order;
they describe international law as a network of government officials, legislators and
judges;2 as a deterritorialized ‘system of rule’ that has transformed the state;3 or as a
flexible, horizontal structure of production of legitimacy spread throughout world
space.4 And yet, while the network model accounts for the diversification of legal
regimes and the multiplication of actors in the international legal process, many cent-
ral building blocks of the traditional international legal system, such as the rules on
diplomatic protection or state responsibility, have remained remarkably stable. Thus,
François Ost and Michel van de Kerchove have aptly noted that ‘de la crise du modèle
pyramidal, émerge progressivement un paradigme concurrent, celui du droit en réseau, sans
que disparaissent pour autant des résidus importants du premier’.5 A central, and increas-
ingly difficult task of the jurist consists in allocating authority in a system composed
of both elements of hierarchical unity and multiple network structures in diverse
issue areas.
The present article focuses on the multiplication of norms in the field of state
responsibility. Numerous recent treaty instruments contain tailor-made rules on the
legal consequences of breach. On the one hand, such special treaty regimes provide
innovative procedures of dispute settlement, surveillance and reporting, leading to
remedies that did not previously exist under general international law. On the other
hand, the secondary rules of special regimes are often less comprehensive than the
established canon of rules on state responsibility. Hence, international lawyers can-
not escape the question of how such special secondary rules concerning the conse-
quences of breach of international legal obligations relate to the general international
law of state responsibility. Are the rules on state responsibility to apply residually?
Such a residual fallback on the rules on state responsibility is controversial with
regard to subsystems that have attained a particularly high degree of regulatory
1
The metaphor is attributed to Arnold Wolfers. A. Wolfers, Discord and Collaboration: Essays on Interna-
tional Politics (1962), at 19–24.
2
A.-M. Slaughter, A New World Order (2004).
3
S. Sassen, Losing Control? Sovereignty in an Age of Globalization (1995), at xv, 25.
4
M. Hardt and A. Negri, Empire (2000), at 13.
5
F. Ost and M. van de Kerchove, De la pyramide au réseau? Pour une théorie dialectique du droit (2002), at 14.
‘From the crisis in the pyramid model, a competing paradigm – that of law as network – is progressively
emerging, although important residues of the former do remain.’
Of Planets and the Universe: Self-contained Regimes in International Law 485
‘thickness’ and, thus, autonomy. The more a system’s operation is ‘closed’ towards its
international law environment, the less likely it is to fall back on the rules on state
responsibility.
More than 20 years ago, in the 1985 Netherlands Yearbook of International Law, one
of the present authors raised the question whether self-contained regimes in the area
of state responsibility were conceivable.6 The regimes of diplomatic law, the law of the
European Communities and human rights instruments, among others, contain spe-
cial rules concerning the consequences of breaches of their respective primary norms.
They thus present potential candidates for autonomous systems decoupled from gen-
eral international law. That earlier study concluded, however, that completely closed
regimes of secondary rules were neither conceivable nor desirable. Since 1985, the
problem has lost not a shade of its topicality. While some old trenches have continued
virtually unchanged for 20 years (as regards, for example, the debate on the concep-
tual possibility of self-contained regimes), other, new academic battlefronts have
opened up. The increasing academic interest in the diversification of the international
legal order has given the debate additional fresh impetus. Thus, we consider that it is
time to revisit the topic of self-contained regimes in the light of this broader discussion
on the alleged fragmentation of international law.
This article probes the role of the lex specialis maxim as a tool for networking ‘tradi-
tional’ international law and ‘new’ subsystems of international law. As international
law has evolved into an elaborate, but fragmented, structure, in which multiple
regimes govern the legal consequences of breach, the conceptual distinction between
general and special laws remains important for maintaining systemic cohesion. In
principle, the special secondary rules of the regime will prevail. Yet, to the extent that
such rules are inexistent or ineffective, the general rules on state responsibility will
remain applicable. Sociological regime differentiation does not preclude normative
compatibility with general international law. It would be too simple, however, to
assert that a fallback on general international law follows ‘automatically’ from a
mechanical application of the lex specialis maxim. Rather, normative considerations
are ultimately decisive. What has been laid down in a special treaty must be deemed
to embody a particularly strong commitment. The principle of effective interpretation
thus requires that special primary obligations be interpreted as being enforceable. If
the rules and procedures of special systems fail, a fallback on general international
law, including resort to countermeasures, is justified.
2 Special Rules in the Field of State Responsibility

A The lex specialis Principle
In the 1953 British Yearbook of International Law, William Jenks observed that ‘the
conflict of law-making treaties . . . must be accepted as being in certain circumstances
an inevitable incident of growth’ of international law. Jenks urged the international
6
Simma, ‘Self-Contained Regimes’, XVI Netherlands Ybk (1985) 111.
486 EJIL 17 (2006), 483–529
lawyer ‘to formulate principles for resolving such conflict when it arises’.7 In the five
decades following Jenks’ visionary article, international law has witnessed an even
more radical process of functional specialization. As a reaction to such specialization,
one conflict rule in Jenks’ sense that has increasingly moved to the centre of (mostly
academic) attention is the maxim lex specialis derogat legi generali.8
The Articles on Responsibility of States for Internationally Wrongful Acts adopted
by the International Law Commission (ILC) in 2001 equally conceptualize the rela-
tionship between general international law and secondary norms contained in the
proliferating new regimes of international law in terms of a general/special distinc-
tion. The provision of Article 55, titled lex specialis, is designed to open the door to
such special sets of secondary rules:
Article 55 (lex specialis)
These articles do not apply where and to the extent that the conditions for the existence of an
internationally wrongful act or the content or implementation of the international responsi-
bility of a state are governed by special rules of international law.9
While the wording of Article 55 is short and straightforward, it is both one of the
most important and most debatable provisions of the ILC’s Articles.
1 Rationale of the lex specialis Rule
The option of complementing international obligations with a specific set of second-
ary rules is a prerogative inherent to the idea of sovereignty. No state is forced to
adhere to a ‘one-size-fits-all’ approach to state responsibility.10 In most instances, the
sovereign will of states will be expressed in special treaty provisions. It is not to be
7
Jenks, ‘Conflict of Law-Making Treaties’, 30 BYbIL (1953) 401, at 405.
8
To name but a few recent book publications, W. H. Wilting, Vertragskonkurrenz im Völkerrecht (1996);
J. Pauwelyn, Conflict of Norms in Public International Law: How WTO Law Relates to Other Rules of Interna-
tional Law (2003); J. Neumann, Die Koordination des WTO-Rechts mit anderen völkerrechtlichen Ordnungen:
Konflikte des materiellen Rechts und Konkurrenzen der Streitbeilegung (2002). Cf. also the International Law
Commission’s treatment of the notion of lex specialis, M. Koskenniemi, ‘Study on the Function and Scope
of the lex specialis Rule and the Question of “Self-Contained Regimes”’, Preliminary Report by the Chair-
man of the Study Group submitted for consideration during the 2004 session of the International Law
Commission, Doc. ILC(LVI)SG/FIL/CRD.1 and Add. 1, available from the Codification Division of the UN
Office of Legal Affairs.
9
International Law Commission, Report on the Work of its Fifty-third Session, Official Records of the General
Assembly, Fifty-Sixth Session, Supplement No. 10 (A/56/10), at 58.
10
In terms of sources doctrine, the lex specialis principle has been categorized either as a general principle of
international law, as a rule of customary international law, or as a principle of legal logic. As the Inter-
national Tribunal for the Law of the Sea has pointed out, the lex specialis principle was referred to as an
example of a general principle in the drafting process of Article 38 of the Statute of the PCIJ, cf. Southern
Bluefin Tuna, ITLOS Order of 27 August 1999, para. 123; cf. also Neumann, supra note 8, at 86. Scholars
who refer to the lex specialis rule as a principle of treaty interpretation tend to consider it as a rule of cus-
tomary international law, cf. the dissenting opinion of Judge Hsu in the Ambatielos case, ICJ Reports
(1952), at 87 et seq.; R. Jennings and A. Watts, Oppenheim’s International Law (1992), at 1280. Other
authors speak of a principle of legal logic, Pauwelyn, supra note 8, at 388. Finally, G. Schwarzenberger
has referred to the principle as merely a ‘tool in aid of the jus aequum rule’, International Law, vol. I, (3rd
ed. 1957) 496.
excluded, however, that special secondary rules may be established by particular (for
example, regional) custom.11
Either way, the raison d’être of special secondary norms remains the same. Partic-
ular secondary rules are crafted to enhance the efficacy of the primary rules. An early
expression of this rationale dates back to Hugo Grotius: ‘Inter eas pactiones quae supra-
dictis qualitatibus pares sunt ut praeferatur quod magis est peculiare, & ad rem propius
accedit: nam solent specialia efficaciora esse generalibus.’12 Emer de Vattel, a century
later, had the following to say on the superior effectiveness of special norms:
De deux Loix, ou de deux Conventions, toutes choses d’ailleurs égales, on doit préférer celle qui est la moins
générale, & qui approche le plus de l’affaire dont il s’agit. Parceque ce qui est spécial souffre moins d’excep-
tions que ce qui est général; il est ordonné plus précisément, & il paroît qu’on l’a voulu plus fortement.13
Efficacy, for de Vattel, is a consequence of fewer exceptions, more regulatory preci-

sion and, most interestingly, the assumption that what has been laid down in a more
specific manner carries a stronger expression of state will.
The rule lex specialis derogat legi generali has been referred to as a well-recognized prin-
ciple of international law.14 Occasionally, the lex generalis/lex specialis distinction has
played a prominent role in international jurisprudence. The International Court of Justice
had occasion to affirm that a treaty concluded between Hungary and Czechoslovakia
governed the relations between the parties as a lex specialis vis-à-vis the rules of state
responsibility.15 In its Nicaragua judgment, the Court ruled that ‘[i]n general, treaty
rules being lex specialis, it would not be appropriate that a State should bring a claim
based on a customary-law rule if it has by treaty already provided means for settle-
ment of such a claim’.16 In its INA Corporation decision, the Iran-US Claims Tribunal
11
Case concerning Right of Passage over Indian Territory (Merits), ICJ Reports (1960), 6. Arangio-Ruiz, by
contrast, suggested that a derogation from the rules on state responsibility should only be permissible by
way of ‘contractual instruments’, G. Arangio-Ruiz, Fourth Report on State Responsibility, in ILC Year-
book (1992), Vol. II Part One, 42.
12
H. Grotius, De jure belli ac pacis, libri tres (1653) Liber II, Caput XVI, § XXIX. ‘Among those treaties,
which, in the above named respects, are equal, the preference is given to such as are more particular,
and approach nearer to the point in question. For where particulars are stated, the case is clearer, and
requires fewer exceptions than general rules do.’ (Trans. A. C. Campbell, 1814, available at http://
www.constitution.org/gro/djbp.htm).
13
E. de Vattel, Les droit des gens ou principes de la loi naturelle, Reproduction of the original 1758 edition
(1916) Liv. II, Ch. XVII, § 316. ‘Of two laws or two conventions, we ought (all other circumstances
being equal) to prefer the one which is less general, and which approaches nearer to the point in ques-
tion: because special matter admits of fewer exceptions than that which is general; it is enjoined with
greater precision, and appears to have been more pointedly intended.’ (Trans. J. Chitty, 1883, available
at http://www.constitution.org/vattel/vattel.htm).
14
Cf. the dissenting opinion of Judge Hsu in the Ambatielos case, ICJ Reports (1952), at 87 et seq.: ‘It is a
well recognized principle of interpretation that a specific provision prevails over a general provision.’; cf.
also Jennings and Watts, supra note 10, at 1280.
15
Gabdíkovo-Nagymaros Project (Hungary v. Slovakia), ICJ Reports (1997), at para. 132.
16
Military and Paramilitary Activities in and against Nicaragua (Nicaragua v. United States of America), Merits,
ICJ Reports (1986), para. 274. The Court continued to argue, however, that Nicaragua was not required
to initiate consultations under a bilateral treaty before seizing the Court since such requirement would
be ‘wholly artificial’ and ‘excessively formalistic’ in the concrete case.
488 EJIL 17 (2006), 483–529
decided that the standard of full compensation laid down in a Treaty of Amity pre-
vailed over the more liberal standard of compensation for nationalization of property
under general international law.17
Nonetheless, case law which makes express reference to the lex specialis principle as
a tool for determining the relationship between ‘general’ international law and ‘spe-
cial’ regimes is relatively scarce. One possible explanation would be that reference to
a special rule presupposes the perspective of general international law. Thus, both the
International Court of Justice and the Claims Tribunal, in a first step, examined the
content of the rule of general international law and considered, in a second step,
whether states in the particular case had derogated from this standard by creating a
more special set of rules. Tribunals established under a special legal subsystem – such
as WTO panels or the European Court of Justice – generally follow the reverse order of
examination. They are primarily concerned with the content of ‘their’ special law.
Only in a second step, if this special regime proves insufficient to resolve a case, is
resort had to general international rules. While tribunals working on the basis of gen-
eral international law resort to the lex specialis principle to justify the non-application
of general international law, special tribunals are not required to provide a compara-
ble justification for applying the special rules under which they were created.
2 The Relative Character of the lex specialis Maxim
At a first, superficial glance, Article 55 of the ILC Articles appears to be a relatively
straightforward provision requiring nothing more than an exercise of legal logic.
However, the lex specialis principle has a few major built-in problems.
Firstly, when exactly can it be said that one rule is more special than another, and
how far does that specialness extend? As Sir Gerald Fitzmaurice has observed, ‘[t]he
generalia rule can only apply where both the specific and general provision concerned
deal with the same substantive matter’.18 Whether a special norm relates to the same
subject matter as the state responsibility rules can be problematic. For instance, does
a non-compliance procedure under a multilateral environmental agreement (MEA)
concern the same subject-matter as the regime of state responsibility? The subject
matter of the rules of state responsibility is, in essence, the legal consequences of
unlawful conduct. State responsibility, thus deals with consequences of ‘breach’,
while the MEA’s regime regulates procedures for the case of ‘non-compliance’. If non-
compliance does not necessarily imply unlawfulness, while ‘breach’ does, it would be
17
INA Corporation v. Iran, Iran-US Claims Tribunal, Award No. 184-161-1 of 12 Aug. 1985, 75 ILR 1987,
para 378.
18
Fitzmaurice, ‘The Law and Procedure of the International Court of Justice 1951–4: Treaty Interpretation
and Other Treaty Points’, 23 BYbIL (1957) 203. In the writings of Anglo-Saxon authors, the maxim is
predominantly expressed as generalia specialibus non derogant (Jennings and Watts, supra note 10, at
1280; Fitzmaurice, ibid., at 236). In substance, both ways of stating the principle refer to the same rule.
As Fitzmaurice has pointed out in ibid., the generalia rule ‘does not merely involve that general provisions
do not derogate from specific ones, but also, or perhaps as an alternative method of statement, that a mat-
ter governed by a specific provision, dealing with it as such, is thereby taken out of the scope of a general
provision dealing with the category of subject to which that matter belongs, and which therefore might
otherwise govern it as part of that category’.
conceivable to apply the ‘hard’ state responsibility regime in parallel, because the two
sets of norms do not purport to regulate the same subject-matter. Alternatively, it
could be argued that both the regime of state responsibility and an MEA’s non-
compliance procedure spell out consequences of a deviation from normative expecta-
tions. Then, the relevant provisions of the MEA could be considered leges speciales.
Even in the event that it can be established that a special norm concerns the same
subject-matter as the Articles on state responsibility, the question remains as to how
far the specialness of that particular norm extends. To give some contentious exam-
ples, the question has been raised whether a state can claim retrospective compensa-
tion pursuant to the general rules of state responsibility for breaches of WTO law,
although the WTO Agreements do not authorize (nor forbid) such compensation.
Another question would be whether a violation of EC law can ultimately be addressed
by unilateral countermeasures, although the EC Treaty contains a comprehensive
dispute settlement machinery.
A second problem with respect to the lex specialis principle is that it is based on a
particular fiction of unified state conduct. It presumes that states act with a unified
legislative will when they conclude treaties or enact customary rules. In other words,
the maxim is premised on the ‘billiard-ball model’ of international law that many the-
orists consider inadequate or dated. It is simply inconceivable that a state’s ‘intent’
could be directed to both A and non-A at the same time. As Karl Engisch put it classi-
cally, ‘nobody can perform an action that is respectively an omission and the contra-
dictory or contrary opposite at the same time. Hence, the will of a commanding
person can never be consciously directed at once towards the execution of an action
and its opposite, that it be A and non-A.’19
Far from corresponding to the fiction of a unified legislative intent, the reality of
treaty-making appears to present a rather heterogeneous picture. Treaty negotia-
tions in different subject-matter areas may fall within the competence of different
domestic ministries or, at the European level, Directorates General. Moreover,
as Martti Koskenniemi noted in his 2004 Preliminary Report for the International
Law Commission, the creation of a new international norm may not be as deliberate a
process as legal theory suggests. ‘There is no single legislative will behind interna-
tional law. Treaties and custom come about as a result of conflicting motives and
objectives – they are “bargains” and “package-deals” and often result from spontane-
ous reactions to events in the environment’.20
This observation leads to a third, more general problem concerning the lex specialis
principle as a rule of treaty interpretation. From a strict rules perspective, the lex spe-
cialis principle has been characterized as a rule of legal (deontic) logic.21 However, if
19
K. Engisch, Die Einheit der Rechtsordnung (1987, facsimile of the original 1935 edition), at 54: ‘Niemand
kann zugleich eine Handlung bezw. Unterlassung und ihr kontradiktorisches oder konträres Gegenteil prästieren.
Deshalb kann auch der Wille eines Befehlenden niemals bewusst darauf gerichtet sein, dass jemand eine Hand-
lung und zugleich ihr Gegenteil vollziehe, daß zugleich A und non A sei.’
20
Koskenniemi, supra note 8, at para. 28.
21
Cf. supra, note 10 and accompanying text.
490 EJIL 17 (2006), 483–529
we look at how legal decisions are made, this assumption appears doubtful. Lawyers
make use of various (and sometimes contradictory) ‘tools’ of interpretation, including
the lex specialis principle, to reconcile competing rationalities expressed in different
rules of law. As Koskenniemi has noted,
[i]nterpretation refers to contested and conflicting principles, none of which can be held supe-
rior to the others in a general way. There are no rules on when to apply a literal and when a
dynamic interpretation; when to have recourse to party will and when to the instrument’s
object and purpose. . . . The arbitrator can resolve the dispute only by leaving the ground of
legal interpretation altogether.22
From a realist perspective, Schwarzenberger thus referred to the principles of

treaty interpretation as merely ‘tool[s] in aid of the jus aequum rule’.23 In that sense,
their function may resemble what Vaughan Lowe labelled as interstitial norms: ‘The
choice is made by the judge not on the basis of the internal logic of the primary
norms, but on the basis of extraneous factors’.24 The application of principles of
treaty interpretation quite generally is not merely an exercise in legal logic. Nor is
the characterization vel non of two norms, identified as dealing with the same subject
matter, in terms of a special law/general law distinction, a schematic exercise.
Whether a prescription is too general to govern a particular case is equally ‘a matter
of harmony with what, for want of a better word, one might term experience and
common sense [: . . . ] an unsystematized complex of moral, cultural, aesthetic, and
other values and experiences.’25 The true function of the lex specialis principle lies
precisely in its capacity to give articulation to such values and experiences of the
international decision-maker.
B From leges speciales to So-called Self-contained Regimes

