EXCEPTIONS IN INTERNATIONAL LAW
Jaap Hage, Antonia Waltermann, Gustavo Arosemena1
University of Maastricht
Abstract
Exceptions to rules play an important role in law, and in particular in international law. A proper
understanding of exceptions is therefore of crucial importance for legal practice, legal doctrine and
legal theory. The aim of this article is to contribute to this understanding; it is not to describe the law
or the present usage with regard to rules and exceptions.
As background for the development of a theory about exceptions to rules, sections 3 and 4 go into
some detail concerning rules, rule-formulations, legal sources, reasons and their logic. In section 5,
we will elaborate on the distinction between the applicability and the application of a rule. A rule is
applicable to a case if the rule is valid, and if its ordinary and scope conditions are satisfied by the
case. If a rule is applied to a case, the rule attaches its legal consequences to the facts of the case.
Normally the applicability of a rule to a case is a contributory reason why the rule should be applied
to the case. An exception to a rule in a case is defined as the situation in which a rule is applicable to,
but nevertheless not applied to the case.
In section 6, two main grounds for making an exception to a rule are identified. First, the maker of
the rule may use the rule-exception construction to create a division in the burden of proof. As
argued in section 9, this leads to a so-called undercutting defeater. Second, there may be reasons
why the legal consequences of the rule in the case are undesirable. This leads to a reason against
applying the rule, which needs to be balanced to the applicability of the rule as reason for
application. In this situation we speak of rebutting defeaters (see section 9.2).
An important reason why it is undesirable to apply a rule to a case is that application would generate
a conflict with another applicable rule. Section 8 discusses a number of tools/techniques that can be
used to avoid rule conflicts and which would in that way make exceptions superfluous.
Finally, the question may be raised whether legal rules really have exceptions. Is it not the case that if
a rule is well-fo ulated, it e tio s all e eptio s as egati e ule o ditio s? “u h a ellfo ulated de i ed ule a the e applied dedu ti el to ases that satisf the ule o ditio s. In
sections 10 and 11, this possibility is discussed and it is shown that this use of derived rules blocks
the possibility to model the division of burden of proof which is implicitly given with the ruleexception model. The article is summarized and concluded in section 12.
Keywords: applicability, application, conflict of compliance, conflict of imposition, derogation,
exception, incorporation, interpretation, reason, rebutting defeater, reference, rule, scope,
undercutting defeater
1.
Introduction
The exception proves the rule is a commonplace statement in every day language. Logically
speaking, matters are more complicated than that, however: if a u i e sal state e t su h as All
*
The authors thank the participants in the Cambridge seminar on Exceptions in International Law for useful
comments on a presentation of an earlier version of this contribution.
1
o ls a fl has a e eptio , this ea s that the state e t is false because it turns out that not
all owls can fly after all. In legal theory, meanwhile, rules attach consequences to fact situations, and
the t pi all do so he thei o ditio s a e satisfied
a fa t situatio a ase . “o eti es,
however, they do not attach consequences to a case even though the case satisfies the rule
conditions: the rule is not applied to the case. Then we speak of an exception to the rule. Such an
exception, as long as it is exceptional, does not make the rule false or – better - invalid. The rule
remains valid, but if there is an exception to it, the rule is not applied.
1.1
Exceptions in international law
Rules take a central place in law, and accordingly, rule exceptions play an important role. This is
particularly the case in international law, because international law has many features that make
conflicts of rules frequent and recalcitrant, and rule conflicts are a major reason for exceptions.
There are several reasons why international law is ripe for rule conflict. Firstly, international law is
fragmented : it is not clear whether it constitutes a single legal system, or an archipelago of mutually
interacting international legal regimes. This, together with the growing interaction between
international, domestic, and regional legal orders, implies that any discussion of rule conflicts in
international law will have to account for conflicts across different legal orders (Koskenniemi & Leino
2002). Secondly, the importance of the formal sources of international law has decreased: the list of
formal sources has little power to control what counts as law and what does not. In fact, many
academics suggest that having a clear view of international law requires us to bypass the doctrine of
sources and identify as law those rules that have real world effectiveness and/or political legitimacy
(cf. Franck 1990 and Arend 1999). The resulting obscurity regarding what counts as international law
is also a potential cause of rule conflicts. Thirdly, and related to the second point, international law
operates without a central legislator. Treaties and customs are generated through the agency of
more or less uncoordinated actors, which means that no central authority can attempt to minimize
rule conflicts or prevent them from arising.
Apart from these three reasons which apply especially to international law, there is also the wish of
the rule creator - e.g. the parties to a treaty - to divide the burden of proof, which can underlie the
existence of rules with exceptions (see sections 6 and 9.1). In short, international lawyers as well as
academics will inevitably be confronted with exceptions in international law, whether these
exceptions have their grounds in international, domestic or other (possibly non-legal) sources.
1.2
Aim
Because exceptions play a crucial role in (international) law, a thorough understanding of what goes
on when we make exceptions to rules is an essential precondition for international legal science. Our
purpose in this contribution is to provide clarity with regard to the nature of exceptions to rules, in
particular as concerns international law, by creating a precise conceptual framework in which
important notions are interconnected. As will become clear from our argument, this conceptual
framework must account for exceptions, rules, their applicability and application, and shifts in the
u de of p oof hi h, as e ill a gue, a k off e eptio s to ules f o
e e egati e ule
conditions.
Two caveats are in place here. The first one is that a conceptual framework is - and should be neutral with regard to the content of the law. Conceptual jurisprudence, the theory according to
2
which the content of the law is to some extent determined by the concepts used in creating and
describing that content, has rightly attracted much criticism (Marx 1977). The flipside of this
neutrality is that readers should not expect that the conceptual framework that is developed here
can provide them with the solutions for legal cases. Clear concepts can contribute to clear legal
thinking, but it is the substantive law itself, and not the concepts by means of which law is created or
described, that provides cases with their solutions.
The second caveat is that, in their standard usage, the concepts and the terms used to express them
are not always very clear and that a theory which aims to provide clear concepts and precisely
defined terms cannot and should not be in complete accordance with actual usage. A copy of how
words are used in actual legal discourse would also copy all the ambiguities and vagueness of this
discourse. The conceptual and terminological proposals of this contribution should therefore be
judged on their usefulness for the production of legal science, and not – at least not in the first place
– on their conformity with the actual practice. Our aim is to improve, not to describe.
1.3
Roadmap
As mentioned, the purpose of this contribution is to provide clarity with regard to the nature of
exceptions to rules, in particular as concerns international law, by creating a precise conceptual
framework in which important notions are interconnected. In order to do so, we will consider the
notions of rules and reasons in sections 3 and 4 respectively. However, before doing so, we will
present two examples that will often be referred to in our argument in section 2.
Section 5 will consider the applicability and application of rules, two notions that are highly relevant
to our definition of an exception. Section 6 takes a closer look at the grounds for making exceptions
to rules. One of the main grounds for making exceptions is that rules conflict with each other. Section
7 distinguishes two ways in which rules can conflict, and section 8 discusses a number of ways in
which rule conflicts can be avoided, thereby taking away the need for exceptions to rules.
In section 9, we take a closer look at the shape in which exceptions to rules can occur. Sections 10
and 11 are devoted to the alleged necessity of allowing exceptions to rules. Section 10 discusses a
technique to make exceptions superfluous, while section 11 argues that exceptions are unavoidable
if one wants to maintain the possibility to divide the burden of proof in legal argumentation.
The argument of this contribution is summarized in section 12.
2.
Recurring examples
We will use two examples repeatedly in the arguments to follow. In this section, we will briefly
introduce these recurring examples and the provisions on which they rest.
2.1
Prohibition of force
The first example concerns the use of force in international law. Article 2, Section 4 of the UN Charter
prohibits the use of force:
All Me e s shall ef ai i thei i te atio al elatio s f o the th eat o use of fo e agai st
the territorial integrity or political independence of any State, or in any other manner inconsistent
ith the Pu poses of the U ited Natio s.
3
An exception to this general prohibition of force can be based on Article 42 of the Charter:
“hould the “e u it Cou il o side that easu es p o ided fo i A ti le 41 ould e
inadequate or have proved to be inadequate, it may take such action by air, sea, or land forces as
may be necessary to maintain or restore international peace and security. Such action may
include demonstrations, blockade, and other operations by air, sea, or land forces of Members of
the U ited Natio s.
