© The Author(s) 2018. Oxford University Press and New York University School of Law.

All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Proportionality, constitutional
law, and sub-constitutional law:
A reply to Aharon Barak

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

Robert Alexy*

Aharon Barak and I agree on many questions of proportionality analysis. But there are also
points of disagreement. One point concerns the relation between constitutional rights and
proportionality. Barak argues that my connection between constitutional rights and propor-
tionality is too narrow. The connection I defend implies that proportionality has effects at the
constitutional level. In contrast to this, Barak defends the thesis that proportionality only
operates at the sub-constitutional level. In this article, I attempt to show that proportionality,
already at the constitutional level, must have effects. Naturally, these constitutional effects
have, in turn, effects at the sub-constitutional level.

A theory can be criticized from outside as well as from inside. Examples of an external
critique of principles theory include Larry Alexander’s and Ralf Poscher’s respective
theses that principles do not exist,1 Habermas’s objection that balancing is a non-
rational or irrational method,2 and Böckenförde’s reproach that principles theory
leads to “a gradual transition from a parliamentary, legislative state to a constitutional
court-based, jurisdiction state.”3 Here, however, an internal critique of principles the-
ory shall be considered, the critique of Aharon Barak.
In the introduction to his book Proportionality, which appeared in English in 2012,
Barak adduces four points of disagreement with me.4 One of these points concerns the

* Professor of public law and legal philosophy, Christian Albrechts University Kiel. Email: alexy@law.uni-kiel.de.
I should like to thank Stanley L. Paulson for suggestions and advice on matters of English style. The author kindly
acknowledges the financial support of a research grant on The Role of Proportionality in Constitutional Adjudication
(15-23955S) awarded by the Grant Agency of the Czech Republic.
1
Larry Alexander, Legal Objectivity and the Illusion of Legal Principles, in Institutionalized Reason. The
Jurisprudence of Robert Alexy 115, 122–123 (Matthias Klatt ed., 2012); Ralf Poscher, The Principles
Theory. How Many Theories and What Is Their Merit?, in Institutionalized Reason. The Jurisprudence of Robert
Alexy 218, 230–232, 235 (Matthias Klatt ed., 2012).
2
Jürgen Habermas, Between Facts and Norms. Contributions to a Discourse Theory of Law and Democracy 259
(William Rehg trans., 1996) (1992).
3
Ernst-Wolfgang Böckenförde, Constitutional and Political Theory. Selected Writings 259 (2017).
4
Aharon Barak, Proportionality. Constitutional Rights and Their Limitations 6 (Doron Kalir trans.,
2012) (2010).

I•CON (2018), Vol. 16 No. 3, 871–879 doi:10.1093/icon/moy084

872 I•CON 16 (2018), 871–879

relation between proportionality, the constitutional level, and the sub-constitutional

level.5 In a recent article, Barak has elaborated this point further.6 The elaboration
comprises four arguments,7 and they shall be considered here.
Barak emphasizes that on decisive points his position is in accordance with mine.
He agrees with the distinction between “constitutional rights as principles and con-
stitutional rights as rules,” with the definition of principles as “optimization require-
ments,” and with the thesis that constitutional rights as principles “are subject to the

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

doctrine of proportionality.”8 Before this background the central point of his critique
is described by Barak with the following words:
My critique is therefore focused on the Gordian knot—a knot formed by Alexy’s very definition
of principle—tying constitutional rights as principles to the rules of proportionality.9

That is, the disagreement concerns “the relationship between the theory of consti-
tutional rights and the theory of proportionality.”10 Barak is right when he says that
in my theory constitutional rights are very closely connected with proportionality.
One could even say that my connection is the closest connection possible. Before I dis-
cuss Barak’s objection of a too close or too narrow connection, it is well to have a
look at some basic elements of principles theory from which the as close as possible
connection stems.