1 Self-contained Regimes: A Definition
If we imagine a sliding scale of specialness, one could conceive a rule at one end that
is only designed to replace a single provision of the set of rules on state responsibility,
while leaving the application of this framework otherwise untouched. At the other
end of the scale, a strong form of lex specialis could exclude the application of the gen-
eral regime of state responsibility altogether, either by explicit provision or by impli-
cation, that is, by virtue of a regime’s particular structure or its object and purpose.
This latter concept of a strong lex specialis designed to exclude completely the general
international law of state responsibility is what we denote as a ‘self-contained
regime’. ILC Article 55 is meant to cover all kinds of special rules, from weaker forms of
specialness that only modify the general regime on a specific point to strong forms
22
Koskenniemi, ‘Hierarchy in International Law: A Sketch’, 8 EJIL (1997) 566, at 575–576.
23
Schwarzenberger, supra note 10, at 496.
24
Lowe, ‘The Role of Law in International Politics’, in M Byers (ed.), The Politics of Law-Making: Are the
Method and Character of Norm Creation Changing? (2000) 216.
25
Ibid., at 220.
such as self-contained regimes that attempt to exclude the application of the general
rules of state responsibility altogether.26
The phrase ‘self-contained regime’ was coined by the Permanent Court of Interna-
tional Justice in the S.S. Wimbledon case. There, the Court was faced with the question
whether the provisions of the Treaty of Versailles relating generally to German water-
ways also applied to the Kiel Canal. The Court pointed out that the drafters of the
Treaty had devoted a special section to the Kiel Canal, which differed substantially
from the rules relating to other watercourses.27 The Court concluded that
[t]he provisions relating to the Kiel Canal in the Treaty of Versailles are therefore self-
contained; if they had to be supplemented and interpreted by the aid of those referring to the
inland navigable waterways of Germany in the previous Sections of Part XII, they would lose
their ‘raison d’être’ . . . The idea which underlies [the specific provisions regarding the Kiel
Canal] is not to be sought by drawing an analogy from these provisions but rather by arguing
a contrario, a method of argument which excludes them.28
In the Wimbledon case, the Court applied the concept of self-containment to resolve a
question of treaty interpretation concerning the relationship between two sets of
primary international obligations.
More recently, the International Court of Justice in its Tehran Hostages judgment
transposed the concept of self-contained regimes to the level of secondary norms. The
Court asserted that the regime of specific legal consequences contained in the Vienna
Convention on Diplomatic Relations was self-contained vis-à-vis the customary interna-
tional law of state responsibility. Consequently, in the event of violations of the Vienna
Convention, no resort may be had to any of the remedies provided for by general inter-
national law, because ‘diplomatic law by itself provides the necessary means of defence
against, and sanction for, illicit activities by members of diplomatic or consular mis-
sions’.29 After exploring in detail the sanctions contemplated by the Vienna Convention
(such as the option of declaring a diplomat persona non grata) the Court concluded:
The rules of diplomatic law, in short, constitute a self-contained regime which, on the one
hand, lays down the receiving State’s obligations regarding the facilities, privileges and immu-
nities to be accorded to the diplomatic missions and, on the other, foresees their possible abuse
by members of the mission and specifies the means at the disposal of the receiving State to
counter any such abuse. These means are by their nature, entirely efficacious.30
The concept of self-contained regimes attracted scholarly attention only after the
Tehran Hostages ruling. Lack of uniform terminology has probably contributed a good
deal to the controversial character of the discussion addressing the alleged self-
containment of legal sub-systems. Various levels of autonomy have been associated
with the term ‘self-contained regimes’.
26
ILC, supra note 9, at, 359 para 5.
27
S.S. Wimbledon, PCIJ, Ser. A, No. 1, at 23.
28
Ibid. at 24.
29
United States Diplomatic and Consular Staff in Tehran, ICJ Reports (1980), at 38.
30
Ibid., at 40.
492 EJIL 17 (2006), 483–529
First, the notion of ‘self-contained regimes’ has been misconceived as an argument

in favour of entirely autonomous legal subsystems. Social systems cannot exist in
splendid isolation from their environment. This point is conceded even by ardent pro-
ponents of regime specialization. According to Niklas Luhmann’s Systemtheorie, for
example, all systems are to some extent interlinked by structural coupling.31 Simi-
larly, legal subsystems coexisting in isolation from the remaining bulk of interna-
tional law are inconceivable. There will always be some degree of interaction, at least
at the level of interpretation. In the words of the Chairman of the International Law
Commission’s Study Group on ‘Fragmentation of International Law’:
No treaty, however special its subject-matter or limited the number of its parties, applies in a
normative vacuum but refers back to a number of general, often unwritten principles of cus-
tomary law concerning its entry into force and its interpretation and application. Moreover,
this normative environment includes principles that determine the legal subjects, their basic
rights and duties, and the forms through which those rights and duties are modified or
extinguished.32
Without the ‘omnipresence of “general law”’33 a special legal subsystem may, as

Georges Abi-Saab put it, mutate into ‘a legal Frankenstein’ that ‘no longer partakes in
the same basis of legitimacy and formal standards of pertinence’.34 Even the European
Court of Justice has asserted that principles of general international law are applic-
able residually within the context of EC law, and has indicated its willingness to defer
to the interpretation of an international agreement by a court established under such
an agreement.35 In the case of the WTO, the Appellate Body has acknowledged that
the GATT remains firmly imbedded in general international law, stating that the
Agreement ‘is not to be read in clinical isolation from public international law’.36
Thus, to avoid confusion, the term ‘self-contained regime’ should not be used to cir-
cumscribe the hypothesis of a fully autonomous legal subsystem.
Nor should the term be used to describe leges speciales at the level of primary rules,
although it is precisely in the context of primary rules that the Permanent Court of
International Justice had originally introduced the concept. In its original meaning,
the concept denoted a set of treaty provisions that cannot be complemented
through the application of other rules by way of analogy. After Tehran Hostages,
however, scholarly debate on self-contained regimes has narrowed down to the
specific question of the ‘completeness’ of a subsystem’s secondary rules. Hence, we
reserve the term ‘self-contained regimes’ to designate a particular category of
31
N. Luhmann, Die Gesellschaft der Gesellschaft (1997), at 170, 776, 779. Cf. infra for a discussion of
fragmentation from the perspective of Systemtheorie.
32
Koskenniemi, supra note 8, at 7.
33
Ibid.
34
Abi-Saab, ‘Fragmentation or Unification: Some Concluding Remarks’, 31 NYU J Int’l L Pol (1999) 919,
at 926.
35
Opinion 1/91, EEA I, 14 Dec. 1991, [1991] ECR, I-6079, at paras 39 and 40.
36
US – Standards for Reformulated and Conventional Gasoline, Report of the Appellate Body, 29 Apr. 1996,
WT/DS2/AB/R at 17; cf. also: US – Import Prohibition of Certain Shrimp and Shrimp Products, Report of the
Appellate Body, 6 Nov. 1998, WT/DS58/AB/R, at 154–157.
subsystems, namely those that embrace a full, exhaustive and definitive, set of sec-
ondary rules. Thus, the principal characteristic of a self-contained regime is its inten-
tion to totally exclude the application of the general legal consequences of wrongful
acts as codified by the ILC, in particular the application of countermeasures by an
injured state.37
2 Approaches to Self-contained Regimes by the International Law Commission
The International Law Commission’s stand with regard to the existence of so-called
self-contained regimes concerning state responsibility has varied with each special
rapporteur taking up the subject of legal consequences of internationally wrongful
acts. In a nutshell, the ILC first appeared to embrace the concept of self-contained sub-
systems (Riphagen), then became highly critical of the systematic feasibility of such
isolation from state responsibility (Arangio-Ruiz), and finally adopted the position of a
pragmatic ‘maybe’ (Crawford).
Special Rapporteur Willem Riphagen’s approach was characterized by consider-
able ambiguity.38 On the one hand, Riphagen charted the international legal system
as an order modelled on a variety of distinct subsystems, within each of which prim-
ary rules and secondary rules are closely interlinked.39 The regime of state responsi-
bility was perceived as merely part of one such subsystem. Consequently, in the
Rapporteur’s view, ‘[t]he idea that there is some kind of least common denominator
in the regime of international responsibility must be discarded’.40 On the other hand,
Riphagen presented scenarios in which ‘the subsystem itself as a whole may fail, in
which case a fallback on another subsystem may be unavoidable’.41
In the era of Special Rapporteur Gaetano Arangio-Ruiz, debate concentrated on
one, particularly contentious, aspect of self-contained regimes, namely the question
whether such a ‘so-called self contained regime affect[s], and if so in what way, the
rights of the participating States to resort to the countermeasures provided for under
general international law’.42 Focusing on the admissibility of countermeasures,
Arangio-Ruiz concluded that none of the systems envisaged as self-contained
regimes43 excluded the application of the rules of state responsibility in concreto. The
Rapporteur added that, in any event, the very concept of closed legal circuits of
37
Cf. Simma, supra note 6, at 117. We thus adopt an autonomous ‘international law’ definition, which is
not identical with Krasner’s classical definition of international regimes as ‘a set of implicit or explicit
principles, norms, rules, and decision-making procedures around which actors’ expectations converge
in a given area of international relations’. Krasner, ‘Structural Causes and Regime Consequences:
Regimes as Intervening Variables’, in S. D. Krasner (ed.), International Regimes (1983) 2.
38
For a more extensive critique of Riphagen’s theoretical approach, cf. Simma, supra note 6, at 115–117.
39
Riphagen, ‘Third Report on State Responsibility’, in ILC Yearbook (1982), Vol. II, Part One, at 24
para 16; cf. also Riphagen, ‘State Responsibility: New Theories of Obligation in Interstate Rela-
tions’, in R. St. J. Macdonald, D. M. Johnston (eds), The Structure and Process of International Law: Essays
in Legal Philosophy, Doctrine and Theory (1983) 600.
40
ILC Yearbook (1982), Vol. I, at 201, para. 8.
41
Riphagen, ‘Third Report’, supra note 39, at 30 para. 54.
42
Arangio-Ruiz, supra note 11, at 35. Cf. also ILC Yearbook (1992), Vol. I, at 76.
43
Among them the European Communities, the GATT, the International Covenant on Civil and Political
Rights, the European Convention on Human Rights, and diplomatic law.
494 EJIL 17 (2006), 483–529
responsibility rules was dubious even in abstracto.44 Arangio-Ruiz suggested that the
limitations built in the customary international law of state responsibility, most nota-
bly the proportionality principle, would suffice to take sufficient account of the peculi-
arities of subsystems. According to the Rapporteur, countermeasures ‘outside’ a
special subsystem’s secondary rules would, in principle, be disproportionate. Conse-
quently, only after all available means within the subsystem are exhausted may states
resort to proportionate countermeasures under general international law.45 Since
Arangio-Ruiz perceived the conflict between special secondary rules and the general
rules of state responsibility to be anchored in (and resolved by) the proportionality
principle, he proposed the deletion of the lex specialis clause in the Draft Articles.
Rather than resolving the conceptual clash between the ILC’s previous rapporteurs,
Special Rapporteur James Crawford decided to refer the issue of self-containment of
subsystems to another topic that the Commission intended to take up, namely the
fragmentation of international law. The commentaries adopted in 2001 remain tacit
as to whether a closed responsibility regime outside the customary law of state
responsibility is conceptually feasible and, if so, which subsystems may qualify as
thus closed. The Commission avoided express recognition of self-contained regimes
by diplomatically speaking of ‘“strong” forms of lex specialis, including what are often
referred to as self-contained regimes’.46 Despite such hesitation, the ILC’s final prod-
uct can still be recognized as a conceptual approximation to the view held by Special
Rapporteur Arangio-Ruiz. Admittedly, the Commission preferred the more conven-
tional legislative technique of an express conflict clause over Arangio-Ruiz’s
approach of re-interpreting proportionality. In substance, however, the very inclu-
sion of a lex specialis clause implies a certain concept of the international legal order.
The Commission moved away from Riphagen’s idea of several competing regimes of
equal rank and towards Arangio-Ruiz’s concept of a canon of general law that is
applicable automatically unless states have specifically contracted out by virtue of a
special legal regime.
In its deliberation of the topic ‘The Fragmentation of International Law’, the ILC
appears to follow the course thus charted. In the Working Paper elaborated in 2003
by Koskenniemi, Chairman of the Study Group, special subsystems are described as
firmly embedded within an omnipresent general international law.47
C Concepts of the International Legal Order