2.2
State responsibility
The second example concerns the responsibility of States to compensate damage. This topic is dealt
with by the draft Articles on Responsibility of States for Internationally Wrongful Acts (ARSIWA). The
establishment of this responsibility is dealt with by Article 1:
Every internationally wrongful act of a State entails the international responsibility of that State.
Article 2 provides the main rule defining what counts as an internationally wrongful act:
The e is a i te atio all
omission:
o gful a t of a “tate
he
o du t o sisti g of a a tio o
(a) Is attributable to the State under international law; and
(b) Constitutes a breach of an international obligation of the State.
Exceptional circumstances may make that the breach of an international obligation nevertheless
does not count as an internationally wrongful act. One of those is when a State acted out of selfdefense. Article 21 ARSIWA reads:
The wrongfulness of an act of a State is precluded if the act constitutes a lawful measure of selfdefence taken in conformity with the Charter of the United Nations.
Lastly, article 31 ARSIWA defines what the State responsibility brings about. It holds that:
1. The responsible State is under an obligation to make full reparation for the injury caused by
the internationally wrongful act.
2. Injury includes any damage, whether material or moral, caused by the internationally wrongful
act of a State.
3.
Rules
Before we turn to exceptions to rules, we need to have a solid understanding of what rules are in the
first place. However, rules – or, as some prefer, norms – are hard to characterize.2 They may be seen
as so ial a tefa ts hi h a date, p ohi it, o pe it e tai fo s of o du t. Follo i g Ha t s
distinction of primary and secondary rules (Hart 2012, p. 91-99), the behavior in question can be the
conduct of human affairs generally, or conduct dealing with the creation, modification, application
and extinction of rules. Even this distinction is unsatisfactory, however, because some rules, rather
than govern behavior, constitute states of affairs. In this paper, we will be using the following
circumscription of rules: that rules are a kind of things which attach new facts to already existing
2
For a more extensive analysis than can be provided here, see Hage 2005, p. 159-202.
4
ones. These new facts exist purely because they are the result of rule application, and they are
constituted by the rule (Hage 2015a).
3.1
Kinds of rules
Different kinds of rules can be distinguished, such as
-
Counts as rules, which make that some things also count as (are) other things. For example,
the King of Belgium also counts as (is) the commander in chief of the Belgian army.
- Fact-to-fact rules, which attach new facts to other facts which exist simultaneously. For
instance, the government of a State has the power to conclude treaties on behalf of that
State.
- Dynamic rules, which attach legal consequences to the occurrence of an event. For instance,
if a State commits an internationally unlawful act, and thereby causes damage to another
State, the former State incurs at that moment the liability to compensate the latter State for
the damage it suffered.
Mandatory norms, which prescribe behavior, are either fact-to-fact rules, as when a State has the
duty to respect the human rights it recognized through conclusion and ratification of a human rights
treaty,3 or dynamic rules, as when a State is assigned the obligation to compensate the damage it
unlawfully caused to another State.
3.2
Rules and language
It is important to distinguish between on the one hand rules as entities which are to a large extent
language-independent4 and on the other hand rule formulations and the sources through which rules
are created - treaties, legislation, and judicial decisions - which are both necessarily languagedependent. A rule, as we use the concept here, is the connection between kinds of facts. For
instance, the rule that the King of Belgium counts as the commander in chief of the Belgian army
connects the fact that somebody is the King of Belgium to the fact that somebody is commander in
chief of the Belgian army. The rule is defined by its content, which consists of the kinds of facts the
rule connects. The rule about the Belgian king is defined by the fact that it connects facts of the kind
ei g the Ki g of Belgiu to fa ts of the ki d ei g the o
a de i hief of the Belgia a
.
This connecting rule may be formulated in any of the three official languages of Belgium – Dutch,
French and German – but they are all formulations of the same rule. Therefore all formulations of the
rule must mention the same rule conditions and the same rule conclusion, be it not necessarily in the
same language. We also see that the rule itself is independent from the rule formulation when
o side i g that the p ohi itio of sla e a e e p essed i
a diffe e t la guages, su h as la
es la itud está p ohi ida o sla e ij is e ode .
Equally, we see the difference between rule and rule source when considering that the prohibition of
slavery is, at once, brought about by Article 4 of the European Convention of Human Rights and
3
4
Notice that the rule imposing the duty to respect a particular human right came about as the result of a
dynamic rule, namely a rule of change in the Hartian sense (Hart 2012, p. 95/6), but that the rule itself is a
fact-to-fact rule, which attaches the fact that a State has a duty to respect the human right to the fact that
the State in question is party to the treaty.
Rules are only to a large extent language-independent, because every rule needs to have a formulation,
and this formulation requires a language to formulate the rule in. This means that the conditions and the
conclusion of a rule, although not linguistic entities themselves, must be expressible in some language.
5
Article 8 of the International Covenant on Civil and Political Rights, and is also part of customary
international law, and of ius cogens.5
Furthermore, the formulation of a rule is not identical to the official text by means of which the rule
was created: the formulation mentions only the conditions and the conclusion of the rule. One-onone correspondence is not required, as the example of Article 330 of the Treaty on the Functioning of
the European Union (TFEU) shows. This article, which is not subdivided into sections, is arguably the
source of four different rules. The conditions and conclusions of these rules can be formulated in
different ways, and in different languages, but to the extent that they all convey the same conditions
and conclusion, they are all formulations of that same rule. Moreover, none of the rule formulations
actually coincides with Article 330 TFEU.
We emphasize this distinction between rule formulations and the sources by means of which rules
are created, because interpretation is the step from a rule source to a rule formulation. Sometimes,
interpretation is a means to avoid rule conflicts or exceptions to rules, and if it is, it can fulfil this
function because it leads to a different rule (formulation) on the basis of the same source. This is
something we elaborate on in section 8.3. On this terminology, only sources for rules are interpreted,
but rules themselves never are.6 This also means that when we speak of exceptions, we mean
exceptions to rules, rather than something which takes place at the level of rule formulations.
4.
Reasons
Another basic notion that plays a central role in our theory about rule exceptions is the notion of a
reason. Perhaps even more than rules, reasons are the topic of an overwhelming amount of
literature, both in legal (e.g. Raz 1999 and Redondo 1999) and in ethical theory (e.g. Scanlon 2014
and Alvarez 2016). Apodictically brief, reasons might be circumscribed as facts that are relevant for
some conclusion. So there are two aspects to each reason:
1. the reason is a fact;
2. this fact is relevant for some conclusion.
4.1
Classification of reasons
Reasons can plead for or against different kinds of conclusions. For example, the fact that a witness
declared that Alice stole perfume from the shop is a reason – not necessarily a decisive one – to
believe that Alice is a thief. In this case, the reason is a reason to believe and the conclusion is that
somebody is justified in believing something.
The fact that the United Nations Security Council (UNSC) has authorized intervention in Libya is the
reason why (it was the case that) foreign troops were allowed to intervene in the Libyan civil war.
This reason is a reason why something is the case, a constitutive reason.
5
6
We assume here, for the sake of argument, that these sources all underlie the same rule. Of course, this
may be disputed, as it may be disputed for any concrete example. However, disputing the example amounts
to recognition of the point we want to make, namely that there may be more than one source for the same
rule.
The determination whether a particular fact situation falls under a rule depends on the conditions of the
actual rule – e.g. was this act internationally unlawful? – but this step in the application of a rule is better
labelled as classification of the facts than as interpretation of the rule. See Hage 1997, p. 95 – 97.
6
And, finally, the fact that a State is liable to compensate the damage of some other State, is a reason
for the International Court of Justice (ICJ) to sentence the former State to payment of damages. The
last example illustrates a reason for action, a reason for or against doing something.
4.2
Universalizability of reasons
Reasons are concrete facts, such as the fact that Alice stole perfume from a shop. Concrete fact can
only be reasons for particular conclusions if similar facts are also reasons for similar conclusions.
Authorization by the UNSC can only be a reason why armed intervention is permissible in a particular
case if authorization by the UNSC tends to be a reason that allows armed interventions in general.