1. Some basic elements of principles theory

The basis of principles theory is the norm-theoretic distinction between rules and
principles.11 Rules are norms that require something determinate. They are defini-
tive commands. Their form of application is subsumption. By contrast, principles are
optimization requirements. As such, they demand “that something be realized to the
greatest extent possible given the legal and factual possibilities.”12 Rules aside, the
legal possibilities are determined essentially by opposing principles. For this reason,
principles, each taken alone, always comprise merely prima facie requirements. The
determination of the appropriate degree of satisfaction of one principle relative to the
requirements of other principles is brought about by means of balancing. Thus, bal-
ancing is the specific form of application of principles.
The nature of principles as optimization requirements leads straightaway to a nec-
essary connection between principles and proportionality. The principle of propor-
tionality consists of three sub-principles: the principles of suitability, necessity, and

5
Id. at 37–42.
6
Aharon Barak, A Critical Review of Alexy Regarding the Relationship Between Constitutional Rights as
Principles and the Theory of Proportionality, in Rechtsphilosophie Und Grundrechtsdogmatik. Robert Alexys
System 347, 347–357 (Martin Borowski, Stanley L. Paulson, & Jan Reinard Sieckmann eds., 2017).
7
Id. at 349.
8
Id.
9
Id.
10
Id. at 347.
11
Robert Alexy, A Theory of Constitutional Rights 47–48 (Julian Rivers trans., 2002) (1985).
12
Id. at 47.
 Proportionality, constitutional law, and sub-constitutional law: A reply to Aharon Barak 873

proportionality in the narrower sense. All three sub-principles express the idea of
optimization. For this reason, the nature of principles implies the principle of propor-
tionality and vice versa. Each logically follows from the other.13 A closer connection
between principles and proportionality is not conceivable. The next step consists of
the inclusion of constitutional rights in this connection. This is achieved by the theses
that one who accepts the necessity of proportionality analysis in the application of
constitutional rights14 must accept the character of constitutional rights as principles,

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

and that one who accepts the character of constitutional rights as principles must
accept the principle of proportionality.
Principles as optimization requirements demand an optimization relative to the
factual possibilities as well as relative to the legal possibilities. The principles of suit-
ability and necessity refer to optimization relative to the factual possibilities.
Optimization relative to the factual possibilities concerns the question of whether
one position can be enhanced without detriment to the other.15 Thus, the first two
sub-principles require Pareto-optimality.
Optimization relative to the factual possibilities consists of avoiding avoidable costs.
Costs, however, are unavoidable when principles collide. Balancing then becomes neces-
sary. Balancing is the subject of the third sub-principle of the principle of proportion-
ality, the principle of proportionality in the narrower sense. This principle expresses
what optimization relative to the legal possibilities means. It is identical with a rule that
can be called the “Law of Balancing,”16 which is, on the side of the “Law of Competing
Principles,”17 one of the two basic laws of principles theory. The Law of Balancing states:
The greater the degree of non-satisfaction of, or detriment to, one principle, the greater must be
the importance of satisfying the other.

The Law of Balancing excludes, inter alia, an intensive interference with principle P1
that can only be justified by assigning a low importance to the satisfaction of the com-
peting principle P2. To allow the interference under these conditions would not be an
optimization of P1 together with P2.
Nearly everywhere in constitutional adjudication, the Law of Balancing is found
in various different formulations. It expresses the essence of balancing and is of
great practical importance. If, however, one wishes to achieve a precise and complete
­analysis of the structure of balancing, the Law of Balancing has to be elaborated fur-
ther. The result of such a further elaboration is the weight formula.18 It runs as follows:

I i ⋅ Wi ⋅ Ri
Wi,j =
I i ⋅ Wj ⋅ R j

13
Id. at 66.
14
On this, see Robert Alexy, The Absolute and the Relative Dimension of Constitutional Rights, 37 Oxford J. Legal
Stud. 31, 36–42 (2017).
15
Alexy, supra note 11, at 67–69.
16
Id. at 102.
17
Id. at 54.
18
See Robert Alexy, The Weight Formula, in Frontiers of the Economic Analysis of Law 9, 25 (Jerzy Stelmach,
Bartoz Brożek, & Wojciech Załuski eds., 2007).
874 I•CON 16 (2018), 871–879