Far from being merely caprices by a handful of theory-enthused academics, the dis-
tinct theoretical positions taken by the various ILC rapporteurs entail different
44
Arangio-Ruiz, supra note 11, at 40. Cf. also ILC Yearbook (1992), Vol. I, at 77.
45
Ibid., at 41, para. 116: ‘“External” unilateral measures should thus be resorted to only in extreme cases. . . .
In other words, the principle of proportionality will have to be applied in a very special way – and very
strictly – whenever the measures resorted to consist in the suspension or termination of obligations
deriving from an allegedly self-contained regime.’
46
ILC, supra note 9, at para. 5.
47
Koskenniemi, supra note 8, at 7.
answers to the question whether self-contained regimes in the field of state responsibil-
ity are conceivable. The principal characteristic of a self-contained regime is its inten-
tion totally to exclude the application of the general international law on state
responsibility, in particular resort to countermeasures by an injured state. The question
that immediately follows in practice is whether such a complete exclusion of all second-
ary rules of general international law is in fact intended by the regime in question.
Whether this question is answered in the affirmative, in turn, may depend on
whether international law is conceived as a unified legal order or as the sum total of
loosely interrelated subsystems. Scholars adhering to the first – universalistic – con-
cept choose as their starting point the perspective of general international law. Dero-
gation from the general international law on state responsibility is only accepted to
the extent that the states parties have clearly stated such an intention. Proponents of
this position tend to deny the existence of self-contained regimes. By contrast, schol-
ars adhering to the second – particularistic – concept of international law draw nor-
mative conclusions from a somewhat sociological analysis of certain particular
regime characteristics. Their analysis is mostly conducted from a perspective within a
particular regime. In their view, special regimes tend to appear as self-contained.
Hence, the claim of self-containment is intrinsically linked to a particular outlook on
the international legal order.
1 Leges speciales in a Unified Legal Order
ILC Article 55 is based on a perception of international law shared by many general
international lawyers: international law is a unified and, to a certain extent, hierar-
chical legal order. ‘L’unité de l’ordre juridique’, as Pierre-Marie Dupuy named his Gen-
eral Course at the Hague Academy,48 is a sociological fact or, at least, a normative
postulate.49 Unity, according to Mireille Delmas-Marty, is an inherent characteristic
of law, since ‘[l]aw does not like multiplicity; it represents order, unified through hier-
archy and symbolized by Kelsen’s pyramid of norms, built for eternity – not by clouds,
even if they are organized’.50
In domestic legal systems, the unity of the legal order is generally considered a conse-
quence of the law’s systemic character.51 Yet, strong notions of systemic unity are diffi-
cult to transpose to the international plane. Is international law systemic because of its
emerging normative hierarchies?52 Is it by virtue of a (more and more) complete set of
48
P.-M. Dupuy, ‘L’unité de l’ordre juridique international’, 297 RdC (2002) 1.
49
Similarly, K. Zemanek, ‘The Legal Foundations of the International System’, 266 RdC (1997) 62: ‘The
‘unity’ of international law in the sense of its homogeneity is therefore a construct.’ Cf. also the discus-
sion in Section 2D of this article.
50
M. Delmas-Marty, Global Law: A Triple Challenge (2003), at 74. For better readability, we have chosen to
quote the English translation. Yet, the connotations of the original text are somewhat different: ‘[l]e droit
a l’horreur du multiple. Sa vocation c’est l’ordre unifié et hierarchisé, unifié parce que hierarchisé. Et l’image qui
vient à l’esprit des juristes, c’est la pyramide des normes, construite pour l’éternité, plutôt que celle des nuages,
fussent-ils ordonnées.’ M. Delmas-Marty, Trois défis pour un droit mondial (1998), at 104.
51
For a general discussion of positivist approaches to the notion of a legal system, cf. J. Raz, The Concept of a
Legal System: An Introduction to the Theory of Legal System (2nd ed., 1980).
52
Delmas-Marty, supra note 50.
496 EJIL 17 (2006), 483–529
secondary rules?53 Is the ‘glue’ of overriding global values expressed in the law the deci-
sive factor?54 Is the systemic character of international law dependent on its efficacy?55
We find it difficult to confidently respond in the affirmative to any of these questions. Nei-
ther formal hierarchies nor a normative ‘overlapping consensus’ of values nor centralized
enforcement mechanisms appear to be sufficiently developed to guarantee legal unity.
While legal reasoning necessarily involves operationalizing hierarchical catego-
ries,56 such hierarchies in international law rarely take the shape of relationships of
inferiority and superiority between particular norms – relationships that could be
called formal hierarchies. As Jean Combacau noted,
Rien de tel en droit international, où le principe de l’égalité souveraine des Etats exclut toute forme de
différenciation hiérarchique des normes puisqu’elles reposent toutes en dernier ressort sur un acte de
volonté, au moins supposée, d’Etats pairs entre eux (c’est l’égalité) et ne connaissant pas de supérieur
commun (c’est la souveraineté).57
In light of the emergence of the concept of jus cogens, Combacau’s insistence on the
equal status of all international norms may already have looked somewhat dated in
1986, when Archives de philosophie du droit published a special edition on the notion of
the système. Certainly today, as jus cogens has emerged into a relatively uncontrover-
sial concept (rumour has it that even France may soon give its blessing to jus cogens
by acceding to the Vienna Convention58), a certain degree of hierarchization of inter-
national norms cannot be denied. As Andreas Paulus has convincingly argued, how-
ever, ‘the indeterminacy of the content and the precise legal effect of jus cogens has
largely condemned it to practical irrelevance’.59 The Al-Adsani judgment by the Euro-
pean Court of Human Rights60 (decided on the relevant point by nine votes to eight) is
a case in point: while the jus cogens character of the prohibition against torture is, in
principle, uncontroversial, it is not at all clear under which conditions a state is
barred from pleading immunity; and under which conditions domestic courts must
consequently entertain civil claims by torture victims to meet states’ obligations
under Article 6 of the European Convention. Hence, what remains open is precisely
53
According to H. L. A. Hart, the existence of such secondary rules, ‘rules of recognition’ and ‘rules of change
and adjudication’, is a necessary precondition for the existence of a legal system. H. L. A. Hart, Concept of
Law (1961), at 113. While reference to Hart’s seminal book is common, one should nonetheless keep in
mind that international legal scholarship has adopted a somewhat different definition of secondary rules.
54
Paulus, ‘Commentary to Andreas Fischer-Lescano & Gunther Teubner: The Legitimacy of International
Law and the Role of the State’, 25 Mich J Int’l L (2004) 1047, at 1050.
55
H. Kelsen, The Pure Theory of Law (1970), at 33 and 320: ‘a coercive order’.
56
Cf. Koskenniemi, supra note 22. Delmas-Marty, supra note 50.
57
Combacau, ‘Le droit international: bric-à-brac ou système?’, 31 Archives de philosophie du droit (1986)
85, at 88. ‘This is not the case in international law, in which the principle of the sovereign equality of
States excludes all forms of hierarchical differentiation of norms, as they are all ultimately founded upon
an act of will, at least supposed, of States that are all on a par with each other (equality) and that recog-
nize no common superior (sovereignty).’
58
Cf. Paulus, ‘Jus cogens Between Hegemony and Fragmentation: An Attempt at a Re-appraisal’, 74 Nordic
J Int’l L (2005) 297, at 298.
59
Ibid., at 330.
60
Al-Adsani v. UK, Application No. 35763/97 [2001] ECHR 761 (21 November 2001).
the question of how a jus cogens rule affects the international legal order in general.
The emergence of jus cogens can thus hardly be regarded as the key factor that would
turn the ‘bric-à-brac’61 of international law into a systemic whole. Depicting jus cogens
as a fundamental, quasi-constitutional canon of values, to which all other interna-
tional law would refer back, means asking for a promise that the concept cannot (yet)
hold. In Paulus’ words, jus cogens ‘substitutes rather than complements a truly public,
quasi-constitutional order in international affairs’.62
This brings us to another, related point: If formal hierarchies in international law
are too weak to constitute or secure the systemic cohesiveness of international law,
can universal values do the job? For a German-speaking lawyer, the classical treatise
on the idea of unity of the legal order is Engisch’s Einheit der Rechtsordnung, an influ-
ential monograph on legal theory published in the 1930s. With remarkable, almost
troubling enthusiasm for the ‘legal science’, Engisch highlights the central role that
the notion of a unitary legal order plays:
Dogmatics is the unity of the legal order! Indeed: If there is one thing that may give the science
of the law the impetus to rise above a mere non-‘dogmatic’ grasping and interpreting of cer-
tain individual provisions of the law, as laymen are equally capable of doing, if there is any-
thing that may be able to secure legal dogmatics the rank of a veritable science – we may even
say more demandingly: the rank of the science of value judgments par excellence, it is the over-
riding significance of the principle of the unity of the legal order . . .63
At the same time, Engisch admits that the empirical reality of the legal order may
look somewhat less unitary than its intellectual construction by jurists: ‘One may even
go as far as to say that the unity of the legal order is nothing but a regulative idea or the
product of particular juridical methods’,64 Engisch notes. The question remains as to
whether we can construe this ‘product of particular juridical methods’ on the interna-
tional terrain. According to a theory developed by Claus-Wilhelm Canaris:
the attempt to conceptualize the system of a particular legal order as a formal one or as an
axiomatic-deductive one is doomed to failure from the very beginning. For the inner unity of
meaning of the law, which is supposed to be grasped in the system, is – in accordance with its deriva-
tion from the notion of justice – not of a logical kind but rather of a value-driven, i.e. axiological kind.65
61
J. Combacau, note 57.
62
A. Paulus, note 58, at 331.
63
K. Engisch, note 19, at 1: ‘Die Dogmatik ist die Einheit der Rechtsordnung! In der Tat: Wenn irgend etwas
der dogmatischen Wissenschaft vom Recht Antrieb zur Erhebung über bloßes nicht’ dogmatisches’ auch dem
Laien mögliches Aufgreifen und Deuten einzelner Rechtsvorschriften zu geben vermag, wenn irgend etwas der
Rechtsdogmatik den von ihr begehrten Rang einer echten Wissenschaft – wir dürfen sogar noch anspruchs-
voller sagen: den Rang der Wertungswissenschaft par excellence – zu sichern geeignet ist, so ist es die
beherrschende Bedeutung des Prinzips der Einheit der Rechtsordnung. ..’
64
Ibid., at 3. ‘Man mag sogar davon sprechen, dass die Einheit der Rechtsordnung nichts weiter als eine regulative
Idee oder das Produkt ganz bestimmter juristischer Methoden sei.’
65
C.-W. Canaris, Systemdenken und Systembegriff in der Jurisprudenz (1969), at 21: ‘In der Tat ist der Versuch,
das System einer bestimmten Rechtsordnung als formal-logisches oder axiomatisch-deduktives zu konzipieren,
von vornherein zum Scheitern verurteilt. Denn die innere Sinneinheit des Rechts, die es im System zu erfassen
gilt, ist entsprechend ihrer Ableitung aus dem Gerechtigkeitsgedanken nicht logischer, sondern wer-
tungsmäßiger, also axiologischer Art.’
498 EJIL 17 (2006), 483–529
The notion of justice according to Canaris notably includes ‘the requirement of the
inner consistency of value judgments, which follows from the principle of equality’.66
Proceeding from the assumption of a ‘value-driven unity’, the central question for
Canaris is of course which values – in the sense of Wertungen rather than Werte, per-
haps more aptly translated by ‘value judgments’ – constitute the inner unity of the
legal order. In Canaris’ words, the ‘essential question remains to be answered: the
question of the constituting elements in which the inner unity and consistency of the
legal order becomes visible’.67 Canaris aptly points out that one cannot reduce this
question to mere conflict resolution – Konfliktentscheidung. Rather, the jurist must
penetrate to the fundamental value judgments, the overarching ratio juris of a given
legal order.68 It is with these fundamental value judgments in mind that a lawyer
must approach a concrete legal problem. Hence, to quote another theorist of the
1930s, Philipp Heck, ‘while [the judge] must decide the individual case before him, he
does so by applying the entire legal order’.69
It is precisely at this point that the international lawyer tends to lose her footing.
First of all, it is doubtful whether a theory that is based on the constitutional principle
of equality can work on the international level where such a general principle is
absent. A second problem concerns Canaris’ recourse to the ‘fundamental value judg-
ments’ of the legal order. To be sure, there are undeniable trends towards a ‘constitu-
tionalization’ of international law, denoting the development of a vertically
integrated, institutionalized, community-oriented and value-laden global order.70
However, despite the emergence of certain global values,71 a fundamental structural
difference remains between the domestic context and the international plane: besides
global values, international law continues to build upon the sovereignty of states as a
major constitutional principle.72 And at the heart of this sovereignty-based interna-
tional order, even today, a multiplicity of values and political convictions compete for
universal recognition. What about the role of jus cogens? Can the concept at least
serve (if not as a tool of formal hierarchization) as a kind of systemic value glue for the
international legal order? The canon of jus cogens rules can certainly be said to
embody ‘fundamental value judgments’ (as posited by Canaris). For the time being,
however, jus cogens rules are few and their contents opaque. To most concrete day-to-
day questions of legal fragmentation, jus cogens simply provides no guidance. As a
66
Ibid., at 16: ‘. . . deshalb bildet das aus dem Gleichheitssatz folgende Gebot wertungsmäßiger Folgerichtigkeit den
ersten entscheidenden Ansatz für die Verwendung des Systemgedankens in der Jurisprudenz . . . ’
67
Ibid., at 46: ‘ist indessen [. . . die] wesentliche Frage noch nicht beantwortet: die nach den tragenden
Elementen, in denen die innere Einheit und Folgerichtigkeit der Rechtsordnung sichtbar wird.’
68
Ibid.
69
P. Heck, Begriffsbildung und Interessenjurisprudenz (1932), at 107.
70
Cf. Simma, ‘Fragmentation in a Positive Light’, 25 Mich JIL (2004) 845.
71
On this cf. Bederman, ‘Globalism and International Law: Values and Choices’, Halle Institute Occasional
Paper (2001).
72
On the undiminished significance of the notion of sovereignty, even and particularly in a globalizing
world, cf. Nolte, ‘Sovereignty as Responsibility’, 99 ASIL Proceedings (2005) 389; Nolte, ‘Zum Wandel
des Souveränitätsbegriffs’, Frankfurter Allgemeine Zeitung, 6 Apr. 2004, at 8; Kingsbury, ‘Sovereignty
and Inequality’, 9 EJIL (1998) 599.
consequence, the international lawyer is sometimes bound to accept systemic incon-

sistencies or incongruence73 between various norms of international law. Sometimes
she simply cannot go beyond what Canaris terms Konfliktentscheidung.
Rather than analogizing strong conceptions of legal systems developed in a
domestic context, international jurists must settle for a more minimal conception.
The idea of system is a central element of legal discourse – perhaps part of what David
Kennedy has called the ‘styles’ of legal argument.74 International law certainly pos-
sesses the basic characteristics to partake in a specifically legal discourse, in the sense
that ‘the various decisions, rules and principles of which the law consists are not ran-
domly related to each other’.75 Leading contemporary scholars in international law,
from Christian Dominicé76 to Jean Combacau,77 from James Crawford78 to Karl
Zemanek79 and Gerhard Hafner,80 have thus affirmed the systemic character of inter-
national law on a rather pragmatic tone.
Yet, can we justifiedly call this minimal system, resembling in many respects a bric-à-
brac rather than an organized whole, a legal order, an ordre juridique in which all
73
Canaris differentiates between Systembrüche (frictions in the system) and systemfremde Normen (norms
alien to the system), supra note 65, at 112 and 131.
74
Similarly, MacCormick, ‘Beyond the Sovereign State’, 56 Modern LR (1993) 1, at 10. For Kennedy’s
characterization of elements of styles of legal argument, see Kennedy, ‘Theses about International Law
Discourse’, 23 GYIL (1980) 354.
75
Koskenniemi, supra note 8, at para. 27. See the similar definition of a system in a weak sense in M. van
Hoecke, Law as Communication (2002), at 109–110: ‘System in the weak sense only requires some struc-
ture, some relationship among the elements of the system, which makes it possible to identify it as some-
thing that exists on its own and can be distinguished from a number of unordered elements and from
other systems.’ (emphasis in original).
76
According to Dominicé, international law is a system, because actors (states as well as non-state actors)
behave accordingly: ‘This conclusion is buttressed by the finding that there is a sort of collective opinio
juris, a conviction that international law exists and that states could not do without it. There is no need
to seek a theoretical foundation to justify this assertion, which results from a mere observation of reality
and is expressed by the maxim ubi societas ibi ius.’ Dominicé, ‘Methodology of International Law’, in
R. Bernhardt et al., Encyclopedia of Public International Law, Vol. III (1997).
77
Combacau’s answer is equally grounded on the empirical reality of international law: ‘Tout se passe
comme si les Etats admettaient que ce qui, structurellement, se présente comme une norme, c’est-à-dire ce qui
peut se formuler verbalement en termes de devoir-être, était en effet obligatoire, par application d’une exigence –
je ne dis pas ‘d’une norme’ – supérieure à celle qu’énonce la norme en question.’ Combacau, supra note 57, at
90. ‘Everything happens as if States assume that that which, structurally, presents itself as a norm (that is,
capable of verbal formulation in terms of an ought) is in effect obligatory, through the application of a
requirement – I won’t say “of a norm” – superior to that which the norm in question expresses.’
78
James Crawford has examined what makes international law an open system. While he proves his case
as far as the openness of international law is concerned, its systemic character is more assumed than
explained. According to Crawford, it suffices to regard ‘international law in the modern period as provid-
ing a formal structure, based on sovereignty, negotiation and consensus, on which we are building in a
variety of ways.’ J. Crawford, International Law as an Open System (2002) 17, at 28.
79
Zemanek, supra note 49, at 61–65.
80
Hafner notes that international law lacks comprehensive organization. Such shortcomings, however, do
not preclude that international law can at least be described as an ‘unorganized system’. Hafner, ‘Risks
Ensuing from the Fragmentation of International Law’, in International Law Commission, Report on the
Work of its Fifty-third Session, Annex, Official Records of the General Assembly, Fifty-Second Session, Sup-
plement No. 10 (A/55/10) 321.
500 EJIL 17 (2006), 483–529
norms are related to each other in a deliberate and meaningful way? Anthony Carty
has sceptically noted that ‘there is no legal system which defines comprehensively the
rights and duties of States towards one another’.81 And he has added elsewhere:
My own post-modern approach to the sources of international standards of behaviour does
not merely leave open the possibility that there is no overarching system whose signification
could be unravelled by a sufficiently subtle and tactful hermeneutic. It insists that such is the
primary problematic with which international lawyers have to work. The LAW is a tapestry of
lacunae with occasional densities of normativity.82
While international law certainly is systemic, it does not necessarily constitute a

comprehensive and organized legal order. Rather, the structure of the international
system still strongly builds on informal hierarchies. Perhaps the most important
among these informal hierarchies (though not strictly speaking a relationship of
higher and lower rank) is the concept of general international law, in which all
‘special’ law is embedded.83 It is only when the role of general international law as a
cornerstone of the international legal system is appreciated that the significance of
the lex specialis principle becomes apparent. The maxim validly gives effect to an
informal hierarchy inherent in the organization of international law. Special
regimes of international law can be conceptually related to general international
law, including the regime of state responsibility, by invocation of the maxim lex
specialis derogat legi generali: the rules on state responsibility, as codified by the ILC,
are part of ‘general’ international law; they can be derogated from by all other sec-
ondary norms, which are by definition ‘special’.
Yet, as a practical matter, the application of the lex specialis maxim meets with diffi-
culties. How far does the specialness of the special treaty extend? To what extent did
the state parties intend to exclude the application of the rules on state responsibility?
Formally, the answer can be found in two simple steps. First, the rules of the special
regime, ordinarily codified in a treaty instrument, must be interpreted according to
Article 31 of the Vienna Convention on the Law of Treaties in order to establish
81
A. Carty, The Decay of International Law? (1986), at 1 (emphasis added).
82
Carty, ‘Critical International Law: Recent Trends in the Theory of International Law’, 2 EJIL (1991) 1,
at 25.
83
There is surprisingly little in-depth discussion as to how the notion of general international law came
about. For a classical discussion of the concept, cf. P. Reuter, ‘Principes de droit international public’,
103 RdC (1961) 469. Reuter rightly distinguishes between two functions of the notion of ‘généralité: on
the one hand, the ‘fonction du nombre abstrait d’Etats liés par une norme’, on the other hand, ‘le degré
d’abstraction d’une norme’ (ibid.). ‘A function of the abstract number of States bound by a norm’ . . . ‘the
degree of abstraction of a norm’ Typically, textbook discussions of the concept of general international
law fail to take account of the multiple aspects of generality. Reuter concludes that ‘pour la mise en ordre
des normes internationales la généralité ne donne que des indications diverses qui doivent être com-
posées avec d’autres considérations.’ Ibid. ‘the notion of generality only provides us with differing indica-
tions as to how international norms should be ordered, which must be combined with other
considerations.’ It is hardly disputed that the rules on state responsibility form part of the body of rules
denoted as general international law. Since these rules can be very specific in content, their ‘generality’
appears to be a function of their universal application and their fundamental character, to take up
Reuter’s distinction.
whether the states parties intended the regime’s secondary rules to be exhaustive and
complete. Second, resort must be had to general international law to verify whether
the latter permits such derogation.84
Practically speaking, however, reference to treaty interpretation is hardly satisfac-
tory. None of the treaty regimes presented as candidates for self-contained regimes
contains an explicit provision regarding the applicability vel non of the rules of state
responsibility. For example, the Treaty establishing the European Communities is
silent as to whether general international law is excluded when no procedures before
the European Court of Justice are available. The WTO’s Dispute Settlement Under-
standing (DSU) contains provisions similar to the rules on cessation of breach and on
countermeasures in the ILC’s regime on state responsibility. The DSU fails to state,
however, whether the regime on state responsibility should apply for all remaining
questions. Human rights treaties spell out particular enforcement mechanisms (in
particular, reporting procedures and individual complaint procedures). Yet, the trea-
ties leave open the question whether such mechanisms affect states’ rights to invoke
state responsibility according to general international law.
Textual arguments in favour of either the open or closed nature of the three
regimes just mentioned are as numerous as they are inconclusive. To cope with this
problem, universalists operate with a presumption in favour of the applicability of the
‘general’ state responsibility regime; a presumption that can be refuted only by con-
tracting out of general international law with sufficient clarity. With regard to the
local remedies rule, an all-time classic of general international law, the International
Court of Justice pointed out in the ELSI case:
The Chamber has no doubt that the parties to a treaty can therein either agree that the local
remedies rule shall not apply to claims based on alleged breaches of that treaty; or confirm that
it shall apply. Yet the Chamber finds itself unable to accept that an important principle of cus-
tomary international law should be held to have been tacitly dispensed with, in the absence of
any words making clear an intention to do so.85
Similarly, Special Rapporteur James Crawford spoke of a ‘presumption against the cre-
ation of wholly self-contained regimes in the field of reparation’.86 According to such a
presumption, the rules on state responsibility apply unless the treaty clearly provides
otherwise. In Joost Pauwelyn’s words, to quote another ‘universalist’, ‘[i]t is for the party
claiming that a treaty has “contracted out” of general international law to prove it’.87
84
Special Rapporteur Arangio-Ruiz had maintained that no derogation was allowed ‘from those essential
rules and principles on the consequences of internationally wrongful acts that are inherent to interna-
tional relations and international law’ (Arangio-Ruiz, supra note 11, at 40, para. 112.). However, state
practice does not unequivocally indicate the recognition of a jus cogens character of the rules on state
responsibility. In most cases, the International Court’s statement in North Sea Continental Shelf appears to
also hold true with regard to secondary norms of general international law: ‘[I]t is well understood that,
in practice, rules of international law can, by agreement, be derogated from in particular cases or as
between particular parties.’ (North Sea Continental Shelf cases, ICJ Reports (1969), at 42 para. 72).
85
Elettronica Sicula (ELSI), ICJ Reports (1989), at 42, para. 50.
86
J. Crawford, Third Report on State Responsibility, UN Doc. A/CN.4/507 (2000) para. 147.
87
Pauwelyn, supra note 8, at 213.
502 EJIL 17 (2006), 483–529
2 Self-contained Regimes in a Fragmented Legal Order