This move from particular cases to more general statements can be expressed by saying that reasons
are universalizable, or – which boils down to the same thing – by saying that underlying every
concrete reason for a particular conclusion is a rule (or principle7) which makes facts of that kind into
reasons for that kind of conclusion.8
Every reason has an underlying rule, and that is important because the rules underlying reasons are
amenable to exceptions. If there is such an exception, a fact which normally would be a reason
exceptionally does not count as a reason anymore. For instance, the rule that thieves are punishable
makes the fact that somebody is a thief into a legal reason why this person is punishable. However, if
there is an exception to the rule, for instance because the person in question is only three years old,
the fact is not a legal reason for punishment. As we will see later (section 9 , this exclusion of a
reason (Raz 1999, p. 35-48) should be distinguished from the situation that the reason is outweighed
by some colliding reason.
The reasons discussed above are contributory reasons. This means that they plead for (pro-reasons)
or against (con-reasons) a particular conclusion, but are not necessarily decisive by themselves. If
there are contributory reasons both for and against a particular conclusion, these reasons need to be
eighed , o ala ed . I this o e tio , eighi g a d ala i g a e e e etapho s; hat eall
goes on is that some decision is made which set of reasons wins against the other set. Such a
decision, which can itself be based on one or more reasons, can be laid down as a premise of an
argument.
An example may illustrate this. The fact that the publication of a photograph is an exercise of the
freedom of press is a reason against prohibiting the publication. The fact that this publication would
violate the privacy of the person on the photo is a reason for prohibiting its publication. The fact that
the publication of the photograph does not serve a public interest is a reason why the contributory
reason based on privacy outweighs the contributory reason based on the freedom of press.
Therefore, the former reason outweighs the latter reason, and the court should prohibit the
publication of the photograph in question (cf. ECtHR 24/09/2004; application 59320/00; Hannover).
If the sets of reasons are in balance, which includes the situation that there are neither pro- nor conreasons, nothing follows, unless the burden of proof decides the issue. We will consider this in
section 11.
7
8
At this place we do not distinguish between rules and principles. A distinction between the two can well be
made, however. For instance, while rules can have exceptions, it does not make sense to speak of
exceptions to principles. More about the difference between rules and principles in Hage 1997, p. 110-113.
The universalizability of reasons is discussed more extensively in Hage 2015b, with references to earlier
literature.
7
5.
Applicability and application
We define an exception to a rule as the situation where a rule is applicable to a case, but is
e e theless ot applied to it. I this o e tio , appli a ilit a d appli atio a e te h i al te s
with a precise meaning, that we will explore in this section.
If a rule is applied to a case, it attaches its consequence to the case. So if the rule which makes a
State responsible for internationally wrongful acts is applied to the case that Outopia breached an
international obligation, it makes Outopia responsible for that breach.
Applicability of a rule to a case is determined by three factors:
1. the rule must exist;
2. the case must fall within the – territorial, temporal, and personal – scope of the rule; and
3. the case must satisfy the ordinary conditions of the rule.
Let us talk about each of these factors, starting first with the requirement that a rule must exist. It
may seem obvious that a rule needs to exist in order to be applied, yet the existence of rules
becomes the subject of legal debate at times. In a legal dispute, one party may argue that a rule
invoked by the other party does not exist. This is usually formulated as a claim that the rule is not
valid, as a rule that lacks validity cannot be applied to any case. This is different for exceptions, which
only make that the rule is not applied to this particular case. There is a further difference between
the two arguments that has to do with the burden of proof; we will consider this in section 11.
Secondly, a case must fall within the scope of the rule in order for it to be applicable. Typically, rules
have certain scope conditions: the rule applies only on a particular territory, during a particular time
frame, or to particular persons. We will talk about scope conditions in more depth in section 8.2.
Thirdly, the case must satisfy the ordinary conditions of the rule for it to be applicable. These
conditions are given with the rule formulation. For example, if it were binding, Article 36 ARSIWA
would create a rule that defines a State s lia ilit fo particular damage. It mentions four conditions
for this liability:
1.
2.
3.
4.
there must have been an internationally wrongful act,
a State must have been responsible for this act,
the damage was caused by this act, and
the damage was not (yet) made good by restitution.
These four factors determine the applicability of the rule. At times, it happens that a rule is not
applicable to a case, but is nevertheless applied to it. This is most often a case of rule application by
analogy. We will not pay attention to this possibility here.9
If a rule is applicable to a case, this is a contributory reason for applying the rule to that case. Since
applicability is only a contributory reason for application of the rule, this reason may have to be
balanced against reasons against application. Because we define an exception to a rule as nonapplication of an applicable rule, reasons against the application of an applicable rule are ipso facto
reasons for making an exception to the rule.
6.
Grounds for exceptions
Before continuing our argument it is useful to take a step back and look at the grounds for making an
exception to a rule. We have defined an exception as the situation in which a rule is applicable to a
9
The interested reader may consult Hage 1997, p. 118 – 121.
8
case, but is nevertheless not applied to this case. There are several reasons for making an exception
to a rule in a particular case:
First, the rule-exception structure can be used in order to create a division in the burden of proof. For
instance, Article 2 ARSIWA defines what counts as an internationally wrongful act, while Article 21
says that this wrongfulness is precluded if, amongst others, the act constitutes a lawful measure of
self-defense taken in conformity with the Charter of the United Nations. Article 21 is meant to make
an exception to Article 2, and by using this rule-exception construction, the maker of these rules has
created a division in the burden of proof. The State that wants its damage to be compensated bears
this burden for the conditions of the rule defining international wrongfulness, while the State that
wants to avoid being held responsible has the burden of proof for the exception. We will return to
the relation between exceptions and the burden of proof in section 11.
Second, it may be deemed undesirable that the legal consequences of the rule come into being for a
particular case, although the rule is applicable to this case. There are three variants of this possibility:
a. Application of the rule in this case would violate the purpose of the rule. We return to this
possibility in section 9.
b. Application of the rule would lead to legal consequences that are incompatible with the
consequences of some other rule that is also applicable to the case. These so- alled ule
o fli ts ill e dis ussed i the se tio s 7 , 8 and 9.2.
c. Application of the rule in this case would harm values or goals that were not, or
insufficiently, taken into account when the rule was created. When the application of a rule
to a particular case has bad consequences, this seems at first sight to be a reason against
applying the rule, and therefore for making an exception to the rule. This may be different if
these bad consequences were sufficiently taken into account when drafting the rule, because
then, apparently, the bad consequences were deemed to be outweighed by the advantages
of the rule. This reason against applying an applicable rule is also discussed in section 9.2.
7.
Rule conflicts
As we mentioned in the previous section, rule conflicts, when they arise, are a frequent cause of
exceptions to rules. As such, exceptions are a tool to deal with, rather than to avoid, rule conflicts.
There are at least two different types of rule conflicts, namely conflicts of imposition and conflicts of
compliance. The former are conflicts whereby the conflicting rules impose incompatible states of
affairs upon the world,10 which cannot co-exist. The latter are conflicts whereby the conflicting rules
prescribe incompatible forms of behavior.
7.1
Conflicts of imposition
Consider the following scenario: a certain territory has been under the factual possession of State A
for some time now, despite a treaty between State A and State B agreeing that the territory falls
within the borders of State B. According to the principle of effective possession, this territory is
deemed to be part of State A. According to the principle of stability of borders, however, this
territory is deemed to belong to State B, reflecting legally agreed upon borders. Due to the nature of
10
We use the e p essio state of affai s to de ote a possi le fa t. A fa t is the a state of affai s that
actually obtains. Cf. Wittgenstein 1921/2, thesis 2 and Textor 2014.
9
territory, it cannot belong to both A and B at the same time, and so the effects of the rules are
incompatible.11
It is important here to note that the incompatibility of the consequences of these rules, and
therefore also the rule conflict in this case, depends on the existence of other rules, namely those of
how territory works and that territory cannot belong to two or more States at the same time.
Generally speaking, whether rules conflict can depend on a presupposed background which makes
that some facts cannot go together, that some states of affairs are incompatible with one another.12
Another example of a conflict of imposition is the following: the UN Charter prohibits the use of
force, but military action with authorization by the UNSC is permitted. One rule prohibits an armed
intervention into another sovereign State, while the other permits precisely this behavior. The state
of affairs that armed intervention is prohibited and the state of affairs that it is permitted are
incompatible, and therefore these rules are in conflict. Note that the focus here is on the states of
affairs imposed on the world: given that a permission is not a duty to intervene, there is no conflict of
compliance in this case.13
7.2. Conflicts of compliance
Consider a State which is obliged, on the basis of two human rights treaties it has signed and ratified,
to invest money in both education and health care, while the State has only sufficient money and
resources to invest in one of them. Alternatively, think of the example of the journalist who is obliged
to reveal her sources for a controversial publication, while she promised her informant not to reveal
his identity. In both cases, a conflict arises because the agent in question cannot comply with both
obligations.