Wi,j represents the concrete weight of the principle Pi relative to the colliding principle Pj.
The weight formula defines this concrete weight as the quotient of three factors stand-
ing, so to speak, on each side of balancing. Ii and Ij are of special importance. Ii stands
for the intensity of interference with Pi. Ij represents the importance of satisfying the
colliding principle. Ij, too, can be understood as intensity of interference, that is, as the
intensity of interference with Pj through non-interference with Pi. Wi and Wj stand for
the abstract weights of the colliding principles Pi and Pj. When the abstract weights are

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

equal, which is often the case in collisions between constitutional rights, they cancel
each other out. Ri and Rj represent two factors that are receiving ever greater attention
in recent discussions on constitutional rights. They refer to the reliability of the empiri-
cal and normative assumptions19 concerning the question of how intensive the interfer-
ence with Pi is, and how intensive the interference with Pj would be if the interference
with Pi were omitted. This is a factor that does not refer to the things, that is, it is not an
ontic factor. Rather, it is a factor that refers to the knowledge of things, that is, it is an
epistemic factor. Beyond that, the reliability of the empirical and normative assumptions
can also relate to the classification of the abstract weights, that is, to Wi and Wj.20
A formula like the weight formula, which expresses a quotient of two products, is
sensible only if all of the factors can be represented by numbers. This is the problem of
graduation. Elsewhere21 I have proposed a discrete, that is, a non-continuous, triadic
scale in which geometric sequences are implemented. This scale assigns the values
“light” (l), “moderate” (m), and “serious” (s) to the intensity of interference and to the
abstract weights. These values are expressed by the numbers 20, 21, and 22, that is,
by 1, 2, and 4. Where the epistemic side is concerned, that is, Ri and Rj, one can work
with the stages “reliable” or “certain” (r), “plausible” (p), and “not evidently false”
(e), to which the numbers 20, 2-1, and 2-2, that is, 1, ½, and ¼, are to be assigned.22 By
means of these triads, most of the decisions of constitutional courts can be grasped.
Where they do not suffice, that is, where one has to introduce more attenuated gradu-
ations, they can be extended to double-triadic scales.23
19
That Ri and Rj refer to empirical as well as to normative assumptions can be expressed by the following
equation:
Ri = Rie ⋅ Rni
This equation might be called “reliability equation.” In cases in which both empirical and normative reli-
ability are in question, Ri und Rj are to be substituted by the respective products on the right side of the
reliability equation. In this way, the following refined version of the weight formula enters the stage:

I i ⋅Wi ⋅ Rie ⋅ Rni

Wi, j =
I j ⋅ WJ ⋅ Rej ⋅ Rnj

On this, see Robert Alexy, Formal Principles: Some Replies to Critics, 12(3) Int’l J. Const. L. 511, 514
(2014).
20
It might be possible that this can be captured by the following refinement of the reliability equation,
which would lead to a refinement of the refined weight formula:
Ri = ReIi ⋅ RnIi ⋅ RWi
e ⋅ Rn
Wi
But this shall not be pursued further here.
21
Alexy, supra note 18, at 20–26.
22
Id. at 25.
23
Id. at 22–23.
 Proportionality, constitutional law, and sub-constitutional law: A reply to Aharon Barak 875

2. The externality argument

Much more could be said about principles theory in general. But what has been
exposed here suffices to begin with a discussion of Barak’s four arguments against a
connection between proportionality and constitutional rights that is too close. His first
argument is the externality argument. Barak puts it in the following words:
According to my approach, the scope of the constitutional right as a principle should be dis-
tinguished from the limitations imposed under the rules of proportionality. These limitations