Thus far the story of fragmentation has been based on the assumption of a unified
legal order. However, such an approach had both friends and foes in the Interna-
tional Law Commission. In stark contrast to what turned out to be the majority view,
Riphagen suggested that international law should be seen as the aggregate of differ-
ent regimes, co-existing without any pre-defined hierarchy:
A theoretical answer might be that a system was an ordered set of conduct rules, procedural
rules and status provisions, which formed a closed legal circuit for a particular field of factual
relationships. A subsystem, then, was the same as a system, but not closed in as much as it had
an interrelationship with other subsystems.88
Focusing on the variety of regimes involved in international law (rather than on the
international legal system), Riphagen advocated a ‘particularistic’ vision of the inter-
national legal order. More recently, the International Criminal Tribunal for the
Former Yugoslavia (ICTY), in its Tadic judgment, expressed a similar conception of
international law when it held that, ‘[i]n International Law, every tribunal is a self-
contained system (unless otherwise provided)’.89
Riphagen’s theory draws on Stephen Krasner’s classical definition in international
relations theory of a ‘regime’ as ‘a set of implicit or explicit principles, norms, rules,
and decision-making procedures around which actors’ expectations converge in a
given area of international relations’.90 While the modern interest in ‘regimes’91 is
probably inspired by the work of theorists such as Krasner, the idea of a polycentric
international law can be traced back (at least) to Georges Scelle. If we allow Antonio
Cassese to lend his succinct words to Scelle, ‘[t]he world community swarms with
myriad legal orders (in today’s parlance we would call them “sub-systems”)’.92
It is only a small step from Scelle’s socio-legal analysis to contemporary sociolog-
ical accounts of international law. Andreas Fischer-Lescano and Gunther Teubner
have fruitfully deployed a version of Luhmann’s Systemtheorie to shed light on the
fragmentation of international law.93 According to these two authors, the debate
on the fragmentation of international law suffers from two reductionisms: if frag-
mentation is equated with a problem of collision of norms, we are faced with a legal
reductionism. If, by contrast, the focus is on the political foundations of norm colli-
sion, a political reductionism is inevitable.94 Claiming to avoid the limitations
88
ILC Yearbook, Vol. I (1982), at 202, para. 16.
89
35 ILM (1996), at 32, para. 11.
90
Krasner, supra note 37.
91
Cf. for example, Winham, ‘International Regime Conflict in Trade and Environment: The Biosafety Pro-
tocol and the WTO’, 2 WTR (2003) 131.
92
Cassese, ‘Remarks on Scelle’s Theory of ‘Role Splitting’ (dédoublement fonctionnel) in International
Law’, 2 EJIL (1990) 210. Cf. G. Scelle, Manuel de droit international public (1948) 19–21.
93
Teubner and Fischer-Lescano, ‘Regime-Collisions: The Vain Search for Legal Unity in the Fragmentation
of Global Law’, 25 Mich J Int’l L (2004) 999. The authors’ analysis touches on numerous interesting
issues. In the following, however, we will limit our summary of the article to what appears to be of par-
ticular relevance for our case study on self-contained regimes.
94
Ibid., at 1003.
inherent in conventional approaches, Fischer-Lescano and Teubner describe the pro-

liferation of international legal regimes as the epi-phenomenon of the multidimen-
sional fragmentation of world society95 – a global space which has differentiated itself
into myriad autonomous systems. ‘Legal fragmentation is merely an ephemeral
reflection of a more fundamental, multi-dimensional fragmentation of global society
itself.’96
The sticking point of their analysis is the observation that all law, whatever its
conventional label (domestic or international, private or public), is closely coupled
with a sectoral system dedicated to maximizing a particular rationality. Within
the legal system, geographical borders have been replaced by functional, systemic
delimitations: ‘the national differentiation of law is now overlain by structural
95
Ibid., at 1017.
96
Ibid., at 1004. This proposition, while plausible at first glance, is not beyond contestation. The somewhat
radical title chosen by the authors for their article in the Michigan Journal, ‘The Vain Search for Legal
Unity’, risks misguiding the reader. For the authors do concede that ‘[u]nity of the legal system has been
achieved at global level’ (ibid., at 1008). In reality, they do not dispute the notion of some kind of ‘legal
unity’, but rather the possibility of a quasi-constitutional global law, integrating in an orderly manner
the most diverse values and rationalities. It is the ‘multitude of internal contradictions’ (ibid.) within
unity that the authors highlight in their piece. A fundamental point of critique of Fischer-Lescano and
Teubner’s approach concerns their apparent equation of differentiation of issue regimes with the fragmen-
tation of international law. Despite the complexity of the authors’ theoretical framework, their analysis
does not pay due account to another important systemic distinction drawn by systems theory: the dis-
tinction between the legal system and its environment (constituted by other systems such as politics,
religion or economics). It is precisely systems theorists who have emphasized that the legal system
achieves a particular inner cohesiveness by virtue of its binary operations, which autopoietically refer
back to the system’s previous operations. Hence, it does not suffice, for example, to point out that a
‘rationality conflict’ between trade concerns and biosafety can be described as a collision of the issue sys-
tems of commerce and health (each of which comes with a particular institutional and legal arrange-
ment). At the same time, the legal system (comprising lawyers within both institutions for trade
facilitation and health protection) remains operatively closed vis-à-vis its political and economic envir-
onment. In an attempt to maximize its own rationality, the legal system attempts to resolve conflicts
within its own operation, at the expense of, say, the political system. Trade lawyers and health law-
yers simultaneously act as elements of two systems: as part of the legal system (which is eager to dis-
tinguish itself from its non-law environment); and as part of the subject-matter systems ‘trade’ or
‘health’ (which aim at maximizing their respective rationality at the expense of other rationalities
prevalent in the environment). In our view, the crux of the fragmentation of international law lies pre-
cisely in this inner tension, the ‘role strain’ that international lawyers are exposed to. Mark Van
Hoecke, who himself strongly draws on systems theory, consequently arrives at a different conclusion:
The more radically a society’s issue systems diversify, the stronger are in turn the forces towards
autonomy of and unity within the legal system. According to Van Hoecke, ‘[a] stronger ‘juridification’
of our societies has led to an increased autonomy of law, but at the price of an increased interdepend-
ence of legal systems.’ (M. Van Hoecke, ‘Legal Orders Between Autonomy and Intertwinement’, in
K.-H. Ladeur (ed.), Public Governance in the Age of Globalization (2004) 177, at 193. Cf. also ibid. at 186:
‘Here, one could also state that the decrease of autonomy vis-à-vis other legal systems increases the
autonomy towards the non-legal systems.’ Given the increase in ‘intersystemic interdepence of legal
systems’ (ibid., at 194), a stronger role of legal doctrine is called for, which achieves a ‘rational struc-
turing of all those data into one coherent whole.’ (Id.) In short: ‘Increasing fragmentation of law is
thus coupled with an enhanced intertwinement of the different areas of law.’ (Id. at 187). Cf. for other
points of critique – in particular concerning a lack of legitimacy of decentralized law-making by non-
state actors – Paulus, supra note 54.
504 EJIL 17 (2006), 483–529
fragmentation’.97 If all international law necessarily follows the logic of a particular

issue area,98 it comes as no surprise that the establishment of ‘genuinely self-
contained regimes’99 appears conceivable. Fischer-Lescano and Teubner explain:
Since such [self-contained] regimes are structurally coupled with the independent logic of the
social sectors, they inevitably reproduce, albeit in a different form, the structural conflicts
existing between the various functional systems within the law.100
Many such genuinely self-contained regimes, namely those that have a strong
institutional structure,101 evolve into auto-constitutional regimes.102 In line with the
logic of Luhmann’s Systemtheorie, such regimes remain, in principle, autonomous. At
best, a selective networking of various regimes can be achieved.103 As a result, ‘a
strange legal esperanto of regimes’ will emerge, ‘within which national, international
and trans-national legal acts clamor for attention’.104
If we look at the various approaches discussed in this section under the heading of
‘particularistic approaches’, the following common baseline can be established: even
proponents of a particularistic theory concede that regimes ‘do not live by them-
selves, each in its own area, but intersect and overlap with each other’.105 To borrow
the language of Luhmann’s Systemtheorie, every system is linked with its environ-
ment through ‘structural coupling’.106 Yet, shifting attention from the systemic
whole to the system’s particular components, from the universe to the planets, does
not remain without consequence. Not surprisingly, life on the planet becomes more
interesting than the fate of the universe. The particularistic approach entails empha-
sizing a regime’s political purposes (rather than the politics of the international com-
munity); the tailor-made character of a regime’s legal rules (rather than the claim to
universal justice of general international law); and the legal culture corresponding to
the politics of the system (rather than the longstanding expertise of international law
generalists).
97
Ibid., at 1008. For a balanced discussion of the move from ‘territoriality’ to ‘functionality’, cf. Paulus,
‘From Territoriality to Functionality? Towards a Legal Methodology of Globalization’, in I. F. Dekker and
W. G. Werner (eds), Governance and International Legal Theory (2004) 59.
98
For this term, cf. Alvarez and Leebron, ‘Linkages’, 96 AJIL (2002) 5, at 6.
99
Ibid., at 1013. It should be noted, however, that Fischer-Lescano and Teubner adopt a rather broad
and somewhat vague definition of self-contained regimes (borrowed from Koskenniemi, supra note 8,
Addendum para. 105.
100
Fischer-Lescano and Teubner, supra note 93, at 1013.
101
The authors refer to other factors, quoting a definition proposed by Walker, ‘The EU and the WTO: Con-
stitutionalism in a New Key’, in G. de Búrca and J. Scott (eds), The EU and the WTO: Legal and Constitu-
tional Issues (2001) 31, at 33.
102
Fischer-Lescano and Teubner, supra note 93, at 1015. Such regimes are characterized by the fact that
they establish a structural coupling between their primary and secondary rule-making with the creation
of substantive social norms in a specific societal sector, ibid., at 1016.
103
Ibid., at 1018 et seq. For a similar conclusion, cf. K.-H. Ladeur, Postmoderne Rechtstheorie: Selbstreferenz –
Selbstorganisation – Prozeduralisierung (1992) 159 et seq. (particularly at 161).
104
Fischer-Lescano and Teubner, supra note 93, at 1022.
105
Ibid.
106
Luhmann, supra note 31, at 776 and 779.
Hence, there is a certain scepticism towards any attempt to ‘smuggle’ alien ele-
ments into the regime. Making use of norms outside the regime is more of an ‘emer-
gency operation’ than a desirable practice. Tribunals established under particular
regimes thus tend to apply a presumption in favour of complete and exhaustive regu-
lation in the respective regime.107 Most notably, such a presumption underlies the
rulings of the European Court of Justice. Since the Community Treaty has ‘created its
own legal system’,108 lacunae are filled by analogies within the system or by recourse
to general principles inherent in the Community legal order instead of falling back on
general international law. The Francovich principles may be cited as one example
where resolving shortcomings of a regime are resolved within its own operations and
in accordance with its own logic.109
As a consequence of a presumption in favour of completeness, the threshold for resort-
ing to rules outside the regime is much higher. Applied to our case study on the relation-
ship between special regimes and general international law, any resort to the rules on
state responsibility requires special justification. State responsibility is not a general fall-
back option, which automatically fills the gaps of the regime. Rather, the framework on
state responsibility is an aliud only to apply in exceptional cases. Once a presumption in
favour of a regime’s completeness is accepted, tribunals established under the regime
may easily conclude (in accordance with the maxim expressio unius est exclusio alterius)
that resort may not be made to remedies other than those specified. Hence, from the par-
ticularistic perspective, regimes are likely to appear as self-contained.
D Of Planets and the Universe: The Level-of-Analysis Problem

Who then is right, the universalists or the particularists? Treaty interpretation in
accordance with the Vienna Convention apparently fails to produce a definite res-
olution of the problem. It seems that the wording of the special treaties (WTO, EC,
human rights . . .) can simply be read both ways: as permitting or precluding
application of the rules on state responsibility. In such a situation, lawyers have
attempted to avoid indeterminacy by recourse to presumptions. A presumption in
favour of general international law is widely shared among public international
lawyers. The contrary presumption prevails among Community lawyers and many
WTO specialists. Both presumptions reflect a certain perspective on the interna-
tional legal order (universalistic v. particularistic view).
Both presumptions are difficult to solidly ground in theory. As Sandifer observes,
‘[b]y its very nature the law of presumptions belongs primarily to the realm of muni-
cipal law, rather than to international law. It is dependent upon a superior authority
with power to define the presumptions and the inferences to be drawn from them. . . ’ .110
107
Cf. in the context of WTO law, McRae, ‘The WTO in International Law: Tradition Continued or New
Frontier?’, JIEL (2000) 27. Similarly, Dunoff, ‘The WTO in Transition: Of Constituents, Competence and
Coherence’, 33 Geo Wash Int’l L Rev (2001) 979, at 1009.
108
Costa v. ENEL, 6/64, [1964] ECR, 585 at 593.
109
Francovich and Bonifaci v. Italy, joined cases C-6/90 and C-9/90, [1991] ECR, I-5357.
110
D. V. Sandifer, Evidence Before International Tribunals (1975), at 141 et seq.
506 EJIL 17 (2006), 483–529
The international legal system, largely dependent on ‘horizontal’ processes of law-

making, lacks such authority. Absent a reliable presumption in favour of the univer-
sal or the particular, the universe or the planet, the analysis appears to be left in a
limbo – between unity and particularity – without any clear indication of which way
to tilt the balance.
We have juxtaposed two approaches to self-contained regimes that proceed from
fundamentally different premises, a universalistic one and a particularistic one. The
former takes the ‘whole’ of the legal order as a starting point. The latter proceeds from
a detailed examination of one of its particular subsystems, followed by a discussion as
to how the particular fits into the wider body of law and institutions. Depending on
whether we choose a universalistic or a particularistic perspective, whether we first
see the universe or the planets, the analysis tends to yield different results. If we focus
on the universe, the law of the universe (general international law) governs the plan-
ets. If we focus on the planets, planetary rules (the rules of the subsystem) leave little
room for universal law. As the ILC’s Special Rapporteur Arangio-Ruiz has observed
with regard to the EC:
Generally, the specialists in Community law tended to consider that the system constituted a self-
contained regime, whereas scholars of public international law showed a tendency to argue that
the treaties establishing the Community did not really differ from other treaties . . . 111
The disparity of the conclusions of ‘specialists in community law’ and ‘scholars of

public international law’ is a consequence of what has been called the level-of-
analysis problem. In international relations theory, Singer warned that analytical
results of one and the same political issue varied depending on whether an analysis
was carried out at the level of the international system or at the level of the national
state.112 By analogy, Singer’s observations are relevant to the present level-of-ana-
lysis problem. ‘Universalist’ public international lawyers perceive special regimes as
integral parts of a larger international system, while their ‘particularist’ colleagues
view the international system through the prisms provided by a particular regime.
The ‘system-oriented model’ (i.e., the universal approach) tends to lead the observer
into a position that exaggerates the impact of the system upon the particular regimes
and, conversely, discounts the impact of the regimes on the system.113 The regime
model, on the other hand, avoids an inaccurate homogenization that may flow from
the systemic focus. However, it may also lead to the opposite type of distortion – a
marked exaggeration of the differences among sub-systemic actors and a certain
blindness to their common systemic construction.114
111
G. Arangio-Ruiz, ‘Summary Records of the Meetings of the Forty-fourth Session’, ILC Yearbook Vol. I
(1992) 76.
112
Singer, ‘The Level-of-Analysis Problem in International Relations’, in K. Knorr and S. Verba, The Interna-
tional System: Theoretical Essays (1960) 77. Singer’s article constitutes the classical articulation of the
level-of-analysis problem. Contemporary international relations theory routinely analyses international
relations at various theoretical levels.
113
Ibid., at 80.
114
Ibid., at 83.
3 The Fallback on General International Law