Conflicts of compliance can arise because the obligations themselves are in conflict (such as a
prescription and a prohibition of the same behavior14), or because factual circumstances make
compliance impossible, as was the case for the State with too limited resources. These kinds of
conflicts can only exist between mandatory rules (prescription and prohibitions) and should be
avoided. When they occur, they force the obligated agent to choose which obligation to violate.
Conflicts of imposition, meanwhile, can exist between all rules, including those that assign statuses
(e.g. the status of a piece of land as belonging to one or the other State).
11
12
13
14
For a discussion of this type of conflict, see Frontier Dispute, Judgment, I.C.J. Reports 1986, p. 554.
In Hage 2000 and Hage 2005, p. 135-157 the theme is explored how rules themselves can be part of the
a kg ou d that auses ule o fli ts the e: ule i o siste
.
The conflict may seem to some to be a conflict of compliance, because the conflicting norms deal with
actions. However a conflict between a prohibition and a permission is not one of compliance. It is possible
to comply with both norms by abstaining from the prohibited behavior, since such abstinence does not
violate the permission. (Permissions cannot be violated at all.) However, there is a conflict of imposition,
since the states of affairs that some kind of action is prohibited and permitted are not compatible.
If the same kind of behavior is both mandatory and prohibited because of two conflicting rules, there is a
conflict of imposition, because the states of affairs that some kind of behavior is prohibited and mandatory
are incompatible. If a kind of action is a sub-kind of two different more general kinds of action, and one of
these more general kinds is mandatory while the other kind is prohibited, the conflict is one of compliance.
10
8.
Making exceptions superfluous
Exceptions are, amongst others, a way to deal with rule conflicts when they arise. In this section, we
will consider a number of tools and techniques that can keep such conflicts from coming into
existence in the first place, thereby taking away the need to make an exception in the first place. The
list of tools and techniques in the following is not meant to be exhaustive.
8.1. Subscripting
Rule conflicts can arise between rules of one system, or between rules of different normative
systems. Think again of the example of a journalist who promised not to reveal her source. This
example illustrates a conflict between a legal requirement to reveal, and a moral obligation not to do
so. The distinction between normative systems becomes visible in the necessity to add subscripts to
legal judgments. For instance, it is not the case anymore that State A may not use military force to
intervene in State B to prevent gross human rights violations. If normative systems are distinct, the
judgement must be that legally, State A is prohibited from intervening (in the absence of UNSC
authorization), but morally, it should. In this example, the subscripts distinguish between the legal
and the moral point of view. However, it is also possible to distinguish between different legal points
of view. For instance, according to European Union law, certain forms of positive discrimination are
not permissible, whereas according to the Committee on the Elimination of Discrimination against
Women, they are required (cf. Waddington and Visser 2012).15
If the legal and other judgments are subscripted, seemingly conflicting judgments are rendered
logically consistent because we see that for a legal decision, the moral rules are simply not
applicable. A legal permission is consistent with a moral prohibition, and a competence that exists
according to the law of one country may be absent according to the law of another country.
Nevertheless, an agent who is confronted with mandatory rules from different legal systems that
both demand obedience but the rules of which cannot both be complied with, is still burdened with a
conflict of compliance.
8.2. The scope of rules
One of the factors determining the applicability of a rule is its scope. Most legal rules identify by
means of their conditions to what kind of cases and to which persons they are applicable. This can be
everybody, as in Art. 2 of the European Convention on Human Rights, or sets of agents such as the
judges in the European Court of Human Rights, as in Art. 21 of the same Convention. However, there
are also limitations on the cases and persons to which a rule applies that are not mentioned in the
ordinary conditions of the rule. These are scope conditions, which combine with the ordinary rule
conditions to determine to which cases or persons rules are applicable.
There can be personal, spatial (territorial), and temporal scope conditions. Personal scope limitations
occur when a rule only applies to a certain class of persons even though this is not necessarily
mentioned in the rule conditions. So, for instance, Talmudic law applies to members of the Jewish
people, but not to non-Jews. Spatial or territorial scope limitations, meanwhile, refer to a distinction
15
This example presupposes that different legal regimes created by different legal instruments, such as the
human rights regimes of the European Union and of the Convention on the Elimination of All Forms of
Discrimination against Women, constitute different legal systems. Whether and to what extent this is the
case is a difficult question, which cannot be dealt with in the present contribution.
11
depending on the place where certain events takes place. So for instance, the penal laws of States
typically apply to events that took place in their own territory, but not in the territory of another
State. Temporal scope limitations postulate that the rules apply in different time periods. For
example the customary rules of treaty interpretation and the rules found in the Vienna Convention
on the Law of Treaties 1969 are different and thus they seem bound to conflict, but the conflict can
be avoided by postulating that the rules of the Vienna Convention apply only to treaties that entered
into force after 1969, as in fact stated in the convention itself.
By limiting the applicability of rules, scope conditions prevent rule conflicts, because two rules can
only conflict if they are both applicable to the same case.
8.3. Interpretation
Another technique to avoid conflicts is to interpret a legal source in such a way that the resulting rule
is not applicable to the case in question. The following example16 illustrates this.
Assume that the Constitution of a federal State prescribes that the State respect the constraints of
federalism, while international law prescribes that the federal State stop one of its constituent States
from engaging in internationally wrongful behavior. These actions cannot be performed both. The
rule conflict is clear, and a breach of either an international or domestic duty seems unavoidable.
Given this impasse, it may be possible to interpret an obligation of a specific type - e.g. stop damage
causing behavior - as an obligation of a more general type - e.g. avoid causing lasting damage. It is
possible to comply with this latter obligation without violating the former obligation, for instance by
compensating the damage that results from the behavior. If this is done, the demands of both legal
systems are deemed to be satisfied, and the conflict is avoided.
A related technique is that a new type of action is created, which can solve the impasse between two
colliding norms for an important range of cases. On a straightforward interpretation, the UN Charter
rules out all acts of aggression that are non-defensive and not authorized by the UNSC. Arguably
there is a developing rule of international that suggests that States have a duty to intervene in cases
of genocide, war crimes and crimes against humanity to defend the civilian population, and this duty
exists irrespective of whether UNSC authorization has been given or not. If a case of genocide breaks
out, and the UNSC does not authorize action, the two rules will conflict. One way to ease the conflict
is to de ise e a tio t pes su h as pea ekeepi g o hu a ita ia i te e tio that do ot fall
within the concept of aggression that is prohibited by the UN Charter (cf. Thompson 2008).
8.4. Derogation
At times, rules conflict with factual necessity, or with rules made to cope with factual circumstances
such as in states of emergency. Human rights on privacy might for instance conflict with measures
taken to prevent terrorist attacks or investigate them. One tool to prevent such conflicts from arising
is derogation. Derogation allows a State to take measures derogating from its obligations under a
treaty, to the extent necessary to handle certain situations (see e.g. Article 15 of the European
Convention on Human Rights). Logically speaking, there is no conflict in cases of derogation between
the rule of the treaty and the rule on the basis of which the measures are taken, because derogation
means that the treaty rule is inapplicable for the duration of the derogation. In short, derogation
16
The example was inspired by, but is not identical to, the case Avena and Other Mexican Nationals (Mexico
v. United States of America ; ICJ 09/01/2003 General List No. 128)
12
involves the temporary suspension of applicability of a potentially conflicting rule, thereby avoiding
the conflict.17
8.5
Incorporation and reference
The easiest way to avoid the dilemma of inter-systemic rule conflicts is to ensure that such conflicts
do not occur. We have seen that interpretation and derogation, but also scope limitations, can fulfil
this function, in that e.g. the national law of one State is limited in its application to the territory of
that State only. With regard to international law in particular, however, scope limitations do not
manage to avoid all conflicts. Methods such as incorporation and reference can prevent intersystemic conflicts from arising as well.
Rules of a foreign system can be used in a legal system through a technique which may be called
efe e e . The fo eig ules a e ot i o po ated i the legal s ste , ut thei e iste e a d
content is considered by the system as facts that are legally relevant from the point of view of the
legal system. Reference avoids conflicts between the rules of the referring system and the rules of
the system to which reference is made, because the content of the referring system is adapted to the
content of the referred system. Private International Law provides many examples of this, because it
contains meta-rules that determine which national legal system provides the applicable rules. For
example, the judgment whether a couple has divorced is given in country A on the basis of the rules
of country B, the validity of which is from the perspective of country A a matter of fact. These rules
are not incorporated in some international set of object-level rules, and neither is there an
independent international system. As a consequence, the rules of the national systems determine
the outcomes of cases, without a potential conflict with rules of an international system. In this way,
rule conflicts can be avoided.