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

are external to the scope of the rights. They affect the ability to realize the right at the sub-
constitutional level. They do not affect the scope of the constitutional right.24

I completely agree with Barak that the limitation of a constitutional right as a

principle does not affect its scope. It is a central thesis of my Theory of Constitutional
Rights that the scope of a constitutional right as a principle, that is, the scope of a
constitutional right as a prima facie position, is not reduced by a limitation.25 This is
the so-called external theory of constitutional rights which is opposed to the so-called
internal theory.26 But Barak seems to understand the external character of limita-
tions as saying more than this. He connects the idea that the scope is not affected with
the idea that proportionality analysis concerns not constitutional rights as such but
only “the realization of [their] scope at the sub-constitutional level.”27 With this, he
attempts to separate propositional analysis from constitutional rights.
At exactly this point Barak critically addresses my theory of derivative constitu-
tional rights norms.28 The idea of derivative constitutional rights norms is closely con-
nected with the first basic law of principles theory, the Law of Competing Principles.
For this law, two concepts are constitutive. The first is the relation of precedence. This
relation is symbolized by “P.” The second is the condition of precedence. It is symbol-
ized by “C.” The result of each application of the weight formula can be expressed on
this basis.29 The standard form is:

(P PP )C
i j

24
Barak, supra note 6, at 349.
25
Alexy, supra note 11, at 178–179.
26
Id. at 179.
27
Barak, supra note 6, at 350. See also Aharon Barak, A Research Agenda for the Future, in Proportionality.
New Frontiers, New Challenges 322, 325 (Vicki C. Jackson & M. Tushnet eds., 2017). Here Barak classi-
fies my approach as an “internal model” and his approach as an “external model.” The internal model
is said to be an internal model because proportional limitations have effects at the constitutional level.
They operate, so to speak, inside the constitutional level. In contrast to this, in Barak’s external model
“[p] roportional limitation has no effect at the constitutional level but at the sub-constitutional level,”
that is, outside the constitutional level.
28
Barak, supra note 6, at 351.
29
Three possibilities exist:
(1) Wi,j > 1 → (PiPPj)C
(2) Wi.j ˂ 1 → (PjPPi)C
(3) Wi,j = 1 → ¬ (PiPPj)C & ¬ (PjPPi)C (stalemate)
These three rules establish the relation between the weight formula and the Law of Competing Principles.
876 I•CON 16 (2018), 871–879

This formula expresses a conditional relation of precedence. It is to be read as follows:

Principle Pi precedes principle Pj under the condition C. The decisive point is that this
conditional relation of precedence implies a rule. Precisely this is expressed by the Law
of Competing Principles, which runs as follows:
If principle Pi takes precedence over principle Pj in circumstances C: (PiPPj)C, and if Pi gives rise
to legal consequences Q in circumstances C, then a valid rule applies which has C as its protasis
and Q as its apodosis: C → Q.30

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

This shall be illustrated with a case. It is the Lebach Judgment of the German Federal
Constitutional Court from 1973. A television channel planned to broadcast a doc-
umentary, “The Soldiers’ Murder at Lebach.” This concerned a crime in which four
sleeping soldiers at the munitions depot of the Federal Army at Lebach were murdered
and weapons were stolen for the purpose of committing further crimes. At the time
of the intended broadcast, a person who has been convicted as an accomplice was
shortly to be released from prison. He was of the view that the broadcast of the pro-
gram in which he was named and in which his picture was given would violate his
personality right, mainly because it would endanger his resocialization. The Federal
Constitutional Court established a competition between the protection of personal-
ity (Pi) and the freedom of media (Pj) and resolved the matter by balancing these two
principles in light of the facts of the case. The result is the precedence of the protection
of personality over the freedom of media reporting under a condition that connects
four elements or sub-conditions: first, a repeated report, second, of a serious criminal
act, which is, third, no longer covered by the interest in up-to-date information, and
which, fourth, endangers the resocialization of the criminal.31 This conditional rela-
tion of precedence implies, according to the Law of Competing Principles, a rule that
runs as follows:
[A] repeated media report, no longer required by the interest in current information, concern-
ing a serious criminal offence, which endangers the resocialization of the offender is constitu-
tionally prohibited.32