A ‘. . . & il paroît qu’on l’a voulu plus fortement’
The debate on self-contained regimes is a case in point for the relative character of
maxims of treaty interpretation. Two valid techniques of interpretation yield diamet-
rically opposed results: if we assume that states have to ‘contract out’ of general inter-
national law, the rules on state responsibility apply to the extent that there is no
express derogation. If we emphasize the particular character of special systems, the
maxim expressio unius est exclusio alterius permits the conclusion that no other rules
than those specified apply. Yet, we do not purport to suggest that the question
whether a fallback on the rules on state responsibility is conceivable is fatally indeter-
minate. The preceding section has demonstrated, however, that a formal, almost
mechanical application of the lex specialis principle, based on a presumption in favour
of the ‘general’ provision, does not suffice to establish the applicability of the rules on
state responsibility. The lex specialis maxim is merely an argumentative tool to articu-
late further, normative considerations.
In special systems, the general international law on state responsibility may be rel-
evant in various ways, including problems of attribution,115 circumstances preclud-
ing wrongfulness,116 or the irrelevance of internal law.117 The hard case, however,
remains the question whether recourse to countermeasures in the event of a continu-
ous violation of a treaty obligation should be permitted. Can states fall back on gen-
eral international law, after they have exhausted the special rules and procedures of a
special regime? Countermeasures under general international law provide an
enforcement mechanism that special systems may lack. Thus, the policy question is
whether primary rules contained in special subsystems ‘deserve’ the additional ‘bite’
that enforcement through countermeasures can deliver.
In other words, are there reasons why obligations in special regimes should be con-
sidered ‘softer’ than other international law? If de Vattel is right, the contrary is the
case: ‘il est ordonné plus précisément, & il paroît qu’on l’a voulu plus fortement’.118 The
fact that states decide to go through the cumbersome process of multilateral treaty-
making suggests that the rules elaborated in this process are of particular import-
ance. It goes without saying that, hypothetically, states are free to negotiate special
norms of international law precisely with a view to softening their obligations. How-
ever, we should not presume lightly that states would be less willing to live up to spe-
cial obligations than to duties under general international law. Absent a clear
indication, special rules must be deemed to embody a particular commitment.
115
Articles 4–11 of the ILC Draft Articles on Responsibility of States for Internationally Wrongful Acts,
supra note 9.
116
Articles 20–27 of the ILC Draft Articles on Responsibility of States for Internationally Wrongful Acts,
ibid.
117
Article 32 of the ILC Draft Articles on Responsibility of States for Internationally Wrongful Acts, ibid.
118
de Vattel, supra note 13. ‘it is enjoined with greater precision, and appears to have been more pointedly
intended’.
508 EJIL 17 (2006), 483–529
B Effective Treaty Interpretation and the ‘Fallback’ on General

International Law
One need not go as far as Kelsen, who stated that international law is true law only if
it is a ‘coercive order’,119 that is, if reprisals are available,120 to justify a fallback on the
general international law of state responsibility.121 A fallback on the general interna-
tional law of state responsibility as a measure of last resort follows from the principle
of effective treaty interpretation. The line of reasoning operates in three steps:
(1) If states create new law – whether it be in the field of human rights, trade or
regional cooperation – there is a presumption that such rules embody a particu-
larly strong commitment.
(2) General international law vests a state with certain capacities to ensure that its
rights be respected, including a restricted right to unilateral enforcement action
(countermeasures).
(3) Among several possible constructions, the principle of effective interpretation122
requires adopting the interpretation that best gives effect to the norm in
question. Effectiveness includes the notion of enforceability. Consequently, it
cannot be easily inferred that a state was willing to give up ‘the rights or facultés
of unilateral reaction it possessed under general international law’123 by comple-
menting special primary obligations with a specific set of secondary obligations.
If states create new substantive obligations along with special enforcement
119
Kelsen, supra note 55, at 320: ‘international law is “law”, if it is a coercive order, that is to say, a set of
norms regulating human behaviour by attaching certain coercive acts (sanctions) as consequences to
certain facts, as delicts, determined by this order as conditions, and if, therefore, it can be described in
sentences which – in contradistinction to legal norms – may be called “rules of law”.’
120
Ibid. at 321: ‘It is easy to demonstrate that this assumption is correct with respect to reprisals. For it is a
principle of general international law that a state which considers some of its interests violated by
another state, is authorized to resort to reprisals against the state responsible for the violation. A
“reprisal” is an interference – under normal circumstances forbidden by international law – in the
sphere of interest of a state; it is an interference that takes place without and against the will of the state
concerned and is in this sense a coercive act, even if it is executed without physical force (i.e., without
force of arms) when the affected state does not resist.’
121
In fact, we would tend to side with Rosalyn Higgins, who stated that the existence of effective sanctions
does not predicate the existence of a norm of international law. R. Higgins, Problems and Process: Interna-
tional Law and How We Use It (1994), at 16. This is particularly true in light of contemporary compliance
theories, which have suggested that coercion is merely one out of several factors that encourages com-
pliance with international obligations, cf. Pulkowski, ‘Testing Compliance Theories: Towards the United
States Obedience of International Law in the Avena Case’, 19 Leiden J Int’l L (2006), issue 2.
122
Despite its frequent use, the systematic foundation and the actual content of the principle of effective
interpretation have remained somewhat opaque. Taking the Vienna Convention on the Law of Treaties
as a starting point, the effectiveness principle (or the maxim ut res magis valeat quam pereat) is reflected
either in the duty to interpret ‘in the light of [a treaty’s] object and purpose’ or in the notion of good faith.
The notion of good faith may be the more convincing solution: the principle of effective interpretation
precludes a state from frustrating the obligations assumed by invoking a formal circumvention of the
conditions under which the norm would apply. Cf. I. Sinclair, The Vienna Convention on the Law of Treaties
(2nd ed. 1984), at 115 et seq.; R. Jennings and A. Watts, Oppenheim’s International Law (9th ed. 1992), at
1280 et seq.
123
Arangio-Ruiz, supra note 111, at 77.
mechanisms, they merely relinquish their facultés under general international

law in favour of a special regime’s procedures to the extent that and as long as
those procedures prove efficacious. When such procedures fail, enforcement
through countermeasures under general international law becomes an option.
C Scenarios for a Fallback on State Responsibility

Under this premise, a fallback on state responsibility – and particularly on the
regime of countermeasures – is conceivable in three scenarios:124 in the case of con-
tinuous violation of an obligation under a special system, despite a decision to the
contrary by the system’s competent dispute settlement body; in the case of an
injured state’s failure to obtain reparation, despite a respective decision by the sys-
tem’s competent dispute settlement body; if unilateral action is necessary as a defen-
sive measure.
Recourse to countermeasures under international law, critics contend, may jeop-
ardize the integrity of the particular regime in question. Such critics close their eyes
to an important truth: while the unilateral enforcement of obligations undoubtedly
constitutes an ‘emergency operation’, countermeasures contribute to creating
future expectations of effective enforcement in the international community. As
Michael Reisman has noted, ‘the expectation of the effectiveness of enforcement
mechanisms is a factor inducing compliance. Enforcement becomes a “self-fulfilling
prophecy”.’125 Once countermeasures have contributed to the peaceful resolution
of one of the above ‘emergency scenarios’, ‘expectations of effectiveness are gener-
ated, permitting enforcement machinery subsequently to fulfil its function by sym-
bolic presence rather than by active intervention’.126 Countermeasures may thus
ultimately preserve (rather than jeopardize) the integrity of a special regime’s
enforcement mechanism.
In the application of countermeasures, due respect must obviously be paid to the
characteristics of the special regime in question. It would be wrong, however, to
assume from the outset that the delicate structures of environmental law or human
rights could not be adequately treated by the general international law practitioner
with her state responsibility toolbox. The Articles take account of sensitivities of spe-
cial regimes in various ways:
(1) The differentiation between injured states and non-injured states entitled to a
limited range of remedies prevents states whose interests are not affected from
meddling into other states’ affairs.
(2) The test of proportionality allows accommodation of the needs of special fields
of international law and the interest of states in the integrity of an institutional
system.
124
Cf. Simma, supra note 6, at 111; Arangio-Ruiz, supra note 11, at para. 115.
125
Reisman, ‘The Enforcement of International Judgments’, 63 AJIL (1969) 1, at 7 (footnote omitted).
126
Ibid.
510 EJIL 17 (2006), 483–529
(3) Except where no other response can be envisaged, no remedy may be resorted to
that would impair the continued performance of the obligation breached. This
refers above all to countermeasures. Thus, the spectre of state responsibility loses
some of its alarm once a closer look is had at how the Draft Articles operate in
practice.
D A Broader Picture: Negotiating the Balance of Effectiveness

and Legitimacy
Hence, we suggest that, in principle, the particular rules of a special system govern
the consequences of breaches of substantive obligations. If the need arises, however,
the invocation of the ‘rules of the universe’ (the general international law on state
responsibility) remains a valid lawyerly technique to enhance the effectiveness, the
effet utile of the rules of the planet (the substantive obligations contained in the
respective subsystem). The lex specialis principle is the methodological tool in the
hand of the jurist to connect a particular regime with general international law in
this case.
Such ‘oscillation’ between the invocation of unity and particularity is a legitimate
and commonly employed technique of legal argument. In strong regimes, general
international law may serve as a source of legitimacy, while the rules of the regime
provide the kind of operational effectiveness that advances the goals of the regime. In
weak regimes, the rules of the regime often embody a superior legitimacy. In this
case, lawyers reach out for the law of the universe of general international law to
increase the effectiveness of the regime’s rules.
Let us illustrate this point with two examples. The WTO is generally perceived as a
‘strong’ regime. However, its effectiveness is impeded by its – partial – incomplete-
ness. The regime does not contain any specific rules on interpretation, standard of
review, burden of proof: in short, rules that assist panellists in administering the sub-
stantive law. In these areas, the WTO regime cannot but refer to rules and principles
developed under general international law. The Appellate Body confirmed the openness
of the regime when it held ‘that the General Agreement is not to be read in clinical isola-
tion from public international law’.127 It may be safely concluded that both the parties
and the judicial organs of the WTO do not hesitate to invoke the ‘unity’ of international
law, when rules outside the regime appear to enhance its effective operation.128
At the level of substantive obligations, WTO dispute settlement presents a different
picture. Every non-trade concern is suspected of giving rise to potential protectionism
127
US – Standards for Reformulated and Conventional Gasoline, WT/DS2/AB/R, 17. Cf. generally, Pauwelyn,
‘The Role of Public International Law in the WTO: How Far Can We Go?’, 95 AJIL (2001) 535.
128
As R. Howse has pointed out, reference to the general international law on treaty interpretation may, in
addition, be seen as enhancing the legitimacy of the WTO: ‘[T]he very decision to follow these general
interpretative rules of public international law enhances the legitimacy of the dispute settlement organs
in adjudicating competing values, because these norms are common to international law generally,
including to regimes that give priority to different values, and are not specific to a regime that has tradi-
tionally privileged a single value, that of free trade.’, Howse, ‘From Politics to Technocracy – and Back
Again: The Fate of the Multilateral Trade Regime’, 96 AJIL (2002) 94, at 110.
impeding the regime’s effectiveness. Hence, panels (and to a lesser extent the Appel-
late Body) tend to stress the particularity of WTO law, refraining from unnecessar-
ily cross-linking WTO provisions with other rules of the ‘universe’. However, this
isolationist discursive strategy can only be pursued as far as the point where the
effectiveness-legitimacy balance begins to tilt. Once the legitimacy of the decision
comes under fire, the invocation of ‘unity’ rather than ‘particularity’ becomes an
interesting discursive option. By relying on rules outside the WTO regime that, in
the view of many, embody legitimate concerns or internationally recognized ethical
positions, the WTO’s judicial bodies have attempted to import the legitimacy offered
by the ‘universe’ to the ‘planet’.
In the US–Shrimp case,129 the Appellate Body referred to international envi-
ronmental instruments outside the WTO to counter the image of the WTO as a
cold-hearted trade-over-everything institution. Adopting such a unitary dis-
course did not even require the Appellate Body to reverse the recommendation of
the panel in substance. While the Appellate Body reproved the panel for its ‘particu-
laristic’, trade-focused approach,130 it nonetheless concluded that the US ‘fails to meet
the requirements of the chapeau of Article XX, and, therefore, is not justified under
Article XX of the GATT 1994’.131 The submissions by the US, Canada and the Euro-
pean Communities in the recent EC-GMO case reflect a similar discursive pattern:
while the complainants (US, Canada) suggested an isolationist reading of the relevant
WTO provisions,132 the EC made a unitary argument, asking the panel to increase the
decision’s legitimacy by recourse to non-WTO rules and standards.133
Similar discursive patterns can be observed in institutionally weaker regimes, such as
human rights. In good times, human rights lawyers tend to stress the particularity of
their regime. Human rights obligations follow a different logic from reciprocal, universal
international law; they constitute ‘objective regimes’.134 As the European Court of
Human Rights stated in the Northern Ireland case, human rights treaties are ‘[u]nlike
international treaties of the classical kind’ since they go beyond reciprocal obligations.135
129
US – Import Prohibition of Certain Shrimp and Shrimp Products, WT/DS58/R, WT/DS58/AB/R.
130
The panel briefly discussed the various international instruments presented by the US. However,
it was not convinced that such instruments would distract from the parties’ substantive WTO
obligations. The Appellate Body, by contrast, deferred to rules and standards outside the WTO
regime to some extent, interpreting the terms of WTO provisions in accordance with such rules
and standards.
131
Ibid., para. 187.
132
The US’ first submission is available at http://www.ustr.gov/Trade_Agreements/Monitoring_
Enforcement/Dispute_Settlement/WTO/Dispute_Settlement_Index_-_Pending.html.
133
The EC’s first submission is available at http://trade-info.cec.eu.int/wtodispute/search.cfm?code=2.
134
Riphagen, ‘Fourth Report’, in ILC Yearbook, Vol. II, Part One, (1983), at 17, para. 89.
135
Ireland vs. UK, ECtHR, Ser. A Vol. 25, para. 239. Cf. similarly the statement by the European Commission of
Human Rights in the Pfunders case: ‘It clearly appears from these pronouncements that the purpose of the
High Contracting Parties in concluding the Convention was not to concede to each other reciprocal rights
and obligations in pursuance of their individual national interests but to realise the aims and ideals of the
Council of Europe, as expressed in its Statute, and to establish a common public order of the free democra-
cies of Europe with the object of safeguarding their common heritage of political traditions, ideals, freedom
and the rule of law.’ 4 Yearbook of the European Convention of Human Rights (1961) 116, at 138.
512 EJIL 17 (2006), 483–529
By contrast, when the effectiveness of human rights regimes faces a severe chal-
lenge, lawyers are inclined to stress the unity of international law, arguing that
human rights are merely part of the larger international legal order. Such unity is a
precondition for enforcing human rights by the use of countermeasures under gen-
eral international law. For example, in response to the 1998 humanitarian crisis in
Kosovo, the European Communities adopted legislation that required the freezing of
Yugoslav funds and an immediate flight ban.136 And when, in 1981, the Polish gov-
ernment suppressed demonstrations and interned dissidents, several Western coun-
tries responded by suspending treaties providing for landing rights.137
4 Case Studies on Self-contained Regimes