In case of reference, the content of a foreign system is treated by the own legal system as a matter of
fact that co-determines the application of the domestic law. In case of incorporation, meanwhile,
foreign law becomes part of domestic law. The typical example of this phenomenon is the
incorporation of international law in a national legal system in so- alled o ist legal s ste s. The
Dutch legal system nicely illustrates incorporation. Provisions from international treaties ratified by
the Netherlands and rules created by international organizations in which the Netherlands
participate (in particular the European Union) automatically become part of the Dutch legal system
(Article 93 Grondwet). The foreign rules are not foreign anymore, except in the sense that they were
not created by native Dutch legislative bodies. They are part of the Dutch legal system to the same
extent as home-made rules.
Strictly speaking, incorporation is not a technique to deal with conflicts between rules of different
systems, but a way to ensure that only one legal system is relevant.18 If EU regulations become
17
18
It may be argued that derogation is making a temporary exception to the derogated rule. On this
interpretation the derogated rule is still applicable, but should temporarily not be applied. Derogation
would then not be a way to avoid rule conflicts, but a way to deal with them.
This holds at least from the perspective of the incorporating system. However, the mere incorporation of
rules of international law into a national legal system does not make any statement about the place of the
incorporated rules in the hierarchy of norms of that legal system. International law will hold itself to be
above the constitution, while national law might give the incorporated rules a different status. This brings
us back to the issue of subscripting, whereby the national legal system holds that it is the only relevant
system because it has incorporated rules of international law, while international law might nevertheless
claim relevance.
13
automatically part of the Dutch national law, there is no need any more to pay attention to EU law,
because the relevant rules are already part of Dutch national law. In the case of the EU one may even
ask whether there exists such a thing as the EU legal system, because arguably the EU only provides
organs which can create (uniform) law that becomes part of the national legal system of the Member
States. If all counties would have monist systems with regard to the relation between laws of
domestic and laws of non-domestic origin, the same might be said about the provisions of human
rights treaties.19 These treaties would then create uniform human rights in different legal systems
and it might be argued then that there is no separate international human rights system. However,
theoretically it is imaginable that some legal system incorporates part of a foreign legal system, while
that foreign system has independent existence. The situation is then comparable to one country that
uses the national currency of some other country.
If fo eig ules a e i o po ated i a atio al legal s ste the a e ot fo eig ules a
o e ut
merely rules with a foreign origin. Such rules may still conflict with rules of a national origin, or with
other rules of foreign origin. However, because of the incorporation, such conflicts are not conflicts
between legal systems anymore. What is avoided by incorporation is not a conflict of rules, but a
conflict of legal systems.20 If there is still a conflict of rules, the techniques used within a single legal
system to deal with conflicts, such as making exceptions, can be used to deal with possible conflicts
between rules from national and international sources.
8.6
Limitation of rule-creating powers
A common way to avoid inconsistencies within a single legal system is to avoid rule conflicts by
preventing conflicting rules from entering into existence at all. A national legislator for instance
might make an exception to the general right of free speech for cases of hate speech. This will not
allow a local legislator to make an exception to the exception for hate speech against people of a
particular origin, such as French speaking people from Walloon. If a local legislator nevertheless
attempted to do so, its rules would simply not be recognized as valid law: the local legislator does not
ha e the po e to ake ules that o fli t ith the highe ules of the atio al legislato . This
limitation of power avoids conflicting rules by impacting the first of the three factors determinative
of applicability.
With regard to international law, however, limitations of power occur less frequently. This is because
such limitations suggest an overarching organization of the distribution of rule creating power that
can divest certain actors of their ability to create rules. Such organization is absent in the
international plane. As mentioned in the introduction, rule creation in international law is
decentralized and tends towards anarchy as even the doctrine of the sources of law has only relative
weight. Instead of limitation of power, international law usually creates prohibitions, which can be
violated. The typical sanction is not invalidity, but the need to perform some sort of compensating
behavior, which may range from monetary compensation to restitution. One notable exception to
this is Eu opea U io la , hi h akes it i possi le fo Me e “tates to a o d p e ede e to a
19
20
Since not all countries use a monist system, this exercise is quite theoretical. However, it is useful to see
what the effects of incorporation might be.
As a matter of fact, incorporation makes more rule conflicts possible, because rules from different systems
can only lead to conflicts of compliance, while rules that belong to the same system can also lead to
conflicts of imposition.
14
u ilate al a d su se ue t
ECR 585).21
9.
easu e o e a legal s ste
a epted
the
. (Case 6/64, Costa v ENEL,
Classification of exceptions
In the previous section, we have considered tools to prevent rule conflicts. Given that rule conflicts
are a frequent cause of exceptions, these tools can negate the need for exceptions. However, not all
rule conflicts can be avoided, and we have seen in section 6 that there are other grounds for
exceptions as well. In short, exceptions cannot always be avoided. In this section, we will have a
closer look at exceptions and how to classify them.
An exception is made to a rule if the rule is not applied to a case, although it was applicable to that
case. its applicability is a contributory reason for the application of a rule, and this has two aspects:
1. If a rule is not applicable to a case, and if – as usual – there are no other reasons to apply the
rule, then the rule should not be applied to that case. In other words: the default situation
with regard to rule application is that an inapplicable rule should not be applied.
2. If a rule is applicable to a case, this is normally a contributory reason for applying the rule to
that case. Barring additional contributory reasons against application, this means that the
rule should be applied to the case.
This last clause indicates how exceptions to a rule may come into play. First, it may be the case that,
because of exceptional circumstances, the applicability of the rule does not count as a contributory
reason to apply the rule. This possibility is studied in section 9.1. Second, it may be the case that
although the applicability of the rule counts as a contributory reason for application, this reason is
outweighed by one or more reasons against application. This possibility will be explored in section
9.2.
9.1. Exceptions in the shape of undercutting
So far, we have said that the applicability of a rule to a case is a contributory reason for its
application to that case. Sometimes, however, it turns out that this reason for applying the rule is
exceptionally not a reason at all. Why, however, should we make such an exception to the general
rule that applicability is a contributory reason for application?
In section 6, we identified one possible reason to do so, namely that a rule-exception construction
was used by the maker of the rule in order to create a division in the burden of proof. The party in a
dispute which wants the rule to be applied has the burden to prove that the rule conditions are
satisfied; the party which does not want the rule to apply has the burden to prove that the
exception-creating rule applies. When the maker of the rules has chosen for this mechanism for
distributing the burden of proof, and if it has been proven that the exception-generating rule applies,
there is no need any more to balance the applicability of the main rule against the ground for the
exception. This balance was already made by the creator of the rules and normally this decision
should be respected.
21
Another exception may be the invalidity of soft law, when it does not complement, but contradicts hard
law. The precise nature of soft law and its relation to hard law are too complicated, however, to deal with
here.
15
A othe possi ilit is that the appli atio of the ule ould e agai st the ule s pu pose i this ase.
An example by Fuller (1958) illustrates this point: there is a rule that forbids sleeping in railway
stations, which has as its purpose to keep tramps from occupying the station as a place to spend the
night. An old lady who wants to meet a friend at the station dozes off when the evening train turns
out to be delayed. If the prohibition were applied to this old lady, this would arguably be against the
purpose of the rule. One might argue that because the case of the old lady is not caught by the
purpose of the rule, the fact that the rule is technically applicable to her case is irrelevant and
therefore loses its function as a contributory reason for application. If this argument is made, we are
not weighing reasons for applying the rule (applicability) against reasons against applying the rule.
Instead, we are saying that there is no reason to apply the rule at all.22
When we have a reason or reasons why a fact, such as the applicability of a rule, that would normally
count as a reason should exceptionally not cou t as a easo afte all, e speak of u de utti g
defeate s i episte olog Pollo k a d C uz 1999, p. 196 f.)23 o of e lusio a easo s i ases
of reasons for actions (Raz 1975, p. 35-48). It is important to notice, however, that undercutting
defeaters are not only relevant in connection with reasons to believe (Pollock and Cruz) or reasons
for action (Raz); they are relevant for all reasons.