This is a highly specific rule that is justified by the constitutional court on the basis of
the competing constitutional rights Pi and Pj in accordance with the two laws of prin-
ciples theory, the Law of Balancing, as explicated by the weight formula, and the Law of
Competing Principles. A rule, justified in this way, is, if the justification is correct, a valid
derivative constitutional rights norm.33 This implies that it is, first, a norm belonging
to the constitutional level and, second, a norm stemming from proportionality analy-
sis. As such, it expresses a definitive limitation of the prima facie right of the freedom
of media (Pj) at the constitutional level.34 This bluntly contradicts Barak’s second argu-
ment which says that “proportionality only operates at the sub-constitutional level.”35

30
Alexy, supra note 11, at 54.
31
BVerfGE 35, 202 (237).
32
Alexy, supra note 11, at 56 (symbols omitted).
33
Id. at 35.
34
Id. at 60, 181.
35
Barak, supra note 6, at 353.
 Proportionality, constitutional law, and sub-constitutional law: A reply to Aharon Barak 877

3. The sub-constitutional level argument

The truth of the sub-constitutional level thesis turns on the import of the claim that
proportionality or the derivative constitutional norm produced by it36 operates only at
the sub-constitutional level. If under “operates” one understands “is applied,” the sub-
constitutional level thesis is true, albeit with one small reservation. This small reserva-
tion concerns the question of whether there can be unconstitutional constitutional
norms. If such norms can exist, then at that juncture unconstitutionality can stem

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

from disproportionality. But this shall not be further elaborated here. In our context,
only one thing is important. This small reservation aside, only general and individual
norms at the sub-constitutional level are subjects of the proportionality test. But how
can the derivative norm, established by proportionality analysis, have the power to
invalidate norms at the sub-constitutional level without belonging to the constitu-
tional level? Belonging to the constitutional level, however, implies having effects on
this level, too. The constitutional level is changed when a derivative norm is added to
it. That is, where “operates only” is understood as “having effects only,” as per Barak’s
reference when he says that derivative constitutional rights norms have “no effect at
the constitutional level,”37 the sub-constitutional level thesis is wrong.
Barak’s main argument for the thesis that proportionality and, with it, derivative
constitutional rights norms have no effect at the constitutional level is that “according
to Alexy’s approach, any law that limits a constitutional right as a principle while sat-
isfying the proportionality doctrine narrows the scope of the constitutional right.”38
My reply is this: The scope of a constitutional right as a principle, that is, the scope of
a prima facie constitutional right is never changed by the derivative constitutional
rights norm corresponding to a proportional limitation. The scope always remains the
same. What happens in such cases is that the prima facie constitutional right is con-
nected at the level of the constitution with a definitive constitutional rights norm that
expresses that the prima facie constitutional right is outweighed under certain condi-
tions (C).
Barak argues that the location of this connection at the constitutional level leads to
serious problems. Under “the principle of coherence and equality” the same derivative
constitutional rights norm “must apply to any similar case. Had we a strong enough
computer, we could have significantly reduced the areas of conflict. Human rights
would have shrunk. We all would have been poorer for it.”39 We would lose “the rich-
ness of constitutional rights.”40 The motto “let a thousand flowers bloom” would no
longer apply in cases of “conflict between rights as principles on the constitutional
level.”41

36
Barak, supra note 4, at 40.
37
Id.
38
Barak, supra note 6, at 353.
39
Id. at 351.
40
Id. See also Barak, supra note 27, at 326.
41
Id.
878 I•CON 16 (2018), 871–879