The latter examples take us right back to the problem of special rules in the field of state
responsibility: the human rights regime falls back on general international law when its
effectiveness is jeopardized. After having established in the foregoing sections that the
existence of self-contained regimes is doubtful conceptually, it is time to take a closer look
at some ‘candidates’ for such self-contained regimes. Four regimes that are often referred
to as quasi-autonomous systems merit closer consideration: diplomatic law, the European
Communities legal system, WTO law and human rights. All four regimes contain special
rules governing the consequences of breach of a primary obligation. To take account of
the special characteristics of the respective regime, the consequences of breach must, in
principle, be deemed to flow from the regime’s special treaty provisions – at the exclusion
of the general international law on state responsibility. Yet, once these rules prove ineffi-
cacious, a unitary counter-argument remains feasible and the rules on state responsibility
apply. None of the regimes in question can be considered as fully self-contained.
A Diplomatic Privileges
The rules of international law most commonly associated with the notion of self-con-
tainment, the rules of diplomatic law, are at the same time the least convincing
example of a closed system of secondary rules. In what may well be referred to as a
jurisprudential overkill (or, to put it mildly, an unnecessarily broad statement), the
International Court of Justice in its Tehran Hostages judgment ruled that diplomatic
law ‘specifies the means at the disposal of the receiving State to counter any such
abuse’138 of diplomatic privileges and that, therefore, diplomatic law constitutes a self-
contained regime. It cannot be contested, of course, that the possibility of declaring a
diplomat persona non grata is a special remedy of diplomatic law, recourse to which
will usually be the appropriate reaction to an abuse of diplomatic privileges. However,
the Court’s dictum goes too far:
136
Common positions of 7 May and 29 June 1998, OJ 1998, L 143 and L 190; implemented through EC
Regulations 1295/98 (L 178, 33) and 1901/98 (L 248, 1).
137
Rousseau, ‘Chroniques des faits internationaux’ – Subsection on ‘Mesures adoptées par les Etats occiden-
taux a l’égard de la Pologne’, 86 RGDIP (1982) 605.
138
United States Diplomatic and Consular Staff in Tehran, ICJ Reports (1980) 40.
(1) The Court was only concerned with the question whether a receiving state is
entitled to resort to countermeasures as a reaction to a breach of the Vienna Con-
vention. However, if one perceives diplomatic law as a closed system, recourse to
the rules of state responsibility would equally be excluded for the sending state.
This is neither logical from a systematic point of view nor is it necessary for the
safeguarding of diplomatic relations. There is no reason why the sending state
should be precluded from initiating countermeasures (not affecting diplomatic
personnel) if the receiving state fails to respect the immunities of the sending
state’s diplomatic personnel.139 For instance, international law did not preclude
the US in the situation of Tehran Hostages from resorting to countermeasures out-
side the Vienna Convention, including the suspension of a treaty of amity, in
order to induce Iran to free US personnel.
(2) Even assuming an abuse of diplomatic privileges similar to the conduct of which
Iran accused the US diplomats in Tehran Hostages, there may be a need to resort
to remedies under general international law. The Vienna Convention does not
contain a provision on reparation. If a diplomat, however, by abusing his privi-
leges inflicts economic damage on the receiving state, there is no reason why the
receiving state should be precluded from recovering that damage, in addition to
declaring the perpetrator persona non grata. In Tehran Hostages, the Court itself
affirmed Iran’s duty to make reparation to the US.
(3) Diplomatic law has always been the classic playground of reciprocity. At least at
the time when the Vienna Convention was drafted, the ILC appears to have been
of the view that the symmetry of its obligations would allow reciprocal repris-
als.140 Given the traditional acceptance of countermeasures as a ‘fact of life’, it is
to be strongly doubted whether diplomatic law has since evolved into a non-
reciprocal regime justifying the categorical exclusion of all forms of unilateral
reaction.141 Paul Reuter’s comment in the ILC plenary of 1984 still merits
consideration:
139
Concurring: Zemanek, ‘The Unilateral Enforcement of International Obligations’, 47 ZaöRV (1987) 40:
‘The Court wanted to protect diplomats from reprisals, not their sending States.’
140
Cf. the Commentary to the ILC Draft Article 44 of 1958 (now Article 47 of the Vienna Convention), ILC
Yearbook, Vol. II (1958), 105, which presupposes the existence of reprisals in diplomatic law: ‘It is
assumed that the restrictive application in the sending State concerned is in keeping with the strict terms
of the rule in question, and within the limits allowed by the rule. Otherwise there is an infringement of
the rule and the action of the receiving State becomes an act of reprisal.’
141
The ILC’s 2001 Commentaries on State Responsibility appear to envision a more far-reaching protection
of diplomatic immunities. ‘[M]easures may be taken affecting diplomatic or consular privileges, not pre-
judicing the inviolability of diplomatic or consular personnel or of premises, archives and documents....
On the other hand, the scope of prohibited countermeasures under article 50 (2) (b) is limited to those
obligations which are designed to guarantee the physical safety and inviolability (including the jurisdic-
tional immunity) of diplomatic agents, premises, archives and documents in all circumstances, includ-
ing armed conflict.’ ILC, supra note 11, at 339. Even Barnhoorn, otherwise a supporter of the self-
containment theory, admits that ‘the exact determination of the scope of the [inviolability] rule has
always given problems.... [I]t is not entirely clear what immunities are precisely covered.’ Barnhoorn,
‘Diplomatic Law and Unilateral Remedies’, XXV Netherlands Yearbook IL (1994) 39, at 80.
514 EJIL 17 (2006), 483–529
He was of the view that, in so far as more general obligations such as humanitarian
obligations were not involved, the injured State could respond in kind to a manifest
violation of the rules on privileges and immunities. For instance, in the event of the
violation of the unanimously accepted rule concerning the diplomatic bag, the injured
State should be entitled to act in the same way as the State responsible for the viola-
tion. In such circumstances, the regime of privileges and immunities did not seem to be
particularly self-contained.142
(4) Finally, resort to countermeasures may be necessary under the fallback scenarios
outlined above – if the sending state fails to withdraw a diplomat who was declared
persona non grata; and, as an immediate defensive measure, if the receiving state
catches the diplomat in flagrante delicto. By the same token, the ILC had stated in the
commentary to its draft convention on diplomatic relations143 that diplomatic invio-
lability ‘does not exclude . . . either measures of self-defence or, in exceptional circum-
stances, measures to prevent [a diplomat] from committing crimes or offences’.144
Given the ‘permeability’ of diplomatic law, in fact its dependence on the reme-
dies of the general rules of state responsibility, there is no ground for arguing that
diplomatic law contains an exhaustive set of secondary rules. Rather, the exemp-
tion of diplomatic immunity from the scope of lawful ‘targets’ of countermeasures
is a limitation within the general law of state responsibility, which follows from the
purposes for which the diplomatic immunities are granted. Diplomatic channels
may be useful, even, and a fortiori, in times of crisis.145 Moreover, diplomatic rela-
tions would be severely impeded if diplomatic personnel were to potentially ‘con-
stitute resident hostages’.146 Thus, the definition of certain limitations to unilateral
reprisals is all that the Tehran Hostages ruling was about in substance. It is unfortu-
nate that the Court’s unnecessarily broad dictum is suggestive of a closed-circuit
system of legal consequences. In Dominicé’s words, ‘pour affirmer qu’une violation
initiale du droit diplomatique ne peut en aucune manière autoriser l’Etat qui en est la vic-
time à le transgresser à son tour, l’argument du régime se suffisant à lui-même n’est pas
nécessaire’.147
The ILC’s Articles on State Responsibility confirm by implication that the theory
that the exemption of diplomatic immunity from countermeasures is merely a ques-
tion of the proper application of the general rules of state responsibility. The Commis-
sion’s earlier debate mirrors the confusion caused by the Court’s Tehran Hostages
pronouncement. Special Rapporteur Riphagen, speaking, more in passing, of a ‘self-
contained regime of diplomatic law’,148 seemed to be inclined to construe diplomatic
142
Reuter, Summary Records of the Meetings of the Thirty-sixth Session, ILC Yearbook, Vol. I (1984), at 264.
143
Which formed the negotiating basis of the Vienna Convention on Diplomatic Relations.
144
ILC Yearbook, Vol. II (1958), at 97.
145
Barnhoorn, supra note 141, at 63.
146
ILC, supra note 9, at 240.
147
Dominicé, ‘Représailles et droit diplomatique’, in Recht als Prozeß und Gefüge. Festschrift für Hans Huber
zum 80. Geburtstag (1981) 551. ‘It is not necessary to have recourse to the notion of a self-contained
regime in order to affirm that an initial violation of diplomatic law cannot in any way authorize the
wronged State to transgress that law in turn.’
148
ILC Yearbook, Vol. II, Part 1 (1983), at 18, para. 97.
law as a closed circuit containing its own conclusive secondary rules.149 Special Rap-
porteur Arangio-Ruiz then questioned whether limitations on the use of counter-
measures really flow from any self-contained nature of diplomatic law. Instead, he
suggested that such limitations were a consequence of the application of the general
rules and principles constituting the regime of countermeasures.150 Consequently, he
introduced a provision in the regime of countermeasures that specifically prohibits
measures affecting diplomatic inviolability.151
The commentary adopted in 1995, however, does not fully reflect Arangio-
Ruiz’s criticism. In the commentary, the ILC suggests that the prohibition of
reprisals is either derived from the self-containment of diplomatic law or from an
alleged peremptory character of some of its essential norms.152 Special Rapporteur
Crawford and, following him, the Commission discarded the latter hypothesis.153
Article 50 in the ILC’s final product now makes an explicit distinction between jus
cogens (paragraph 1) and other obligations whose performance may be particu-
larly useful in times of conflict – diplomatic immunities and contractual dispute
settlement obligations (paragraph 2). While such an exemption of diplomatic
immunity from countermeasures ‘on functional grounds’154 was generally
endorsed, the Commission avoided grounding it in any self-contained character
of diplomatic law.
Systematically, the most convincing solution is to perceive the limitations to coun-
termeasures in Article 50(2) of the ILC text simply as an expression of the proportion-
ality principle laid down in its Article 51.155 Countermeasures affecting diplomatic
inviolability or the suspension of dispute settlement obligations are, in principle,
deemed disproportionate. If Article 50(2) is understood as a presumption of dispro-
portionality, however, this presumption can be refuted in an individual case if a state
demonstrates that overriding interests are at stake. The International Court of Justice
has had occasion to confirm that insistence on dispute settlement duties may be
149
Riphagen’s report, however, offers an unfortunate example of a misleading use of terminology, using
‘self-contained regimes’ as a synonym for any subsystem, cf. supra fn 88 and accompanying text.
150
ILC Yearbook, Vol. II, Part 1(1992), at 39, para. 110.
151
Article 14 as proposed in 1995, International Law Commission, Report on the Work of its Forty-seventh
Session, Official Records of the General Assembly, Fiftieth Session, Supplement No. 10 (A/50/10) 149.
152
International Law Commission, Draft Commentaries to Articles 13 and 14 of Part Two of the Draft Arti-
cles, A/CN.4/L.521, with further references.
153
International Law Commission, supra note 26, at 339: ‘The justification in each case concerns not so
much the substantive character of the obligation but its function in relation to the resolution of the dis-
pute between the parties which has given rise to the threat or use of countermeasures.’
154
Ibid., at 341.
155
There is some indication that the ILC considered the admissibility of targeting diplomatic personnel by
countermeasures to be a question of proportionality because the Commission notes that, after the reme-
dies provided under the Vienna Convention have been exhausted without success, ‘[a]t a second level,
measures may be taken affecting diplomatic or consular privileges, not prejudicing the inviolability of
diplomatic or consular personnel or of premises, archives and documents. Such measures may be lawful
as countermeasures if the requirements of this chapter are met.’ Ibid., at 339. We go only one step fur-
ther in the sense that even the inviolability of diplomatic personnel should not be sacrosanct provided
that sufficiently grave state interests are at stake.
516 EJIL 17 (2006), 483–529
‘wholly artificial’ and ‘excessively formalistic’:156 in short, disproportionate in individ-

ual cases. Similarly, in the cases sketched out above, insistence on the inviolability of
diplomatic personnel may be a disproportional limitation to a state’s right to unilat-
eral reaction. The threshold for infringing the physical freedom of diplomats, how-
ever, would obviously have to be very high.
B The European Communities Legal System

While diplomatic law is a rather unlikely candidate for self-containment (if it were not
for an unfortunate dictum of the Hague Court), the legal system set up by the ‘Treaty
establishing the European Community’ bears very strong characteristics of self-
containment. The point of view of the European Court of Justice is as well known as it
is unequivocal: the Community Treaty has ‘created its own legal system’,157 which
‘constitutes a new legal order of international law’.158 Therefore, ‘the basic concept of
the Treaty requires that member States shall not fail to carry out their obligations and
shall not take the law into their own hands’,159 irrespective of considerations of reci-
procity. ‘A member State cannot under any circumstances unilaterally adopt, on its
own authority, corrective measures or measures to protect trade designed to prevent
any failure on the part of another member State to comply with the rules laid down by
the Treaty’.160 The continuous assertion of the Community’s sui generis character,
however, does not by itself create an ‘own legal order’. From a public international
law perspective, the EC legal system remains a subsystem of international law.161
Since any amendment of the constituting treaties still depends on the consent of the
Union’s sovereign member states, ‘the Community legal order is still dominated by
the spirit of international law’.162
From a purely systematic point of view, the rules on state responsibility are, hence,
residually applicable.163 But is there still a need for such residual application? We
have posited that states only intend to relinquish their facultés under general interna-
tional law in favour of a special regime’s procedures to the extent that and as long as a
subsystem’s procedures prove efficacious. Consequently, a fallback on the general
rules of state responsibility presupposes that the mechanisms under the EC Treaty fail
156
Military and Paramilitary Activities in and against Nicaragua (Nicaragua v. United States of America), Merits,
ICJ Reports (1986), para. 274.
157
Costa v. ENEL, 6/64, ECR [1964], 585 at 593.
158
Van Gend en Loos, 26/62, ECR [1963], 1 at 12.
159
Commission v. Luxemburg and Belgium, joint cases 90/63 and 91/63, ECR [1964], 625.
160
Mutton and Lamb, 232/78, ECR [1979], 2729.
161
Cf. Pellet, ‘Les fondements juridiques internationaux du droit communautaire’, V Collected Courses of the
Academy of European Law (1994) 249; A. Marschik, Subsysteme im Völkerrecht (1997) 210; Conway,
‘Breaches of EC Law and the International Responsibility of Member States’, 13 EJIL (2002) 679.
Hartley, ‘International Law and the Law of the European Union – A Reassessment, 72 BYbIL (2001) 1.
162
De Witte, ‘Rules of Change in International Law: How Special is the European Community?’, XXV
Netherlands Yearbook IL (1994) 299, at 331.
163
Cf. also the analysis by Hartley, supra note 161, who concludes that the EC has not constituted its separ-
ate legal system by virtue of some kind of constitutional metamorphosis subsequent to the EC’s founda-
tion in the realm of international law.
to give effect to the obligations members have assumed under the Treaty. The repara-
tion regime under the EC Treaty is comprehensive and mostly effective. As Gerard
Conway has correctly pointed out,164 most of the gaps still existing 20 years ago165
have now been filled up. The Francovich166 and Brasserie du Pêcheur167 principles have
vested individuals with a remedy to recover damages resulting from breaches of EC
law by Member States; Article 228 EC Treaty has been introduced to give the declara-
tory judgments of the Court rendered according to Articles 226 and 227 the ‘bite’ of a
pecuniary sanction, if a member state disregards them; and Article 7 of the EU Treaty
provides for an institutionalized procedure for the suspension of membership in case
of a grave and continuous violation of Community law.
As far as we can see, only two hypothetical scenarios can be identified in which
lacunae in Community law may prompt a fallback on state responsibility.
(1) The first scenario would involve a continuous violation of Community law by an
EC Member State. Obviously, the lex specialis principle would require another
Member State (respectively the Commission) to exhaust the more ‘special’ mech-
anisms under the EC Treaty. A starting point would be for an affected state to
bring the matter before the Commission and, if the Commission does not take up
the case itself, to the European Court of Justice pursuant to Article 227 of the EC
Treaty. If the violating state chooses not to change its conduct despite a judg-
ment of the Court, the Commission has the authority to seize the Court once
more with an application for a pecuniary penalty pursuant to Article 228(2) of
the Treaty.168 If the violation persists, the Member State may attempt to have the
membership of the violating state suspended pursuant to Article 7 EU Treaty.
The latter remedy, which operates on the political rather than the legal plane,
will not prove to be successful, however, unless the ‘injured’ state can obtain a
declaration of a grave and continuous breach by unanimous vote of the Council.
Arguably, in this extreme scenario, the only option to induce compliance that
remains for the ‘injured’ state is a fallback on unilateral countermeasures.
(2) The second scenario revolves around the issue of state-to-state reparation for
breaches of EC law. Claims for compensation of damages suffered by nationals,
that is, cases of diplomatic protection, do not fall under this category; following
Brasserie du Pêcheur and Francovich, individuals are vested with a cause of action
of their own. Cases where economic losses may be incurred by a Member State’s
domestic economy as such, rather than by certain individuals, may resemble the
so-called guerre de moutons. In this instance, the United Kingdom indicated that
164
Conway, ‘Breaches of EC Law and the International Responsibility of Member States’, 13 EJIL (2002)
679, at 688.
165
Cf. Simma, supra note 6, at 126 et seq.
166
Francovich and Bonifaci v. Italy, joined cases C-6/90 and C-9/90, ECR [1991], I-5357.
167
Brasserie du Pêcheur v. Germany and Factortame v. UK, joined cases C-46/93 and C-48/93, ECR [1996], I-
1029.
168
If the Commission fails to do so, the ‘injured’ state could consider action against the Commission
pursuant to Article 232.
518 EJIL 17 (2006), 483–529
the damage caused by French non-compliance with a Court judgment resulted

in economic losses of £20 million.169
Yet, the EC Treaty makes no explicit provision for a mechanism that would allow
inter-state claims of reparation. The literature has proposed several dogmatic bridges
to cover up this lacuna and accommodate inter-state claims for reparation within the
EC system. The first is to resort to an extensive interpretation of Articles 227 and 228
of the EC Treaty, according to which a declaration of breach by the European Court
entails the duty ‘to re-establish the status quo ante (restitutio in integrum) and, where
necessary, pay compensation for any injury suffered’.170 Some authors, however,
doubt that the Commission’s involvement under Article 228 is suitable for reparation
cases and consequently suggest grounding claims for compensation directly on
Article 220, which emphasizes the Court’s role as a guardian of the law.171 A third
dogmatic foundation of inter-state claims would be an application mutatis mutandis of
the Francovich principles.172 Recourse to the rules on state responsibility will not be
necessary if the European Court of Justice accepts one of these dogmatic bridges to
accommodate inter-state claims for damages within the European legal system.
Nonetheless, a fallback on countermeasures would still be conceivable to enforce a
Member State’s claim of reparation. Member States may well have to resort to the
‘mechanisms of state responsibility . . . as traditionally understood’,173 if the mecha-
nisms within the EC do not prove effective.
Thus, systematically, a fallback on countermeasures remains conceivable, albeit
only in two narrow ‘emergency’ scenarios. Some authors claim, however, that the
specific teleology of the Community order precludes such unilateral action. One argu-
ment points to the exclusive jurisdiction of the European Court of Justice in adjudicat-
ing disputes on Community law (Article 292 EC Treaty). Yet, the fallback scenarios
envisaged take due account of the Court’s central role, since unilateral action is
reserved for cases of non-compliance with the Court’s decision.174
A more weighty argument relates to the alleged lack of reciprocity in legal relations
among the Member States. The European Court has asserted that the new legal order
created by the EC Treaty leaves no room for reciprocity considerations. Reciprocity,
however, is not something states can choose to ‘abolish’; it is an inherent structural
169
Cf. Schwarze, ‘Das allgemeine Völkerrecht in den innergemeinschaftlichen Beziehungen’, 18
Europarecht (1983) 24. The UK government initially requested that the Commission order France to pay
£20 million in damages. The request, however, was withdrawn for political reasons.
170
A. G. Toth, Legal Protection of Individuals in the European Communities, Vol. II (1978), at 181; cf. for fur-
ther references, A. Marschik, Subsysteme im Völkerrecht (1997) 231 et seq.
171
Tomuschat, ‘Völkerrechtliche Schadensersatzansprüche vor dem EuGH’, in J. F. Baur, P. Müller-Graff
and M. Zuleeg (eds), Europarecht – Energierecht – Wirtschaftsrecht: Festschrift für Bodo Börner zum 70.
Geburtstag (1992) 458.
172
Marschik, supra note 161, at 233, who points out that the Court in Commisson v. Italy, 39/72, ECR
[1973], 112 explicitly mentioned that one Member State may be liable towards another Member State.
173
J. H. H. Weiler, The Constitution of Europe: The Transformation of Europe (1999) 29.
174
As Conway points out, Article 292 EC Treaty only excludes other judicial or institutional methods of dis-
pute resolution; it is not concerned with the question of unilateral enforcement. Conway, supra note
161, at 686.
characteristic of obligations that operate on a do ut des basis. Many of the substantive

provisions of the Treaty, such as the fundamental freedoms, are structurally recipro-
cal guarantees. Whether Member States have renounced their rights to enforce these
guarantees by reciprocal measures is quite a different question. They certainly have
done so to a great extent. But here we return to the basic proposition underlying the
unitary argument: Member States only intended to relinquish their reciprocal facultés
under general international law to the extent that the EC’s special procedures proved
effective. Consequently, the non-reciprocity argument does not preclude counter-
measures in the two ‘emergency’ scenarios outlined above, in which the procedures
inherent in the EC system fail.
Developing a third argument against unilateral action, Joseph Weiler has pointed to
the adverse effects of reprisals on individuals. ‘[T]he recourse to counter-measures
would inevitably affect individuals removed from the dispute, militating against the
very notion of a “new legal order of international law . . . the subjects of which comprise
not only Member States but also their nationals”.’175 It is true that countermeasures
will usually impair certain individual benefits. Weiler seems to disregard, however, that
this is equally true for the continuous violation of EC law that provokes unilateral reac-
tion. A grave and enduring breach of Community law constitutes not only a measure of
disrespect for the other Member States’ rights but at the same time inevitably affects the
benefits of ‘individuals removed from the dispute’ in those Member States.
To sum up: it is reasonable to conclude that EC law operates as a closed system of
secondary rules for most practical purposes. Conceptually, however, it is not a self-
contained regime since there remain scenarios in which a fallback on state responsibil-
ity remains feasible and necessary, and since such a fallback is not precluded by pecu-
liar characteristics of the Community order.176 Only after the European Union has
attained such a degree of integration that recourse to general international law is not
conceivable without putting into question the whole raison d’être of the treaty frame-
work might the conclusion of self-containment be justified. But this, it is submitted, is
not (yet) the state of European integration.
C The WTO System