9.2. Exceptions in the shape of rebuttal
As we have mentioned, if a rule is applicable to a case, this is usually a contributory reason to apply
the rule to this case. However, a contributory reason is not as such decisive. If there are contributory
reasons against applying the rule as well, then the reasons for and against application must be
balanced. If the balance favors the con-reasons, the rule does not apply. Since the rule was, ex
hypothesi, applicable, the result would be non-application of an applicable rule and therefore an
exception to the rule.
Because application of a rule leads to the facts that the rule attaches to the case to which it is
applied, reasons against applying a rule are usually reasons why the attached facts undesirable. They
may be undesirable for reasons which have nothing to do with law, as when, because of exceptional
circumstances, the application of a rule about free trade leads to a decrease rather than an increase
of welfare. They may also be undesirable because the consequences of one applicable rule conflict
with the consequences of another rule which is applicable too, as we have seen in section 7. And
finally, they may be undesirable because the actual o se ue es ould iolate the ule s pu pose.24
This brings us back to the example of the old lady who falls asleep at the train station. While the
argument can be made, as e did i the p e ious se tio , that the iolatio of the ule s pu pose
negates the function of applicability as a contributory reason for application, a different
interpretation is equally possible.25 On an alternate account, we can still consider applicability of the
rule a contributory reason for its application to the case of the old lady, while we consider the
22
23
24
25
We will see in the following section, however, that the same example can be construed differently.
These undercutting defeaters are also discussed in Prakken 1997, p. 102/3 and Sartor 2005, p. 682-685.
This distinction between three categories, which harks back to section 4, does not claim that the three
categories are mutually exclusive.
This also shows that logic and legal theory can be of great help in classifying possibilities, but that it
depends on substantive arguments or reasons which choices should be made. We briefly talk about the role
of legal theory in this context in section 11.
16
violation of purpose a reason against application. If, in balancing the two, the reason against
application wins out, we make an exception to the rule in case of the old lady.
If application of a rule to a particular case leads to undesirable consequences, this is a reason against
the application of the rule to that case. This reason still needs to be balanced against the applicability
of the rule as reason for application. Usually the balancing of these reasons is conceptualized as the
determination which rule prevails over the other rule. The reason based on the prevailing rule
outweighs the reason based on the other rule. Several contributory reasons can be – and in fact are –
recognized in this connection:26 one option is that the rule that better fits in the overall legal system
prevails over the less fitting rule (coherence). Another option is that the rule that was made by the
highe autho it p e ails o e the ule ade
the lo e autho it lex superior). Equally, the
more specific rule could prevail over the more general rule (lex specialis), or the more recent rule
over the older one (lex posterior).27
There is much more to be said about what should become the outcome of this operation of
balancing competing rules, and in particular about the authority of law in this connection,28 but this is
not the place to go into the details of these substantive reasons. We will confine ourselves to the
logical aspects of these so- alled e utti g defeate s Pollo k a d C uz 1999, p. 196). They are as
follows. If a rule applies to a case, its consequences are attached to the case. If these consequences
are for some reason undesirable, the same reason is also a reason why the rule should not be
applied. This reason must be balanced against the applicability of the rule as reason for application,
and depending on the outcome of this balancing, it may be the case that the rule should not be
applied. If the rule does not apply, even though it is applicable, there is an exception to the rule.
9.3. Two observations concerning exceptions
We would like to make two observations in connection with exceptions to rules. The first concerns
the difference between exceptions based on undercutting defeaters and exceptions based on
rebutting defeaters. In the former situation there are no reasons for applying the rule, because
applicability exceptionally does not count as a reason for application. In the latter situation there is at
least one reaso fo appl i g the ule, the ule s appli a ilit , ut this easo is out eighed
reasons against application, typically reasons involving the undesirability of the rule conclusion.
The second observation concerns the role of exceptions in relation to the non-application of a rule.
That there is an exception to a rule is, in the terminology proposed here, not a reason against the
application of the rule, but a conclusion based on the premises that the rule is applicable and that
the rule does not apply. There must be a reason why the rule does not apply, and this can be an
undercutting defeater or a rebutting defeater. However, neither one kind of defeater is the
exception to the rule. It is at best what brings about the exception to the rule. So we would not call
the fact that causes an exception to a rule an exception, and neither would we call the rule which
makes this fact into a defeater an exception. In our terminology, an exception to a rule is never itself
a rule.
26
27
28
The following is a non-exhaustive list.
Taking the lex posterior rule even one step further is the tool of implied repeal, whereby it is presumed that
if the later rule conflicts with the earlier rule, the later rule not only prevails, but the earlier rule is in fact
epealed. Ho e e , if o e of the o fli ti g ules counts as repealed, there is no real conflict, since the
repealed rules does not exist anymore and can for that reason not be applicable.
See for instance Schauer 1995, p. 38-111.
17
An example may clarify this point.
Substance
Effect
According to Article 2 ARSIWA, there is an
internationally wrongful act if the act constitutes
the breach of an international obligation.
Article 2 ARSIWA is applicable.
In case X, State A acted in a way that breached
an international obligation.
According to Article 21 ARSIWA, the
wrongfulness of an act of a State is precluded if
the act constitutes a lawful measure of self- Article 21 ARSIWA is applicable.
defense taken in conformity with the UN Charter.
State A acted in self-defence.
Due to Article 21, the applicability of Article 2
does not count as a reason for application of that
“tate A s act is not internationally wrongful.
rule (undercutting defeater).
Article 2 is not applied despite its applicability.
An exception is made.
In this example, the rule of Article 2 does not apply even though it is applicable, which means that
there is an exception to that rule. This exception was brought about by the self-defense of the State,
which in turn derived its legal relevance from the rule of Article 21. However, neither the selfdefense, nor the rule of Article 21 is the exception to the rule of Article 2. They both contributed in
their own way to that exception, but neither one of them should be identified with it.
10. Derived rules
If an exception to a rule is made – be it for reasons undercutting the application of the rule or for
reasons rebutting it – some cases or persons are excluded from the application of the rule. The same
effect can be brought about
ea s of s ope o ditio s o egati e ule o ditio s. The ordinary
conditions of Article 10 ECHR, fo i sta e, o tai su h egati e o ditio s , since they explicitly
exclude amongst others the licensing of broadcasting from the protection of this article.
When we ask which cases are governed by rules of international law, there are at least three kinds of
reasons why some cases are not governed by particular rules, and they are negative rule conditions,
scope conditions, and exceptions. How can exceptions be distinguished from ordinary negative rule
conditions?29 The answer is provided in this and the following section, and the burden of proof plays
a central role in it.
The idea that legal rules are open to exceptions is somewhat controversial. This has to do with a
certain ambiguity in the notion of a legal rule. In section 3.1, we talked about the distinction between
rules and rule formulations. A legal rule may be seen as something that can be found in a legal
source, such as a treaty, legislation, international custom, or case law. It may also be seen as a
general connection between operative legal facts and legal consequences. The latter is the
u de sta di g of legal ules ith hi h e ope ate he e. A si ple e a ple a agai illust ate
that these two understandings are not identical.
29
To withhold this contribution from becoming too lengthy, we ignore scope conditions from here on.
18
According to Article 2, Section 4 of the UN Charter, the use of force is prohibited. Article 42 of the
Charter holds that the UNSC may authorize military intervention, in which case it is deemed to be
permitted. These two rules seem to conflict, and since the second rule is a lex specialis with regard to
the rule that the use of force is prohibited, it would normally prevail over it. Application of the
second rule then makes an exception to the first.
However, it comes naturally to state that there is just one rule, namely that the use of force is
prohibited unless authorized by the UNSC. We ill all this a de i ed ule . For this derived rule,
there is no need to make an exception. In fact, it has the absence of the exception as one of its
conditions, and hence the prohibition would not be applicable if the UNSC has authorized the use of
force. Since exceptions are made when an applicable rule is not applied, no exception would be
necessary – or even possible – in case of authorized use of force.
Let us have a closer look at this phenomenon of derived rules.30 If a rule is applicable to a case this
normally means that the rule is applied to that case and attaches its legal consequences to it. This is
so normal that the logic of rule application seems to be nothing else than an ordinary syllogistic
argument (Alexy 1983, 273-283; MacCormick 1978, 19-53). The facts of a case are subsumed under a
general rule, and the conclusion that describes the legal consequences of the case follows
deductively. This deductive application of rules seems so natural that it requires explanation that
exceptions to rules are possible. If a rule seems to have an exception, why not say that the rule was
not formulated properly, and that it actually has an additional condition namely that the exceptional
circumstances are absent?