This impoverishment argument, however, has two shortcomings. The first is that it
has only limited power. The constitutional rights as principles, that is, as prima facie
rights, do not vanish. They remain present at the constitutional level with their great
power. Derivative constitutional rights norms as precedents remain exposed to distin-
guishing and overruling.42 The reason for this is that the validity of derivative consti-
tutional rights norms is based on argumentation.43 New arguments can undermine
the power of old arguments and, with this, the validity of a derivative constitutional

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

rights norm. The old derivative constitutional rights norm is overruled. And new cases
always display new characteristics that can be put forward as reasons for distinguish-
ing. With this, the constitutional level remains rich enough. The second shortcom-
ing of the impoverishment argument is that it has the consequence that too much
is left open at the constitutional level. It provides, so to speak, a constitutional level
that is too rich. Principally, in every conflict everything is possible. The door to an ad
hoc adjudication is wide open. With this, too little weight is attributed to the principle
of legal certainty. This principle requires a constitutional development at the consti-
tutional level, just as it can be achieved by derivative constitutional rights norms as
constitutional precedents.

4. The definitive right argument

Barak’s third argument says that “every constitutional right as a principle is a defini-
tive right.”44 This contradicts my thesis that every constitutional right as a principle is
a prima facie right. A prima facie right expresses an ideal “ought.” An ideal “ought”
is an “ought” that abstracts from any countervailing “ought,” be it a right or a duty.
This abstraction means that it is not yet brought to bear on the limited possibilities of
the real world. By contrast, a definitive right expresses a real “ought.” A real “ought” is
an “ought” that results from an ideal “ought” of a principle when this principle takes
precedence over a competing principle according to the Law of Competing Principles
and the Law of Balancing as expressed in the weight formula.45
Barak seems to be using a different concept of a definitive right. In the context of
his definitive right argument, he emphasizes that proportionality “does not affect
the scope of the constitutional right.”46 I completely agree with this. The scope of a
constitutional right always remains the same. Barak stresses that our approaches are
both “based on a similar two-stage constitutional test: at the first stage the scope of
the constitutional right is determined; at the second, the permitted limitations on it
are determined.”47 The “fundamental difference” is said to be that in my approach

42
Alexy, supra note 11, at 376
43
Id. at 36.
44
Barak, supra note 6, at 354.
45
On this, see Robert Alexy, Ideal “Ought” and Optimization (2017) (unpublished manuscript) (on file with
author).
46
Barak, supra note 6, at 354.
47
Id. at 354–355.
 Proportionality, constitutional law, and sub-constitutional law: A reply to Aharon Barak 879

“both stages operate at the constitutional level,” whereas in his approach only “the
first stage operates at the constitutional level —it determines the scope of the consti-
tutional right (which is always definitive) and the infringement of this right.”48 This
suggests that Barak’s concept of a definitive right refers to the unlimited character of
rights at the constitutional level, that is, to constitutional rights as expressing an ideal
“ought,” whereas my concept of a definitive right refers to the structural counterpart
of a constitutional right expressing an ideal “ought,” that is, to constitutional rights

Downloaded from https://academic.oup.com/icon/article/16/3/871/5165840 by guest on 17 March 2025

as expressing a real “ought.”

5. The broader scope argument

Barak’s fourth argument is that “the scope of a right as a principle at the constitutional
level is different from the scope of the same right at the sub-constitutional level.”49
This difference is explained as follows: “The scope of the right in the constitution,
however, is broader than its scope in the statute.”50 Now a statute, if it complies with
all constitutional requirements, especially with those of the proportionality principle,
can indeed limit a constitutional right. But the scope of the limited constitutional right
is the very same scope as the scope of the unlimited constitutional right. For this rea-
son I would prefer not to connect the difference between the constitutional and the
sub-constitutional level with different scopes. The difference between an ideal “ought”
and a real “ought,” both found at the constitutional level, seems to be preferable.

48
Id. at 355.
49
Id.
50
Id. at 356.