In an article for the Netherlands Yearbook of International Law, Pieter Kuyper con-
trasted the GATT, ‘which was a self-contained regime of international law only in
aspiration but not in reality’, with the (then) newly created WTO, which ‘has moved
decisively in the direction of such a self-contained regime’.177 From a legal point of
view, the WTO appears to be more independent of and ‘closed’ towards general inter-
national law than the GATT. At the same time, however, the WTO has increasingly
‘opened’ to a common culture of international lawyers. The conclusion only appears
175
Weiler, supra note 173, at 29.
176
Concurring: Conway, supra note 161, at 691 and Marschik, supra note 161, at 173, 259 et seq.
177
Kuyper, ‘The Law of GATT as a Special Field of International Law’, XXV Netherlands Yearbook IL (1994)
227, at 252.
520 EJIL 17 (2006), 483–529
paradoxical at a first glance. In reality, as Weiler has put, it ‘[j]uridification is a pack-

age deal. It includes the Rule of Law but also Rule of Lawyers.’178
In the GATT era, there was a clear need for unilateral enforcement, since a single
contracting party, including the defeated respondent, was able to block the adoption
of a panel report (or the establishment of a panel in the first place).179 The Marrakesh
Agreement with its Dispute Settlement Understanding (DSU) has eliminated this
loophole by establishing a negative consensus procedure. Moreover, Article 23 DSU
obliges members to submit (all of) their disputes to the WTO’s dipute settlement
machinery: ‘When Members seek the redress of a violation . . . they shall have
recourse to, and abide by, the rules and procedures of this Understanding’.
The ILC, in its commentaries on state responsibility, interprets this provision as an
explicit derogation from the rules of state responsibility. ‘In some cases it will be clear
from the language of a treaty or other text that only the consequences specified are to
flow. . . . An example . . . is the World Trade Organization Dispute Settlement Under-
standing as it relates to certain remedies.’180 Other authors have affirmed the con-
trary, namely that ‘[n]othing in Article 23 . . . amounts to an express derogation’181 of
the rules of state responsibility. It is true that the wording of Article 23 establishes the
primacy of WTO dispute settlement. On the other hand, the wording neither specifi-
cally permits nor prohibits recourse to additional remedies under general interna-
tional law beyond those provided by WTO law. However, are there remedies under
state responsibility that WTO law would not contain in a more tailor-made form?
It is clear that the DSU contains a lex specialis on cessation of the breach and on
continued performance of the obligation.182 Pursuant to Article 19.1 of the DSU, pan-
els are empowered to order183 a member to ‘bring the measures into conformity’ with
WTO law. However, it is far from clear whether the WTO system contains remedies
akin to reparation184 and countermeasures.185 Article 22 of the DSU addresses ‘com-
pensation’ and the ‘suspension of concessions’ (also known as ‘sanctions’ or ‘retalia-
tion’). From a systematic point of view, however, both remedies can be regarded as a
178
Weiler, ‘The Rule of Lawyers and the Ethos of Diplomats: Reflections on the Internal and External Legiti-
macy of WTO Dispute Settlement’, 35 JWT (2001) 191, at 197.
179
Before the creation of the WTO, there was severe disagreement between the US and the EC, however, on
whether unilateral action was precluded. The EC took the view that the GATT was essentially a political
arrangement rather than a legal one and that the positive consensus system was an integral part of the
GATT. Cf. M. Hahn, Die einseitige Aussetzung von GATT-Verpflichtungen als Repressalie (1995).
180
ILC, supra note 11, at 357.
181
M. Garcia-Rubio, Unilateral Measurs as a Means of Enforcement of WTO Recommendations and Decisions
(2001). Cf. also Mavroidis, ‘Remedies in the WTO Legal System: Between a Rock and a Hard Place’, 11
EJIL (2000) 763.
182
Articles 29 and 30 of the ILC Draft Articles on Responsibility of States for Internationally Wrongful Acts,
supra note 9.
183
‘Recommend’ in the DSU’s language. The recommendation formally assumes a binding character once
it is adopted by the Dispute Settlement Body.
184
Articles 31 et seq. of the ILC Draft Articles on Responsibility of States for Internationally Wrongful Acts,
supra note 9.
185
Articles 49 et seq. of the ILC Draft Articles on Responsibility of States for Internationally Wrongful Acts,
ibid.
special variant of the inadimplenti non est adimplendum principle of the law of trea-
ties.186 In case of a breach, the general rules of the law of treaties permit the tempor-
ary suspension of a treaty to restore the symmetry of obligations. The compensation
provision of Article 22.1 introduces an element of flexibility. It allows a state to tem-
porarily uphold a violation in exchange for concessions in other areas, provided that
the injured state agrees.187 Similarly, suspension of concessions pursuant to Article
22.4 constitutes merely a partial non-performance of the treaty, subject to the prior
authorization and the subsequent monitoring by the Dispute Settlement Body. Since
‘compensation’ and ‘retaliation’ restore the reciprocity balance within the same
Agreement,188 they systematically ‘deal with the same substantive matter’189 as Arti-
cle 60 of the Vienna Convention;190 being essentially ‘remedies’ under the law of
treaties, they would not necessarily derogate as leges speciales the rules of state
responsibility.
Systematic arguments aside, and seen from a functional perspective, sanctions pursu-
ant to Article 22 DSU operate in a very similar way to countermeasures. As Pauwelyn
has pointed out, concessions are suspended primarily with a view to inducing compli-
ance. Once the principal objective is compliance (and not the restoration of contractual
equity), however, the WTO sanctions ‘have shifted into the area of State responsibil-
ity’.191 The US – Foreign Sales Corporations and US – Steel disputes between the EC and the
US confirm the finding that states employ suspensions of concessions similarly to coun-
termeasures under general international law. In the Steel case, the mere threat of sanc-
tions worth more than $2 billion192 persuaded the US administration to cut back
domestic steel subsidies after the Appellate Body had confirmed their illegality.193 Conse-
quently, strong grounds exist for regarding the WTO rules on retaliation as leges speciales
vis-à-vis countermeasures under general international law. Recourse to reprisals parallel
to or in lieu of WTO-authorized sanctions is thus precluded. Similarly, the panel in US –
Section 301 concluded that ‘WTO members [are prevented] from unilaterally resolving
their disputes in respect of WTO rights and obligations’.194
186
Cf. Article 60 of the Vienna Convention on the Law of Treaties; Simma and Tams, ‘Commentary on Arti-
cle 60’, in O. Corten and P. Klein (eds), Commentaire article par article des Conventions de Vienne de 1969 et
1986 sur le droit des traités (forthcoming).
187
This agreement to compensation has thus been described as ‘a temporary renegotiation of the WTO
treaty’, Pauwelyn, supra note 8, at 219.
188
However, it could be argued that the option of cross-retaliation, which allows a member under certain
conditions to suspend the performance of other covered Agreements, introduces an element of reprisals.
189
Fitzmaurice, supra note 18, at 203.
190
Cf. Sicilianos, ‘The Relationship between Reprisals and Denunciation or Suspension of a Treaty’, 4 EJIL
(1993) 341; J. Crawford, Second Report on State Responsibility, A/CN.4/498 Add 2 (1999) paras 314–
329. Crawford and Olleson, ‘The Exception of Non-Performance: Links between the Law of Treaties and
the Law of State Responsibility’, 21 Australian Yearbook IL (2000) 1.
191
Pauwelyn, supra note 8, at 230.
192
Cf. http://www.eubusiness.com/afp/031124173342.qe9cs5uw.
193
United States – Definitive Safeguard Measures on Imports of Certain Steel Products, Report of the Appellate
Body, 10 November 2003, WT/DS248/AB/R.
194
United States – Sections 301–310 of the Trade Act of 1974, Report of the Panel, 22 December 1999, WT/
DS152/R, 7.46.
522 EJIL 17 (2006), 483–529
Most contentious, however, is the question whether a fallback on reprisals

under state responsibility remains an option once the suspension of concessions
proves ineffective. The ILC commentaries are critical of such a fallback from the
WTO type of countermeasures to the state-responsibility type. Having discussed
WTO sanctions, the ILC adds that ‘[t]o the extent that derogation clauses . . . are
properly interpreted as indicating that the treaty provisions are “intransgressible”,
they may entail the exclusion of countermeasures’.195 Yet, that the DSU’s clauses
are to be ‘properly interpreted’ as indicating that their provisions are ‘intrans-
gressible’, is a doubtful proposition. Of course, the procedures under the DSU take
precedence. Article 22.7 of the DSU vests a body of arbitrators with the compe-
tence to determine the amount of sanctions to be authorized, thus depriving WTO
members of the right to unilaterally determine the level of countermeasures. Once
these procedures have been exhausted, however, countermeasures under general
international law remain the ultima ratio, only to apply in cases of continuous
non-compliance and provided that there is a prospect that unilateral action will
effectively induce compliance.
To the large trading powers, such reprisals will hardly be an option. WTO-autho-
rized trade sanctions will generally be the most effective means of instigating compli-
ance. If trade sanctions worth $4 billion annually cannot persuade the United States
to amend its Foreign Sales Corporations subsidies legislation, it would be surprising if
the suspension of a bilateral treaty could. By contrast, for smaller and developing
countries, recourse to countermeasures outside the WTO system may be a necessary
last resort. If a large trading nation fails to comply with a WTO ruling, the temporary
non-performance of a non-WTO treaty may well strengthen the position of a smaller
injured state. General international law may, thus, be a useful tool in the hand of a
smaller or developing country to address what is generally perceived as a deplorable
asymmetry in the WTO’s enforcement process. As the arbitrators in the Bananas dis-
pute noted with respect to Ecuador and the EC:
One may ask whether this objective [of encouraging compliance] may ever be achieved in a
situation where a great imbalance in terms of trade volume and economic power exists
between the complaining party seeking suspension and the other party which has failed to
bring WTO-inconsistent measures into compliance with WTO law.196
To give only a rough indication of the economic asymmetry involved, the volume of
trade in goods of the EC, the potential target of Ecuadorian retaliation, exceeded Ecua-
dor’s trade volume by a factor of two hundred.
The question of monetary compensation remains. The WTO Agreements con-
tain no functional equivalent of retrospective reparation. Can states, conse-
quently, fall back on the rules of state responsibility and demand restitutio in
integrum and monetary compensation, either in panel proceedings or outside the
dispute settlement machinery? Member states have argued partly in favour of,
195
ILC, supra note 9, at 357; cf. also ibid., at 330. Critical also Pauwelyn, supra note 8, at 233.
196
Para. 73.
partly against, an obligation to make reparation for injury.197 Academic literature is

similarly divided.198 WTO panels have shown an inclination to recognize an obliga-
tion to grant restitutio in integrum, where restitution was feasible. This was particu-
larly the case with respect to illegal subsidies199 (which the beneficiary could pay
back) and unlawfully imposed anti-dumping duties200 (which the collecting state
could reimburse to the unjustly targeted company).
For the bulk of trade-restricting measures, however, restitution is not a feasible
option.201 Monetary compensation, by contrast, raises a number of problems. Which
damages would states claim? Claiming losses of nationals would introduce an ele-
ment of diplomatic protection, which may be alien to the ‘objective’ trade regime.202
Moreover, total damages would easily amount to billions if the breach extends over
several years. Had the negotiating states during the Uruguay round envisaged an
obligation to make such retrospective compensation, they would have defined the
terms of operation of such a massive reparation regime (calculation of damages, stat-
utory limitation, etc.). Given the current state of WTO law, it seems reasonable to
infer that the treaty-law mechanisms of Article 22 DSU are indeed exhaustive when it
comes to monetary compensation.
With the introduction of ‘suspensions of concessions’ as a countermeasures equival-
ent, the WTO system has indeed ‘moved decisively in the direction of . . . a self-contained
regime’.203 The object and purpose of the DSU do not permit a state to have parallel
recourse to claims for compensation or countermeasures under the general international
law of state responsibility. As an ultima ratio in cases of continuous non-compliance, how-
ever, countermeasures outside the WTO framework remain an option. Thus, arguably,
not even the WTO is completely decoupled from the secondary rules governing the con-
sequences of breaches under general international law.204
197
In favour or reparation, e.g., Ecuador, WT/DSB/M/89, para. 8. Against reparation Australia, the EC, the
US and Canada, WT/DSB/M/75, 5 et seq.
198
In favour: Mavroidis, supra note 181, at 765; Pauwelyn, supra note 8, at 226. Against: Crawford, supra
note 87, at para. 420.
199
Australia – Subsidies Provided to Producers and Exporters of Automotive Leather, Panel Report, 21 January
2000, WT/DS126/RW, 6.48. Even in subsidies cases it is dubious whether restitutio is always a feasible
option. In the Foreign Sales Corporations case, e.g., the US would potentially be under an obligation to
demand thousands of recipients of FSC tax credits to pay unjustly credited taxes for a period of two dec-
ades. Such an obligation would almost certainly conflict with the prohibition of retroactive legislation of
most states’ constitutions.
200
Guatemala – Cement II, Panel Report, 17 November 2000, WT/DS156.
201
Since it is ‘materially impossible’ or involves ‘a burden out of all proportion to the benefit deriving from
restitution instead of compensation’, Article 35(b) of the ILC Draft Articles on Responsibility of States for
Internationally Wrongful Acts, supra note 9.
202
Petersmann, ‘GATT Dispute Settlement Proceedings in the Field of Anti-Dumping Law’, 8 CMLR (1991)
101; Marschik, ‘Too Much Order? The Impact of Special Secondary Norms on the Unity and Efficacy of
the International Legal System’, 9 EJIL (1998) 212 at 226.
203
Kuyper, supra note 177, at 252.
204
Cf. also Marceau, ‘WTO Dispute Settlement and Human Rights’, 13 EJIL (2002) 753 at 766 et seq. Perez
has suggested that countermeasures may, furthermore, be necessary if a state relies in an abusive manner
on the ‘self-judging security exception’ of Article XXI GATT (Perez, ‘WTO and U.N. Law: Institutional
524 EJIL 17 (2006), 483–529
D Treaties for the Protection of Human Rights