The insight that rules can have exceptions can be reconciled with the impression that rules can be
applied in deductive arguments by means of so- alled case-legal consequence pai s (CLCPs).31
CLCPs describe the effects of rules such as the prohibition of the use of force and the permission of
military action when authorized by the UN Security Council. The two inconsistent rules are combined
i to a si gle ule that leads to a si gle o siste t esult. We use uotatio
a ks he e to i di ate
that this ule diffe s f o the t o ules that e e used to o st u t it. Both the ule p ohi iti g the
use of force and the rule permitting it in cases of Security Council authorization are based on an
offi ial legal sou e, i this ase the UN Cha te . The de i ed ule , ho e e , a ot e t a ed a k
directly to such a source, but is the result of combining the two original rules in light of their
apparent purposes, thereby creating a CLCP.
It is possible to characterize a legal system as defined by an exhaustive set of such CLCPs: for every
kind of case that has legal consequences, there exists a CLCP that gives the characteristics of the kind
of case and the legal consequences attached to it. These CLCPs are the outcome of the original rules
(including rights and legal principles or incorporated or referred rules) of the system, interpretation
and solutions of potential rule conflicts by means of prevalence (such as lex superior) or any other
technique the system in question employs to resolve rule conflicts.
The CLCPs are constructed in such a way that no particular case can fall under two different cases to
which incompatible consequences are attached. For example, the e ill e a ase fo
ilita a tio
ithout “e u it Cou il autho izatio a d o e fo
ilita
a tio
ith “e u it Cou il
autho izatio , ut ot o e fo ilita a tio i ge e al, e ause the latte ight gi e diffe e t legal
30
31
The question whether it is possi le to de i e ules o norms , which is the more frequently used term)
from other rules is highly debated. For an overview, see Navarro and Rodriguez 2014, chapter 2.
Hage 2005, 27. This idea of CLCPs was inspired by the theory of Alchourrón and Bulygin about the Universe
of Cases (Alchourrón and Bulygin 1971, p. 24-30).
19
consequences in a concrete cases of military actions with and without Security Council authorization.
Understanding a legal system as an exhaustive set of CLCPs, it is not possible that a case has
inconsistent legal consequences. Imminent inconsistencies are filtered out in the step from the
original (conflicting!) rules to the CLCP. Moreover, there are no exceptions to CLCPs. If there seems
to be an exception, this means that the CLCP was formulated too broadly: there should be two
different CLCPs, one for the normal cases and one for the exceptional cases.
If a legal system is interpreted as an exhaustive set of CLCPs, exceptions to rules only play a role in
this step f o the o igi al ules ased o legal sou es to the de i ed ules – the CLCPs - that define
the outcome of all the interacting original rules. It is this step that cannot be handled well by means
of deductive logic. The derivation of the legal consequences of a case by applying the relevant CLCP
to that case can be purely deductive, because all exceptions have already been filtered out in
constructing the CLCP. When we discuss exceptions to rules, we are not necessarily talking about the
immediate application of rules to cases; we may also be talking about the construction of CLCPs
which can in turn be used for legal justification is a deductively valid manner.
11. Burden of proof
It is possible, as we have seen in section 10, to maintain the syllogistic form of legal argument and to
remove exceptions entirely from a legal system by viewing a legal system as an exhaustive set of
CLCPs. In this section, however, we argue that while possible, this view neglects an important
function of exceptions, which is connected to the burden of proof.
In order to understand this function, we must replace our understanding of legal arguments as
ordinary syllogisms with a view of the legal system as a dialogic practice.32 Imagine a dialogue
between two parties: the one party, the proponent, wants to establish a particular legal consequence
for a case and to do so it invokes a legal rule and wants it to apply. The other party, the opponent,
does not want that consequence and therefore does not want the rule to be applied. Both proponent
and opponent can adduce reasons: the proponent reasons why the rule should be applied, the
opponent reasons why the rule should not be applied. For instance, the proponent should adduce
that the rule conditions are satisfied by the case to which he wants the rule to apply. The opponent
might adduce reasons why the rule should nevertheless not be applied and which would lead to an
exception if they outweigh the reasons adduced by the proponent.
If all reasons for and against application of the rule have been adduced, there are three possibilities:
1. the reasons for application outweigh the reasons against application;
2. the reasons against application outweigh the reasons for application;
3. the reasons are, at least for the purpose of decision making, in balance; there is a draw.
In the first case, the rule should be applied and its consequences are attached to the case as the
proponent would want. In the second case, the rule should not be applied and its consequences are
not attached to the case as the opponent would want. But what to do with the third case, when the
reasons are in balance? Here is where the burden of proof comes in. The third case should either
lead to an outcome that favors the proponent, or to an outcome that favors the opponent. In the
former case we say that the burden of proof lies with the opponent, because if the opponent does
not want the rule to apply, he must make sure that the balance of reasons does not end in a draw. In
32
A o e ie of this dialogi al ap oa h to logi
p. 227-264.
a
e fou d i Bath and Krabbe 1982. See also Hage 2005,
20
the latter case we say that the burden of proof lies with the proponent, because then the proponent
has an interest in avoiding a draw in the balance of reasons.
The burden of proof translates into the burden of production, with which the burden of proof is
easily confused.33 If the balance of the reasons that have been adopted at a certain moment during a
legal dialogue would lead for an outcome that favors the proponent, it is up to the opponent to
produce more reasons.34 At that moment, the opponent has the burden of production. It is possible
that the opponent succeeds in doing so, and then the burden of production shifts to the proponent.
The process of the production of more reasons, with a shift in the burden of production as result,
may continue for some time, and the burden of production may shift several times during a dialogue.
However, it is the situation at the end of the dialogue, and in particular which side in the dialogue
benefits from a draw, which determines who has the burden of proof. The burden of proof never
shifts, but it determines, together with the state of the dialog, which party has at a certain moment
the burden of production.
This connection between exceptions and the burden of production (and ultimately of proof) cannot
be accounted for if we abolish exceptions by viewing the legal system as an exhaustive set of CLCPs.
To further illustrate the view of a legal system as dialogic and to show how dialogs can be used to
model legal reasoning and the operation of exceptions, we will consider an example. Our example
will feature State O and State V. State O breaches an international obligation, and State V, which
suffered damage as result, wants reparations from State O.
Step 1: Straightforward rule application
A party who wants a rule to apply must prove that the normal conditions of the rule are satisfied.35
Given the rule of Article 31 ARSIWA, V must prove that O committed an internationally wrongful
act.36 Given the rule of Article 2 ARSIWA, V must therefore prove that O conduced itself in a way that
was either an act or omission to act, conduct that can be attributed to O, and that breached an
international obligation of O. We ignore the attribution aspect, to find that, according to the rule of
Article 12 ARSIWA, V has to prove that the act of O was not in conformity with an international
obligation of O. And so on...
We see a chain of rules, where the conclusion of the one rule makes that some condition of another
rule is satisfied, and the chain ends with an obligation to repair for State O. State V must prove that
the conditions of all these rules are satisfied, and can sometimes fulfil this burden by justifying the
application of another rule in the chain.
33
34
35
36
We took the distinction between the burden of proof and the burden of production from the contribution
of Joost Pauwelyn to this volume.
Actually there are more possibilities to make the burden of production shift, including showing that the
seeming reasons adduced by the other party cannot withstand criticism. However, this is not the place to
discuss these possibilities.
In these examples we will ignore the possibility of analogous rule application, and also the role of scope
conditions.
We will assume, for the sake of exposition, that the validity of the rule does not have to be proven, because
the court knows the law (ius curia novit). Strictly speaking this introduces a third party (the court) into the
dialog, but, again for the sake of exposition, we ignore this complication.
21
Step 2: Preclusion
Suppose that State V in Example 1 has proven that State O breached an international obligation and
that this can be attributed to O. Normally this means that the rule of Article 2 ARSIWA is applicable
and that would be a reason to apply that rule with as consequence that the conduct of State O
counts a wrongful. However, State O can block the step from the applicability of the rule to its
application by invoking the rule of Article 21 ARSIWA. To do so it must prove that it acted in selfdefense, that therefore the rule of Article 21 ARSIWA. is applicable, that therefore this rule applies
and that that the wrongfulness of what State O did is precluded. If O succeeds in doing so, the rule of
Article 2 is applicable, but does not apply, and there is an exception to this rule.