The internationalization of human rights has turned the state ‘inside out’. To the
extent that the treatment of individuals has been elevated to the level of international
law, an important area of state regulation has been carved out of the traditional
domaine reservé. The trend towards internationalizing the protection of human rights
has led to the adoption of various international conventions such as, at the regional
level, the European and American Conventions on Human Rights (ECHR, ACHR) and
the Banjul Charter; the two ‘global’ Covenants on Civil and Political Rights (ICCPR)
and on Economic, Social and Cultural Rights (ICESCR); and sectoral human rights
regimes such as the Convention on the Elimination of all Forms of Racial Discrimina-
tion (CERD).
All these treaties spell out particular enforcement mechanisms. However, given the
centrality of human rights in 21st-century international relations, it is not surprising
that the spirit of human rights has transcended these specific instruments, entering
the formerly state-oriented area of ‘general’ international law. For instance, the rules
on state responsibility codified in 2001 contain a specific regime on responsibility for
breaches of obligations in the collective interest. The crucial question in the context of
this paper is: To what extent do these rules on state responsibility apply to breaches of
obligations under a human rights treaty?
The International Court’s Nicaragua judgment illustrates in an exemplary way the
oscillation of legal argument between particularity and unity. On its face, the judg-
ment appears to suggest that remedies under human rights treaties should be consid-
ered exclusive (acknowledgment of particularity).
[W]here human rights are protected by international conventions, that protection takes the
form of such arrangements for monitoring or ensuring respect for human rights as are pro-
vided for in the conventions themselves.205
However, a closer look qualifies the hasty conclusion of exclusiveness. Remedies

under human rights treaties, according to the Court, are not exclusive per se. Rather,
the Court found that the special treaty mechanisms excluded unilateral enforcement
mechanisms under general international law because ‘the mechanisms provided for
have functioned’.206 Bearing this qualifier in mind, the International Court of Justice
Comity in National Security’, 23 Yale J Int’l L (1998) 310, at 324). In our view, however, the security
exception should be subject to limited panel review. Applying a good faith test, which would require that
the measure taken must appear at least somewhat reasonable and understandable from an objective
point of view, seems the most balanced and practical option. It takes account of states’ interests to avoid
presenting a detailed case on what is a national security matter, while maintaining a certain level of
transparency and accountability (cf. the overview in Pulkowski, ‘Coalition Procurement for Iraq in the
Crosshairs of WTO Law’ 5 German LJ (2004) 257, at 279). This approach was essentially developed by
Schloeman and Ohlhoff, ‘‘Constitutionalization’ and Dispute Settlement in the WTO: National Security
as an Issue of Competence’, 93 AJIL (1999) 424, at 444. Consequently, there is no remedial law deficit
as diagnosed by Perez.
205
Military and Paramilitary Activities in and against Nicaragua (Nicaragua v. United States of America),
Merits, ICJ Reports (1986), 134.
206
Ibid.
turns to a unitary counter-argument, with a view to preventing a potential ‘effective-

ness gap’ in human rights treaties: treaty mechanisms exclude the rules on state
responsibility as long as they function effectively. Where this is no longer the case,
states may have recourse to the remedies provided for under the general law of state
responsibility. Once the exclusiveness theory has been thus discarded, we can turn to
two familiar questions: Which of the mechanisms under human rights treaties are
leges speciales to the rules on state responsibility in the first place? And how far does
their specialness extend?
Human rights treaties have created innovative procedures for monitoring and
enforcing compliance with the obligations under the convention. Such procedures
include periodic reporting, which forces parties to justify their human rights record
before treaty bodies, individual complaint procedures as well as political and judicial
inter-state proceedings. To what extent are these procedures leges speciales to the rules
on state responsibility? Conceptually, the responsibility of states arises irrespective of
who has the right to invoke it and in which fashion.207 The various procedures have
no bearing on the question whether a state is responsible for a breach. Instead, they
operate on the level of invocation of responsibility, dealing with the question of who
may present a claim against the responsible state and request which remedies. Thus,
they compete with Articles 42 to 48 of the ILC’s codification.
However, most of these procedures do not constitute leges speciales to Articles 42 to
48. Individual claims procedures are not ‘concerned with the same substantive mat-
ter’ as the rules of state responsibility – to pick up on Fitzmaurice’s definition of special
rules. The scope of the ILC Articles is limited to the right of states to invoke the respon-
sibility of other states. They have no bearing on the question whether, and under
which conditions, individuals are entitled to present claims or to request remedies.
Vice versa, the mere existence of an individual claims procedure cannot warrant the
conclusion that inter-state invocation is consequently precluded. Rather, the invoca-
tion of state responsibility then rests on two pillars.208
Moreover, reporting procedures cannot be deemed leges speciales vis-à-vis the rules
on state responsibility. The Articles on state responsibility are concerned with the
legal consequences of concrete breaches. Reporting procedures before treaty bodies
are not. Their function is to provide a comprehensive critique of the human rights
situation in a particular member state. They are designed to help improve the overall
human rights record of states by using an innovative mix of judicial assessment akin
to a tribunal, political pressure and legal assistance. It is true that both reporting pro-
cedures and inter-state invocation of responsibility for breaches of human rights
serve the overarching goal of improving compliance with human rights standards.
However, they involve different actors and operate with different techniques of
207
The distinction between responsibility and its invocation underlies the ILC’s Articles. According to Arti-
cle 1, every internationally wrongful act of a state entails that state’s responsibility. Article 2 defines as a
wrongful act any conduct that is attributable to a state and constitutes a breach of an international
obligation. ILC, supra note 9.
208
Cf. K. J. Partsch, Rassendiskriminierung. Die UN-Konvention und ihre Wirkungsweise (1971) 14.
526 EJIL 17 (2006), 483–529
persuasion. Moreover, reporting procedures do not result in an order of specific reme-

dies. Thus, such procedures complement the right to invoke state responsibility
rather than replacing it. The fact that reporting procedures do not involve ‘the same
substantive matter’ as inter-state complaints is reaffirmed by the conventions them-
selves. Where inter-state complaint procedures are available, such procedures exist
alongside a periodic reporting system.
To reach a first conclusion: only state-to-state procedures of a judicial character
that address concrete violations of the treaty are leges speciales to the rules on invoca-
tion of state responsibility. To the extent that a human rights treaty contains such
procedures for inter-state claims, states are barred from invoking the responsibility of
another state through other channels. The decision-making power of the respectively
competent body must not be bypassed by unilateral auto-interpretation. Where no
such procedures are provided, states are free to invoke the responsibility of a violator
state pursuant to the provision of Article 48 of the ILC’s Articles as a ‘State other than
an injured State’. Such a state can claim only a limited range of remedies including
cessation, and guarantees of non-repetition. Moreover, it may request reparation in
the interest of the persons whose rights were violated.
So far so good: What if a state fails to cease an ongoing violation of human rights or
denies the victims of a human rights violation adequate reparation? The question of
enforceability of human rights has remained most controversial – so much so that it
almost jeopardized the adoption of the final Draft Articles by the ILC during its 2001
session. Human rights treaties do not contain special provisions regarding the right of
states to enforce obligations under these treaties. In accordance with the proposition
outlined above, it could be easily concluded that states have a right to ensure that
international law be respected, including a restricted right to unilateral enforcement
action (countermeasures). However, it has been suggested that ‘human rights are dif-
ferent’ and, therefore, decoupled from the general regime of legal consequences of
internationally wrongful acts. The arguments advanced in support of such a conten-
tion fall within two principal streams. The first argument is a teleological one: unilat-
eral enforcement of human rights may create ‘disorder’ in international relations,
especially in light of the fact that such enforcement rights are prone to political
abuse.209 The second argument is a doctrinal one. Given the structural non-reciproc-
ity of human rights obligations, it is being argued, bilateral inter-state enforcement
through countermeasures would be systematically flawed.210
Human rights obligations have been characterized as ‘objective regimes’.211 As the
European Court of Human Rights stated in the Northern Ireland case:
209
Frowein, ‘Die Verpflichtungen erga omnes im Völkerrecht und ihre Durchsetzung’, in Festschrift für
Hermann Mosler (1983) 242, at 255; Koskenniemi, ‘Solidarity Measures: State Responsibility as a New
International Order?’, 72 BYbIL (2001) 337, at 347.
210
Cf. the references cited in B. Simma, Das Reziprozitätselement im Zustandekommen völkerrechtlicher
Verträge (1972), at 176 et seq.
211
Riphagen, supra note 134, 17, para 89.
Unlike international treaties of the classical kind, the Convention comprises more than mere
reciprocal engagements between contracting States. It creates, over and above a network of
mutual, bilateral undertakings, objective obligations which, in the words of the Preamble,
benefit from a ‘collective enforcement’.212
In our view, the Court’s statement should not be interpreted as suggesting that
human rights are of a non-reciprocal character. While human rights have an objec-
tive, public-law-like, perhaps even constitutional,213 character, technically, they
nonetheless formally remain ‘reciprocal engagements between contracting States’.214
It is crucial to distinguish between reciprocity as a formal characteristic of a norm on
the one hand, and reciprocity as a substantive do-ut-des relationship on the other.
Human rights treaties do not involve such a substantive exchange, since their ulti-
mate beneficiaries are individuals under the jurisdiction of the state undertaking the
obligation. However, since human rights remain ‘mutual, bilateral undertakings’
owed to the other state parties to the respective convention, there is no compelling
systematic reason why states should be precluded from bilateral enforcement of
human rights.215 In Michel Virally’s words: ‘Chaque Etat partie à un tel traité a le droit
d’exiger des autres parties qu’elles respectent ces engagements, pris envers lui-même.’216
As to the second strand of argument, are countermeasures prone to creating par-
ticular disorder if they are employed with the purpose of inducing compliance with
human rights obligations? From a developed country perspective, Jochen Frowein
expressed the concern that unilateral enforcement would create considerable legal
uncertainty.217 Koskenniemi added (with a surprising touch of Morgenthau-style
realism) that elevating the enforcement of human rights from politics to the level of
law ‘would vest diplomacy with pressures and expectations it cannot carry’.218
Authors from developing countries, by contrast, stress the danger of political coercion
of developing countries by rich and powerful states.219 While it is true that counter-
measures for the enforcement of human rights have so far remained a domain of
212
Ireland v. UK, European Court of Human Rights, Ser. A Vol. 25 para. 239.
213
In its decision in Chrysostomos and others, the European Commission spoke of a ‘constitutional instru-
ment’, 34 Ybk European Commission on Human Rights (1991) 35, at 52 et seq.
214
Cf. generally, Simma, supra note 123; B. Simma, ‘From Bilateralism to Community Interest in Interna-
tional Law’, 250 RdC (1994, VI) 217, at 373.
215
Provided, of course, that the human rights treaty’s rules on invocation of responsibility have been
exhausted, cf. supra Sections 3C and 3D.
216
M. Virally, ‘Le principe de réciprocité dans le droit international contemporain’, 122 RdC (1967, III) 1, at
26. ‘Every State Party to such a treaty has the right to require the other parties to respect these commit-
ments, which have been made to the State itself.’
217
Frowein, supra note 209, at 256. Frowein’s analysis is strongly influenced, however, by the conflicting
interpretations of human rights by capitalist countries on the one hand and socialist countries on the
other hand. His critique may be largely obsolete after the collapse of the East Bloc.
218
Koskenniemi, supra note 122, at 347. However, the focus of Koskenniemi’s critique is the attempt to cod-
ify countermeasures in the collective interest, rather than the existence of such measures.
219
Cf. the discussion in Crawford, supra note 86, Add.4, 16 and 18, who concedes that there is ‘a due pro-
cess issue’ from the perspective of the targeted state if a few powerful states decide to act as a kind of
human rights police for the world, based on their auto-interpretation of human rights law; cf. Koskenniemi,
supra note 131, at 347.
528 EJIL 17 (2006), 483–529
‘Western’ industrialized states,220 Christian Tomuschat has argued that the regime of
countermeasures ‘does not place new States in a position of inferiority’,221 pointing
out as an example that developing countries could freeze foreign assets to induce
compliance.
The ILC drafting committee had proposed in the year 2000 a provision pursuant to
which ‘any State may take countermeasures, in accordance with the present Chapter
in the interest of the beneficiaries of the obligation breached’.222 To accommodate the
concerns of developing countries, the draft contained various safeguards to prevent
political abuse, including a provision that would restrict such countermeasures to
‘gross and systematic’ violations of human rights obligations.223 However, the pro-
posal was received critically by a majority of states in the Sixth Committee of the Gen-
eral Assembly. Thus, in its 2001 session, the ILC sought a compromise that would
eliminate the risk that the Assembly would not ‘pass’ the Draft Articles. The final
Article 54, a general saving clause, is an attempt to leave the question open without
prejudicing the evolution of a future opinio juris regarding the permissibility of coun-
termeasures in the collective interest.224
In our view, the ILC draft of 2000 may still provide valuable guidance as to what
the evolving law on countermeasures in the collective interest could look like.
There is no return to an international law that puts on an indifferent face to human
rights. Human rights can no longer be fenced in an exclusive domaine reservé; once
their genie was out of the bottle, human rights necessarily transcended to the realm
of general international law. As Reisman put it, human rights are ‘more than a
piecemeal addition to the traditional corpus of international law’. They bring about
‘changes in virtually every component’.225 Against this backdrop, it is only logical
not to deny states the means to induce compliance with obligations under human
rights treaties, once the collective enforcement mechanisms of the treaty have
failed.226 The concerns raised by some authors (and states) neglect the realities of
human rights protection: so far, states have hardly shown the excessive human
220
Cf. the cases compiled in the ILC commentary on Article 54, supra note 11, at 351 et seq., and the cases
compiled by K. Weschke, Internationale Instrumente zur Durchsetzung der Menschenrechte (2001), at 98 et
seq.
221
Tomuschat, ’Are Counter-measures Subject to Prior Recourse to Dispute Settlement Procedures?’, 5 EJIL
(1994) 77, at 78.
222
Article 54 of the Draft Articles provisionally adopted by the Drafting Committee on second reading, A/
CN.4/L.600.
223
Article 41 of the Draft Articles provisionally adopted in 2000 restricted countermeasures to wrongful
acts ‘that constitute a serious breach’, meaning a ‘gross and systematic failure by the responsible State to
fulfil its obligation, risking substantial harm to the fundamental interests protected thereby’. Draft Arti-
cles 51 to 53 spelled out limitations similar to 49 to 53 of the ILC Articles on Responsibility of States for
Internationally Wrongful Acts, supra note 9.
224
Cf. Simma, ‘Staatenverantwortlichkeit und Menschenrechte im ILC-Entwurf 2001′, in J. A. Frowein,
K. Scharioth, I. Winkelmann, R. Wolfrum (eds), Verhandeln für den Frieden – Negotiating for Peace, Liber
Amicorum Tono Eitel (2003) 423.
225
Reisman, ‘Sovereignty and Human Rights in Contemporary International Law’, 84 AJIL (1990) 866, at
872.
226
Concurring, L. A. Sicilianos, Les réactions décentralisées à l’illicité (1990) 167; Weschke, supra note 220.
rights ‘vigilantism’227 dreaded by some. Far from obsessively policing human rights
violations across the world,228 the attitude of states towards human rights violations
is all too often characterized by a remarkable lack of vigour to counter such treaty
breaches.
5 The Dialectics of ‘Multiple Unity’

While international law has in many respects transformed ‘de la pyramide au réseau’,229
from a small but organized body of general rules to a spread-out web of normativity, this
transformation has been premised on the stability of the systemic framework of the
international system as a whole. Unity and multiplicity, stability and change mutually
condition each other. On the one hand, new special regimes of international law have
evolved as a response to insufficiencies or the perceived lack of flexibility of the general
system. On the other hand, the very same regimes that have thus emancipated them-
selves from general international law remain dependent on this body of rules in a
number of ways, including, as a measure of last resort, the enforcement of primary obli-
gations. In other words, the fragmentation of the international legal order requires a
simultaneous affirmation of its central systemic building blocks, including general
international law. As Matthew Craven put it, ‘[a]s a diachronic narrative, therefore,
progressive systematization appears in tension with that of fragmentation. . . . A system
is a system by reason of its ability to unify a diversity of experience – it must be simulta-
neously homogeneous and heterogeneous, unified but multipolar.’230
Legal practice involves operationalizing the dialectics inherent in the multiple
unity of international law. Jurists need to acknowledge the particular logic of a vari-
ety of special regimes. The sophisticated institutional arrangements of such regimes
will often render the expertise of the general international law practitioner largely
futile. At the same time, however, general international law provides a systemic fab-
ric from which no special legal regime is completely decoupled. And, by framing a
prescription in legal terms, states have opted to subordinate a particular issue to the
logic of international law as a whole. Framing policy choices in terms of international
treaties inevitably involves a shift from the ‘ethos of the diplomat’ to the ‘rule of
the lawyer’.231 General international lawyers should not be too reluctant to assert
their role.
227
McCaffrey, ‘Lex Lata or the Continuum of State Responsibility’, in J. H. H Weiler, A. Cassese and M. Spinedi
(eds), International Crimes of State (1989) 244.
228
For instance, in the history of the European Convention, states have generally lodged inter-state com-
plaints almost exclusively in situations where their interests were specially affected. The two exceptions
were the application against Greece in 1967 (11 Yearbook European Commission on Human Rights (1968)
691; 12 (II) Yearbook European Commission on Human Rights (1969)), and the application against Turkey
in 1982 (applications no. 9940-9940/82).
229
Cf. the title of a recent book by Ost and van de Kerchove, supra note 5. ‘From the pyramid to the network’.
230
Craven, ‘Unity, Diversity and the Fragmentation of International Law’, XIV Finnish Yearbook IL (2003)
3, at 12.
231
To play on a theme once developed by Weiler, supra note 178.

Of Planets and The Universe: Self-Contained Regimes in International Law

Uploaded by

Copyright:

Available Formats

Of Planets and The Universe: Self-Contained Regimes in International Law

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Of Planets and The Universe: Self-Contained Regimes in International Law

Uploaded by

Copyright:

Available Formats

The European Journal of International Law Vol. 17 no.

3 © EJIL 2006; all rights reserved

Of Planets and the Universe:

1 Self-contained Regimes: A Classic Revisited

2 Special Rules in the Field of State Responsibility

Efficacy, for de Vattel, is a consequence of fewer exceptions, more regulatory preci-

From a realist perspective, Schwarzenberger thus referred to the principles of

B From leges speciales to So-called Self-contained Regimes

First, the notion of ‘self-contained regimes’ has been misconceived as an argument

Without the ‘omnipresence of “general law”’33 a special legal subsystem may, as

C Concepts of the International Legal Order

consequence, the international lawyer is sometimes bound to accept systemic incon-

While international law certainly is systemic, it does not necessarily constitute a

2 Self-contained Regimes in a Fragmented Legal Order

inherent in conventional approaches, Fischer-Lescano and Teubner describe the pro-

fragmentation’.97 If all international law necessarily follows the logic of a particular

D Of Planets and the Universe: The Level-of-Analysis Problem

The international legal system, largely dependent on ‘horizontal’ processes of law-

The disparity of the conclusions of ‘specialists in community law’ and ‘scholars of

3 The Fallback on General International Law

B Effective Treaty Interpretation and the ‘Fallback’ on General

mechanisms, they merely relinquish their facultés under general international

C Scenarios for a Fallback on State Responsibility

D A Broader Picture: Negotiating the Balance of Effectiveness

4 Case Studies on Self-contained Regimes

‘wholly artificial’ and ‘excessively formalistic’:156 in short, disproportionate in individ-

B The European Communities Legal System

the damage caused by French non-compliance with a Court judgment resulted

characteristic of obligations that operate on a do ut des basis. Many of the substantive

C The WTO System

paradoxical at a first glance. In reality, as Weiler has put, it ‘[j]uridification is a pack-

Most contentious, however, is the question whether a fallback on reprisals

partly against, an obligation to make reparation for injury.197 Academic literature is

D Treaties for the Protection of Human Rights

However, a closer look qualifies the hasty conclusion of exclusiveness. Remedies

turns to a unitary counter-argument, with a view to preventing a potential ‘effective-

persuasion. Moreover, reporting procedures do not result in an order of specific reme-

5 The Dialectics of ‘Multiple Unity’

You might also like