Notice that State V has the burden of proof with regard to the question whether State O must repair
the damage, but that State O has the burden of proof for the presence of self-defense. If State V
would have had the burden to prove that there was no self-defense, the absence of self-defense
would have been a negative condition of the rule of Article 2. Apparently the burden of proof is
relative to the issue at stake. V bears the burden for the obligation to repair, while O bears the
burden for self-defense.
Because there are two burdens of proof, for two issues, there is no shift in the burden of proof.
However, there is a shift in the burden of production, because O does not have to do anything as
long as V did not prove that O breached an international obligation. Only when V succeeded in
proving that, the burden of production shifts to O, which must prove self-defense.
The Role of Legal Theory
Whether the applicability of Article 21 ARSIWA makes for an exception, with a shift in the burden of
production, or whether this Article combines with Article 2 (and some other articles) as sources for a
complicated rule that has some negative conditions, is something that must be established by
interpretation of the ARSIWA. There is no simple test for it, and the difference between negative rule
conditions and exceptions is therefore a matter of interpretation too. The only thing that legal theory
has to offer in this connection is that if there is no shift in the burden of production, we have a
negative rule condition, and that if there is such a shift, we have an exception.37
12. Conclusion
Exceptions to rules play an important role in law, and in particular in international law. A proper
understanding of exceptions is therefore of crucial importance for legal practice, legal doctrine and
legal theory. The aim of this article is to contribute to this understanding; it is not to describe the law
or the present usage with regard to rules and exceptions.
As background for the development of a theory about exceptions to rules, sections 3 and 4 have
gone into some detail concerning rules, rule-formulations, legal sources, reasons and their logic. In
section 5 the distinction between the applicability and the application of a rule was elaborated. A
rule is applicable to a case if the rule is valid, and if its ordinary and scope conditions are satisfied by
the case. If a rule is applied to a case, the rule attaches its legal consequences to the facts of the case.
Normally the applicability of a rule to a case is a contributory reason why the rule should be applied
37
Again, we ignore possible third ways, such as scope conditions.
22
to the case. An exception to a rule in a case is defined as the situation in which a rule is applicable to,
but nevertheless not applied to the case.
In section 6, two main grounds for making an exception to a rule were identified. First, the maker of
the rule may use the rule-exception construction to create a division in the burden of proof. As
argued in section 9, this leads to a so-called undercutting defeater. Second, there may be reasons
why the legal consequences of the rule in the case are undesirable. This leads to a reason against
applying the rule, which needs to be balanced to the applicability of the rule as reason for
application. In this situation we speak of rebutting defeaters (see section 9.2).
An important reason why it is undesirable to apply a rule to a case is that application would generate
a conflict with another applicable rule. Section 8 discussed a number of tools/techniques that can be
used to avoid rule conflicts and which would in that way make exceptions superfluous.
Finally, the question may be raised whether legal rules really have exceptions. Is it not the case that if
a rule is well-fo ulated, it e tio s all e eptio s as egati e ule o ditio s? “u h a ellfo ulated de i ed ule a the e applied dedu ti el to ases that satisf the ule o ditio s. I
sections 10 and 11, we discussed this possibility and showed that this use of derived rules blocks the
possibility to model the division of burden of proof which is implicitly given with the rule-exception
model.
23
References
Alchourrón and Bulygin 1971.
Carlos E. Alchourrón and Eugenio Bulygin, Normative Systems, Wien: Springer Verlag 1971.
Alexy 1983.
R. Alexy, Theorie der juristischen Argumentation, 7e Auflage, Frankfurt a/M: Suhrkamp 1983.
Alvarez 2016.
Ma ia Al a ez, ‘easo s fo A tio : Justifi atio , Moti atio , E pla atio , The Stanford Encyclopedia
of Philosophy (Summer 2016 Edition), Edward N. Zalta (ed.),
http://plato.stanford.edu/archives/sum2016/entries/reasons-just-vs-expl/.
Arend 1999.
Anthony Clark Arend, Legal rules and international society. Oxford: Oxford University Press 1999.
Bath and Krabbe 1982.
E.M. Barth and E.C.W. Krabbe, From Axiom to Dialogue, Berlin: Walter de Gruyter 1982.
Franck 1990.
Thomas M. Franck, The power of legitimacy among nations. Oxford: Oxford University Press 1990.
Fuller 1958.
Lo L. Fulle , Positi is
(1958), 630-672.
a d Fidelit to La . A ‘epl to P ofesso Ha t , i Harvard Law Review 71
Hage 1997.
Jaap C. Hage, Reasoning with Rules, Dordrecht: Kluwer 1997.
Hage 2005.
Jaap C. Hage, Studies in Legal Logic, Dordrecht: Springer 2005.
Hage 2015a.
Jaap Hage, “epa ati g ‘ules f o No ati it , in Mi hal A aszkie i z, Pa el Ba aś, To asz
Gizbert-Studnicki and Krzystof Pleszka (eds.), Problems of Normativity, Rules and Rule-Following,
Cham: Springer 2015, 13-30.
Hage 2015b.
Jaap Hage, The Justifi atio of Value Judg e ts. Theo eti al Fou datio s fo A gu e ts a out the
Best Level to Regulate European Private La , i Bram Akkermans, Jaap Hage, Nicole Kornet and Jan
Smits (eds.), Who Does What? On the allocation of regulatory competences in European Private Law,
Cambridge: Intersentia 2015, p. 15-56.
Hart 2012.
Herbert L.A. Hart, The Concept of Law, 3rd ed. Oxford: Oxford University Press 2012, 1st ed. 1961.
Koskenniemi & Leino 2002.
M. Koskenniemi, and P. Lei o, F ag e tatio of i te atio al la ? Post ode
Journal of International Law, 15 (2002), p. 553-579.
MacCormick 1978.
Neil MacCormick, Legal Reasoning and Legal Theory, Oxford: University Press 1978.
24
a ieties. Leiden
Marx 1977.
Mi hael Ma , “ ste e de 19e Jah hu de ts , i A. Kauf a a d W. Hasse e eds. , Einführung
in Rechtsfilosophie und Rechtstheorie der Gegenwart, 1st ed., Heidelberg: Müller 1977, p. 89-96.
Navarro and Rodríguez 2014.
Pablo E. Navarro and Jorge L. Rodríguez, Deontic Logic and Legal Systems, Cambridge: Cambridge
University Press 2014.
Pollock and Cruz 1999.
John L. Pollock and Joseph Cruz, Contemporary Theories of Knowledge, 2nd ed. Lanham: Rowman and
Littlefield 1999.
Prakken 1997.
Henry Prakken, Logical Tools for Modelling Legal Argument. A Study of Defeasible Reasoning in Law,
Dordrecht: Kluwer 1997.
Redondo 1999.
Cristina Redondo, Reasons for Action and the Law, Dordrecht: Kluwer 1999.
Raz 1999.
Joseph Raz, Practical Reason and Norms, 2nd edition, Oxford: Oxford University Press 1999 (1st edition
London: Hutchinson 1975).
Sartor 2005.
G. Sartor, Legal Reasoning, A Cognitive Approach to the Law, Dordrecht: Springer 2005.
Scanlon 2014.
T.M. Scanlon, Being realistic about reasons, Oxford: Oxford University Press 2014.
Schauer 1991.
F. Schauer, Playing by the Rules, Oxford: Clarendon Press 1991.
Textor 2014.
Ma k Te to , “tates of Affai s , The Stanford Encyclopedia of Philosophy (Summer 2014 Edition),
Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/sum2014/entries/states-ofaffairs/>.
Thompson 2008.
M. Thompson, Life and action, Cambridge: Harvard University Press 2008.
Waddington and Visser 2012.
L. Waddington and Laura Visser, Te po a “pe ial Measu es u de the Wo e s Co e tio a d
Positi e A tio u de EU La : Mutuall Co pati le o I e o ila le? i I. Weste do p ed. , The
Wo e s Co ve tio Tur ed 30: A hieve e ts, Set a ks, and Prospects, Cambridge: Intersentia
2012.
Wittgenstein 1921/2.
Ludwig Wittgenstein, Tractatus logico-philosophicus, Frankfurt a/M: Suhrkamp 1984 (1st edition
1921/2).
25